genomewide characterization of host-pathogen interactions by ...

G E N O M E W I D E C H A R A C T E R I Z A T I O N
O F H O S T - P A T H O G E N I N T E R A C T I O N S
B Y T R A N S C R I P T O M I C A P P R O A C H E S
I N A U G U R A L D I S S E R T A T I O N
ZUR
ERLANGUNG DES AKADEMISCHEN GRADES
DOCTOR RERUM NATURALIUM (DR. RER. NAT.)
AN DER MATHEMATISCH-NATURWISSENSCHAFTLICHEN FAKULTÄT
DER ERNST-MORITZ-ARNDT-UNIVERSITÄT GREIFSWALD
VORGELEGT VON MAREN DEPKE
GEBOREN AM 11. 01. 1978
IN HAMBURG
GREIFSWALD, DEN 24. 09. 2010

Dekan:
Prof. Dr. rer. nat. Klaus Fesser
1. Gutachter : Prof. Dr. rer. nat. Uwe Völker
2. Gutachter: Prof. Dr. rer. nat. Christiane Wolz
Tag der Promotion: 22. 12. 2010

Maren Depke
C O N T E N T S
GENOMEWIDE CHARACTERIZATION OF HOST-PATHOGEN INTERACTIONS
BY TRANSCRIPTOMIC APPROACHES 1
CONTENTS 3
ZUSAMMENFASSUNG DER DISSERTATION 5
SUMMARY OF DISSERTATION 9
INTRODUCTION 13
INFECTIOUS DISEASES AND IMMUNE SYSTEM 13
Aspects of Innate Immunity 13
Aspects of Adaptive Immunity 16
Modulation of Immune Reactions 18
STAPHYLOCOCCUS AUREUS 20
General Features of S. aureus 20
S. aureus, a Commensal and Opportunistic Pathogen 21
S. aureus Virulence Factors 22
Mechanisms of S. aureus Adaptation to its Environment 28
Increasing Importance and Danger of S. aureus Infections 29
STUDIES OF HOST-PATHOGEN INTERACTIONS 31
Model Systems for Studies of Host Reactions Potentially Influencing the Outcome of Infections 31
Model Systems for Studies of Host-Pathogen Interactions 35
Questions and Aims of the Studies Described in this Thesis 37
MATERIAL AND METHODS 39
LIVER GENE EXPRESSION PATTERN IN A MOUSE PSYCHOLOGICAL STRESS MODEL 39
KIDNEY GENE EXPRESSION PATTERN IN AN IN VIVO INFECTION MODEL 43
GENE EXPRESSION PATTERN OF BONE-MARROW DERIVED MACROPHAGES AFTER INTERFERON-GAMMA TREATMENT 47
HOST CELL GENE EXPRESSION PATTERN IN AN IN VITRO INFECTION MODEL 51
PATHOGEN GENE EXPRESSION PROFILING 55
Growth Media Comparison Study 55
In vitro Infection Experiment Study 57
Tiling Array Expression Profiling 60
3

Maren Depke
Contents
RESULTS 63
LIVER GENE EXPRESSION PATTERN IN A MOUSE PSYCHOLOGICAL STRESS MODEL 63
KIDNEY GENE EXPRESSION PATTERN IN AN IN VIVO INFECTION MODEL 75
GENE EXPRESSION PATTERN OF BONE-MARROW DERIVED MACROPHAGES AFTER INTERFERON-GAMMA TREATMENT 91
HOST CELL GENE EXPRESSION PATTERN IN AN IN VITRO INFECTION MODEL 111
PATHOGEN GENE EXPRESSION PROFILING 131
Growth Media Comparison Study 131
In vitro Infection Experiment Study 136
DISCUSSION AND CONCLUSIONS 171
LIVER GENE EXPRESSION PATTERN IN A MOUSE PSYCHOLOGICAL STRESS MODEL 171
KIDNEY GENE EXPRESSION PATTERN IN AN IN VIVO INFECTION MODEL 178
GENE EXPRESSION PATTERN OF BONE-MARROW DERIVED MACROPHAGES AFTER INTERFERON-GAMMA TREATMENT 182
HOST CELL GENE EXPRESSION PATTERN IN AN IN VITRO INFECTION MODEL 190
PATHOGEN GENE EXPRESSION PROFILING 197
REFERENCES 207
PUBLICATIONS 233
AFFIDAVIT / ERKLÄRUNG 235
ACKNOWLEDGMENTS 237
4

Maren Depke
Z U S A M M E N F A S S U N G D E R D I S S E R T A T I O N
Mensch und Tier sind regelmäßig Mikroorganismen ausgesetzt. In der Interaktion mit pathogenen
Mikroorganismen entwickelten sich Abwehrmechanismen, die entweder Infektionen verhindern
oder sie zu überwinden helfen. Parallel dazu erwarben Pathogene Mechanismen, um der
Abwehr ihres Wirts zu entgehen. Außer der wirtseigenen Immunregulation und den Faktoren der
Pathogene üben weitere Faktoren wie körperliche Anstrengung und die psychische Verfassung
des Wirts Einfluß auf das Immunsystem aus. Während kurze Streßepisoden sogar die Immunantwort
fördern können, kann sie durch zu lang andauernde Streßphasen negativ beeinflußt
werden. Doch nicht nur die Immunantwort, sondern auch metabolische Prozesse unterliegen
Modifikationen durch solche Stressoren. Deshalb können Untersuchungen zu Wirt-Erreger-
Wechselwirkungen helfen, Mechanismen aufzuklären, die entweder für Wirt oder Pathogen von
Vorteil sind, und dazu beitragen, Interventionsstrategien im Fall von Infektionskrankheiten zu
etablieren. Diese Dissertation beschreibt die Ergebnisse aus Transkriptomstudien zu verschiedenen
Aspekten von Wirt-Erreger-Wechselwirkungen.
Zunächst wurde das Lebergenexpressionsprofil aus einem Mausmodell für chronischen,
psychologischen Streß verwendet, um den Einfluß von Streß auf Metabolismus und Immunantwort
der Tiere zu verdeutlichen. Psychische und physische Stressoren können neuroendokrine,
immunologische, verhaltensbezogene und metabolische Funktionen stören. Vor kurzem wurde
von Kiank et al. publiziert, daß BALB/c-Mäuse ein schwere systemische Immunsuppression,
neuroendokrine Störungen und depressionsähnliches Verhalten entwickeln, wenn sie in einem
Modell für starken, chronischen, psychischen Streß über 4,5 Tage periodisch akustischem Streß in
Kombination mit Bewegungseinschränkung ausgesetzt wurden (Kiank et al. 2006; Brain Behav
Immun. 20(4):359). Außerdem litten diese Mäuse unter deutlichem Gewichtsverlust. Um Gründe
dafür aufzuklären, wurde das Genexpressionsprofil der Leber, die eine Hauptrolle im Stoffwechsel
übernimmt, analysiert. Die Leber übt außerdem eine Wächterfunktion zwischen dem
Verdauungstrakt und dem Blutsystem aus. Deshalb wurde in dem Genexpressionsdatensatz
zusätzlich der Einfluß von psychischem Streß auf immunregulierende Prozesse untersucht.
Bereits nach einer einzelnen, akuten Streßphase wies das hepatische Genexpressionsmuster
deutliche Veränderungen auf. Auch wenn zu diesem Zeitpunkt noch keine metabolischen Veränderungen
sichtbar wurden, begann dennoch eine Genexpressionskaskade, die zu den beobachteten
Störungen führte, nachdem der Streß die chronische Phase erreicht hatte. Dort waren
dann besonders stoffwechselbezogene Gene in ihrer Expression verändert. Die differentielle
Expression betraf Kohlenhydrat-, Aminosäure- und Fettmetabolismus. Es wurde gezeigt, daß
chronischer Streß in weiblichen BALB/c-Mäusen zu einem hypermetabolischen Syndrom einschließlich
Auslösung von Gluconeogenese und Hypercholesterinämie und dem Verlust von
essentiellen Aminosäuren führte. Des weiteren deckte diese Analyse eine veränderte Expression
von Genen der Immunantwort auf. Darin war die Auslösung einer Akute-Phase-Antwort, aber
auch von immunsupprimierenden Abläufen und die Unterdrückung von hepatischer Antigenpräsentation
enthalten. In chronisch gestreßten Mäusen wurde gesteigerte Leukozyteneinwanderung,
verstärkter oxidativer Streß, der aber auch mit gegenregulatorischen Expressionsveränderungen
einherging, sowie Apoptose detektiert.
5

Maren Depke
Zusammenfassung der Dissertation
Die Experimente in der Studie am Modell für psychischen Streß wurden noch ohne den zusätzlichen
Einfluß eines pathogenen Erregers durchgeführt, der aber in der zweiten Studie berücksichtigt
wurde. Hierbei wurde der Einfluß einer intravenösen Staphylokokkeninfektion auf die
Nierengenexpression des Wirts in einem weiteren in vivo Mausmodell analysiert, wobei der
Wildtypstamm Staphylococcus aureus RN1HG und seine isogene sigB-Mutante eingesetzt
wurden. S. aureus, ein Gram-positives Bakterium, besiedelt als persistierender Kommensale den
vorderen Nasenraum von ca. 20 % der menschlichen Bevölkerung. Normalerweise bewirkt die
Besiedlung mit S. aureus keine Erkrankung. Andererseits kann S. aureus aber auch für ein breites
Spektrum an Erkrankungen verantwortlich sein, das von schwachen bis schweren lokalen Infektionen
der Haut oder Rachenschleimhaut über Infektionen innerer Organe (z. B. Endokarditis,
Osteomyelitis) bis zu systemischen Erkrankungen wie Sepsis geht. In das Blut kann S. aureus nach
Verletzungen oder durch medizinische Hilfsmittel wie Katheter gelangen. Ein einfaches Modell,
um solche Blutsysteminfektionen zu imitieren, ist die i. v.-Infektion von Mäusen. Die Wirtsreaktion
kann dabei mit physiologischen, immunologischen und molekularen Messungen aufgezeichnet
werden. In dieser Studie wurde die Transkriptomanalyse auf Mausnierenproben angewandt.
Obwohl Literaturangaben oft von ähnlicher Virulenz von sigB-defizienten Mutanten und Wildtypstämmen
berichten, könnte sich der Mechanismus der Pathogenese – zum Teil auch in
Abhängigkeit von dem gewählten Infektionsmodell – zwischen beiden unterscheiden. Deshalb
sollte mit dieser Studie untersucht werden, ob die Deletion von sigB zu einer veränderten Wirtsantwort
während der Infektion führt. Das Genexpressionsprofil in infiziertem Nierengewebe war
sehr gut reproduzierbar. Der Vergleich mit nichtinfizierten Kontrollen zeigte eine starke, proinflammatorische
Reaktion der Niere, die z. B. Signalweitergabe sogenannter „Toll-like“-Rezeptoren,
Komplementsystem, Antigenpräsentation, Interferon- und IL-6-, aber auch gegenregulatorische
IL-10-Signalwege einschloß. Die Studie konnte keine Unterschiede im Mechanismus der
Pathogenese der S. aureus-Stämme RN1HG und seiner isogenen sigB-Mutante belegen, da sich
die Wirtsantwort in beiden Fällen nicht unterschied. Falls solche Unterschiede tatsächlich
existieren sollten, sind sie womöglich transienter Natur und nur zu früheren Zeitpunkten auffällig.
Effekte von SigB könnten in der in vivo Infektion auch von dem verflochtenen Regulationsmuster
anderer Regulatoren überlagert sein. Des weiteren besteht die Möglichkeit, daß SigB in vivo gar
nicht aktiv ist, wodurch die ähnliche Wirtsreaktion auf Wildtyp und Mutante erklärbar wäre.
Möglicherweise besitzt SigB weniger Eigenschaften eines Virulenzfaktors als vielmehr eines
Virulenzmodulators, der in vivo die Feinabstimmung der bakteriellen Reaktionen übernimmt,
oder SigB ist in speziellen Nischen während einer Infektion von Bedeutung. Solch eine Funktion
würde das Fehlen nachweisbarer Unterschiede zwischen der Wirtsreaktion auf S. aureus RN1HG
und seine isogene sigB-Mutante erklären.
Expressionsstudien an Gewebeproben aus in vivo Modellen zeichnen direkt den relevanten
physiologischen Zustand im Zusammenhang mit allen komplexen Interaktionen und Einflüssen
auf und liegen medizinischen Fragestellungen am nächsten. Dennoch ist es schwer, die einzelnen
Anteile zu unterscheiden, da es sich bei Gewebe immer um eine Mischung verschiedener Zelltypen
handelt, die sogar gegensätzliche Reaktionen aufweisen können. Deshalb wurden zusätzlich
in vitro Modelle untersucht, die sich auf einen einzelnen, definierten Zelltyp fokussierten.
Makrophagen sind in die erste Reaktion auf eine Infektion einbezogen. Sie nehmen, zusammen
mit dendritischen Zellen, eine zentrale Position im angeborenen Immunsystem ein. Durch
ihre Funktion als Wächter und Phagozyten sind sie maßgeblich an der Beseitigung von Infektionen
beteiligt. Peptide phagozytierter Antigene werden durch sie auf MHC-II-Komplexen für
6

Maren Depke
Zusammenfassung der Dissertation
Lymphozyten präsentiert, wodurch die Makrophagen auch an der Regulation der adaptiven
Immunantwort teilhaben. Die Präparation von Knochenmarkstammzellen und ihre in vitro
Differenzierung zu sogenannten “bone-marrow derived macrophages“ (BMM) stellt ein Modell
zur Untersuchung der Reaktionen von Makrophagen dar, das immunologische Einflüsse, die vom
Immunzustand des Tieres sogar unter standardisierten Laborbedingungen ausgehen können,
umgeht. Bis vor kurzem wurden weitere unkontrollierbare Einflüsse durch die Verwendung von
Serum, das undefinierte und variierende Substanzen enthält, mit dem Kulturmedium in die
experimentelle Anordnung eingebracht. Um diese Ursache experimenteller Schwankungen zu
vermeiden, wurde durch Eske et al. ein System der serumfreien Kultivierung von BMM eingeführt
(Eske et al. 2009; J Immunol Methods. 342(1-2):13), das nun für die Untersuchung der Reaktion
von BMM als drittem Teil dieser Dissertation angewendet wurde. Dabei wurden BMM verschiedener
Mausstämme mit IFN-γ, einem Modulator der Makrophagenfunktion und einem der ersten
Signale während der Initiation der Immunantwort, behandelt. Publizierte Experimente zeigten,
daß BMM aus den Mausstämmen BALB/c oder C57BL/6 unterschiedlich auf die Konfrontation mit
Burkholderia pseudomallei reagierten, insbesondere, wenn vorher eine Stimulation mit IFN-γ
stattfand (Breitbach et al. 2006; Infect Immun. 74(11):6300). Auch weitere Studien wiesen Unterschiede
zwischen den beiden Mausstämmen in vivo und in vitro nach. Vor dem Hintergrund der
genetisch bedingten Unterschiede in der Reaktion der BMM wurden nun BALB/c- und C57BL/6-
BMM mit IFN-γ stimuliert, um auf molekularer Ebene genomweit die Reaktion auf IFN-γ als
Initialsignal zu bestimmen. Außerdem sollten in dieser Studie mögliche Unterschiede der
Reaktion von BMM beider Stämme auf IFN-γ charakterisiert werden.
Das Genexpressionsprofil zeigte nach der Behandlung mit IFN-γ in BMM beider Stämme hauptsächlich
induzierte Genexpression. Darin wurden bekannte IFN-γ-Effekte wie die Induktion von
Immunproteasom, Antigenpräsentation, Genen der Interferonsignalwege und von GTPasen/GTPbindenden
Proteinen, sowie der induzierbaren Stickstoffmonoxidsynthase bestätigt. IFN-γabhängige
Genexpressionsänderungen waren zwischen BALB/c- und C57BL/6-BMM in hohem
Maße ähnlich. Sogar nur in BMM des einen Stamms signifikant verändert exprimierte Gene
zeigten einen vergleichbaren Trend in BMM des anderen Stamms. Genexpressionsunterschiede
zwischen BMM beider Mausstämme wurden sowohl in unbehandelten Kontroll-BMM als auch
nach IFN-γ-Behandlung bestimmt. Dabei wiesen ca. 55 % bis 60 % der unterschiedlich stark
exprimierten Gene eine höhere Expression in BALB/c-BMM auf, während die verbleibenden Gene
stärker in C57BL/6-BMM exprimiert wurden. Vergleichbar zu der Beobachtung bei den IFN-γ-
Effekten war auch bei den Unterschieden zwischen BMM der beiden Stämme ein ähnlicher Trend
in beiden experimentellen Behandlungen, Kontrolle und IFN-γ-Stimulation, sichtbar. Die stammabhängig
unterschiedlich exprimierten Gene schlossen immunrelevante und zelltodassoziierte
Gene ein, aber die Abdeckung der funktionellen Gruppen war begrenzt. Phenotypische Unterschiede
in der Reaktion von BALB/c- und C57BL/6-BMM wurden nach Literaturangaben zumeist
in der Anwesenheit von IFN-γ und eines weiteren Stimulus wie LPS oder Infektion beobachtet.
Um die molekularen Ursachen für die beobachteten Unterschiede in der Abtötung von
Pathogenen oder der Produktion von Cytokinen zu bestimmen, müssen weitere Proben, die
außer IFN-γ einem weiteren Stimulus ausgesetzt waren, eingeschlossen werden.
Nicht nur Immunzellen bzw. Phagozyten kommen mit Pathogenen in Kontakt, sondern auch z. B.
Epithel- oder Endothelzellen, die für den Erhalt von Struktur und Funktion von Wirtsorganen und
-gewebe verantwortlich sind. Diese Zellen sind tatsächlich unter den ersten, die das Eindringen
von Pathogenen erkennen und darauf reagieren. Auch wenn S. aureus den vorderen Nasen-
7

Maren Depke
Zusammenfassung der Dissertation
bereich des Menschen besiedelt, kann das Bakterium Pneumonie verursachen, wenn es z. B.
durch Einatmung oder medizinische Geräte in die Lunge gelangt. Die Bronchialepithelzellinie S9
wurde als Modell für die in vitro Infektion mit Staphylokokken verwendet. Ein neues experimentelles
System erlaubt die Untersuchung der Wirt-Erreger-Interaktion im Zusammenhang mit allen
bakteriellen Faktoren, ob membrangebunden oder sezerniert, und vermeidet zusätzlich störende
Effekte auf die Physiologie der Bakterien durch längeres Bearbeiten, Zentrifugation und Waschen
(Schmidt et al. 2010; Proteomics. 10(15):2801). Das vierte Kapitel dieser Dissertation beschreibt
S9-Genexpressionssignaturen nach in vitro Infektion mit S. aureus RN1HG für die zwei Zeitpunkte
2,5 h und 6,5 h nach Beginn der Infektion. Zum frühen 2,5 h-Zeitpunkt wurde differentielle
Expression nur für die kleine Anzahl von 40 Genen gemessen. Trotz der geringen Anzahl zeigten
diese Gene eine beginnende pro-inflammatorische Antwort, z. B. durch die verstärkte Expression
von Cytokinen (IL-6, IFN-β, LIF) oder Prostaglandinendoperoxidsynthase 2 (PTGS2), aber auch
gegenregulatorische Prozesse wie die verstärkte Expression von CD274, Ligand für den immuninhibitorischen
Rezeptor PD-1, waren erkennbar. Die Wirtsantwort verstärkte sich zum späteren
6,5 h-Zeitpunkt dramatisch, an dem differentielle Expression für 1196 Gene detektiert wurde.
Darin waren Cytokine, Signalwege der Rezeptoren für bakterielle Strukturen, Antigenpräsentation
und an der Immunantwort beteiligte Gene (z. B. GBP, MX, APOL) enthalten. Negative
Auswirkungen auf Wachstum und Proliferation traten noch stärker auf als schon beim früheren
Zeitpunkt beobachtet (z. B. verringerte Expression von Histongenen), und Zeichen apoptotischer
Prozesse wurden sichtbar (z. B. verstärkte Expression von BCL10, BLID, BAK1, BAG1).
Das letzte Kapitel dieser Dissertation befaßt sich schließlich mit Tiling-Array Genexpressionsanalysen
des Pathogens, die zunächst für S. aureus RN1HG in aeroben Schüttelkulturen für verschiedene
Wachstumsphasen als Beginn und Referenzpunkt durchgeführt wurden. Anschließend
wurde das bereits beschriebene in vitro Infektionsmodell der S9-Zellen eingesetzt, um Staphylokokken
nach einer Phase der Internalisierung in ihren Wirtszellen zu den beiden Zeitpunkten
2,5 h und 6,5 h nach Beginn der Infektion wieder zu extrahieren und ihre Genexpression zu
charakterisieren. Das bakterielle Expressionsprofil zeigte die Aktivität des SaeRS Zwei-Komponenten-Systems
in internalisierten Staphylokokken an. Zum Teil in Abhängigkeit von SaeRS wurde
die stärkere Expression von Adhesinen (z. B. fnbAB, clfAB), Toxinen (hlgBC, lukDE, hla) und
Genen, die der Umgehung des Immunsystems dienen (z. B. chp, eap), beobachtet. Außerdem
wurden Expressionsveränderungen metabolischer Gene deutlich, z. B. verstärkte Expression von
Genen der Aminosäurebiosynthese, des TCA-Zyklus und der Gluconeogenese. Dazu passend
wurden Glycolysegene verringert exprimiert und des weiteren Gene der Purinbiosynthese und
Gene, die tRNA-Synthetasen kodieren. Die Expressionsanalyse erfaßte ein bakterielles Genexpressionsprogramm,
welches Literaturangaben einer spezifischen, von der Bakterienstamm-
Wirtszellinien-Kombination abhängigen, transkriptionellen Adaptation des Erregers bestätigte.
Neben klassischen physiologischen und mikrobiologischen Methoden tragen die modernen molekularbiologischen
Technologien unter der Voraussetzung verfügbarer Genomsequenzen deutlich
dazu bei, das Verständnis der Prozesse während des Zusammentreffens von Wirt und Pathogen
zu verbessern. Auf der ersten Ebene der Reaktion verschaffen RNA-Expressionsuntersuchungen
einen Überblick über die möglichen physiologischen und metabolischen Veränderungen. Aus
diesem Grund sind die Ergebnisse eine wichtige Vervollständigung der Ergebnisse anderer Untersuchungsebenen
und vergrößern die Grundlagenkenntnisse zum Geschehen während der
Interaktion des Erregers mit seinem Wirt. Auf lange Sicht ist das Ziel dieser Grundlagenforschung,
dazu beizutragen, Interventionsstrategien für Infektionskrankheiten zu etablieren.
8

Maren Depke
S U M M A R Y O F D I S S E R T A T I O N
Humans and animals regularly encounter microorganisms like bacteria and need to defend
themselves in order to avoid or limit damage. The host developed defense mechanisms to either
avoid infection or to overcome it, whereas the pathogen gained mechanisms to evade these host
defense strategies. The host’s immune system is not only influenced by its own regulatory
mechanisms and by the pathogens’ factors, but also by additional elements like physical efforts
or the psychological state of the organism. While short stressful episodes might even enhance
the immune response, the immune response can be suppressed when the stress becomes
chronic. But not only the immune system, but also metabolic processes might underlie
modification by such stressors. Therefore, the improvement of knowledge on host-pathogen
interactions helps to elucidate mechanisms which benefit either host or pathogen. It will
contribute to the establishment of new intervention strategies during infectious diseases. This
thesis contains results from transcriptome studies on different aspects of host-pathogen
interactions.
First, liver gene expression profiles from a murine chronic stress model served to elucidate
aspects of the influence of stress on the metabolism and the immune response state of the
animals. Psychological and physiological stressors can disturb neuroendocrine, immunological,
behavioral, and metabolic functions and adaptive physiological processes aim to reconstitute a
dynamic equilibrium. Recently, Kiank et al. described that BALB/c mice developed severe
systemic immune suppression, neuroendocrinological disturbances and depression-like behavior
due to 4.5 days of intermittent combined acoustic and restraint stress which served as a murine
model of severe chronic psychological stress (Kiank et al. 2006; Brain Behav Immun. 20(4):359).
Furthermore, mice subjected to chronic stress suffered from severe loss of body mass. Therefore,
gene expression profiles of the liver, which plays the major role in metabolism, were analyzed to
investigate primary causes for the observed loss of body weight. The liver fulfills an important
function as a gatekeeper between the intestinal tract and the general circulation. Thus, the
influence of psychological stress on immune regulatory processes in the liver was additionally
investigated in the gene expression data set.
Hepatic gene expression profiles of stressed mice exhibited distinct changes even after a
single acute stress session. Although metabolic disturbances were not visible yet, a regulatory
gene expression cascade started at that time point which led to the disturbances observed when
stress had become chronic. Considering chronically stressed mice, genes related to metabolic
diseases were most significantly influenced. Differential expression affected carbohydrate, amino
acid, and lipid metabolism. Chronic stress in female BALB/c mice was shown to lead to a
hypermetabolic syndrome including induction of gluconeogenesis, hypercholesteremia, and loss
of essential amino acids. Additionally, the analysis of liver homogenates of acutely and
chronically stressed mice revealed an altered mRNA expression of immune response genes.
These included the induction of the acute phase response, but also of immune suppressive
pathways. Furthermore, repression of hepatic antigen presentation was observed. In chronically
stressed mice, increased leukocyte trafficking, increased oxidative stress together with counterregulatory
gene expression changes, and an induction of apoptosis were detected.
9

Maren Depke
Summary of Dissertation
The in vivo model of psychological stress in a complex mammalian host was performed
without additional influences of a pathogen. This additional factor was addressed in the second
study: Here, the influence of staphylococcal intra-venous infection on the host kidney gene
expression was analyzed in another murine in vivo model using the wild type strain
Staphylococcus aureus RN1HG and its isogenic sigB mutant.
S. aureus, a Gram-positive bacterium, is a persistent commensal in the anterior nares of
approximately 20 % of the human population. The bacterium is found intermittently in 30 % to
60 % of the population, and in non-carriers, which is the remaining fraction of the population,
nasal swabs are never positive for S. aureus. Normally, such carriage does not cause any
symptoms of illness. On the other hand, S. aureus can be responsible for a broad variety of
diseases: S. aureus can cause mild to severe local infections of skin or pharyngeal mucosa, but
also of inner organs (e. g. endocarditis, osteomyelitis) and systemic disease like sepsis. S. aureus
can be transmitted to the blood after body injury or by medical devices like catheters. An
elementary model to mimic blood stream infection is the i. v. infection of laboratory animals, e. g.
mice. Host reactions can be monitored by physiological, immunological or molecular readout
systems. In this study, transcriptome analysis of murine kidney samples was performed.
Although the virulence of sigB deficient strains is often reported to be similar to that of wild
type strains the pathogenesis or pathomechanism of different infection settings might vary.
Therefore, the rationale of this study was to investigate whether the deletion of sigB will lead to
a different reaction of the infected host. Gene expression profiling indicated a highly
reproducible host kidney response to infection with S. aureus. The comparison of infected with
non-infected samples revealed a strong inflammatory reaction of kidney tissue. This included e. g.
Toll-like receptor signaling, complement system, antigen presentation, interferon and IL-6
signaling, but also counter-regulatory IL-10 signaling. However, the results of this study did not
provide any hints for differences in the pathogenesis or pathomechanism of the S. aureus strains
RN1HG and ΔsigB in the selected model of i. v. infection in mice, since the host response did not
differ between infections with the two strains analyzed. If really existing, such differences might
be transient and only apparent at earlier time points. Effects of SigB might also be compensated
for in in vivo infection by the interlaced pattern of other regulators. There is also the possibility of
missing activity of SigB in vivo which could explain the similarity of host reaction to infection with
S. aureus RN1HG and its sigB mutant in this study. SigB might possess only to a lesser extent
characteristics attributed to virulence factors and might act in vivo more like a virulence
modulator and fine tune bacterial reactions, or SigB might be important in special niches during
infection. Assuming such function failure to detect differences in the host’s reaction to S. aureus
RN1HG and its isogenic sigB mutant could be explained.
Tissue expression profiling from in vivo models has the advantage of directly recording the
relevant physiological state with all its complex interactions and influences and its vicinity to
medical questions in the human. Nevertheless, it is very difficult to distinguish the different
components because the tissue samples are always a mixture of different cell types which might
even feature contrary reactions. Therefore, in vitro models were additionally analyzed in which
only one defined host cell type was studied. Macrophages are an example for an immune cell
type involved in the first steps of the encounter between the host and a pathogen. In the innate
immune system, macrophages, together with dendritic cells, hold a central position. They are
main effectors of the clearance of infections by their sentinel and phagocytic function.
Macrophages present phagocytosed antigen derived peptides on MHC-II to lymphocytes. By this
10

Maren Depke
Summary of Dissertation
function, macrophages take part in regulation of the adaptive immune response. The preparation
of bone marrow stem cells and the in vitro differentiation of the stem cells into so-called bonemarrow
derived macrophages (BMM) is a model to study reactions of macrophages. The
advantage of this approach is that these macrophages have never been under any immunological
influence which might result from the immune status of the animal even under standardized
laboratory conditions. Until recently, BMM experiments were standardized only to a limited
extent because serum-supplemented culture medium was used. To overcome sources of
experimental variation resulting from undefined and varying substances in serum, Eske et al.
introduced serum-free culture conditions for BMM (Eske et al. 2009; J Immunol Methods. 342(1-
2):13) which were now applied to study the reaction of BMM as third part of this thesis. BMM of
different mouse strains were treated with IFN-γ, a modulator of macrophage function, which is
one of the first signals during initiation of the immune response in vivo. Host-pathogen
interaction experiments have revealed differences in reactions of BMM derived from BALB/c and
C57BL/6 mice when confronted with Burkholderia pseudomallei especially after IFN-γ stimulation
and at higher multiplicities of infection (Breitbach et al. 2006; Infect Immun. 74(11):6300). Also
other infection studies uncovered differences between these two mouse strains in vivo and
in vitro. Against the background of genetic influences on the BMM reactions, BALB/c and C57BL/6
BMM were stimulated with IFN-γ to specify on a molecular level the reaction to the priming
signal IFN-γ as basic principle. Furthermore, the study aimed to profile potential differences of
reactions between the BMM of both mouse strains.
Gene expression profiling revealed mainly induction of gene expression after treatment of
BMM with IFN-γ. Gene expression changes confirmed known IFN-γ effects like the induction of
immunoproteasome, antigen presentation and associated genes, interferon signaling related
genes, GTPase/GTP binding protein genes, inducible nitric oxide synthase and others. IFN-γ
dependent gene expression changes were highly similar in BALB/c and C57BL/6 BMM. Even for
genes, which were differentially expressed only in BMM of one strain, a similar trend was
observed in the other strain’s BMM. Gene expression differences between BMM of both strains
were analyzed on the level of untreated controls as well as after IFN-γ treatment. Here,
approximately 55 % to 60 % of the differentially expressed genes exhibited higher expression in
BALB/c BMM, whereas the remaining genes featured higher expression in C57BL/6 BMM.
Equivalent to the IFN-γ effects, a similar expression trend in the strain comparisons was visible at
both treatment levels, even when regulation was significant only in one. Differentially expressed
genes between BMM of both strains included immune-relevant genes as well as genes linked to
cell death, but the coverage of functional groups was limited. The phenotypical differences
between the reaction of BALB/c and C57BL/6 BMM were often determined in the presence of
IFN-γ and a second stimulus like LPS or infection. To elucidate molecular reasons for the observed
differences in killing of pathogens or cytokine production, the inclusion of samples subjected to a
second stimulus in addition to IFN-γ is recommended.
Not only immune cells or specially phagocytes get in touch with pathogens, but also cells
responsible for functional and structural integrity of host organs and tissue, like epithelial and
endothelial cells. Such cells are actually part of the first line of recognition and reaction to a
pathogenic invasion into the host. While S. aureus colonizes humans in the anterior nares it can
be a cause of pneumonia when transferred to the lung e. g. by aspiration or medical devices. The
bronchial epithelial cell line S9 was used as an in vitro model system for the infection with
staphylococci. A new experimental system permits the study of host-pathogen interactions in the
11

Maren Depke
Summary of Dissertation
context of all bacterial factors, membrane-bound and secreted, and additionally prevents effects
on bacterial physiology by prolonged handling, centrifugation and washing of bacteria (Schmidt
et al. 2010; Proteomics. 10(15):2801). The fourth chapter in this thesis includes host gene
expression signatures of S9 cells after in vitro infection with S. aureus RN1HG for the two time
points 2.5 h and 6.5 h after start of infection. At the early time point, only the small number of 40
genes was differentially expressed. Nevertheless, these few genes indicated a beginning proinflammatory
response e. g. by the induction of cytokines (IL-6, IFN-β, LIF) or prostaglandinendoperoxide
synthase 2 (PTGS2), but also counter-regulatory processes were discernible e. g. by
the induction of CD274, ligand for the immunoinhibitory receptor PD-1. Furthermore, negative
effects on cell cycle progression became visible. The host cell response was dramatically
aggravated at the later 6.5 h time point. Differential expression was detected for 1196 genes.
These included induced cytokines, pattern recognition receptor signaling, antigen presentation,
and genes involved in immune defense (e. g. GBPs, MX, APOL). Negative effects on growth and
proliferation were even more enhanced in comparison to the early time point, e. g. repression of
histone genes, and signs for apoptotic processes were revealed, e. g. induction of BCL10, BLID,
BAK1, and BAG1.
Finally, the last chapter addresses the pathogen’s expression profile, which was first recorded
from agitated, aerobic S. aureus RN1HG cultures in different growth phases as a starting and
reference point. Afterwards, the already described S9 cell in vitro infection model was used to
extract staphylococci after an internalization phase inside the host cell. Internalized bacteria
were analyzed at the two time points 2.5 h and 6.5 h after start of infection in comparison to
different control samples by tiling array gene expression analysis.
Pathogen expression profiling revealed the activity of the SaeRS two-component system in
internalized staphylococci. Partly dependent on this regulatory system, the induction of adhesins
(e. g. fnbAB, clfAB), toxins (hlgBC, lukDE, hla), and immune evasion genes (e. g. chp, eap) was
observed. Furthermore, expression changes of metabolic genes were recorded. This included the
induction of amino acid biosynthesis pathways, TCA cycle genes, and gluconeogenesis. In
accordance with these findings, glycolysis genes were repressed as well as purine biosynthesis
and tRNA synthetase genes. Expression analysis recorded a distinct bacterial expression program,
which supported literature results of a specific, bacterial strain and host cell line dependent
transcriptional adaptation of the pathogen.
Different approaches can be applied to define even more and specific interactions between
the host and the pathogen. Besides classical physiological and microbiological methods, the
modern molecular technologies can considerably help to improve the understanding of the
processes acting during encounter of host and pathogen provided that genome sequence
information is available. On the first level of reaction, RNA expression profiling will generate an
overview on the potential changes of physiology and metabolism. Therefore, the results are an
important complementation of results from other analysis levels and increase the fundamental
knowledge on processes during the encounter of pathogen and its host. In the long-term, this
basic research will probably contribute to the establishment of intervention strategies during
infectious disease.
12

Maren Depke
I N T R O D U C T I O N
INFECTIOUS DISEASES AND IMMUNE SYSTEM
Humans and animals regularly encounter microorganisms like bacteria and need to defend
themselves in order to avoid or limit damage of their own organismal integrity.
Before any possibly dangerous interaction could occur, they first aim to prevent the nearer
contact by physical mechanisms. Dry, closely connected skin surfaces impede the settlement of
microorganisms, and inner surface epithelium often uses mucus as trapping and rinsing agent or
creates unfavorable conditions e. g. by low pH in the stomach. Such physical barriers cannot
completely prevent colonization. A commensalism of certain microorganisms has been
established. But this is not a disadvantage: The occupation of interaction surfaces by apathogenic
commensals limits the available habitat for pathogens.
Additionally, the host organism has developed active defense mechanisms. On the one hand,
general defense factors are produced in advance and deposited at potential interaction sites.
Such factors, like lytic enzymes or defensin peptides, will exert their properties without further
assistance of the host’s physiological systems. These factors can additionally be further induced
after initiation of the immune response. On the other hand in case of a successful infection by a
pathogen, the host is able to start reaction cascades, which will be activated and enhanced after
contact with the pathogen. The host possesses receptors which bind conserved structures of
pathogens. This enables the host to recognize the infection, initialize alarm cascades, and
activate further antimicrobial factors, which are prepared as inactive precursors. The host also
activates the blood coagulation cascade and by this means aims to confine infection to specific
sites and to reduce spreading. Also cellular defense processes are initiated in sequence of
pathogen infiltration into the body, where a first step is pathogen-unspecific phagocytosis by
macrophages or neutrophils. In the later phases of infection, the host can adapt even more
specialized defense systems to react to the specific infecting agent on a humoral and cellular
level. This approach needs more time, but is most effective and finally enables the host to
overcome the infectious disease.
Aspects of Innate Immunity
Many pathogens possess conserved structures. These can be related to their outer structure
like peptidoglycan of the bacterial cell wall, but also to the very basic genetic information like the
double-stranded RNA constituting the genome of some viruses. Some of them are essential for
the pathogen and therefore highly conserved during evolution. In their co-evolution, both sides,
13

Maren Depke
Introduction
host as well as pathogen, developed mechanisms to fight or to gain some advantage against the
other. The host developed defense mechanisms to either avoid infection or to overcome it,
whereas the pathogen gained mechanisms to evade these host defense strategies. One part of
the host’s immune defense is the innate immune system, for which the detailed information is
encoded in the genome and which has been passed on and further modified and optimized
during evolution. The host as a complex organism commands two main levels of such immune
defense, non-cellular and cellular. On epithelial barriers, the host deposits antimicrobial peptides,
which are often membrane-active and pore-forming and aim to lyse the pathogen during its
attempt to enter the body (Schröder 1999). Also enzymes like lyzozyme are secreted and serve a
similar function.
A very important and complex defense system of cascade-activated components is the
complement system (Dunkelberger/Song 2010). The complement system consists of several
serum proteins, which interact and build complexes, contain several sequential and
interdependent activation steps and enzymatic activities, and finally achieves three main goals:
First, in a process called opsonization the labeling of pathogens with a molecular tag to render it
recognizable for phagocytosis, and second the lysis of the pathogen by pore-forming
components. The third function during activation of the complement is to pass the information of
occurring infection to the other immune defense systems.
Three activation pathways of the complement system exist (Fig. I.1). The first, called classical
pathway, is linked to the non-innate, i. e. adaptive immune system, because it relies in the first
step on antibodies, which recognize specific pathogenic structures. Pathogen surface bound
antibodies bind and activate the C1 factor of the complement system. The following steps of
proteolytic cleavage produce from factors C2 and C4 the smaller “a” and the bigger “b”
fragments. C2a and C4b together form the first very important complex during complement
activation, the C3-convertase, which cleaves C3 into C3a and C3b. The C3-convertase (via C4b)
and C3b will be covalently bound to the pathogen’s cell surface after reaction of an active
thioester bond and accomplish opsonization.
The second complement activation pathway leads to the same effects, but starts with a
different recognition complex including lectins (Fujita et al. 2004). This lectin-activation pathway
does not require antibodies bound to the surface of the pathogen. Here, mannose-binding lectin
(MBL) or ficolin recognize conserved carbohydrate structures of the pathogen. These molecules
belong to the so-called pattern recognition receptors (PRRs), conserved and non-variable
receptors for conserved pathogenic structures, which in turn are called pathogen-associated
molecular patterns (PAMPs). MBL or ficolin are complexed, among others, with a serine protease
(MASP), which achieves the next activation steps of C2 and C4 and formation of the C3-
convertase as described before in the classical pathway.
The third way of complement activation is named alternative pathway. For activation, it uses
the spontaneous hydrolysis of C3 in the plasma to C3(H 2 O), which exposes the molecular site for
covalent binding to pathogenic surfaces. In the presence of such surface, C3(H 2 O) binds to it and
subsequently complement factor B to C3(H 2 O). Finally, factor D cleaves factor B into the
fragments Ba and Bb. The resulting complex of C3(H 2 O)Bb constitutes the initial C3-convertase of
the alternative pathway and cleaves further C3 molecules. The resulting C3b fragments again
opsonize the pathogen, but also bind factor B and D, and form further C3-convertase complexes
C3bBb. At this step, a strong amplification takes place, also when the C3b fragments are derived
from activation of the classical and lectin pathway.
14

Maren Depke
Introduction
Besides opsonization, the complement system aims to lyse the pathogen. This so-called
effector function is initiated after formation of C3b by any of the activation pathways. When C3b
is complexed with the C3-convertases of classical/lectin or alternative pathway it forms the C5-
convertase complex (C2bC4bC3b or C3bBbC3b). The enzymatic function is the cleavage of
complement factor C5 into C5b and C5a. Now, C5b initiates the formation of the membrane
attack complex (MAC). Sequential binding of factors C6, C7 (membrane binding) and C8
(membrane intrusion) leads to the forming of a pore by several C9 molecules. Such pores finally
can lead to the lysis of the pathogen. Because the complement system is a very powerful defense
system, which could lead to immense damage to the host itself, it is tightly controlled by many
regulatory proteins which especially prevent the activation on the host’s own cellular structures.
Fig. I.1: Complement activation pathways.
Abbreviations: MBL – mannose-binding lectin; MASP – MBL-associated serine protease; sMAP – small MBL-associated protein;
C3(H 2O) – hydrolysed C3
From: Foster 2005.
During the activation of the complement cascade, several small peptide fragments, called
anaphylatoxins, are produced. C3a, C5a, and to a lesser extent C4a have inflammatory properties.
In a localized infection, they lead to increased permeability of blood vessels, induce endothelial
adhesion molecules, influence monocytes and neutrophils to increase attachment, and activate
mast cells which further enforce the effect with their own mediators.
Hence, the small peptides link the non-cellular to the cellular part of the innate immune
system. Phagocytes have a central function in the innate immune defense. These cells clear the
pathogens from the site of infection and kill them intracellularly. Monocytes from the blood
stream migrate into the tissue and develop into long-living macrophages, which reside and are
activated in case of an encounter with pathogens. Several organs, especially those at a higher risk
to be exposed to pathogens, embody special types of macrophages, e. g. Kupffer cells of the liver.
15

Maren Depke
Introduction
Neutrophil granulocytes are another group of phagocytes. They are found in the blood stream
and leave it only at sites of infection. During an infection, the numbers of neutrophils can
increase strongly, but these cells are only short-living. Phagocytes recognize pathogenic
structures by membrane PRRs. Binding to the pathogen initiates phagocytosis. First, the
pathogen is enclosed in the phagosomal compartment, which afterwards fuses to lysosomes.
Lysosomes bring enzymes and antimicrobial mediators to the new phagolysosomal compartment
and finally accomplish killing of the pathogen in a process called respiratory burst, during which
enzymes produce toxic reaction products like H 2 O 2 , O 2 – , and NO under consumption of O 2 , the
name-giving effect. During activation, phagocytes produce cytokines and chemokines, which lead
to a pro-inflammatory reaction and to chemotaxis of further immune cells like monocytes and
neutrophils (Janeway et al. 2002).
Aspects of Adaptive Immunity
A third type of phagocytic cells, dendritic cells, links the innate to the adaptive immune
system. These cells ingest different antigens in the peripheral tissue by phagocytosis and
macropinocytosis and and transport them to the lymph nodes, where non-activated cells of the
apaptive immune response wait for an activation trigger. Dendritic cells are the mediator cells
which relay the information of peripherally present antigens and therefore potential infection to
cells which might bear the specific receptor to these antigens. The information is transferred in a
process called antigen presentation. It involves major histocompatibility complex (MHC)
molecules. The complex is formed either by an α-chain with transmembrane domain and a noncovalently
linked beta-2-microglobulin (B2M) molecule as β-chain in class I type of MHC or by an
α- and a β-chain which are both inserted in the membrane in class II type of MHC (Fig. I.2 A).
MHC-I α-chains and MHC-II α- and β-chains form a binding groove for antigen derived peptides
and are encoded in the genome in a highly polymorph manner. Only B2M is encoded by a single
gene. The two types of this complex, class I and class II, present antigenic peptides from different
origin to subsets of T cells, which in turn possess a receptor for MHC-peptide-complexes, the
T cell receptor (TCR).
A
Fig. I.2:
Antigen presentation to T cells via MHC molecules.
A. T cell receptor (TCR) mediated recognition of a peptide antigen bound
to an MHC-derived molecule.
B. Activation of naïve T helper cells requires the presence of co-stimulatory
molecules, e. g. B7.
Abbreviations:
MHC – major histocompatibility complex; APC – antigen presenting cell
From: Andersen et al. 2006.
B
16

Maren Depke
Introduction
Extracellular antigens are presented by MHC-II molecules. After phagocytosis, these antigens
are degraded in the phagolysosome. Transport vesicles from the Golgi, which carry unloaded
MHC-II complexes, fuse to the phagolysosome, peptides are loaded to the MHC-II, and the
complete complex is transferred to the cell surface.
MHC-I serves the presentation of intracellular antigens. Here, the cleavage of proteins is
performed in the cytosol by the proteasome, a high molecular weight protease complex. Peptide
transporters transfer the peptides to the ER, where the loading of MHC-I molecules takes place.
Antigen presentation via MHC-II is restricted to professional antigen presenting cells (APC) of
the immune system, whereas presentation of intracellular antigens via MHC-I is additionally
present in many non-immune cells. Interestingly, natural killer (NK) cells, which belong to the
innate immune response and do not possess antigen specific receptors, sense the missing of
MHC-I presentation, which occurs in some virus-infected or cancer cells, and combat these cells
by inducing apoptosis (Jensen 1999).
The TCR of T cells is only able to bind MHC-peptide complexes tightly when the TCR features
the corresponding specificity to the peptide antigen. A very efficient system uses combination of
different TCR chains and somatic recombination, which is the removal of certain genomic
sections and recombining of the remaining parts, to generate an immense repertoire of different
T cell clones with different antigen specificity. These cells are additionally selected for nonreactivity
to the host’s own antigens and build a reservoir of defense cells for pathogenic
antigens which might interfere with the host during life-time. This is the distinctive feature of the
adaptive immune system: It carries the potential for the defense against almost every pathogen,
but it realizes in an adaptive manner with high specificity only the defense needed in the really
occurring case of infection, which makes it a time-consuming, but highly effective defense
mechanism against pathogens.
T-cells are subdivided into populations differing by the expression of surface antigens and by
the secreted cytokines, which in summary links them to distinct functions. The main two groups
are characterized by expression of CD4 or CD8. CD8 + T cells are cytotoxic effector cells. Their TCR
binds MHC-I molecules and therefore is able to recognize pathogenic intracellular antigens.
Specific binding in case of intracellular infection activates processes to destroy the cell and kill the
pathogen. CD4 + T cells are also called helper T cells. Their TCR binds to MHC-II molecules and thus
recognizes antigens derived after phagocytosis. In consequence, CD4 + T cells initiate mechanisms
fighting extracellular infection or disposing extracellular antigens. This is on the one hand a link
back to the innate immune cells, e. g. macrophages, which can be activated, but on the other a
link to a further aspect of the adaptive immune response. Besides the cellular adaptive immune
response the host commands an adaptive humoral immunity mediated by antibodies, which
originate from B cells (plasma B cells). In a process similar to the generation of specific TCR in
T cells, the B cell produces its own B cell receptor (BCR), which already includes the specificity of
the later antibodies. B cells are able to bind antigens via their BCR, phagocytose and degrade
them and finally present them on their surface as MHC-II-peptide complexes. An activated CD4 +
T cell with a specificity to this peptide can activate the B cell, which in turn starts further
differentiation steps and the production of antibodies. These antibodies autonomously bind their
specific antigen, and constitute an opsonization factor for the clearance of antibody-antigen
complexes by phagocytosis. Furthermore, they are the activation complex for the complement
system as described before (Janeway et al. 2002).
17

Maren Depke
Introduction
As T cells have a central position in the immune defense as cytotoxic effectors or potent
activators of other immune cells, MHC-molecule binding alone does not lead to their activation,
but co-stimulatory signals are necessary. These signals originate from co-stimulatory surface
receptors like the B7-family of proteins expressed by antigen-presenting cells (APC), e. g. the
dendritic cells. Only when both signals initiate signal transduction at the same time, the T cell
activation takes place (Fig. I.2 B).
Modulation of Immune Reactions
The immune system is not only influenced by its own regulatory mechanisms and by the
pathogens’ factors, but also by additional elements like physical efforts (Gleeson et al. 1995) or
the psychological state of the organism (Glaser et al. 1999). While short stressful episodes might
even enhance the immune response, the immune response can be suppressed when the stress is
lasting too long. Immune suppression is mainly mediated by glucocorticoids (Dallman 2007).
Thus, different stressors might affect the demands of time for fighting an infection or even the
success of immune defense mechanisms (Fig. I.3, Peterson PK et al. 1991, West et al. 2006). But
not only the immune system, but also metabolic processes might underlie modification by such
stressors. Increasing demands and overwhelming environmental stimuli in the modern society
continuously heighten the stress level of humans and escalate the pathogenesis of stressassociated
illness such as the metabolic syndrome or depression and increase the risk of
infections (Bartolomucci 2007, Leonard 2006, Lundberg 2005).
STRESSOR
type
intensity
timing
duration
PATHOGEN
species, strain (virulence)
inoculum size
HOST
species (genetics)
age
sex
concomitant disease
nutritional status
previous experience
with pathogen (immunity)
with stressor (tolerance)
INFECTIOUS DISEASE
Fig. I.3: Various factors can modify the impact of a stressor on the
pathogenesis of an infectious disease.
From: Peterson et al. 1991.
symptomatic
infection
DEATH
asymptomatic
infection
HEALTH
A physiological stress response is short lasting and physiologically important for survival to
cope with a changing environment or to deal with potentially life-threatening situations.
Adaptive processes are very rapidly mounted to reconstitute a balanced allostatic system in the
stressed body which primarily include the neuroendocrine and immune system. Stress-induced
neuroendocrine alterations include activation of the sympathetic nervous system with increased
secretion of catecholamines, and stimulation of the hypothalamus-pituitary-adrenal (HPA) axis
18

Maren Depke
Introduction
with heightened release of glucocorticoids. Prolonged and increased release of catecholamines is
associated with cardiovascular diseases such as hypertension, myocardial infarction, or stroke.
Excessive secretion of glucocorticoids was linked to diabetes, dyslipidemia, cardiovascular
alterations, immunosuppression, and mood disorders (Lundberg 2005, McEwen 2004,
Viswanathan/Dhabhar 2005, Wrona 2006).
Recently, Kiank et al. described that BALB/c mice developed severe systemic immune
suppression, neuroendocrinological disturbances and depression-like behavior due to 4.5 days of
intermittent combined acoustic and restraint stress which serves as a murine model of severe,
chronic psychological stress (Kiank et al. 2006, 2007a). Immunosuppression was substantiated by
a heightened anti-inflammatory cytokine bias, an apoptotic loss of lymphocytes leading to
lymphocytopenia, T cell anergy, and impaired phagocytic and oxidative burst responses. The
immunodeficient state increased the animals’ susceptibility to experimental infection with E. coli
(Kiank et al. 2006, 2007b). Repeatedly stressed mice actually developed spontaneous bacterial
infiltrations into the lung measurable even 7 days after the last chronic stress cycle, which was
associated with a reduced inducibility of IFN-γ, a cytokine that was shown to be important to
prevent spreading of translocated commensals from the gut (Kiank et al. 2008). On the one hand,
stress-induced immunosuppression was accompanied with a reduced clearance of experimental
infections in the long term, but on the other hand, with attenuation of a hyperinflammatory
septic shock (Kiank et al. 2006, 2007a). Furthermore, behavioral alterations with increased
depression-like behavior and neuroendocrine alterations such as prolonged activation of the HPA
axis and increased turnover of catecholamines were documented.
Finally, a prominent stress-induced loss of body mass without significant changes of food and
water intake during the observation period became detectable (Kiank et al. 2006). It is known
that stress exposure is linked to changes of body weight. There is evidence that hypothalamic
control of food intake is influenced by stress, which in consequence alters metabolism. In such a
situation, some people lose and others gain weight in response. However, the molecular
mechanisms of the stress to body weight connection remain to be elucidated.
It is important for the understanding of stress-induced immunoregulatory mechanisms that
stress hormones like catecholamines and glucocorticoids can modulate several liver functions
including carbohydrate, protein and lipid metabolism or affect the immune response. Especially
corticosteroids can suppress inflammatory processes by preventing the release of
proinflammatory mediators, by diminishing immune cell trafficking, phagocytosis and radical
production or by down-regulating antigen presentation, by inhibiting lymphocyte proliferation,
and by inducing apoptosis of immune cells. Thus, prolonged increase of glucocorticoid levels
during chronic stress can activate hepatic catabolic pathways but also effectively suppress the
local immune response (Bartolomucci 2007, Chida et al. 2006, Elenkov 2004, Swain 2000).
19

Maren Depke
Introduction
STAPHYLOCOCCUS AUREUS
General Features of S. aureus
Staphylococci are Gram-positive bacteria of approximately 1 µm diameter, which often
aggregate in grape-like structures. The golden color of Staphylococcus aureus colonies – derived
from staphyloxanthin and other C 30 triterpenoid carotenoids (Marshall/Wilmoth 1981) which
function as antioxidant protection molecules (Liu GY et al. 2005) – led to its naming. Several
S. aureus properties like the ability to perform hemolysis and expression of catalase and
coagulase allow differentiation of the species from other bacteria. S. aureus can produce energy
by aerobic respiration, but is facultatively able to survive in anaerobic environments. In situations
without oxygen it can apply lactate fermentation. S. aureus seems to be naturally auxotrophic
and requires e. g. amino acids as growth supplement. The staphylococcal cell wall peptidoglycan
features a special crossbridge structure formed by multiple glycine residues. These are the target
structures for lysostaphin, a staphylocidal enzyme from Staphylococcus simulans first described
by Schindler and Schuhardt in 1964, which was taken over by scientist for several laboratory
research purposes and which might even be applied as mupirocin substitute for eradication of
staphylococcal nasal colonization (Recsei et al. 1987, Kokai-Kun et al. 2003). Staphylococci are
resistant to lysozyme, a muramidase, which is present as antimicrobial enzyme in body fluids and
at sites of infection as part of the innate immune defense. The resistance is mediated by O-
acetylation in the muramic acid of peptidoglycan (position C6-OH) which is conducted by the
peptidoglycan-specific O-acetyltransferase OatA. OatA mutants are susceptible to lysozyme (Bera
et al. 2005).
The S. aureus genome has 32 % GC content. Its chromosome is about 2.7E+06 to 2.9E+06
nucleotides in length, contains about 2600 to 3000 genes, and has been sequenced for several
strains until now (Table I.1). Additionally, sequence information for several plasmids is available
(http://www.ncbi.nlm.nih.gov/sites/entrez?db=Genome).
For sub-division purposes, the S. aureus genome has been classified into two parts. The core
genome, which occupies approximately 75 % of the complete genome, includes genes with basic
functions e. g. in house-keeping and central metabolism. These genes are present in the genomes
of most strains and exhibit a high degree of conservation in the individual genes’ sequences in
different strains. The core part also features a high similarity to the phylogenetically related
Bacillus species genomes (Hiramatsu et al. 2004). The remaining 25 % belong to the accessory
part of the genome. This part consists of further genes which have been acquired by some
S. aureus strains e. g. from mobile genetic elements. Here, variation in the presence of genes is
observed in different strains. Many virulence genes are classified as fraction of the accessory
genome (Lindsay/Holden 2004).
S. aureus has many options of varying its accessory genome part by different mobile genetic
elements (Plata et al. 2009). Staphyloccoccal pathogenicity islands (SaPI) are located at constant
positions of the genome, contain superantigen genes and sequences with similarity to prophages
(Lindsay et al. 1998, Novick RP 2003). Similarly, genes of the νSa-family of pathogenicity genomic
islands, of which several types (1-4, α, β, γ) exist, code for virulence factors and toxins (Gill et al.
20

Maren Depke
Introduction
2005). SaPIs and some νSa-members can be excised from the genome and are involved in
horizontal gene transfer. S. aureus strains often harbor prophages, which also incorporate
virulence factors and are an aid for their transfer between strains and in general influence
S. aureus evolution (Lindsay/Holden 2004). The Panton-Valentine leukocidin is a phage-encoded
staphylococcal toxin. S. aureus NCTC8325 carries three lysogenic phages, φ11, φ12, and φ13, of
which φ13 introduces the virulence factor staphylokinase (sak) into the strain (Iandolo et al.
2002). Kwan et al. reported in 2005 the genome sequences and predicted proteins of 27
bacteriophages of S. aureus (Kwan et al. 2005). Additionally, gene transfer can be accomplished
by shorter insertion sequences or by longer transposons which bring along their own transposase
genes and recognition sites (Hiramatsu et al. 2004, Plata et al. 2009). Plasmids are a further place
of encoding useful, although not essential information, e. g. for antibiotic resistances. S. aureus
strains include different plasmids of varying size and copy number. For example, the methicillinresistant
S. aureus strain COL owns the 4440 nt plasmid pT181 (accession number NC_006629;
http://www.ncbi.nlm.nih.gov/sites/entrez?db=Genome), whose sequence became available in
2005. Some S. aureus strains have also acquired the vanA operon for vancomycin resistance from
an Enterococcus spp. transposon via a conjugative plasmid (Péricon/Courvalin 2009).
Table I.1: Complete, RefSeq genome sequences for bacterial chromosomes of Staphylococcus aureus strains from NCBI Entrez
Genome database (http://www.ncbi.nlm.nih.gov/sites/entrez?db=Genome) of August 25 th 2010.
accession number strain length / nt date created genes
NC_002745 Staphylococcus aureus subsp. aureus N315 2814816 2001/04/21 2664
NC_002951 Staphylococcus aureus subsp. aureus COL 2809422 2001/09/14 2723
NC_002758 Staphylococcus aureus subsp. aureus Mu50 2878529 2001/10/04 2774
NC_002952 Staphylococcus aureus subsp. aureus MRSA252 2902619 2001/11/06 2839
NC_002953 Staphylococcus aureus subsp. aureus MSSA476 2799802 2001/11/06 2715
NC_003923 Staphylococcus aureus subsp. aureus MW2 2820462 2002/05/31 2704
NC_007622 Staphylococcus aureus RF122 2742531 2005/11/24 2663
NC_007793 Staphylococcus aureus subsp. aureus USA300_FPR3757 2872769 2006/02/11 2648
NC_007795 Staphylococcus aureus subsp. aureus NCTC8325 2821361 2006/02/18 2969
NC_009487 Staphylococcus aureus subsp. aureus JH9 2906700 2007/05/23 2816
NC_009632 Staphylococcus aureus subsp. aureus JH1 2906507 2007/07/03 2870
NC_009641 Staphylococcus aureus subsp. aureus Newman 2878897 2007/07/06 2687
NC_009782 Staphylococcus aureus subsp. aureus Mu3 2880168 2007/09/06 2768
NC_010079 Staphylococcus aureus subsp. aureus USA300_TCH1516 2872915 2007/12/03 2799
NC_013450 Staphylococcus aureus subsp. aureus ED98 2824404 2009/11/19 2752
S. aureus, a Commensal and Opportunistic Pathogen
Staphylococcus aureus is a persistent commensal in the anterior nares of approximately 20 %
of the human population (persistent carriers). The bacterium is found intermittently in 30 % to
60 % of the population, and in non-carriers, which is the remaining fraction of the population,
nasal swabs are never positive for S. aureus (Kluytmans et al. 1997, Wertheim et al. 2005). Some
authors subdivide the group of intermittent carriers into intermittent and occasional carriers
(Eriksen et al. 1995). More recent publications classify the carriage status only into two groups:
persistent carriers and others (van Belkum et al. 2009).
The occurrence of the two contrary groups of persistent and non-carriers strongly hints for an
influence of the host’s immune system on the possibility for S. aureus to colonize persistently
21

Maren Depke
Introduction
(Foster 2009). Others have already published polymorphisms associated with each phenotype,
linking variants enhancing the immune response with reduced colonization risk (van den Akker et
al. 2006, Emonts et al. 2007; Foster 2009). In the main site of colonization, the anterior nares,
S. aureus uses its adhesins like ClfB, IsdA, SdrC, SdrD, and SasG to attach to the host tissue (Foster
2009).
Normally, such carriage does not cause any symptoms of illness. On the other hand, S. aureus
can be responsible for a broad variety of diseases: S. aureus can cause mild to severe local
infections of skin or pharyngeal mucosa, but also of inner organs (e. g. endocarditis,
osteomyelitis) and systemic disease like sepsis. The pathogenicity of S. aureus is a world-wide
problem. By the “SENTRY Surveillance Program”, S. aureus was described to be leading cause of
bloodstream, lower respiratory tract and skin/soft tissues infections (Diekema et al. 2001).
S. aureus produces several toxins that are responsible for toxin related diseases like toxic shock
syndrome or scalded skin syndrome. Staphylococcal endotoxin can cause food poisoning in case
of consumption of contaminated meals. S. aureus carriage is a risk factor for bacteremia. In the
majority of bacteremia cases, the strain isolated from blood is identical with the nasal colonizing
strain. But paradoxically, carriers have a lower mortality than non-carriers in case of bacteremia
(van Eiff et al. 2001, Wertheim et al. 2004).
S. aureus has been known to be an extracellular pathogen for long time (Finlay/Cossart 1997).
But it has been also proven that S. aureus is not only an extracellular pathogen but that it can be
internalized by non-professional phagocytes like epithelial cells (Almeida et al. 1996, Hudson et
al. 1995). Staphylococcal fibronectin-binding proteins (FnBP) are important mediators for the
uptake of bacteria by the host cell, but it is not clear whether the uptake is an attribute of
bacterial pathogenicity or of host defense (Sinha et al. 2000). The process of internalization
implies an active participation of host cells (Hudson et al. 1995).
S. aureus Virulence Factors
S. aureus has the genetic information to produce a wide range of virulence factors, which help
to improve the bacterium’s survival e. g. in infection settings and to fight against the defense of
its host. Some are even required for its ability to establish an infection (Fig. I.4).
S. aureus secretes several enzymes and exotoxins. A direct mode of impairing the host’s
defense is to lyse host cells, which additionally renders nutrients accessible for the pathogen. As
first virulence factors secreted enzymes (proteases, lipases, elastases, hyaluronidase and others)
degrade host tissue molecules and therefore disintegrate tissue structure and defense barriers
and facilitate first invasion and later spread of bacteria (Gordon RJ/Lowy 2008). All bacteria need
iron for their survival. This essential nutrient is not freely available in the host but is sequestered
by host proteins. S. aureus owns iron uptake promoting proteins, which are encoded by the ironresponsive
surface determinant locus isd. This system includes binding proteins for different ironcontaining
host proteins, transfer and transporter proteins, and proteins which release the iron
into the bacterial metabolism (Maresso/Schneewind 2006).
S. aureus carries different hemolysins which not only act on erythrocyte membranes, but also
on that of other cell types. Alpha-hemolysin, which is also named alpha-toxin (encoded by the
gene hla), is one of the prototypes of pore-forming toxins. This toxin has surprising characteristics
of being water-soluble, but also able to form pores in hydrophobic membrane surrounding, and it
22

Maren Depke
Introduction
contains all properties for its own insertion into membranes and therefore needs no assistance
by proteins like chaperones (Menestrina et al. 2001). Other cytolytic toxins of S. aureus are betahemolysin
(hlb), gamma-hemolysin (hlgA, hlgB, hlgC), delta-hemolysin (hld), and the groups of
leukocidins (luk). The Panton-Valentine leukocidin (PVL, lukF-PV and lukS-PV), a long-known and
well-studied bicomponent toxin, depends in its pore-forming activity on two components, which
first form themselves a heterodimer (Kaneko/Kamio 2004). Also the other leukocidins and Hlg are
bicomponent lysins. Hlg can form either the AB toxin or the CB toxin depending on the
association of the subunits to heterodimers.
Fig. I.4: Staphylococcal structural and secreted virulence factors.
A. Overview on surface and secreted proteins. TSST-1 – toxic shock syndrome toxin 1.
B. Cell envelope structure and localization of selected proteins.
C. Domains of cell wall anchored proteins.
From: Gordon/Lowy 2008.
Other important staphylococcal exotoxins are the so-called superantigens (SAg). These
proteins disturb the host’s cellular adaptive immune response. Normally, T cells bind with their
antigenic-peptide specific receptor (T-cell-receptor, TCR) to MHC-molecules, which present
peptides derived from proteins inside or outside the antigen-presenting cell. This forms the socalled
“immunological synapse” between antigen-presenting cell (APC) and T cell. In case of both
specific recognition of the peptide by TCR as well as presence of further pro-stimulatory signals,
the T cell will be activated, start to proliferate and secrete cytokines, and in general pursue their
function in immune response. Superantigens interfere in this interaction. They can bind MHC-IImolecules
on APC (with different preferences for MHC class II alleles) and also the TCR, in detail a
subset of possible recombined TCR molecules characterized by certain Vβ-fragments in the β-
chain (Herman et al. 1991). By this means, they build a bridge linking both receptors
independently of an antigenic peptide and receptor specificity. This cross-linking in the
trimolecular complex TCR/MHC-II/SAg induces a massive T cell proliferation and cytokine
production leading to shock-like syndromes and generalized life-threatening damage. But the
23

Maren Depke
Introduction
triggered host reaction does not help to fight the causative pathogen, because the T cell
activation is antigen-independent and therefore unspecific to the pathogen. This, together with
induction of immunosuppression, might be one of the advantages which superantigens grant the
bacteria, although the question why bacteria harbor superantigens has not been completely
clarified yet (Herman et al. 1991).
Twenty superantigens of S. aureus have been described so far: toxic shock syndrome toxin-1
(TSST-1), staphylococcal enterotoxins (SE A-E, G-J), and staphylococcal enterotoxin-like toxins (SEL
K-R, U, U2, V). They cause diseases like toxic-shock syndrome and are linked to atopic dermatitis
disease mechanism and allergic rhinitis (Fraser/Proft 2008). Mobile genetic elements of S. aureus
encode the superantigens (Novick RP 2003).
Further immunomodulation is achieved by extracellular adherence protein (Eap), which is also
called MHC class II analogous protein (Map). This protein has similarity to MHC class II molecules
(Jönsson et al. 1995) and influences T cell response during infection in favor of the infecting
S. aureus (Lee et al. 2002). This inhibitory effect on T cells is dose-dependent and occurs only at
high concentrations of Eap, while Eap stimulates T cell proliferation when present in low
concentrations (Haggar et al. 2005).
Virulence of S. aureus is promoted by the ability to attach to the host’s tissue, i. e. to the
host’s extracellular matrix or the host’s cells, by its own surface components. Microbial surface
components recognizing adhesive matrix molecules (MSCRAMMs) mediate the attachment,
which can be regarded as the first initializing step for the establishment of infection
(Gordon RJ/Lowy 2008). Molecules with similar function, but secreted in contrary to the
microbial surface bound MSCRAMMs, are called secretable expanded repertoire adhesive
molecules (SERAMs).
Fibronectin binding proteins A and B (encoded by fnbA and fnbB) have binding properties for
fibronectin, a high molecular weight protein of the host’s extracellular matrix (ECM). S. aureus
also possesses adhesins binding to other ECM components like collagen (collagen binding protein
Cna), elastin (elastin-binding protein EbpS). Clumping factor A and B (encoded by clfA and clfB)
are fibrinogen binding proteins. This property results on the one hand in adhesin function of
clumping factor (Clf). But as Clf can not only bind to the blood coagulation precursor fibrinogen,
but also activate fibrinogen to form fibrin, Clf has on the other hand agglutination properties.
Structurally similar proteins to Clf exist. These are serine-aspartate (SD) repeat proteins, which
are encoded by the genes sdrC, sdrD and sdrE, but their function is still unknown (Foster/Höök
1998).
A MSCRAMM-function of protein A is hypothesized from its ability to bind von-Willebrandfactor
(vWF) in the environment of damaged epithelium where vWF originally enables the
adhesion of platelets (Hartleib et al. 2000). Interestingly, a novel vWF binding protein (vWBP,
encoded by vwb) of S. aureus has been discovered more recently (Bjerketorp et al. 2002).
Fibronectin-binding protein (FnBP), clumping factor A (ClfA), protein A (Spa), and collagen
binding protein (Cna) belong to the same group of MSCRAMMs. They contain a LPXTG motif
which is cleaved during protein secretion through the cellular membrane with sequential
anchoring of the remaining protein via the now accessible carboxylgroup of threonin (T) to the
cell wall peptidoglycan (Foster/Höök 1998). Further MSCRAMMs in S. aureus are laminin-,
vitronectin-, thrombospondin- and bone sialoprotein-binding proteins (Patti et al. 1994).
MSCRAMMs do not only mediate the adhesion to the host’s extracellular matrix, but they are
also involved in invasion into the host cells. Fibronectin-binding proteins attach the bacterial cell
24

Maren Depke
Introduction
via fibronectin as linker to the host’s α 5 β 1 integrin molecules, which initiates the internalization of
the bacterial cell into the host cell (Hauck/Ohlsen 2006, Peacock et al. 1999, Schwarz-Linek et al.
2004, Sinha et al. 1999). S. aureus strains harboring the SCCmec type I also possess the pls gene
located therein. This gene encodes the plasmin sensitive surface protein Pls, which reduces,
contrarily to the MSCRAMMs, the adherence of S. aureus to host structures and its cellular
invasiveness. Very recently, it has been published that Pls causes this effect by steric hindrance
and blocking adhesin and host factor interaction (Hussain et al. 2009).
An example for a SERAM is the staphylococcal coagulase. The extracellular enzyme is able to
bind prothrombin, which is the enzyme precursor for the coagulation activation by conversion of
fibrinogen to fibrin. Thrombin activation by staphylococcal coagulase does not occur by
proteolysis but by forming an active complex of both molecules in a stoichiometric ratio of 1:1,
which is called staphylothrombin (Kawabata et al. 1985, Kawabata/Iwanaga 1994). Moreover,
coagulase molecules can be attached to the staphylococcal cell surface where it can bind directly
to fibrinogen also without presence of prothrombin (Bodén/Flock 1989).
S. aureus is aiming to evade the immune response, and thus, it has gained several immunemodulating
properties. After the very first steps of infection, the immune response has to be
initiated by chemoattractants which recruit defense cells like neutrophils or macrophages to the
site of infection. The bacterial protein chemotaxis inhibitory protein of staphylococci (CHIPS)
blocks with two different domains the host cell receptors for the host chemoattractant C5a from
the activated complement system and for bacterial formylated peptides, which are secreted by
the pathogen (de Haas et al. 2004, Murdoch/Finn 2000). Tightly associated with interference in
the chemotaxic process is the already mentioned bacterial extracellular adherence protein (Eap).
This protein blocks the endothelial cell receptor ICAM-1. During the normal immune response,
ICAM-1 has receptor function for the leucocyte’s membrane protein LFA-1 and thus allows
leucocyte adhesion at sites of infection. Adhesion is followed by entry of leucocytes into the
tissue in the two steps of diapedesis and extravasation. In cases when ICAM-1 is already occupied
by Eap, not only leucocyte adhesion but also the following steps are impeded (Chavakis et al.
2002).
S. aureus protects and hides its treacherous typical bacterial cell surface structures by a
capsule of polysaccharides. The capsule inhibits binding of opsonic host factors and thus reduces
the phagocytosis of the bacterial cell. An increased production of capsular polysaccharides
increases the virulence of S. aureus (Nilsson et al. 1997, Thakker et al. 1998).
Besides their already mentioned function of MSCRAMMs, clumping factor and protein A also
have anti-phagocytic functions. Clumping factor binds the host molecule fibrinogen which then
serves as camouflage net on the bacterial cell surface. Protein A binds immunoglobulin G (IgG) in
a way beneficial for the bacterium, but not for the host. While the normal antibody-binding with
the antigen-specific F ab part will opsonize the bacterium for recognition by immune defense
mechanisms, the binding of antibodies by protein A occurs via the F c part and thus prevents the
mediation of opsonization signals (Uhlén et al. 1984).
Another protein which interferes with opsonization for phagocytosis is Staphylococcus
complement inhibitor (SCIN). SCIN binds to the C3-convertase complexes from the three
complement activation pathways and inhibits C3b deposition on the pathogen’s surface.
Therefore, the following steps of complement opsonization are impaired. Interestingly, the
complement-inhibitory function of SCIN is specific for the human host (Rooijakkers et al. 2005b).
In a similar function, complement factor C3 is the target of the highly conserved, staphylococcal
25

Maren Depke
Introduction
extracellular fibrinogen-binding protein Efb. Efb most probably inhibits the binding of C3b, the
bigger fragment of C3 after cleavage during complement activation, to the bacterial surface and
therefore reduces opsonization. Efb is able to bind both C3 and fibrinogen at the same time,
because the C3-binding region is located C-terminal and the fibrinogen-binding region N-terminal
in the Efb protein (Lee et al. 2004a, 2004b). More recently, a homologous and 44 % identical
protein to Efb was identified. The protein called “Efb homologous protein”, Ehp, features the
same C3b-deposition inhibitory function as Efb does, but in an even more potent manner. An
additional C3-binding domain is proposed to cause this stronger inhibitory effect (Hammel et al.
2007).
S. aureus does not only use its own proteins for immune-modulatory purposes, but also
engages and manipulates host factors to perform a protective effect for the bacterium. Host
plasminogen is the inactive precursor of the serine protease plasmin, which is the main enzyme
involved in fibrinolysis, the degradation of fibrin clots after injury and blood coagulation. Bacterial
staphylokinase binds plasminogen to the staphylococcal cell surface and mediates the activation
of the precursor into active plasmin. The plasminogen activation mechanism is not mediated via
cleavage by staphylokinase as the bacterial protein does not possess enzymatic activity. More
precisely, staphylokinase binds plasminogen in a 1:1 stoichiometric ratio and renders the enzyme
precursor by formation changes more susceptible for activation. Activation finally occurs by other
already active proteases, e. g. trace amounts of plasmin, which are able to start a reinforcing
activation cascade (Silence et al. 1995). In a simple mode, active plasmin will help the pathogen
to spread from site of infection, either by enzymatic fibrinolysis of blood coagulation clots or by
degradation of extracellular matrix molecules. More sophisticated, the activated plasmin
protease molecules degrade IgG and C3b from the bacterial surface and thus reverse
opsonization (Rooijakkers et al. 2005a).
When S. aureus is finally phagocytosed it has to deal with reactive oxygen species (ROS).
Superoxide dismutase enzymes and non-enzymatic superoxide dismutase Mn 2+ inactivate
superoxide anion O 2 – . Deletion mutants in superoxide dismutase or manganese cation uptake
systems are less virulent than wild type strains. When ROS react with protein residues to
methionine sulfoxide, these residues are reduced by methionine sulphoxide reductases, which
also have impact on in vivo virulence (Foster 2005, Horsburgh et al. 2002a, Karavolos et al. 2003,
Singh/Moskovitz 2003).
Modulation of complement, opsonization and phagocytosis is not the only mechanism
S. aureus applies for immune evasion. It also developed resistance mechanisms against killing by
antimicrobial peptides, especially after phagocytosis, when the four groups of antimicrobial
substances interfere with the bacterial integrity: small anionic peptides (e. g. dermicidin in airway
surfactant), small cationic peptides (e. g. cathelicidins in neutrophils), cationic disulphide bond
forming peptides (e. g. α- and β-defensins), and anionic and cationic peptide fragments derived
from larger proteins (Foster, 2005). In this context, staphylokinase has a further function. It binds
α-defensins, which are secreted by neutrophils, and inhibits most of the defensins’ bactericidal
potency (Jin et al. 2004). Vice versa, defensins inhibit the staphylokinase’s ability of activating
plasminogen. This effect can be relevant in clinical settings when recombinant staphylokinase
(sakSTAR) is planned to be developed into a thrombolytic pharmaceutical and might be applied in
inflammation paralleled disorders like vascular occlusive diseases (Bokarewa/Tarkowski 2004).
S. aureus secretes different proteases. One of them, aureolysin, cleaves the bactericidal peptide
26

Maren Depke
Introduction
cathelicidin LL-37 at a special site, which is not recognized by the V8 protease. After aureolysin
cleavage, LL-37 loses its antistaphylococcal activity (Sieprawska-Lupa et al. 2004).
S. aureus cell wall is subjected to modifications of the teichoic acid structure. Dlt proteins
(encoded by the dltABCD operon) lead to D-alanine incorporation. The alanine residues are
further esterified. Without these alanine esters the bacterial surface would carry a strongly
negative net charge, which would attract positively charged cationic antimicrobial molecules.
Therefore, Dlt decreases the vulnerability of the bacterial cell wall via reduction of molecule
charge (Peschel et al. 1999). On the other hand, a stronger negative charge would reduce
S. aureus’ capacity to adhere to polystyrene or glass. Dlt mutants are thus impaired in their
capability of biofilm formation (Gross et al. 2001). Also the MprF protein neutralizes charges of
the cell wall, in this case by linking lysine to phosphatidylglycerol (Peschel et al. 2001).
Biofilm formation, which is often associated with intra-vascular medical devices like catheters,
is another option for S. aureus to protect the bacterial cells from the immune system and
antimicrobials. The material of the medical devices will be covered with host (serum) molecules
like fibrinogen or fibronectin soon after implantation. Therefore, it serves as an ideal attachment
site for the various adhesins of staphylococci (Lowy 1998). In biofilms, S. aureus aggregates in a
multi-cellular structure which already hinders contact between immune cells and substances with
bacterial cells by physical means. Bacterial cells are surrounded by a polysaccharide matrix,
composed of poly-N-acetylglucosamine (PNAG). The ica operon, which is present in almost all
S. aureus strains although not all of them form biofilms in in vitro assays, encodes the proteins
necessary for PNAG biosynthesis. Its expression is regulated by IcaR and TcaR and the
environmental conditions, and SarA and its homologues influence the regulation. Quorum
sensing mechanisms are important in coordination of the individual bacterial cell’s gene
expression, which enables the establishment of the biofilm (Fitzpatrick et al. 2005, Grinholc et al.
2007).
For bacteria, the most effective way of evading the immune response is to hide inside the host
cell. Intracellularly, S. aureus can mask its presence by changing its phenotype into the so-called
small colony variant (SCV). These variants are characterized by physiological adaptation like slow
growth and reduced toxin production, which is caused by alterations of the electron transport
chain. They contribute to persistence and recurrence of staphylococcal infections
(McNamara/Proctor 2000, von Eiff et al. 2006)
Not all staphylococcal strains harbor the information for all virulence factors in their genome,
and they do not necessarily express all virulence factors which are encoded. Virulence factor
expression is limited by strict regulation to avoid wasting of energy or nutrient resources.
Furthermore, the long list of different virulence factors and mechanism underlines the fact that
S. aureus virulence factors are both redundant and multiple in function, with several factors
performing the same or similar function and also one factor harboring several functions
(Gordon RJ/Lowy 2008).
27

Maren Depke
Introduction
Mechanisms of S. aureus Adaptation to its Environment
For its survival in changing environments, the bacterium needs an adaptive response and thus
an effective regulation of all cellular processes including gene expression, protein synthesis, and
turnover allowing adaptation to different conditions. Such regulation not only controls cell
division and metabolism, but also systems for exploiting limited nutrients, adaptation of the
composition of the cell wall as outer boundary, surface factors like adhesins, and systems
rendering stress resistance and guaranteeing endurance in situations non-favorable for growth or
survival. In bacteria, the regulation often affects the transcription of genes as starting point for
changes in the following levels like protein synthesis. Two-component systems (TCS) mediate the
sensing of environmental signals with the sensor component and the adequate reaction with the
response regulator component. An example for such TCS is the agr system. Additionally,
transcription factors like SarA modulate staphylococcal gene expression (Cheung et al. 2004).
Another well-conserved system is the use of alternative sigma factors. Transcription in
bacteria is only initiated when the catalytic core complex of RNA polymerase, formed of the five
subunits α 2 ββ’ω, is associated with a sigma factor recognizing the promoter (-10/-35-region). A
so-called house-keeping sigma factor (in S. aureus: σ A ; Deora/Misra 1995) maintains the baseline
expression of a “standard” set of genes that is generally needed by the cell. Alternative sigmafactors
are activated in specific situations, e. g. after heat shock or salt stress. They recognize a
specific set of promoters of genes whose function is required to encounter the corresponding
physiological conditions. This set of genes includes further transcriptional regulators which
themselves positively or negatively influence the expression of genes depending on the need of
the bacterial cell. S. aureus possesses three alternative sigma factors: σ B , σ H , and − most recently
discovered − σ S (Wu et al. 1996; Morikawa et al. 2003; Shaw et al. 2008). SigB is homologous to
sigB of Bacillus subtilis which has been subjected to intensive studies. Despite the similarity of
sigB in B. subtilis and S. aureus, only half of the gene cluster coding for proteins belonging to the
control cascade of SigB activation/inactivation (rsbU-rsbV-rsbW-sigB) is conserved in S. aureus. Of
special importance is the gene rsbU, coding for the phosphatase that is the starting point for
activation of the alternative sigma factor SigB. After dephosphorylation by RsbU the anti-antisigma
factor RsbV becomes active and binds RsbW leading to liberation of SigB from its antisigma
factor RsbW. Contrarily to B. subtilis where further rsb gene products are regulators of
RsbU activity, in S. aureus an increase in RsbU already leads to SigB activation (Senn et al. 2005).
Deletion of sigB in S. aureus leads among other things to a loss of pigmentation, increased
sedimentation rate and increased sensitivity to hydrogen peroxide in the stationary growth
phase. Different effects are observed on the expression of genes and on the abundance of
proteins. On the one hand some proteins are missing in sigB deletion mutants (e. g. Asp23) that
are directly regulated and transcribed by SigB. On the other hand, other proteins have a higher
abundance in the sigB deletion mutants (e. g. lipase, thermonuclease). For such genes a negative
regulation by SigB itself or a SigB controlled regulator is basis of the regulation observed (Kullik et
al. 1998). Bischoff et al. (2004) propose YabJ, SpoVG, SarA, SarS and ArlRS as regulators possibly
responsible for the indirect SigB effects.
The SigB regulon in S. aureus has some similarities to that of B. subtilis, but the main function
to obtain a broad-range stress resistance to the triggering stimulus and in advance also to other
stressors is not conserved in S. aureus. Energy and ethanol stress do not trigger a SigB response
in S. aureus. Nevertheless, some stress conditions such as heat shock, MnCl 2 , NaCl and alkaline
28

Maren Depke
Introduction
shock are also effective in activating SigB in S. aureus. The entry into stationary growth phase
additionally leads to the activation of the SigB regulon. Only a minor fraction of B. subtilis SigB
dependent genes is found as orthologue in S. aureus and vice versa. In an approach to
functionally characterize the SigB regulon of S. aureus, physiological aspects like cell envelope
composition, membrane transport processes, and intermediary metabolism emerged as affected
(Pané-Farré et al. 2006).
The SigB regulon contains several virulence associated genes like coa, fnbA, ssaA, clfA, hla,
hlgABC, lip, nuc, and sak. In a general view, a pattern of induction of cell surface virulence factors
and a repression of exoproteins and toxins by SigB is recognized. Therefore, SigB effects are
inverse to the effects of the active agr system via RNAIII (Bischoff et al. 2004).
Taken together, S. aureus exhibits a complex control of its virulence factors that includes
several interlaced pathways of central regulators like agr, sarA, and sigB which are interesting
targets for knock-out mutants to be tested in infection models.
Increasing Importance and Danger of S. aureus Infections
S. aureus is one of the main causes of hospital-and community-acquired infections, and its
prevalence and antibiotic resistance have been studied intensively (Diekema et al. 2001, Goto et
al. 2009).
S. aureus as pathogen causing infection could be treated for a long time with classical
antibiotics. But the pathogen has a potent ability to develop or acquire antibiotic resistances.
With today’s increasing cases of antibiotic-resistant staphylococcal strains the question whether
the era of untreatable infections has arrived is raised. This accentuated question emphasizes the
recently enhanced importance and danger of S. aureus and complementarily, the imparative of
strict infection control and the importance of discovering new antibiotics and vaccination targets
and developing these agents for medical purposes (Livermore 2009).
Staphylococcal strains which are not only resistant to long- and often-used antibiotics, but
also to those kept as reserve emerged and increase in prevalence in many countries (Livermore
2000, Fig. I.5). This took place for methicillin (methicillin-resistant S. aureus – MRSA; MRSA might
also be used in the context of multi-resistant S. aureus) since the 1960s, only some years after
start of clinical application, or more recently for vancomycin (vancomycin-resistant
S. aureus − VRSA) after S. aureus strains have obtained the vanA operon from an Enterococcus
spp. transposon on a conjugative plasmid (Péricon/Courvalin 2009). The Staphylococcal Cassette
Chromosome mec (SCCmec), a genetic region which is situated on a mobile genomic island and is
thought to originate from Staphylococcus sciuri (Wu et al. 2001), is responsible for resistance to
methicillin and all other ß-lactam antibiotics. This region exists in form of several different types
(I-VIII) and can be used along with protein A typing (spa), multilocus sequence typing (MLST), and
pulse-field gel electrophoresis (PFGE) to distinguish different staphylococcal lineages on a
molecular level (Deurenberg/Stobberingh 2008, Zhang et al. 2009). Staphylococcal resistance to
methicillin is based on different mechanisms. The main form is the expression of PBP2a, a nonmethicillin-sensitive
variant of the sensitive original penicillin-binding proteins (PBPs). PBPs, of
which S. aureus owns four types PBP1-4, have essential function in cell wall biosynthesis, where
they are responsible for cross-linking of peptide chains in the peptidoglycan and additionally
stretch the glycan part via their transglycosylase activity. Methicillin inhibits PBP’s peptide cross-
29

Maren Depke
Introduction
linking. In the presence of methicillin, MRSA induce the expression of PBP2a from their mecA
gene, and survive antibiotic treatment because PBP2a resumes peptide cross-linking instead of
the inhibited standard PBPs. Other resistance mechanisms including a methicillin-specific
lactamase (methicillinase) and PBP2 mutations with reduced methicillin binding are discussed
(Chambers 1997, Stapleton/Taylor 2002).
First, hospitals were the primary setting of transmission of and infection with such strains,
certainly also favored by the gathering of people who might be regarded as pre-disposed victims
for infection because of pre-existing illnesses or poor health state which additionally have
frequent injections and insertions of medical devices (Lindsay/Holden 2004).
But nowadays, MRSA has left the hospital as its transmission site and so-called “communityassociated”
MRSA arise. Existence of this new group has been published beginning from 1990 for
different countries. Interestingly, hospital-acquired / healthcare-associated (HA) and communityassociated
(CA) MRSA consist of distinct strains, although of course CA-MRSA can cause
nosocomial infections when transmitted by a carrier into hospital.
CA-MRSA strains have characteristics which distinguish them from HA-MRSA. They often
colonize non-classical niches in the human body and not the nares, and the distribution of the
Staphylococcal Cassette Chromosome mec (SCCmec) types is different between HA- and CA-
MRSA. Additionally, in CA-MRSA a higher frequency of Panton-Valentine leukocidin (PVL)-positive
strains is found, although the PVL virulence factor is not a universal marker for CA-MRSA.
Although the prevalence of CA-MRSA in Europe seems to be low, the number is already
increasing and additionally might be underestimated because of difficulties in monitoring
(symptom-free non-nasal carriage). Such unrecognized and hidden reservoir of MRSA is
challenging with respect to attempts in controlling staphylococcal infectious disease
(Otter/French 2010).
A B C
Fig. I.5: Proportion of MRSA isolates in EARSS-participating counties in 1999 (A), 2001 (B), and 2008 (C).
Graphics from European Antimicrobial Resistance Surveillance System (EARSS), a european wide network of national surveillance
systems, providing reference data on antimicrobial resistance for public health purposes (http://www.rivm.nl/earss/).
30

Maren Depke
Introduction
STUDIES OF HOST-PATHOGEN INTERACTIONS
Different approaches can be applied to define even more and specific interactions between
the host and the pathogen. Besides classical physiological and microbiological methods, the
modern molecular technologies can considerably help to improve the understanding of the
processes acting during encounter of host and pathogen provided that genome sequence
information is available. On the first level, RNA expression profiling will generate an overview on
the potential changes of physiology and metabolism. This approach has the advantage of
monitoring the whole repertoire of the organism using a single analysis method and only one
small sample, because nucleic acid material can be easily amplified by laboratory methods.
Whether such changes of the RNA messenger really will be substantiated by changes on protein
level must be analyzed by global or specific proteomic approaches. Here, different methods
address the diverse cellular fractions and can provide a comprehensive impression of the
different cellular states referring to protein abundance, but also to protein modification and
subcellular protein localization, which is not accessible with the transcriptomic analysis. A
limitation is in some cases the amount of sample material and the bigger effort needed to
perform proteome studies in comparison to transcriptome studies. Results of both studies
complement one another. Hence, a combination of several different approaches in co-operation
studies is strongly recommended.
Model Systems for Studies of Host Reactions Potentially Influencing the
Outcome of Infections
Liver gene expression pattern in a mouse psychological stress model
Psychological and physiological stressors can disturb neuroendocrine, immunological,
behavioral, and metabolic functions (Harris et al. 1998, Leibowitz/Wortley 2004, Mizock 1995)
and adaptive physiological processes aim to reconstitute a dynamic equilibrium (McEwen 2004,
Viswanathan/Dhabhar 2005).
In a murine model of severe, chronic psychological stress due to 4.5 days of intermittent
combined acoustic and restraint stress BALB/c mice developed severe systemic
immunosuppression, neuroendocrinological disturbances and depression-like behavior. Besides
heightened anti-inflammatory cytokine bias, lymphocytopenia, T cell anergy, impaired phagocytic
and oxidative burst responses, increased susceptibility to experimental infection with E. coli,
spontaneous bacterial infiltrations of gut commensals into the lung, reduced clearance of
experimental infections in the long term, attenuation of a hyperinflammatory septic shock, and
finally, behavioral and neuroendocrine alterations and a prominent stress-induced loss of body
mass without significant changes of food and water intake during the observation period became
detectable (Kiank et al. 2006, 2007a, 2007b, 2008).
31

Maren Depke
Introduction
Several authors propose chronic stress to be a main feature in the pathogenesis of the
metabolic syndrome, which is associated with obesity, type 2 diabetes/insulin resistance,
dyslipidemia, and hypertension (Alberti et al. 2009, Kuo et al. 2007, Lambert et al. 2010). On the
other hand, stressors like injury, infection, traumatic events, or prolonged sleep deprivation
capably induce hypercatabolism and, therefore, may cause cachexia (Alberda et al. 2006,
Koban/Swinson 2005, Morley et al. 2006, Vanhorebeek/Van den Berghe 2004, van Waardenburg
et al. 2006, Wilmore 2000, Wray et al. 2002). Such metabolic-driven wasting may result from
pain, depression, or anxiety, causing malabsorption and maldigestion or morphological and
functional alterations of the gastrointestinal system and is typically seen during repeated
inflammatory processes or during sepsis (Alberda et al. 2006, Hang et al. 2003, Hansen MB et al.
1998, McGuinness et al. 1995, Morley et al. 2006, van Waardenburg et al. 2006, Wilmore 2000).
In the literature, two main pathways leading to massive loss of body mass are discussed: the
response to starvation and the hypermetabolic response. Starvation is induced by inadequate
calorie intake and causes the use of the body’s own tissue. At first, carbohydrate stores are
emptied for energy production. Secondarily, the body switches to usage of lipid and protein
stores for gluconeogenesis. Finally, fatty acids are absorbed into the liver where ketone bodies
are produced that, for example, neurons can use as an energy source to restore functional
homeostasis. In a feedback loop, efferent neurons signal to the periphery so that
gluconeogenesis is lowered, and protein breakdown is diminished that causes adaptation to
starvation. The supply for basal energy production then comes from the calories of adipose
stores (Alberda et al. 2006, Morley et al. 2006).
In contrast, a hypermetabolic response that is often seen during critical illness is
predominantly mediated by hormones and inflammatory mediators. As during the initial phase of
starvation, gluconeogenesis is accelerated by usage of lipids and amino acids. Protein breakdown
is massively increased due to glucagon and glucocorticoid effects, and can be accelerated by
proinflammatory cytokines such as TNF. Amino acids, along with fatty acids and glycerol, are used
for gluconeogenesis in the liver, causing hyperglycemia. This process is mainly mediated by
catecholamines and corticosteroids, and finally results in a rapid loss of lean body mass without
sufficient metabolic adaptation to diminish tissue breakdown (Costelli et al. 1993, Goodman
1991, Harris et al. 1998, Lang et al. 1984, McGuinness et al. 1999, Mizock 1995, Morley et al.
2006, Wilmore 2000, Vanhorebeek/Van den Berghe 2004, van Waardenburg et al. 2006, Weekers
et al. 2002, Wray et al. 2002).
To investigate primary causes for the severe loss of body mass in chronically stressed mice,
gene expression profiles of the liver, which plays the major role in metabolism, were analyzed.
Based on documented changes of the expression profile of hepatic genes involved in
carbohydrate, lipid, and amino acid metabolism, a characterization of metabolic disturbances
was started and biologically relevant alterations of glucose metabolism, dyslipidemia, and
changes in amino acid turnover in repeatedly stressed BALB/c mice were found which revealed a
chronic stress-induced hypermetabolic syndrome (Depke et al. 2008).
The liver fulfills an important function as a gatekeeper between the intestinal tract and the
general circulation (Dietrich et al. 2003). Thus, the influence of psychological stress on immune
regulatory processes in the liver was additionally investigated in the gene expression data set.
Veins drain the intestinal tract and then segment to secondary capillary beds in the liver
where the removal and metabolism of absorbed nutrients or toxic substances takes place. Food
antigens and products of commensal bacteria are abundantly detectable in the portal vein and
32

Maren Depke
Introduction
are effectively cleared by Kupffer cells and sinusoidal endothelial cells (Limmer et al. 2000,
Maemura et al. 2005, van Oosten et al. 2001).
Physiologically, there is no detectable intrahepatic immune activation in response to the low
amount of microbial agent and food antigens derived from the gut because of a tolerogenic
environment (Limmer et al. 2000, Macpherson et al. 2002). However, when the antigenic
challenge is increased, an inflammatory response with a recruitment of neutrophils,
monocytes/macrophages, and lymphocytes may be induced (Wiegard et al. 2005). Then, an
increased production of inflammatory mediators such as reactive oxygen species (ROS) may
cause cell damage and loss of hepatocyte functions (Ott et al. 2007).
To elucidate hepatic immune regulatory pathways that may contribute to chronic
psychological stress-induced immune suppression, the hepatic gene expression profiling was
used to analyze expression changes of immune response and cell survival genes in stressed and
non-stressed mice (Depke et al. 2009).
Gene expression pattern of bone-marrow derived macrophages after interferon-gamma
activation
In the innate immune system, macrophages, together with dendritic cells, hold a central
position. They are main effectors of the clearance of infections by their sentinel and phagocytic
function. Macrophages present phagocytosed antigen derived peptides on MHC-II to
lymphocytes. By this function, macrophages take part in regulation of the adaptive immune
response (Gordon S 2007). Phagocytosis is mediated by different receptors like complement
receptors, mannose receptor or via the interaction between lipopolysaccharide (LPS), LPS-binding
protein (LBP), CD14, and Toll-like receptor TLR (Janeway/Medzhitov 2002, Greenberg/Grinstein
2002). Upon binding of ligands to TLR, macrophages secrete chemokines like CXCL8, CXCL10,
CCL3, CCL4, and CCL5, which mediate chemotaxis of neutrophils, NK cells, and T cells (Luster
2002). Additionally, pro-inflammatory cytokines like TNF-α and IL-1 secreted by macrophages
further contribute to inflammation. Macrophages promote extracellular matrix degradation by
their matrix metalloproteinases, MMPs (Gibbs et al. 1999a, 1999b). They exhibit increased killing
activity towards phagocytosed pathogens during respiratory burst, which is among others
mediated by toxic, reactive defense molecules like NO and O –
2 produced by inducible NO
synthase (iNOS) and NADPH oxidase, respectively (DeLeo et al. 1999, Iles/Forman 2002, Kantari
et al. 2008, MacMicking et al. 1997, Mori/Gotoh 2004, Park JB 2003). This macrophage
phenotype corresponds to the classical activation and initiates primarily a Th1-based immune
response (Fig. I.6). Vice versa, IFN-γ as part of the Th1 cytokine pattern provokes such classical
activation (Duffield 2003).
After a phase of inflammatory activation and phagocytosis of apoptotic cells at the site of
inflammation, macrophages are able to switch their functional profile now aiming to reduce
inflammation by anti-inflammatory cytokines, stop matrix degradation and even assist healing by
production of fibronectin and tissue transglutaminase. A similar phenotyp with reduced
intracellular killing of pathogens is observed after activation by Th2 cytokines like IL-4 and TGF-β.
Activation leading to this phenotype is called “alternative” (Duffield 2003). Macrophages can be
activated in a third way, called type II activation. This activation needs ligand binding to Fcγ
receptors in combination with an activating stimulus like that mediated by TLRs. Afterwards, the
33

Maren Depke
Introduction
phenotype of macrophages changes from production of IL-12 to IL-10 production, whereas TNFα,
IL-1, and IL-6 are still secreted (Mosser 2003). Type II activation initiates a Th2-based immune
response. There is experimental evidence that the phenotype of macrophages is primarily
determined by the first cytokine type to which they are subjected (Erwig et al. 1998).
A naïve macrophage can react to activation stimuli. Nevertheless, when the macrophage
received a priming signal in advance, it had the possibility to prepare for the potentially following
second stimulus by transcriptional regulation and is able to react stronger and faster to
activation. Low-dose IFN-γ is the most important macrophage priming signal. Primed
macrophages are not activated yet and do not secrete proinflammatory cytokines. This occurs in
classical activation after a second signal like IFN-γ or LPS, peptidoglycan, and others (Dalton et al.
1993, Huang et al. 1993, Ma J et al. 2003, Mosser 2003).
Fig. I.6: Inducers and selected functional properties of different polarized macrophage populations.
Abbreviations: DTH – delayed-type hypersensitivity; IC – immune complexes; IFN-γ – interferon-γ; iNOS – inducible nitric oxide
synthase; LPS – lipopolysaccharide; MR – mannose receptor; PTX3 – the long pentraxin PTX3; RNI – reactive nitrogen intermediates;
ROI – reactive oxygen intermediates; SLAM – signaling lymphocytic activation molecule; SRs – scavenger receptors; TLR – Toll-like
receptor.
From: Mantovani et al. 2004.
Studies aiming to analyze the reaction in general and more specific the gene expression
pattern of macrophages are impaired by the influence of immunological factors on the mature
macrophages which can be prepared from animal organs, even under highly standardized
laboratory conditions. This problem can be circumvented when using bone-marrow derived
macrophages (BMM). Bone marrow stem cells are prepared from the animal and cultivated
in vitro for proliferation and differentiation. Under the influence of granulocyte-macrophage
colony stimulating factor (GM-CSF) cells differentiate into macrophages, which have the
advantage of having never received any in vivo stimulus or immunological conditioning that
might influence their cellular reaction.
Until recently, further uncontrollable influences on macrophage or BMM experiments were
introduced by the utilization of serum-supplemented culture medium. Serum contains immune
34

Maren Depke
Introduction
cell stimulating substances like cytokines, hormones or endotoxin, and to worsen the problem,
these substances might vary not only between the sera from different vendors but also between
batches of serum from the same manufacturer. To overcome such sources of experimental
variation, Eske et al. introduced in 2009 serum-free culture conditions for BMM. Standard cell
culture medium RPMI 1640 was supplemented with a defined mixture of proteins, hormones,
and other compounds. Differentiation of stem cells with recombinant murine GM-CSF yielded
BMM expressing macrophage markers F4/80, CD11b, CD11c, MOMA-2, and CD13 after 10 days of
cultivation, and further characteristics of BMM differentiated in serum-containing medium were
additionally retained (Eske et al. 2009).
Host-pathogen interaction experiments have revealed differences in reactions of BMM
derived from BALB/c and C57BL/6 mice when confronted with Burkholderia pseudomallei
especially after IFN-γ stimulation and at higher multiplicities of infection (MOI). Also other
infection studies uncovered differences between these two mouse strains in vivo and in vitro
(Breitbach et al. 2006, Autenrieth et al. 1994, van Erp et al. 2006). The mouse strains BALB/c and
C57BL/6 are characterized by a Th2 and Th1 centered type of immune response, respectively.
Accordingly and possibly causatively, the main type of macrophage activation differs between the
strains, with BALB/c preponderating the alternative macrophage activation pattern and C57BL/6
prevailing classical activation (Mills et al. 2000).
Against the background of genetic influences on the BMM reactions, a combined proteome
(Dinh Hoang Dang Khoa) and transcriptome (Maren Depke) study was initiated. In a first
experiment, BALB/c and C57BL/6 BMM were stimulated with IFN-γ to specify on a molecular level
the reaction to the priming signal IFN-γ as basic principle. Furthermore, the study aimed to
profile potential differences of reaction between the BMM of both mouse strains.
Model Systems for Studies of Host-Pathogen Interactions
S. aureus strain RN1HG
The genetic background of different S. aureus strains influences the reaction of the bacterium
to experimental conditions. Therefore, the strain for experimental analysis must be chosen
carefully because not all observations can be transferred from one to another S. aureus strain.
The number of strains available for studies of S. aureus has grown recently when the group of
Friedrich Götz (Department of Microbial Genetics, Eberhards-Karls University of Tübingen,
Germany) provided a rsbU + repaired RN1-derivative strain called RN1HG(001) and later tcaR + and
rsbU + tcaR + repaired RN1-derivative strains to the scientific community (Herbert et al. 2010). The
parental strain RN1 (NCTC8325) has already been widely used as model organism for diverse
studies on staphylococcal physiology. On the other hand its defect in the important regulator
rsbU was known (Kullik/Giachino 1997), which resulted in compromised conclusions from studies
addressing the regulation of virulence factors (Giachino et al. 2001). Therefore, strain RN1HG in
which the mutation in the regulatory gene rsbU has been complemented has been chosen in the
experimental setup of the studies described in this thesis.
35

infection rate
cfu / organ
Maren Depke
Introduction
Kidney gene expression pattern in an in vivo infection model
S. aureus can be transmitted to the blood after body injury or by medical devices like
catheters. An elementary model to mimic blood stream infection is the intra-venous infection of
laboratory animals, e. g. mice. Host reactions can be monitored by physiological, immunological
or molecular readout systems. In this study, transcriptome analysis was applied.
Strongest accumulation of S. aureus after i. v. infection of mice is observed in kidneys, which is
also accompanied by bacterial proliferation in the time course of infection (Fig. I.7, data courtesy
of Tina Schäfer). Thus, this organ was chosen for host gene expression profiling.
Fig. I.7:
Accumulation and bacterial proliferation of S. aureus
Xen29 in murine kidney tissue after i. v. infection.
Female BALB/c mice were infected with 1.0E+08 cfu
via the tail vein. Data represent median and
interquartile range of n = 3 experiments.
Data courtesy of Tina Schäfer, Würzburg.
1.0E+09
1.0E+08
1.0E+07
1.0E+06
1.0E+05
1.0E+04
1.0E+03
1.0E+02
1.0E+01
0.25 24 48 72
time post infection / h
Although the virulence of sigB deficient strains is often reported to be similar to that of wild
type strains the pathogenesis or pathomechanism of infection might be different. Therefore, the
rationale of this study was to investigate whether the deletion of sigB will lead to a different
reaction of the infected host.
Host cell and pathogen gene expression pattern in an in vitro infection model
While S. aureus colonizes humans in the anterior nares, it can be cause of pneumonia when
transferred to the lung e. g. by aspiration or medical devices. In a longitudinal study with more
than 10000 patients in Japan, S. aureus was identified as one of the leading causative organisms
of pneumonia besides Streptococcus pneumonia and Haemophilus influenzae (Goto et al. 2009).
Here, different cell types first encounter the pathogen: Cells associated with structural and
functional aspects of lung and “guardian” cells of the immune system, e. g. alveolar macrophages.
Epithelial cells form the inner surface of the lung. Cilial epithelial cells transport mucus out of the
organ. The mucus is produced in the lower bronchia and alveoli by type II pneumocytes and
functions as surfactant to reduce surface tension and as protective film to remove inhaled
particles and pathogens. Adjacent type I pneumocytes take part in oxygen exchange of the air
with the blood.
A model to study reactions of epithelial cells to pathogens is the human bronchial epithelial
cell line S9. S9 cells originate from the IB3-1 cell line, which was established about 20 years ago
from a male, white cystic fibrosis (CF) patient’s bronchial epithelial cells after transformation with
adeno-12-SV40 virus (Zeitlin et al. 1991). This cell line contains the non-functional chloride
36

Maren Depke
Introduction
channel CFTR (cystic fibrosis transmembrane conductance regulator), which has been repaired in
the S9 cell line by viral transfection with the wild-type gene (American Type Culture Collection
ATCC, Manassas, VA, USA; www.atcc.org; S9 cell ATCC number CRL-2778).
In vitro infection models can be accomplished by addition of three main types of bacterial
supplement: 1) supernatant of bacterial cultures containing secreted proteins or virulence
factors, 2) PBS-washed bacterial cells which bring only their membrane-bound and intracellular
factors into the infection experiment, and 3) complete bacterial culture with both secreted
proteins from supernatant and whole bacterial cells. This last option is nearest to the in vivo
situation when the pathogen is able to influence its host with secreted factors. On the other
hand, the established bacterial culture media, which are often lysates of protein-rich raw
material (e. g. tryptic soy broth TSB), are highly artificial with reference to in vivo models or even
to eukaryotic cell culture. Therefore, a medium was developed that allows bacterial growth but
additionally has similarity to eukaryotic cell culture media (Schmidt et al. 2010). The authors
describe the use of eukaryotic cell culture medium MEM supplemented with different amino
acids, but without addition of serum. This new experimental system permits the study of hostpathogen
interactions in the context of all bacterial factors, membrane-bound and secreted, and
additionally prevents effects on bacterial physiology by prolonged handling, centrifugation, and
washing of bacteria.
Here, exponential growth phase bacterial cultures were used to infect confluent S9 cell
cultures. The strain S. aureus RN1HG has been chosen for a first insight into the molecular
reactions in this model. RN1HG is a rsbU + repaired RN1-derivative strain (Herbert et al. 2010) with
a SigB-positive phenotype.
In a combined approach of transcriptome (Maren Depke) and proteome (Melanie Gutjahr)
analysis the host reaction to infection and bacterial internalization was recorded. But not only
stood the host cell in the focus of studies, but also the bacterium. In a similar experimental setup,
internalized staphylococci were extracted from their S9 host cells and the bacterial RNA profile
was recorded using a tiling array approach (Maren Depke). Bacterial intracellular proteins were
monitored and quantified after stable isotope labeling with amino acids in cell culture, SILAC
(Sandra Scharf).
Questions and Aims of the Studies Described in this Thesis
This thesis contains results from transcriptome studies on different aspects of host-pathogen
interactions. First, liver gene expression profiles from a murine chronic stress model served to
elucidate aspects of the influence of stress on the metabolism and the immune response state of
the animals. In the experiments for this study, the in vivo model of psychological stress in a
complex mammalian host was performed without additional influences of a pathogen. Such
influence was introduced in the second study: Here, the influence of staphylococcal i. v. infection
on the host kidney gene expression was analyzed in another murine in vivo model using a wild
type S. aureus strain and its isogenic sigB mutant. Tissue expression profiling from in vivo models
has the advantage of directly recording the relevant physiological state with all its complex
interactions and influences and its vicinity to medical questions in the human. Nevertheless, it is
very difficult to distinguish the different components because the tissue samples are always a
37

Maren Depke
Introduction
mixture of different cell types which might even feature contrary reactions. Therefore, in vitro
models were additionally analyzed in which only one defined host cell type was studied.
Macrophages are an example for an immune cell type involved in the first steps of the encounter
between the host and a pathogen. A model to study reactions of macrophages is the preparation
of bone marrow stem cells and the in vitro differentiation of the stem cells into so-called bonemarrow
derived macrophages (BMM). The advantage of this approach is that these macrophages
have never been under any immunological influence which might result from the immune status
of the animal even under standardized laboratory conditions. Thus, the third part of this thesis
focuses on the reaction of BMM. BMM of different mouse strains were treated with IFN-γ, a
modulator of macrophage function which is one of the first signals during initiation of the
immune response in vivo. But not only immune cells or specially phagocytes get in touch with
pathogens, but also cells responsible for functional and structural integrity of host organs and
tissue, like epithelial and endothelial cells. Such cells are actually part of the first line of
recognition and reaction to a pathogenic invasion into the host. The bronchial epithelial cell line
S9 was used as an in vitro model system for the infection with staphylococci. The fourth chapter
in this thesis includes host gene expression signatures of S9 cell after in vitro infection with
S. aureus RN1HG. Finally, the following chapter addresses the pathogen expression profile which
was first recorded from agitated, aerobic S. aureus RN1HG cultures in different growth phases as
a starting and reference point. Afterwards, the already described S9 cell in vitro infection model
was used to extract staphylococci after an internalization phase inside the host cell. Internalized
bacteria were analyzed at two time points in comparison to different control samples by tiling
array gene expression analysis.
38

Maren Depke
M A T E R I A L A N D M E T H O D S
LIVER GENE EXPRESSION PATTERN IN A MOUSE
PSYCHOLOGICAL STRESS MODEL
Animal experiments [performed by Cornelia Kiank]
Female BALB/c mice aged 6 to 8 weeks were randomly grouped into the experimental and
control groups starting at least 4 weeks before being used in experiments. The group size in
different experiments differed from 6 to 12 mice per cage. Animals stayed in their group until the
end of the experiments and were not mixed up to avoid social stress. All animals were
maintained with sterilized food (ssniff R−Z; ssniff Spezialdiäten GmbH, Soest, Germany) and tap
water ad libitum for adaptation under minimal stress conditions. Influences of irregularities of
the estrous cycles of unisexually grouped female mice were not analyzed selectively and may
cause higher SD values in the statistical analyses.
Animal rooms had a 12 h light, 12 h dark cycle and were maintained at a constant
environment before the experiment. To avoid any additional effect, e. g. acoustic or olfactory
effects, the handling of mice during the adaptation period and during the experiments was
restricted to one investigator. All animal procedures were performed as approved by the Ethics
Committee for Animal Care of Mecklenburg-Vorpommern, Germany.
Repeated stress model [performed by Cornelia Kiank]
Mice were exposed to combined acoustic and restraint stress on 4 successive days, for 2 h
twice a day during the physiological recovery phase of rodents (0800–1000 and 1600–1800 h). On
day 5, only one stress session was performed in the morning. For immobilization mice were
placed in 50 ml conical centrifuge tubes with multiple ventilation holes without penning the tail.
Acoustic stress was induced by a randomized ultrasound emission device between 19 kHz and
25 kHz with 0 dB to 35 dB waves in attacks (patent no. 109977; SiXiS, Taipei, Taiwan), allowing
the mice no adaptation to the stressor. Between the stress sessions, mice stayed in their home
cages and had free access to food and tap water. Control mice were kept isolated from stressed
animals during the 4.5 days stress exposure to avoid any acoustic or olfactory communication
between the groups. Therefore, the nonstressed group stayed in the incubator where the
animals were adapted. The stressed mice remained outside the incubator in the same animal
laboratory during the whole period of the stress model. All successive experiments and analyses
were performed starting at 1000 h after the ninth stress exposure. Different in vivo analyses were
performed with 6 to 12 mice per group in at least two experiments according to the experimental
protocol to ensure reproducibility. For array analysis two independent stress experiments were
performed with nine mice per group (first experiment) and eight mice per group (second
experiment).
39

Maren Depke
Material and Methods
Liver Gene Expression Pattern in a Mouse Psychological Stress Model
Acute stress model [performed by Cornelia Kiank]
Mice were exposed to a single 2 h combined acoustic and restraint stress cycle in the morning
(0800–1000 h). In vivo analyses were performed immediately after or 6 h after the stress session
with six to nine mice per group. Two acute stress experiments for array analysis were composed
of eight animals for each control group, and nine animals per stress group in the first and eight
animals per stress group in the second experiment.
Organ harvesting for RNA preparation [Cornelia Kiank / Maren Depke]
Mice were killed by cervical dislocation, and organs were removed immediately to avoid RNA
degradation. For liver samples, a small piece of tissue was immediately homogenized with a
micropestle in 350 µl Buffer RLT / 1 % β-mercaptoethanol (QIAGEN GmbH, Hilden, Germany /
Sigma-Aldrich Chemie GmbH, München, Germany). The liver lysates were shock frozen in liquid
nitrogen. All samples were stored at −70°C.
RNA preparation
Liver sample lysates were thawed and processed at room temperature for RNA preparation
with the RNeasy Mini Kit (QIAGEN GmbH, Hilden, Germany) according to the manufacturer’s
instructions. After ethanol precipitation, the RNA was quantified spectrophotometrically, and its
quality was verified using an Agilent 2100 Bioanalyzer and RNA Nano Chips (Agilent Technologies
Inc., Santa Clara, CA, USA).
DNA array analysis using Affymetrix expression arrays
Pools containing equal amounts of RNA from each individual animal were prepared for each
group and used for subsequent microarray analysis. Five micrograms of pooled total RNA were
used for the synthesis of double-stranded cDNA, and this solution then served as a template for
an in vitro-transcription reaction using GeneChip Expression 3’ Amplification Reagents
(Affymetrix, Santa Clara, CA, USA) according to the manufacturer’s instructions. After spincolumn
based cleanup, concentration and quality of cRNA were measured as described above.
cRNA was fragmented, added to the hybridization cocktail, denatured, and hybridized with
Affymetrix GeneChip Mouse Expression Arrays 430A/430A 2.0 according to the manufacturer’s
instructions.
Washing, staining, and scanning were performed using Affymetrix GeneChip FluidicsStation
and scanners according to standard protocols.
Data analysis for Affymetrix expression arrays
The Affymetrix expression analysis was performed for the livers of repeatedly stressed and
healthy control mice with technical duplicates of two independent biological experimental series
each. For the analysis of the effects of acute stress, array hybridizations were also performed of
two independent biological experiments for both groups (control and acute stress). Affymetrix
array image data generated with MAS 5.0 (repeated stress) were analyzed using the GeneChip
Operating Software 1.2 (Affymetrix, Santa Clara, CA, USA) with default values for parameter
settings. For normalization, a scaling procedure with a target value of 150 was used. Image data
of the acute stress experiment were directly analyzed in GeneChip Operating Software 1.4 with
default settings and normalized by scaling to the target value 500. After data transfer to the
GeneSpring software package (Agilent Technologies Inc., Santa Clara, CA, USA), genes displaying
40

Maren Depke
Material and Methods
Liver Gene Expression Pattern in a Mouse Psychological Stress Model
differential regulation in response to repeated and acute stress were identified based on the
following criteria: 1) the signal of probe sets had to be present in the arrays at least in the control
(for repressed genes) or in stressed mice (for induced genes) in both biological experiments,
2) the difference of mean signals between control and stressed mice had to equal or exceed 100,
and 3) the fold change factors calculated from the signal values in each experimental replicate
had to exceed a cutoff of more than or equal to 1.5 or less than or equal to −1.5 in both biological
experiments.
Functional analysis of gene expression data using Ingenuity Pathway Analysis
For data interpretation in the context of metabolism, lists of probe sets displaying differential
regulation in both acute and repeated stress, or specifically after acute or repeated stress were
uploaded as Excel spreadsheets (Microsoft Corp., Redmond, WA, USA) into the Ingenuity
Pathway Analysis (IPA) Version 5.5 (Ingenuity Systems, Inc., Redwood City, CA, USA;
http://www.ingenuity.com) and used for the interpretation of the array data in the context of
already published knowledge. Biological functions were assigned to the networks based on the
content of the Ingenuity Pathway Knowledge Base. Complete hepatic gene expression data of
stressed and nonstressed mice are available at the NCBIs Gene Expression Omnibus (GEO,
http://www.ncbi.nlm. nih.gov/geo/) database and are accessible through GEO Series accession
no. GSE11126.
For functional analysis in the context of immune response and apoptosis-associated genes,
lists of differentially regulated probe sets were uploaded as Excel spreadsheets into the Ingenuity
Pathway Analysis tool (IPA 6, www.ingenuity.com). In order to compare effects of acute and
chronic stress, two groups of genes were included: (1) genes displaying differential regulation
after acute stress and (2) genes differentially regulated after chronic stress. IPA combined the
uploaded Affymetrix probe set IDs and assigned annotations depending on the content of the socalled
Ingenuity Pathway Knowledge Base (IPKB). The IPKB was used to get further insight into
the relation of differentially regulated genes and immunological functions. Searches for relevant
keywords and meaningful combinations of keywords resulted in lists linked to the functions in
focus. IPA also offers the association of differentially expressed genes with canonical pathways,
which can be rated by a corresponding p-value. This approach was used to depict functional
aspects of differential gene expression.
Real-time PCR
DNA was removed by DNase treatment, and subsequent to purification using a RNeasy Micro
Kit (QIAGEN GmbH, Hilden, Germany) and ethanol precipitation, concentration and quality of
RNA samples were assayed as described above. Validation of expression data by real-time PCR
was separately performed for all individual RNA preparations (n = 9 plus n = 8 mice per group) of
the two biological experiments with repeated stress exposure. For real-time PCR analysis, 1 µg
RNA was reverse transcribed into cDNA using the High Capacity cDNA Archive Kit in the presence
of SUPERase•In RNase inhibitor (Ambion/Applied Biosystems, Foster City, CA, USA). Twenty
nanograms of cDNA served as a template for real-time PCR using the following 20x TaqMan Gene
Expression Assays (Applied Biosystems, Foster City, CA, USA): Asl (Mm00467107_m1), Srebf1
(Mm00550338_m1), Pck1 (Mm00440636_m1), Gadd45b (Mm00435123_m1), Sds
(Mm00455126_m1), and Actb (Mm00607939_s1). Differential regulation in repeatedly stressed
and control mice was confirmed by comparing the ΔCt values (Ct value of the target
gene − Ct value of the reference gene Actb in identical cDNA samples) of all control mice and
repeatedly stressed mice with a Mann-Whitney U test, requiring a p-value of less than or equal to
0.05.
41

Maren Depke
Material and Methods
KIDNEY GENE EXPRESSION PATTERN IN AN
IN VIVO INFECTION MODEL
In vivo infection model, organ harvesting, and group size [performed by Tina Schäfer]
Female BALB/c mice (Charles River, Sulzfeld, Germany) were infected i. v. with 5.0E+07 colony
forming units (cfu; first biological replicate, BR1) and 7.0E+07 cfu (second biological replicate,
BR2) S. aureus RN1HG or 5.0E+07 cfu (first biological replicate) and 8.0E+07 cfu (second biological
replicate) of its isogenic sigB mutant S. aureus RN1HG ΔsigB. The third group in this study
comprised sham-infected mice which received an injection of 100 µl physiological saline solution
(performed only in the second biological replicate of the experiment). After 4 days (first biological
replicate) or 5 days (second biological replicate) mice were sacrificed and kidneys were explanted
immediately afterwards, flash-frozen in liquid nitrogen and stored at −80°C.
Each of the three experimental groups was comprised of 5 independent samples per biological
replicate (originating from kidneys of 5 mice) except for the group of infection with S. aureus
RN1HG in the first biological replicate, which only consisted of 4 samples. In total, 24 samples
were analyzed in this study.
Tissue disruption
In constant submersion in liquid nitrogen in a mortar, both kidneys of the mice were ground
into very small pieces, but not into powder. By this means it was possible to yield a homogenous
tissue mix which allowed accurate estimation of infection rate. As the tissue was not completely
disrupted into powder, it was still possible to handle the frozen tissue mix e. g. during aliquoting
without the risk of sample thawing.
DNA preparation
Small aliquots of disrupted tissue were added to Lysing Matrix D tubes (MP Biomedicals,
Solon, OH, USA) which contain 1.4 mm diameter ceramic spheres. Tissue was completely
disrupted in 190 µl of 42 mM EDTA using a FastPrep FP120 (Thermo Fisher Scientific Inc.,
Waltham, MA, USA) at level 6.5 for 20 s. After short cooling on ice, samples were digested with
250 µg/µl lysostaphin (AMBI PRODUCTS LLC, Lawrence, NY, USA) for 45 min at 37°C to ensure
complete lysis of staphylococci in the infected tissue. The following DNA preparation employed
the Wizard Genomic DNA Purification Kit (Promega Corp., Madison, WI, USA) with minor
modifications to the manufacturer’s protocol. The tissue lysate (200 µl) was mixed with 830 µl of
Nuclei Lysis Solution (Promega) and incubated for 5 min at 80°C. Subsequently, samples were
cooled for 5 min on ice and digested with 19.4 ng/µl RNase A (Promega) for 30 min at 37°C.
Samples were again cooled on ice for 5 min and afterwards processed at room temperature.
Lysates were transferred to new 1.5-ml-tubes, 345 µl of Protein Precipitation Solution were
added, samples were vortexed for 20 s and protein was precipitated for 10 min on ice. The
solution was cleared by two centrifugation steps at 20000 x g for 4 min (room temperature). The
final clear supernatant was divided into two aliquots of 620 µl each and the DNA was precipitated
by addition of 470 µl of isopropanol (2-propanol) and mixing by gentle inversion. After
centrifugation for 2 min at 16000 x g (room temperature) the DNA pellet was washed with 600 µl
43

Maren Depke
Material and Methods
Kidney Gene Expression Pattern in an in vivo Infection Model
of 70 % ethanol, air dried for approximately 2 min and solved over night at 4°C in DNA
Rehydration Solution (Promega, 10 mM Tris/HCl pH 7.4, 1 mM EDTA pH 8.0; 50 µl / pellet). Both
aliquots of each sample were combined on the following day, the concentration was determined
photometrically (NanoDrop ND-1000, NanoDrop Technologies, Wilmington, DE, USA), and the
quality was checked by electrophoresis in 0.8 % agarose-TBE gels.
Molecular estimation of infection rate with qPCR
The DNA from infected tissue consists of a mix of host and pathogen DNA, which contains
almost constant levels of host DNA and a varying proportion of bacterial DNA that very much
depends on the infection rate. Therefore, quantifying the ratio of pathogen to host DNA allows
an estimation of the infection rate, which can also be calibrated with the aid of artificially mixed
standard samples.
The highly conserved staphylococcal genes nuc (coding for thermonuclease) and dapA
(encoding dihydrodipicolinate synthase) were quantified in reference to the mouse cytoskeleton
gene Actb using 20x TaqMan Gene Expression Assays (Applied Biosystems, Foster City, CA, USA;
Actb: inventoried assay Mm00607939_s1; nuc: custom assay, forward primer
GATCCAACAGTATATAGTGCAACTTCAACT, reverse primer ACCGTATCACCATCAATCGCTTT, reporter
AACCTGCGACATTAATT; dapA: custom assay, forward primer CTTTGAAGCGATTGCAGATGCT,
reverse primer TGGTTCAATTGTCATGTTCGTTCTTG, reporter ACGACTGGTAATTTC; each reporter’s
5’ end is labeled with FAM; a non-fluorescent quencher (NFQ) and a Minor Groove Binder (MGB)
molecule is linked to the 3’ end).
The real-time PCR was performed in triplicates with 20 mM Tris/HCl pH 7.4, 50 mM KCl, 3 mM
MgCl 2 , 4 % glycerol, 0.2 mM dNTP mix (dATP, dCTP, dGTP, dTTP, 0.2 mM each), 1 % ROX reference
dye (Invitrogen, Karlsruhe, Germany) and 0.35 U Platinum Taq DNA Polymerase (Invitrogen) using
20 ng of DNA as template in each reaction. The so-called “universal settings” for Applied
Biosystems’ TaqMan Assays were applied: 2 min at 50°C, 10 min at 95°C followed by 40 cycles of
15 s at 95°C and 1 min 60°C.
A calibration curve with DNA prepared from a mixture of non-infected kidney tissue and
in vitro cultivated staphylococcal cells was recorded in analogous reactions. Each standard
contained a defined number of cfu per 10 mg tissue ranging from 1.32E+05 cfu/10 mg to
3.31E+08 cfu/10 mg (Fig. M.2.1).
A
B
8.0
6.0
4.0
2.0
mean
Linear (mean)
y = -3.3265x + 23.246
R² = 0.9946
8.0
6.0
4.0
2.0
mean
Linear (mean)
y = -3.2883x + 22.211
R² = 0.9956
ΔCt
0.0
ΔCt 0.0
-2.0
-2.0
-4.0
-4.0
-6.0
-6.0
-8.0
5.0 6.0 7.0 8.0 9.0
log ( cfu / 10 mg tissue)
-8.0
5.0 6.0 7.0 8.0 9.0
log ( cfu / 10 mg tissue)
Fig. M.2.1: Calibration curves for molecular estimation of infection rate using the staphylococcal genes dapA (A) and nuc (B).
Mean values of ΔCt (ΔCt = Ct bacterial gene – Ct Actb) from n = 2 or n = 3 independent standard samples and the range (minimum to
maximum) of these independent measurements are displayed.
The Ct values of bacterial genes and Actb were used to calculate ΔCt
(ΔCt = Ct bacterial gene − Ct Actb ). This value negatively correlates to the log-transformed values of
cfu/10 mg and can be described by a linear equation. The use of this linear correlation as
calibration curve allows calculation the infection rate of the unknown experimental samples.
44

Maren Depke
Material and Methods
Kidney Gene Expression Pattern in an in vivo Infection Model
RNA preparation
A frozen aliquot of disrupted tissue was disintegrated mechanically at 2600 rpm for 2 min in a
bead mill (Mikrodismembrator S, B. Braun Biotech International GmbH, Melsungen, Germany;
now part of Sartorius AG, Göttingen, Germany) with 0.5 ml TRIZOL (Invitrogen, Karlsruhe,
Germany) using liquid nitrogen cooled Teflon vessels. Thereafter, another 0.5 ml TRIZOL was
added to the still frozen lysate. After thawing, the lysate was incubated for 10 min at room
temperature, flash-frozen in liquid nitrogen and stored −70°C until RNA preparation was
continued. The lysates were again thawed at room temperature. Chloroform was added (200 µl
chloroform / 1 ml TRIZOL), samples were shaken vigorously for 15 s and incubated at room
temperature for 5 min. Organic and aqueous phase were separated by centrifugation (12000 x g,
15 min, 4°C) and RNA was precipitated from the aqueous phase with 500 µl isopropanol (2-
propanol) / 1 ml TRIZOL overnight at −20°C. After two washes with −20°C pre-cooled 80 % ethanol
RNA was dried at room temperature and dissolved in nuclease-free water (Ambion Inc., Austin,
TX, USA, now part of Applied Biosystems, Foster City, CA, USA).
RNA was DNase treated and afterwards purified using the RNA Clean-Up and Concentration
Kit (Norgen Biotek Corp., Thorold, ON, Canada; distributed by BioCat GmbH, Heidelberg,
Germany). The concentration was determined photometrically (NanoDrop ND-1000, NanoDrop
Technologies, Wilmington, DE, USA), and the quality was checked with an Agilent 2100
Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA).
DNA array analysis
The RNA expression profile was analyzed on GeneChip Mouse Gene 1.0 ST arrays (Affymetrix,
Santa Clara, CA, USA) using the Whole Transcript (WT) Sense Target Labeling and Control
reagents according to the manufacturer’s instructions. Arrays were washed and stained in a
GeneChip FluidicsStation 450 and scanned with a GeneChip Scanner 3000 (all: Affymetrix).
DNA array data analysis
The array image files (CEL) were first quality controlled in the Expression Console software
(Affymetrix) and then imported into the Rosetta Resolver software (Rosetta Biosoftware, Seattle,
WA, USA) for data analysis. Signals were generated and normalized using the RMA algorithm.
Groups were compared in log-transformed space using error-weighted one-way ANOVA with
Benjamini-Hochberg False Discovery Rate multiple testing correction and p* < 0.01 was regarded
as significant. The control sequences (e. g. negative, positive, polyA, and hybridization controls)
and sequences that were not expressed on all arrays in the selected group comparison (with
p > 0.01 on intensity profile level in Rosetta Resolver software) were not included in statistical
testing. Afterwards, sequence sets were translated into EntrezGene records using the Rosetta
Resolver annotation of December 2009. Two criteria were used for definition of differential
expression: Significance in ANOVA statistical testing and a minimal absolute fold change of 2.
Genes significant in ANOVA but with a minimal absolute fold change of only 1.5 were considered
to be regulated by trend.
Ingenuity Pathway Analysis (IPA) of gene expression data
EntrezGene identifiers and fold change gene expression data of differentially expressed genes
were imported to the Ingenuity Pathway Analysis tool (Ingenuity Systems Inc.,
www.ingenuity.com) and analyzed using all genes of the GeneChip Mouse Gene 1.0 ST array as
reference set without further restriction.
45

Maren Depke
Material and Methods
GENE EXPRESSION PATTERN OF BONE-MARROW DERIVED
MACROPHAGES AFTER INTERFERON-GAMMA TREATMENT
Stem cell preparation, cultivation, and differentiation to macrophages [performed by
Katrin Breitbach]
Preparation and cultivation of mouse stem cells and their differentiation into bone-marrow
derived macrophages (BMM) was conducted as described by Eske et al. in 2009. Briefly, bone
marrow cells from tibias and femurs of n = 3 or n = 15 BALB/c or n = 4 or n = 15 C57BL/6 mice
were prepared in sterile conditions, pooled and cultivated for ten days using the serum-free
BMM-medium supplemented with GM-CSF as established by Eske et al. (2009). Differentiated
BMM were harvested on day 10 for consecutive experiments.
Interferon-γ activation of bone marrow derived macrophages [performed by Katrin Breitbach]
Mature BMM were seeded in 6-well-plates with 0.8E+06 to 1.5E+06 cells / well. BMM in half of
the wells were treated for 24 hours by addition of 300 units/ml IFN-γ (Roche, Mannheim,
Germany) in serum-free BMM-medium, the other half was cultivated for the same time in the
same medium without IFN-γ. For transcriptome analysis, 1.6E+06 to 4.5E+06 cells / sample were
available, while proteome analysis needed a higher cell number and therefore had a sample size
of 1.0E+07 to 1.5E+07 cells.
Sample [Katrin Breitbach] and RNA preparation [Maren Depke] for transcriptome analysis
Medium was carefully removed from the sample wells, and BMM were lyzed in 1 ml
TriReagent per sample (Sigma, Steinheim, Germany) by repetitive pipetting. After incubation at
room temperature for 15 min, the lysate was flash-frozen in liquid nitrogen and stored at −70°C
until RNA-preparation.
RNA-preparation took place using a combined protocol of phenol-based preparation and
column-based purification. Chloroform was added to TriReagent lysates, samples were vigorously
shaken and incubated at room temperature for 5 min. Organic and aqueous phase were
separated by centrifugation for 15 min at 12000 x g and 4°C. Afterwards, the aqueous
supernatant was mixed with 0.5 ml isopropanol (2-propanol) and transferred to RNeasy Mini
columns (Qiagen GmbH, Hilden, Germany). The following steps of RNA purification were carried
out according to the manufacturer’s instructions including the optional DNase-treatment (RNasefree
DNase Set, Qiagen GmbH, Hilden, Germany). After ethanol precipitation, the RNA was
quantified spectrophotometrically, and its quality was verified using an Agilent 2100 Bioanalyzer
and RNA Nano Chips (Agilent Technologies Inc., Santa Clara, CA, USA).
Affymetrix DNA array analysis
Four strain-treatment combinations were included into this study: 1) medium control BALB/c
BMM, 2) IFN-γ treated BALB/c BMM, 3) medium control C57BL/6 BMM, and 4) IFN-γ treated
C57BL/6 BMM. Three biological replicates were analyzed for each of the four sample groups
described before.
47

Maren Depke
Material and Methods
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
Labeled cDNA for hybridization was prepared for each sample from 200 ng totalRNA using the
GeneChip Whole Transcript (WT) Sense Target Labeling Assay and hybridized to GeneChip Mouse
Gene 1.0 ST Arrays according to the manufacturer’s instructions (Affymetrix, Santa Clara, CA,
USA). Arrays were washed and stained using the GeneChip Hybridization, Wash, and Stain Kit in a
GeneChip Fluidics Station 450 and scanned with a GeneChip Scanner 3000 (all items from
Affymetrix, Santa Clara, CA, USA).
Resulting array image files (CEL-files) were imported to the Rosetta Resolver System (Rosetta
Biosoftware, Seattle, WA, USA). Signal intensities were extracted using the RMA algorithm and
differentially expressed probe sets were accessed with error-weighted one-way Analysis of
Variance (ANOVA) including Benjamini Hochberg False Discovery Rate (FDR) multiple testing
correction at the analysis levels of Intensity Profiles and sequences in log-transformed space.
Values of p* 0.01 on intensity profile level in Rosetta Resolver software) were not included in
statistical testing.
Each run of the statistical test compared two sample groups that consisted of three biological
replicates each. Four comparisons were included into statistical testing: 1) IFN-γ treated BALB/c
BMM versus medium control BALB/c BMM to access IFN-γ effects in BALB/c BMM, 2) IFN-γ
treated C57BL/6 BMM versus medium control C57BL/6 BMM to retrieve IFN-γ effects in C57BL/6
BMM, 3) medium control C57BL/6 BMM versus medium control BALB/c BMM to obtain strain
differences at the non-activated control level, and 4) IFN-γ treated C57BL/6 BMM versus IFN-γ
treated BALB/c BMM to elicit strain differences at the IFN-γ activated level.
The Rosetta Resolver software allows mapping of 28944 records of sequence level
information (i.e. the Affymetrix probe sets) to genes via the EntrezGene nomenclature resulting
in 20074 records. Additionally, the software calculates expression data for genes from the
original sequence level values. This function also combines intensities of two or more probe sets
for genes that are represented by more than one probe set. The lists of statistically significant
differentially expressed probe sets resulting from ANOVA were translated into lists of
differentially expressed genes by using the EntrezGene level annotation included in the Rosetta
Resolver software. In order to exclude biologically irrelevant small changes in expression level
from the statistically significant lists resulting from ANOVA, expression data on EntrezGene level
was restricted to a minimal absolute fold change of 1.5. Nevertheless, only a minority of
statistically significant genes did not pass this fold change cutoff.
Proteome analysis [performed by Dinh Hoang Dang Khoa]
Proteome analysis was performed by Dinh Hoang Dang Khoa using both gel-based and gelfree
approaches. Briefly, BMM protein extracts in urea/thiourea buffer were separated by Two
Dimensional Fluorescence Difference Gel Electrophoresis (2D-DIGE), and spots on the gels were
identified using MALDI-MS-MS. Tryptic peptides of the protein extracts were analyzed with LTQ-
FT-ICR mass spectrometer after liquid chromatographic (LC) prefractionation. Differential analysis
of label-free MS data was performed using the Rosetta Elucidator software package (Rosetta
Biosoftware, Seattle, WA, USA).
Comparison of transcriptome and gel-free LC-MS/MS proteome results
To compare results from transcriptome and gelfree proteome analysis, it was necessary to
map proteins from LC-MS/MS identification to genes available on the GeneChip Mouse Gene 1.0
ST array (Affymetrix). In order to do so, protein IPI identifiers were translated into EntrezGene IDs
using the PIPE (http://pipe.systemsbiology.net/pipe) and UniProt ID (www.uniprot.org) mapping
tools (Lars Brinkmann). Missing EntrezGene IDs were added manually if possible. The resulting list
48

Maren Depke
Material and Methods
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
of proteins was imported into the Rosetta Resolver software using the EntrezGene IDs and
compared with genes on the Affymetrix array based on the annotation of the Rosetta Resolver
software.
Global functional analysis of transcriptomic [Maren Depke] and proteomic [Dinh Hoang Dang
Khoa] results using Ingenuity Pathway Analysis (IPA)
The lists of differentially expressed genes with an absolute fold change of at least 1.5 were
directly uploaded from Rosetta Resolver System to the Ingenuity Pathway Analysis tool Version
7.5 and 8.0 (IPA, Ingenuity Systems, www.ingenuity.com). IPA assigned the EntrezGene IDs to the
corresponding records of the so called Ingenuity Pathway Knowledge Base (IPKB), combined
related genes to networks and conducted a statistical analysis for over-represented functions and
canonical pathways in the imported lists of differentially expressed items in relation to the
sequences available on the GeneChip Mouse Gene 1.0 ST Array as reference set. Generally,
analysis in IPA was not restricted to species, type of relationship, type of molecule or the like.
Only for building user-defined pathways, a restriction to cell type “macrophages” or “RAW cells”
was used.
Correspondingly, protein IPI identifiers, p-values, and fold change of regulated proteins were
uploaded to IPA and also analyzed by the network, function and canonical pathway tools. In this
case, the whole list of identified proteins was uploaded and afterwards restricted to a minimal
absolute fold change of 1.5 and a p-value of at least 0.01. This approach allowed using all
identified proteins as reference set.
In case of interest, automatically generated networks from separate analyses were merged
using the automatic merge-tool from IPA. User-defined networks (called pathways) centered
around the starting node IFN-γ were created using the grow-tool of IPA‘s pathway function. The
addition of further nodes according to information stored in the so-called Ingenuity Pathway
Knowledge Base (IPKB) was restricted to a list of genes/proteins that are differentially expressed
in at least 1 of 4 comparisons 1) IFN-γ effects in BALB/c BMM, 2) IFN-γ effects in C57BL/6 BMM,
3) strain difference at non-treated control level and 4) strain difference after IFN-γ activation.
Networks were built either without further restrictions or with additional restriction to
macrophage/RAW cells (i.e. only genes/proteins in the IPKB described to be linked to IFN-γ in
macrophages or RAW cells were allowed to enter the new network). Lists of genes included in
the IFN-γ centered networks were compared to identify common and unique genes between the
networks without and with restriction to macrophage/RAW cells and finally exported via
spreadsheet.
49

Maren Depke
Material and Methods
HOST CELL GENE EXPRESSION PATTERN IN AN
IN VITRO INFECTION MODEL
Host cell line and conditions of cell cultivation [performed by Melanie Gutjahr]
S9 cells were grown to confluency in 10-cm-diameter cell culture dishes with 10 ml eMEM cell
culture medium. Cell culture medium eMEM consists of 1x concentrated MEM (PromoCell GmbH,
Heidelberg, Germany) supplemented with additional 4 % FCS (fetal calf serum, Biochrom AG,
Berlin, Germany), 1 % non-essential amino acids (PAA Laboratories GmbH, Pasching, Austria), and
2 % L-glutamine (200 mM stock; PAA).
Bacterial growth medium and cultivation [performed by Melanie Gutjahr and Maren Depke]
Staphylococcus aureus RN1HG GFP (plasmid pMV158GFP) was grown in 100 ml pMEM
medium at 37°C with linear shaking of 125 strokes/min (stroke length 28 mm) in 500-ml-
Erlenmeyer bacterial culture flasks. Optical density (OD) was measured at 600 nm.
The adapted cell culture medium pMEM contains 1x concentrated MEM without sodium
bicarbonate (10x concentrate; Invitrogen, Karlsruhe, Germany), 1 % non-essential amino acids
(PAA Laboratories GmbH, Pasching, Austria) and 4 mM L-glutamine (PAA) and is supplemented
with 10 mM HEPES (PAA) and 2 mM of each L-alanine, L-leucine, L-isoleucine, L-valine, L-
aspartate, L-glutamate, L-serine, L-threonine, L-cysteine, L-proline, L-histidine, L-phenyl alanine,
and L-tryptophan (PromoCell GmbH, Heidelberg, Germany). The pH is adjusted to 7.4 with NaOH.
Use of this medium has been established by Sandra Scharf and has been first reported by
Schmidt et al. in 2010.
Cell culture infection model [performed by Melanie Gutjahr and Maren Depke]
When bacterial cultures reached an OD of 0.4 the S9 cells were infected with a multiplicity of
infection (MOI) of 25. Bacterial cultures and eukaryotic cell culture medium were mixed in a final
volume sufficient for inoculation of all cell culture plates processed in parallel. Subsequently, this
mix was added to the cell culture plates to ensure equal distribution of staphylococci on the host
cell layer and reproducible infection of all cell culture plates in each experiment.
6.36E+07 cfu/ml had been determined at OD 0.4 of S. aureus RN1HG GFP in pMEM (Melanie
Gutjahr). Confluent 10-cm-diameter cell culture plates contain 8.0E+06 S9 cells. Therefore, in
these experiments approximately 30 % of bacterial culture had to be included in the infection
medium to obtain a suspension of which 10 ml infect one cell culture plate with a MOI of 25
(3.14 ml bacterial culture in a total of 10 ml infection medium).
S9 cells and staphylococci were co-incubated for 1 h at 37°C in 5 % CO 2 -atmosphere.
Afterwards, the infection medium was replaced by eMEM containing 10 µg/ml lysostaphin (AMBI
PRODUCTS LLC, Lawrence, NY, USA) until harvesting of cells. Two time points were included in
this study: 2.5 h and 6.5 h after start of infection (Fig. M.4.1).
Control samples were treated accordingly. Only the volume of bacterial culture in the
infection mix was substituted by fresh, sterile bacterial cell culture medium pMEM.
51

Maren Depke
Material and Methods
Host Cell Gene Expression Pattern in an in vitro Infection Model
cultivation to OD 0.4
experimental time point / h
0 1 2 3 4 5
6 7
infection
1h
lysostaphin treatment
t 0 2.5 h 6.5 h
infected
eukaryotic
samples
FACS for
RNA and
protein
FACS for
RNA and
protein
Fig. M.4.1:
Time line of
infection
experiments
for eukaryotic
host samples
and control
measurements.
eukarotic
control
samples
control
measurements
controlfor
protein
cfu
host cell
vitality
controlfor
RNA and
protein
cfu
host cell
vitality
controlfor
RNA and
protein
cfu
host cell
vitality
Sample harvesting for transcriptome analysis
After co-incubation of staphylococcal with host cells and subsequent lysostaphin treatment,
the S9 cell layer was washed once with Dulbecco’s PBS without Ca 2+ and Mg 2+ (PAA Laboratories
GmbH, Pasching, Austria). Cells were detached with 1 ml trypsin/EDTA (PAA) supplemented with
0.1 µg/ml actinomycin D (Sigma-Aldrich, Steinheim, Germany) and 80 mM sodium azide (Merck
KGaA, Darmstadt, Germany) for some minutes. Trypsin reaction was stopped with 3 ml cell
culture medium eMEM, and cells were pelleted at room temperature for 5 min at 500 x g with
slightly reduced break. The cells were washed once with Dulbecco’s PBS with Ca 2+ and Mg 2+ (PAA)
and resuspended in FlacsFlow (Becton Dickinson Biosciences, San Jose, CA, USA) for sorting of
infected and non-infected host cells. Both solutions were supplemented with 0.025 µg/ml
actinomycin D (Sigma-Aldrich) and 20 mM sodium azide (Merck). Control samples were treated in
the same way.
Protein samples and control measurements [performed by Melanie Gutjahr]
Samples for transcriptome analysis were harvested in parallel with samples for host proteome
analysis (Fig. M.4.1). For protein samples, trypsin, PBS and FacsFlow solutions were not
supplemented with actinomycin D and sodium azide, and control samples were directly lysed in
UT buffer (8 M urea, 2 M thiourea). One additional control without any treatment was included.
For the bacterial starting culture (OD 0.4), the infection mix, and samples after 2.5 h and 6.5 h
of infection (internalized bacteria) viable cell counts were determined by plating on TSB agar and
incubation for 24-48 h at 37°C.
FACS measurements and cell sorting [performed by Petra Hildebrandt], sample harvest and
disruption [performed by Maren Depke]
Cells were sorted into infected (green fluorescence positive) and non-infected (green
fluorescence negative) cells in a biosafety 2 level FACS Aria high-speed cell sorter (Becton
Dickinson Biosciences, San Jose, CA, USA) with 488 nm excitation from a blue Coherent Sapphire
solid state laser at 18 mW. Optical filters were set up to detect the emitted GFP fluorescence at
52

Maren Depke
Material and Methods
Host Cell Gene Expression Pattern in an in vitro Infection Model
515 nm to 545 nm (FITC filter block). All FSC and SSC data were recorded at linear scale, the
fluorescence data were recorded at logarithmic scale with the FACS Diva v5.03 software (Becton
Dickinson). Prior to measurement of bacteria containing eukaryotic S9 cells, the proper function
of the instrument was determined by using Spherotech’s Rainbow calibration particles
(Spherotech, Lake Forest, IL, USA). Prior to sorting, the drop delay was adjusted to >99 % sorting
with the ACCUDROP beads (Becton Dickinson). Sorting was performed from the gate set in FSC
vs. FITC dot plot after eliminating doublets and cell clumps by gating SSC−W vs. SSC−A and FSC−W
vs. FSC−A at a threshold rate up to 3000 cells/s with sort mode purity, which results in a sorted
sample that is highly pure, at the expense of recovery and yield.
For each transcriptome sample, 5.0E+05 to 7.0E+05 cells were collected, whereas for protein
samples approximately 4.0E+06 cells were sorted. During the whole period of sorting and sample
collection, the sample reservoir and the collection tubes were cooled to 4°C. Control samples
were stored on ice during sorting of infected samples. Sorted and control cell suspensions were
centrifuged for 5 min at 500 x g and 4°C. The supernatants were removed and each cell pellet
devoted to transcriptome analysis was lysed in 1 ml TRIZOL (Invitrogen, Karlsruhe, Germany) by
repeated pipetting. After incubation for 10 min at room temperature, the lysate was flash-frozen
in liquid nitrogen and stored −70°C until RNA preparation. The remaining cells after sorting of
samples for transcriptome and proteome analysis were used to acquire information on cell
vitality by propidium iodide staining.
RNA preparation
The lysates were thawed at room temperature. Chloroform was added (200 µl
chloroform / 1 ml TRIZOL), samples were shaken vigorously for 15 s and incubated at room
temperature for 5 min. Organic and aqueous phase were separated by centrifugation (12000 x g,
15 min, 4°C) and RNA was precipitated from the aqueous phase with 500 µl isopropanol (2-
propanol) / 1 ml TRIZOL and 1 µl 5 mg/ml linear acrylamide (Ambion Inc., Austin, TX, USA, now
part of Applied Biosystems, Foster City, CA, USA) as precipitation aid overnight at −20°C. After
two washes with −20°C pre-cooled 80 % ethanol, the RNA was dried at room temperature and
solved in nuclease-free water (Ambion). The concentration was determined photometrically
(NanoDrop ND-1000, NanoDrop Technologies, Wilmington, DE, USA), and the quality was
checked with an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA).
DNA array analysis
The RNA expression profile of host cells was analyzed on GeneChip Human Gene 1.0 ST arrays
(Affymetrix, Santa Clara, CA, USA) using the Whole Transcript (WT) Sense Target Labeling and
Control reagents according to the manufacturer’s instructions. Arrays were washed and stained
in a GeneChip FluidicsStation 450 and scanned with a GeneChip Scanner 3000 (all: Affymetrix).
DNA array data analysis
The array image files (CEL) were first quality controlled in the Expression Console software
(Affymetrix) and then imported into the Rosetta Resolver software (Rosetta Biosoftware, Seattle,
WA, USA) for data analysis. Signals were generated and normalized using the RMA algorithm.
Groups were compared in log-transformed space using error-weighted one-way ANOVA with
Benjamini-Hochberg False Discovery Rate multiple testing correction and p* < 0.01 was regarded
as significant. The control sequences (e. g. negative, positive, polyA, and hybridization controls)
and sequences that were not expressed on all arrays in the selected group comparison (with
p > 0.01 on intensity profile level in Rosetta Resolver software) were not included in statistical
testing.
53

Maren Depke
Material and Methods
Host Cell Gene Expression Pattern in an in vitro Infection Model
Afterwards, the sequence sets were translated into EntrezGene records using the Rosetta
Resolver annotation of December 2009, and an absolute fold change cutoff of 1.5 after
combining the biological replicates was applied.
Ingenuity Pathway Analysis (IPA) of gene expression data
EntrezGene identifiers and fold change gene expression data of differentially expressed genes
were imported to the Ingenuity Pathway Analysis tool version 8.6 (Ingenuity Systems Inc.,
www.ingenuity.com) and analyzed using all genes of the GeneChip Human Gene 1.0 ST array as
reference set without further restriction.
54

Maren Depke
Material and Methods
PATHOGEN GENE EXPRESSION PROFILING
Growth Media Comparison Study
Bacterial cultivation, growth media, and sampling time points
Staphylococcus aureus RN1HG was grown at 37°C with orbital shaking of 200 rpm in
Erlenmeyer bacterial culture flasks. Culture volume did not exceed 1/5 of culture flask volume.
Optical density (OD) was measured at 600 nm (Fig. M.5.1).
Different media were included in this study in an international co-operation in the settings of the
EU-IP-FP6-project BaSysBio (LSHG-CT2006-037469) consortium, e. g. the complete bacterial
culture medium TSB, cell culture medium, minimal medium, and human serum. Here, only results
from samples of the medium “pMEM” will be presented. The contents of the adapted cell culture
medium pMEM (Schmidt et al. 2010) have already been listed above (see Material and
Methods/Host Cell Gene Expression Pattern in an in vitro Infection Model/Bacterial growth
medium, page 51).
Bacterial samples were taken in the exponential growth phase and 2 h (t 2 ) and 4 h (t 4 ) after
entry into stationary growth.
OD600
10.00
1.00
0.10
TSB
pMEM
Fig. M.5.1:
Example for bacterial growth in TSB and pMEM medium.
0.01
0 2 4 6 8 10 12 14 16 18 20 22 24
time / h
Bacterial cell harvest and disruption
At different growth phases, 5 to 15 optical density units of S. aureus RN1HG were harvested
on ice with addition of at least one third volume of Killing Buffer (20 mM Tris pH 7.5, 5 mM
MgCl 2 , 20 mM NaN 3 ). Pellets were flash-frozen in liquid nitrogen and stored at −70°C until cell
disruption.
For cell disruption, the pellet was resuspended on ice in 200 µl Killing Buffer, transferred to
liquid nitrogen cooled Teflon vessels, and disintegrated mechanically at 2600 rpm for 2 min in a
bead mill (Mikrodismembrator S, B. Braun Biotech International GmbH, Melsungen, Germany;
now part of Sartorius AG, Göttingen, Germany). Frozen cell and buffer powder mix was
resuspendend in 4 ml of 50°C pre-warmed lysis solution (4 M guanidine-thiocyanate, 25 mM
sodium acetate pH 5.5, 0.5 % N-lauroylsarcosinate) until the solution appeared clear and
homogeneous. Four aliquots with 1 ml each were intermittently frozen in liquid nitrogen and
stored −70°C.
55

Maren Depke
Material and Methods
Pathogen Gene Expression Profiling
RNA preparation
RNA was prepared using the acid-phenol-method. The lysates / aqueous phases were
extracted twice with an equal volume of acid phenol solution (Roth, Karlsruhe, Germany) and
once with an equal volume of chloroform / isoamyl alcohol (3-methyl-1-butanol) mix (24 vol. of
chloroform and 1 vol. of isoamyl alcohol equilibrated with 1 M Tris/HCl pH 8.0). RNA was
precipitated from the remaining cleaned aqueous phase with 1/10 volume of 3 M sodium acetate
pH 5.5 and 1 volume of isopropanol (2-propanol) overnight at −20°C. After two washes with
−20°C pre-cooled 80 % ethanol, the RNA was dried at room temperature and solved in nucleasefree
water (Ambion Inc., Austin, TX, USA, now part of Applied Biosystems, Foster City, CA, USA).
To avoid the influence of potential DNA contamination, the RNA was DNase treated and
afterwards purified using the RNA Clean-Up and Concentration Kit (Norgen Biotek Corp., Thorold,
ON, Canada; distributed by BioCat GmbH, Heidelberg, Germany). The concentration was
determined photometrically (NanoDrop ND-1000, NanoDrop Technologies, Wilmington, DE,
USA), and the quality was checked with an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa
Clara, CA, USA).
56

Maren Depke
Material and Methods
Pathogen Gene Expression Profiling
In vitro Infection Experiment Study
Host cell line and culture conditions [performed by Petra Hildebrandt]
S9 cells were cultivated as described above (see Material and Methods/Host Cell Gene
Expression Pattern in an in vitro Infection Model, page 51).
Bacterial growth medium and cultivation [performed by Melanie Gutjahr and Maren Depke]
Staphylococcus aureus RN1HG or RN1HG GFP (plasmid pMV158GFP) was grown in 200 ml
pMEM medium at 37°C with linear shaking of 110 strokes/min (stroke length 28 mm) in 1000-ml-
Erlenmeyer bacterial culture flasks. All other experimental parameters (e. g. medium pMEM)
were identical to those described above (see Material and Methods/Host Cell Gene Expression
Pattern in an in vitro Infection Model, page 51).
Cell culture infection model [performed by Melanie Gutjahr and Maren Depke]
The cell culture infection model was performed as described above (see Material and
Methods/Host Cell Gene Expression Pattern in an in vitro Infection Model, page 51) and also
included the same two time points of 2.5 h and 6.5 h after start of infection (Fig. M.5.2).
cultivation to OD 0.4
experimental time point / h
0 1 2 3 4 5
6 7
infection
1h
lysostaphin treatment
t 0 1 h
2.5 h 6.5 h
internalized
bacterial
samples
internalized
S. aureus
internalized
S. aureus
Fig. M.5.2:
Time line
of infection
experiments
for bacterial
samples.
bacterial
control
samples
exponential
growth
phase
(OD 0.4)
S. aureus
nonadherent
S. aureus
serum/CO 2
control
S. aureus
anaerobic
S. aureus
serum/CO 2
control
S. aureus
serum/CO 2
control
S. aureus
57

Maren Depke
Material and Methods
Pathogen Gene Expression Profiling
Bacterial control samples
Bacterial samples were taken as comparison and to control for additional experimental effects
(Fig. M.5.2 and Fig. M.5.3):
exponential growth phase at OD 0.4
staphylococci after 1 h, 2.5 h, and 6.5 h of incubation in the infection medium in 5 %
CO 2 -atmosphere at 37°C without agitation (i.e. in cell culture dishes in serum
containing medium but without presence of host cells)
non-adherent staphylococci after 1 h of co-incubation with eukaryotic cells and their
products in 5 % CO 2 -atmosphere and at 37°C
staphylococci after 2.5 h of anaerobic incubation in pMEM medium at 37°C
Cells were harvested using Killing Buffer (20 mM Tris pH 7.5, 5 mM MgCl 2 , 20 mM NaN 3 ) as
described above (see Material and Methods/Growth Media Comparison Study/Bacterial cell
harvest, page 55).
exponential growth phase
anaerobic
incubation:
2.5 h
infection
mix
serum control
with CO 2 exposure:
1 h, 2.5 h, 6.5 h
non-adherent staphylococci: 1 h
Fig. M.5.3:
Visualization of bacterial control samples in
the in vitro infection experiments for tiling
array analysis.
internalized staphylococci: 2.5 h, 6.5 h
S9 cells
staphylococci
Preparation of internalized staphylococci
The lysostaphin-containing medium was replaced by 1 ml Killing Buffer (20 mM Tris pH 7.5,
5 mM MgCl 2 , 20 mM NaN 3 ) with 150 mM NaCl. In this isotonic buffer the infected S9 cells were
scraped from the culture dish, resuspended and transferred to a 1.5-ml tube for further
processing while maintaining the integrity of the majority of host cells. All following steps were
performed on ice or at 4°C. Cells were pelleted (600 x g for 5 min) and fixed in ice-cold
acetone/ethanol (50 % v/v) for 4 min as described by Garzoni et al. (2007) followed by a
centrifugation step at 20000 x g for 3 min. The eukaryotic part of the cell pellet was lysed in RLT
buffer (Qiagen, Hilden, Germany) and homogenized twice using QIAshredder (Qiagen, Hilden,
Germany) and centrifugation at 20000 x g for 2 min. In this process staphylococci were not lysed
although they lost their viability. Therefore, the resulting pellet contained staphylococcal cells
and eukaryotic cell debris. Pellets from several plates processed in parallel were combined and
washed once with RLT buffer and four times with TE buffer (10 mM Tris/HCl pH 8, 1 mM EDTA
pH 8) to remove residual contaminations of the host cells. Staphylococcal cell pellets were flashfrozen
in liquid nitrogen and stored at −70°C until cell disruption.
58

Maren Depke
Material and Methods
Pathogen Gene Expression Profiling
Bacterial cell disruption
For cell disruption the bacterial cell pellet was resuspended on ice in 60 µl TE (10 mM Tris/HCl
pH 8, 1 mM EDTA pH 8) supplemented with 20 mM NaN 3 and transferred together with 0.5 ml
TRIZOL (Invitrogen, Karlsruhe, Germany) to liquid nitrogen cooled Teflon vessels. Cells were
disintegrated mechanically at 2600 rpm for 2 min in a bead mill (Mikrodismembrator S, B. Braun
Biotech International GmbH, Melsungen, Germany; now part of Sartorius AG, Göttingen,
Germany). Another 0.5 ml of TRIZOL was added and the frozen lysate was allowed to thaw. In
total, the liquid lysate was incubated at room temperature for 10 min and afterwards flashfrozen
in liquid nitrogen and stored −70°C until RNA preparation.
RNA preparation
The lysates from bead mill disruption were thawed at room temperature. Chloroform was
added (200 µl chloroform / 1 ml TRIZOL), samples were shaken vigorously for 15 s and incubated
at room temperature for 5 min. Organic and aqueous phase were separated by centrifugation
(12000 x g, 15 min, 4°C) and RNA was precipitated from the aqueous phase with 500 µl
isopropanol (2-propanol) / 1 ml TRIZOL overnight at −20°C. After two washes with −20°C precooled
80 % ethanol the RNA was dried at room temperature and solved in nuclease-free water
(Ambion Inc., Austin, TX, USA, now part of Applied Biosystems, Foster City, CA, USA).
DNase treatment was only possible for control samples because the RNA yield of internalized
staphylococci was near the minimum amount needed for tiling array hybridization. For controls,
DNase digestion and RNA purification was performed as described before (see Material and
Methods/Growth Media Comparison Study, page 56). The concentration was determined
photometrically (NanoDrop ND-1000, NanoDrop Technologies, Wilmington, DE, USA) and the
quality was checked with an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA,
USA).
59

Maren Depke
Material and Methods
Pathogen Gene Expression Profiling
Tiling Array Expression Profiling
Tiling array design
The “080604_SA_JH_Tiling” array was designed by Hanne Jarmer (Center for Biological
Sequence Analysis, Department of Systems Biology, Technical University of Denmark, Lyngby,
Denmark) using algorithms, which have recently been published in the context of a B. subtilis
tiling array (Rasmussen et al. 2009). In brief, iso-thermal probes of 45 nucleotides (nt) to 65 nt
were designed to cover the whole genome of S. aureus. Probes were arranged in 22 nt intervals
on each strand and possessed an offset of 11 nt between the strands. The array design for
S. aureus was part of the international cooperation EU-IP-FP6-project BaSysBio (LSHG-CT2006-
037469) which is funded by the European Commission and coordinated by Philippe Noirot (INRA,
Mathématique Informatique et Génome, Jouy-en-Josas, France).
The sequences were synthesized on quartz wafers by NimbleGen (Roche NimbleGen,
Madison, WI, USA) in a custom, high-density DNA array format by applying the Maskless Array
Synthesizer (MAS) technology in combination with photo-mediated synthesis chemistry. The
basic principle of the synthesis is a solid-state array of miniature aluminum mirrors which direct
UV-light at the place where the next reaction steps should occur. An UV-labile protection group is
separated from the nascent oligonucleotide and releases a reactive site for binding of the next
nucleotide. Thus, the synthesis of the oligonucleotide sequences requires m x 4 reaction steps,
where m is the length of the oligonucleotide (number of nucleotides) and at each intermediate
length step the four possible nucleotides A, T, C, and G will be added in a single, separate
reaction.
Tiling array hybridization
The tiling array hybridization was performed at NimbleGen (Roche NimbleGen, Madison, WI,
USA) according to standard protocols. Briefly, 10 µg of high quality RNA samples were reverse
transcribed to cDNA in the presence of actinomycin D with subsequent alkaline RNA hydrolysis.
Precipitated and resolved cDNA was labeled with Cy3 dye via NHS-Ester Dye Coupling Reaction.
After spin-column based purification and quality control, labeled cDNA was hybridized together
with appropriate controls to tiling arrays of the 080604_SA_JH_Tiling design. Arrays were washed
and scanned and raw intensity data of tiling probes were provided to the customer.
Tiling array raw data analysis
The raw data analysis and condensing of probe data to transcripts/genes was performed by
Pierre Nicolas, Aurélie Leduc, and Philippe Bessières (INRA, Mathématique Informatique et
Génome, Jouy-en-Josas, France). Intensity values for annotated genes were derived from the
individual probe data by calculating the median of the probes located within the genomic
coordinates of these genes. Analysis of hybridization signal, identification of transcription start
and end boundaries and by this means segmentation into transcriptional units was executed by a
novel algorithm based on a hidden Markov model, which allows to identify new, formerly
unknown or non-annotated transcripts (Nicolas et al. 2009).
60

Maren Depke
Material and Methods
Pathogen Gene Expression Profiling
Tiling array expression data analysis
Condensed, linear space gene expression values were imported via the Gene Expression Text
Loader (GETL) into the Rosetta Resolver software (Rosetta Biosoftware, Seattle, WA, USA) for
data analysis. Inter-chip median-scaling (normalization) and detrending (Rosetta Resolver
proprietary algorithm) were applied in the experiment definitions (ED) to allow direct comparison
of values from different arrays. Groups were compared in log-transformed space using textbook
one-way ANOVA with Benjamini-Hochberg False Discovery Rate multiple testing correction and
p* < 0.05 was regarded as significant. An absolute fold change cutoff of 2 was applied.
61

corticosterone [pg/ml plasma]
corticosterone [pg/ml plasma]
corticosterone [pg/ml plasma]
Maren Depke
R E S U L T S
LIVER GENE EXPRESSION PATTERN IN A MOUSE
PSYCHOLOGICAL STRESS MODEL
The results of the liver gene expression profiling in a mouse psychological stress model have
been published by Depke et al. in 2008 and 2009. All array data are available at the NCBI’s Gene
Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) database and are accessible
through GEO Series accession no. GSE11126. Furthermore, lists of differentially expressed genes
are included as supplemental data (Depke et al. 2008) at the journal’s homepage
(http://endo.endojournals.org). Stress and infection experiments, physiological measurements,
and histological analyses were performed by Cornelia Kiank. Her data are cited here because they
are essential for the understanding of physiological events and potential dysregulations during
acute and chronic stress in mice.
Repeated stress-induced cachexia accompanied by hypercortisolism, hyperleptinemia, and
hypothyroidism [Cornelia Kiank]
Repeated psychological stress caused a severe loss of body mass in BALB/c mice while mean
food intake and water consumption were unaltered during 4.5 days stress exposure. Food intake
was 95.1 ± 19.9 g/cage in the repeatedly stressed vs. 91.9 ± 17.4 g/cage in the nonstressed
groups, and water consumption was 250 ± 40 ml/cage in stressed vs. 240 ± 40 ml/cage in the
control mice (three independent experiments with nine mice per cage). As the food and water
intake was consistently found to be normal, the question arose whether the severe loss of body
mass after repeated stress exposure was dependent on hormonal changes. Repeatedly stressed
mice showed increased corticosterone concentrations in the peripheral blood (Fig. R.1.1 A) along
with a hypertrophy of the adrenal cortex with decreased size of lipid storage vesicles in the
glucocorticoid-producing zona fasciculata (Fig. R.1.1 B, C).
A
A
A750 *
A
750 *
750 *
B
B
B
B
Fig. R.1.1: Repeated stress-induced activation of the HPA axis in BALB/c
mice.
A. Increased plasma corticosterone levels in repeatedly stressed mice
(black box plot) compared with nonstressed mice (white box plot) (n = 9
mice per group).
B, C. Hypertrophy of the zona fasciculata of the adrenal cortex (white line)
in repeatedly stressed mice (B) compared with nonstressed controls (C)
(HE staining magnification, x100); each picture is representative for nine
mice per group. *, p < 0.05 Mann-Whitney U test; data reproduced in at
least three independent experiments.
500
500
500
250
250
250
0
0
0
CC
C
C
63

change of body weight [g]
leptin [pg/ml]
total T3 [ng/dL]
total T4 [µg/dL]
Maren Depke
Results
Liver Gene Expression Pattern in a Mouse Psychological Stress Model
In fact, the increase in circulating glucocorticoids (GCs) was less pronounced than after acute
stress (Depke et al. 2008: supplemental material 1 at http://endo.endojournals.org), but the
continuing hypothalamic-pituitary-adrenal (HPA) axis response might contribute to the reduction
of body weight of up to 20 % during the time course of repeated stress exposure (Fig. R.1.2 A).
This loss of body mass was not associated with significantly altered growth hormone (GH)
concentrations in the blood of stressed (13.6 ± 7.6 ng/ml) vs. nonstressed mice
(10.8 ± 5.4 ng/ml). However, stress-induced hyperleptinemia was measured (Fig. R.1.2 B). Total
triiodothyronine / T 3 (Fig. R.1.2 C) and total thyroxine / T 4 levels (Fig. R.1.2 D) were reduced in the
plasma of stressed animals, whereas thyroglobulin concentrations remained unchanged (data not
shown).
Fig. R.1.2: Repeated psychological stressinduced
loss of BW, increase of plasma
leptin levels, and hypothyroidism in mice.
A. Loss of body mass during the period of
4.5 days intermittent stress (black box
plots) compared with nonstressed control
mice (white box plots) (n = 12 mice per
group).
B. Plasma leptin levels after nine stress
cycles compared with nonstressed mice
(n = 12 mice per group).
C, D. T 3 (C) and T 4 (D) concentrations in
the plasma of repeatedly stressed and
control mice (n = 12 mice per group).
* p < 0.05; ** p < 0.01 Mann-Whitney
U test; data representative for two
independent experiments.
A B C D
*
0
-2.5
-5
*
750
500
250
0
350
*
300
250
200
7.5
5.0
2.5
0
**
Repeated stress-induced changes of global hepatic gene expression
In an approach to a more comprehensive characterization of the metabolic changes that occur
as a result of repeated acoustic and restraint stress, the expression signatures of liver from
repeatedly stressed BALB/c mice were recorded and compared with those of nonstressed
controls. The Affymetrix-based mRNA expression profiling of the liver of repeatedly stressed vs.
nonstressed animals revealed induction and repression, respectively, of 120 and 50 genes in both
independent stress experiments performed. To discriminate effects of repeated stress from those
of acute stress, the changes in the hepatic gene expression that occurred as a result of a single
stress exposure were additionally analyzed. In this model of acute stress, 192 and 123 genes
displayed stress-mediated induction or repression of expression.
Comparatively analyzing the effects of acute and repeated stress, it became clear that both
types of stress target a common set of 94 genes. Furthermore, 221 and 76 genes were
predominantly regulated by acute and repeated stress, respectively (Fig. R.1.3 A).
To analyze the changes in gene expression within the framework of already accumulated
knowledge, the lists of genes differentially expressed after acute and repeated stress or both
were subjected to an analysis using the IPA software. This software allowed for an intuitive
mining of the data of the 391 differentially expressed genes to gather an impression of the
biological rationale of the expression changes experimentally observed within the context of
published data.
When the IPA software was used to analyze the molecular and cellular functions targeted by
stress, an influence on broad categories such as “cell growth and proliferation” and “cell death”
was noted (Fig. R.1.3 B). However, it was also apparent that genes related to metabolic diseases
were most significantly influenced by the repeated stress exposure (Fig. R.1.3 C). This finding was
in line with the metabolic disturbances observed before. Supporting this notion of a major impact
64

Maren Depke
Results
Liver Gene Expression Pattern in a Mouse Psychological Stress Model
of repeated stress on metabolism, highly significant changes were also noted for more specific
categories such as amino acid metabolism and lipid metabolism. Some of these influences on
metabolism were already noted during acute stress because changes related to metabolic
disease ranked at number six when genes commonly influenced by acute and repeated stress
were analyzed. Genes involved in more specific categories of metabolism such as lipid and amino
acid metabolism were only moderately influenced by acute stress. Thus, the gene expression
profiling favors the idea that acute stress sets into motion a gene regulation cascade that is then
manifested during repeated stress exposure finally leading to the observed metabolic
disturbances.
A. Numbers of differentially expressed genes after acute and chronic stress or in both models
221
94
76
acute stress specific
differential gene expression
differential gene expression
after both
acute and repeated stress
repeated stress specific
differential gene expression
B. Over-represented functions with highest significance in the lists of differentially expressed genes
rank function function function
1 Cell Cycle Cellular Growth and Proliferation Metabolic Disease
2 Cellular Growth and Proliferation Cell Death Hematological Disease
3 Cell Death Connective Tissue Disorders Cell Signaling
4 Gene Expression Inflammatory Disease Immune Response
5 Cellular Development Skeletal and Muscular Disorders Lipid Metabolism
6 Cancer Metabolic Disease Amino Acid Metabolism
C. Selected over-represented metabolic functions in the lists of differentially expressed genes
function rank rank rank
Metabolic Disease 26 6 1
Amino Acid Metabolism 24 62 6
Lipid Metabolism 16 16 5
Fig. R.1.3: Impact of acute and repeated stress on metabolic genes and genes associated with metabolic disease.
A. Graphical display of genes differentially expressed after acute or repeated stress, or both stress types. The numbers given in the
Venn diagram include cDNAs, which are not functionally annotated: 29 of 221 genes regulated specifically in acute stress, nine of 94
genes regulated in both acute and repeated stress, and two of 76 genes regulated specifically for repeated stress.
B, C. The lists of genes differentially expressed either only after acute stress and repeated stress or after both acute and repeated
stress were loaded into IPA version 5.5 (Ingenuity Systems) to interpret the affected genes within the context of the published
literature. Metabolism seemed to be a major target of gene regulation after repeated stress exposure, and, thus, the influence of
acute and repeated stress on metabolism-associated genes was analyzed. The impact of acute or repeated stress alone as well as both
stress types is displayed by showing the rankings within the list of statistically significantly overrepresented functional groups for
genes related to metabolic disease, amino acid metabolism and lipid metabolism.
B. Over-represented functions with highest significance in the lists of differentially expressed genes in acute and repeated stress.
C. Impact of acute and repeated stress on metabolic functions.
To elucidate the reasons for stress-induced cachexia, it was decided to concentrate selectively
on changes of expression of genes whose products are involved in metabolic processes and
regulation of metabolic pathways. Several genes that were regulated in the liver of repeatedly
stressed animals could be linked to hypercatabolism (Fig. R.1.3). Genes relevant for amino acid
metabolism, especially amino acid transporters and enzymes metabolizing glucogenic amino
65

Maren Depke
Results
Liver Gene Expression Pattern in a Mouse Psychological Stress Model
acids [Slc15a4, Slc25a15, Slc3a1, Asl, and glutamate oxaloacetate transaminase 1 (Got1)], were
mostly induced. Moreover, the liver gene expression profiling of repeatedly stressed mice
indicated increased metabolism of lipids [Adh4, Apoa4, Cd74, Chpt1, Cyp17a1, Cyp2b10,
Cyp3a13, Cyp4a10, Cyp8b1, Hsd17b2, Hsd3b2, 4632417N05Rik (Hspc105), Saa2, Slco1a1, Xbp1].
To validate the array data, real-time RT-PCR focusing on chronic stress-induced dysregulation
and its pathophysiological effects was performed. Therefore, genes that were associated with
repeated stress-influenced metabolic processes of carbohydrate metabolism (Pck1), fat
metabolism (Srebf1), and amino acid metabolism (Asl, Sds), and with stress-induced apoptosis
(Gadd45b) were chosen. For all selected genes, the regulation found with the Affymetrix-based
expression profiling was confirmed (Table R.1.1).
Table R.1.1: Real-time PCR validation of array data in repeated stress experiments
target gene
control group a
ΔCt (Ct target − Ct reference Actb)
repeated stress group b
p-value
Asl 2.90 ± 0.33 1.50 ± 0.23 < 0.0001
Gadd45b 9.66 ± 0.64 6.83 ± 1.55 < 0.0001
Pck1 2.53 ± 0.52 0.02 ±0.72 < 0.0001
Sds 5.66 ± 0.41 3.70 ± 0.64 < 0.0001
Srebf1 8.01 ±0.18 8.36 ± 0.13 < 0.0001
a Validation of expression data by real-time PCR was carried out for all individual RNA preparations of the two biological experiments
(n = 9 plus n = 8 mice/group) of the two experiments focusing on effects of repeated stress exposures.
b Differences of ΔCt values were analyzed by Mann-Whitney test.
Induction of gluconeogenesis in repeatedly stressed mice
Stimulated by the observed loss in total body mass and the suspected involvement of
carbohydrate metabolism, the expression profiles for relevant genes were specifically
investigated, even if this category was not part of the first most significant biological functions
according to the IPA categorization. Stress-induced increased expression of Foxo1, Igfbp1, Irs1,
and Pck1, as well as reduced mRNA levels of Srebf1, can induce hyperglycemia because of
activation of gluconeogenic pathways. In contrast, the gene products of Cebpb, Igfbp1, St3gal5,
and Tnfrsf1b are associated with hypoglycemia, and may indicate counterregulatory processes to
decrease blood glucose levels.
In singularly stressed mice, in vivo analysis did not reveal significant changes of carbohydrate
regulation pathways, e. g. of leptin concentrations in the plasma, blood glucose levels, or liver
histology (Depke et al. 2008: supplemental material 1 at http://endo.endojournals.org).
In contrast, repeated stress induced pathophysiologically relevant alterations of protein and
lipid metabolism to provide fuel for gluconeogenesis. Only in chronically stressed mice
disturbances of the carbohydrate metabolism became detectable also in vivo. This included a
transient hypoglycemic period immediately after the termination of the ninth stress session.
However, after resuming food intake in the home cage, blood glucose levels increased rapidly
and resulted in a prolonged hyperglycemia that still was detectable 2 h later (Fig. R.1.4 A). In the
liver of repeatedly stressed mice, an increased usage of carbohydrate reservoirs was assessed by
reduced PAS staining that stains aldehyde groups of carbohydrates in tissue and revealed
reduced storage of carbohydrates in the liver of repeatedly stressed mice compared with healthy
control mice (Fig. R.1.4 B, C). Moreover, after repeated stress, insulin concentrations in the
plasma were slightly increased (272.5 ± 131.4 pg/ml) compared with control mice
(170 ± 149 pg/ml). Resistin, an insulin-resistance inducing adipokine, was significantly increased
in the plasma of repeatedly stressed mice when compared with nonstressed animals
(Fig. R.1.4 D). Analysis of pH in EDTA plasma samples revealed stress-induced acidosis
(Fig. R.1.4 E).
66

lood glucose levels [nmol/l]
resistin (ng/ml)
pH
Maren Depke
Results
Liver Gene Expression Pattern in a Mouse Psychological Stress Model
A B D E
*
*
15
*
60
*
7.5
*
10
*
C
50
7.4
5
40
7.3
0
0 0.5 2 h
30
7.2
Fig. R.1.4: Disturbances of murine carbohydrate metabolism after repeated psychological stress.
A. Kinetics of blood glucose levels immediately after termination of the ninth stress cycle (black box plots) compared with controls
(white box plot) (n = 9 mice per group).
B, C. Reduced storage of carbohydrates in the liver of repeatedly stressed mice (B) compared with nonstressed mice (C) (PAS staining
magnification, x200); each picture is representative for nine mice per group.
D, E. Plasma resistin levels (D) and pH of EDTA plasma (E) of stressed and nonstressed mice (n = 12 mice per group). * p < 0.05 Mann-
Whitney U test; data representative for two independent experiments.
Hypercholesteremia after repeated stress exposure
Global gene expression analysis of the liver of repeatedly stressed mice revealed stressinduced
changes of the gene expression profile of lipid metabolism (Fig. R.1.3). Therefore, the
lipid turnover of these mice was analyzed. After repeated stress exposure, a hepatic steatosis was
observed (Fig. R.1.5 A), whereas no significant numbers of lipid vesicles were detected in the liver
of control mice (Fig. R.1.5 B). A Sudan III staining, which selectively stains triglycerides but not
cholesterol esters, did not indicate any differences between stressed vs. nonstressed mice (data
not shown). Therefore, the lipids that were accumulated in the liver were presumably not
triglycerides but steroids or their precursor molecules. This is supported by the array data that
showed up-regulation of genes for steroid metabolism (Cyp17a1, Cyp2b10, Cyp39a1, Cyp4a14,
and Por).
Analysis of plasma lipid composition in repeatedly stressed mice showed reduced triglyceride
levels (Fig. R.1.5 C) but increased total cholesterol concentrations (Fig. R.1.5 D). Among
lipoproteins, the HDL fraction was increased (Fig. R.1.5 E), whereas VLDL concentrations were
strongly decreased (Fig. R.1.5 F). LDL-cholesterol levels did not change (Fig. R.1.5 G).
In contrast to the repeated stress model, plasma lipid composition or histological alterations
in the liver were not detected when comparing acutely stressed and control mice (Depke et al.
2008: supplemental material 1 at http://endo.endojournals.org). In addition, the expression
profiling of acutely stressed mice did not reveal major changes in genes involved in lipid
metabolism, probably indicating that hepatocytes of stressed mice started an anticipatory gene
expression program during the repeated stress sessions to face the stressful situation whose
physiological impact did not become detectable until stress exposure was repeated.
Loss of essential amino acids in repeatedly stressed mice
The gene expression analysis of the liver of repeatedly stressed animals also showed altered
expression profiles of genes whose products are involved in amino acid metabolism (e. g. Asl,
Got1, Prodh, Slc15a4, Slc25a15, Slc3a1, and Tdo; Fig. R.1.3). Despite the small group size of
analyzed animals, the amino acid composition of fresh plasma samples revealed significantly
67

triglycerides [mmol/l]
cholesterol [mmol/l]
HDL [mmol/l]
VLDL [mmol/l]
LDL [mmol/l]
Maren Depke
Results
Liver Gene Expression Pattern in a Mouse Psychological Stress Model
reduced concentrations of several essential amino acids, e. g. arginine, threonine, methionine,
and tryptophan, whereas nonessential amino acids showed fewer alterations in repeatedly
stressed mice (Depke et al. 2008: supplemental material 5 at http://endo.endojournals.org). In
addition, gene expression profiling of the liver of repeatedly stressed mice showed an induction
of genes for amino acid transporters and amino acid metabolizing enzymes (Sds, Slc15a4,
Slc25a15, Slc3a1, Got1, and Tat), and provided hints for increased activation of amino acid
degradation pathways (Aass, Ahcy, Asl, Prodh, and Tdo2). The induction of Asl expression along
with a loss of arginine and citrulline in the plasma (Depke et al. 2008: supplemental material 5 at
http://endo.endojournals.org) provided hints for altered urea cycle activity. This substantiates
the observations of systemic usage of the body’s protein stores in repeatedly stressed BALB/c
mice. In contrast, after a single acute stress session, mRNA expression of only a few glucogenic
amino acid transporters and metabolizing enzymes was induced in the liver (Sds, Slc38a2, and
Tat), which did not result in altered amino acid levels in the periphery (data not shown).
A
B
C D E F G
4
3.5 **
***
3
3.0
2
2.5
1
2.0
3.0
***
0.3 ***
2.5
0.2
2.0
0.1
0.4
0.3
0.2
0.1
0
1.5
1.5
0
0
Fig. R.1.5: Disturbances of fat metabolism in repeatedly stressed mice.
A, B. Hepatic steatosis in repeatedly stressed mice (A) compared with nonstressed mice (B) (HE staining magnification, x20); each
picture is representative for nine mice per group.
C–G. Plasma fat composition of repeatedly stressed mice (black box plots) and nonstressed controls (white box plots); triglyceride
levels (C), total cholesterol (D), HDL-cholesterol (E), VLDL-cholesterol (F), and LDL-cholesterol (G) were measured immediately after
the ninth stress session (n = 12 mice per group). ** p < 0.01; *** p < 0.001, Mann-Whitney U test.
Stress-induced alterations of hepatic gene expression of immune response genes after acute
and chronic stress
The analysis of liver homogenates of acutely and chronically stressed mice revealed an altered
mRNA expression of some immune response genes. Here, canonical pathway analysis of
Ingenuity Pathway Analysis (IPA, www.ingenuity.com) was applied to gain a deeper insight in the
mechanism of stress-induced alterations of immune functions after acute or repeated
psychological stress. Looking at the effects of acute stress, ranking of canonical pathways by IPA
68

LBP [ng/ml]
CPR [ng/ml]
Maren Depke
Results
Liver Gene Expression Pattern in a Mouse Psychological Stress Model
revealed the following five prime targets: “PXR/RXR activation” (p = 4.04E−06), “Glucocorticoid
Receptor Signaling” (p = 4.26E−05), “LPS/IL-1 Mediated Inhibition of RXR function”
(p = 6.44E−05), “IGF-1 Signaling” (p = 7.29E−05), and “Hepatic Fibrosis/Hepatic Stellate Cell
Activation” (p = 2.96E−04). An IPA driven analysis of the consequences of chronic stress exposure
in turn uncovered a prime influence onto the following five canonical pathways, which partially
overlapped with those influenced by acute stress in hepatic tissue: “LPS/IL-1-Mediated Inhibition
of RXR Function” (p = 4.38E−06), “PXR/RXR activation” (p = 5.41E−06), “Acute Phase Response
Signaling” (p = 2.48E−05), “Antigen Presentation Pathway” (p = 1.78E−04), and “Glucocorticoid
Receptor Signaling” (p = 4.24E−04).
Already after acute stress, many genes which are associated with immune activation were
induced (e. g. Egfr, Fgf1, Jun, Irf2, Lgals4, Lipg, Map3k5). Several of these markers such as Il1r1,
Il6ra, Cxcl1, Lpin1, Tnfrsf1b, or Vcam1 were highly expressed in hepatic tissue also after repeated
stress exposure when compared with non-stressed controls. Interestingly, IPA revealed a high
mRNA expression of acute phase response (APR) genes immediately after acute stress exposure
(canonical pathway affected at rank 8 after acute stress: “Acute Phase Response Signaling” with
p = 8.7E−04), which was even further increased after the ninth stress session (Canonical pathway
“Acute phase response”; Table R.1.2). To validate the biological relevance of an increased APR,
LPS-binding protein (LBP) and C-reactive protein (CRP) concentrations were determined in
plasma. Acutely stressed mice did not show any significantly increased concentration of these
APR proteins after 2 h stress exposure whereas LBP and CRP levels were significantly enhanced in
the plasma of chronically stressed mice (Fig. R.1.6).
Fig. R.1.6: Acute phase proteins in the
plasma after acute and chronic stress and
in control mice.
Plasma level of (A) LPS-binding protein
(LBP) and (B) C-reactive protein (CRP) of
acutely stressed mice (grey box plot),
chronically stressed animals (black box
plot), and of non-stressed controls.
n = 12 mice/group;
summarized from two independent
experiments with comparable results.
** p < 0.01, *** p < 0.001 by one-way
ANOVA and Bonferroni multiple
comparison test.
A
750 *** **
500
250
0
B
500
400
300
200
100
0
***
Moreover, gene expression patterns contained several hints for the induction of immune
suppressive pathways which included an up-regulation of Tsc22d3 (GILZ, glucocorticoid-induced
leucine zipper) or Fkbp5 in both acutely and chronically stressed mice and reduced expression of
interferon gamma inducible target genes such as Ifi47, Iigp1, Stat1, and Socs2 selectively after
chronic stress compared with acutely stressed mice and non-stressed controls. Importantly,
chronically stressed mice showed a reduced hepatic transcription of antigen presentation
pathway molecules such as CD74 antigen (Ia antigen-associated invariant chain) and the
histocompatibility class II antigens H2-Aa and H2-Ea. An overlay of the gene expression data to
the pre-defined IPA-canonical pathway “Antigen Presentation Pathway” depicts that chronic
stress-induced repression especially affected genes of the MHC class II signaling pathway
(Fig. R.1.7) which may significantly reduce the capacity to mount an antibacterial response.
69

Maren Depke
Results
Liver Gene Expression Pattern in a Mouse Psychological Stress Model
Table R.1.2: Tabular overview of stress-regulated hepatic genes, which are included in Ingenuity Pathway Analysis’ (IPA, Ingenuity
Systems, www.ingenuity.com) canonical pathway „Acute Phase Response Signaling“.
information on canonical pathway
„Acute Phase Response Signaling“ in IPA
information on stress regulated hepatic genes
cellular
location and
Function
plasma
membrane
receptor
cytoplasm /
signal
transduction
nucleus /
transcription
regulator
affected
genes of
acute phase
response
name in
canonical
pathway
(IPA) a
members
in IPA, if
group of
more than
one gene
gene
name
(NetAffx)
gene title (NetAffx)
probe
set ID
fold change b
in
chronic
stress
in
acute
stress
IL1R1 IL-1R,
Il1r1 interleukin 1 receptor, type I 1448950_at 4.4 4.0
IL-1R/TLR
IL-6R Il6ra interleukin 6 receptor, alpha 1452416_at 3.3 5.5
TNFR NGFR,
TNFRSF1A,
TNFRSF1B,
TNFRSF11B
Tnfrsf1b tumor necrosis factor receptor
superfamily, member 1b
1418099_at 2.4 3.4
ASK1
Map3k5
(LOC
675366)
mitogen activated protein kinase
kinase kinase 5 /// similar to
mitogen activated protein kinase
kinase kinase 5
mitogen activated protein kinase
3
ERK1/2 MAPK1,
MAPK3
Mapk3
GCR* Nr3c1 nuclear receptor subfamily 3,
group C, member 1
IκBα* BCL3,
Nfkbia nuclear factor of kappa light
NFKBIA,
chain gene enhancer in B-cells
NFKBIB,
inhibitor, alpha
NFKBIE
SOCS*
SOCS1,
SOCS2,
SOCS3,
SOCS4,
SOCS5,
SOCS6
Socs2
suppressor of cytokine signaling
2
1421340_at 2.5
1427060_at -1.6
1460303_at -1.5
1448306_at 1.8 2.0
1449109_at -2.4
c-FOS Fos FBJ osteosarcoma oncogene 1423100_at 2.5
c-JUN Jun Jun oncogene 1448694_at
1417409_at
1.7
2.0
GCR* Nr3c1 nuclear receptor subfamily 3, 1460303_at -1.5
group C, member 1
NF-IL6 Cebpb CCAAT/enhancer binding protein
(C/EBP), beta
1418901_at
1427844_a_at
3.3
4.0
5.1
6.9
CRP Crp C-reactive protein, pentraxinrelated
1421946_at 1.6
IκBα* BCL3,
Nfkbia nuclear factor of kappa light 1448306_at 1.8 2.0
NFKBIA,
chain gene enhancer in B-cells
NFKBIB,
inhibitor, alpha
NFKBIE
ORM ORM1,
Orm2 orosomucoid 2 1420438_at 3.3
SAA
SOCS*
ORM2
SAA1,
SAA2,
SAA4
SOCS1,
SOCS2,
SOCS3,
SOCS4,
SOCS5,
SOCS6
Saa1 serum amyloid A 1 1419075_s_at
1450788_at
23.7
27.9
Saa2 serum amyloid A 2 1449326_x_at 30.5
Saa4 serum amyloid A 4 1419318_at
1419319_at
2.6
2.6
Socs2 suppressor of cytokine signaling 1449109_at -2.4
2
a Some genes and their products marked by * belong to more than one group of the first column.
b Fold change values are only given if the gene was considered as regulated in acute or chronic stress experiments.
2.2
2.5
70

Maren Depke
Results
Liver Gene Expression Pattern in a Mouse Psychological Stress Model
Fig. R.1.7: Hepatic repression of antigen
presentation related genes after chronic
stress.
Canonical “Antigen Presentation Pathway” of
Ingenuity Pathway Analysis (adapted from
IPA, Ingenuity Systems, www.ingenuity.com)
displaying genes specifically repressed during
chronic but not during acute stress.
Repression of gene expression is indicated in
green. The following genes are affected: H2-
Aa and H2-Ea (MHC-IIα); Cd74 (CLIP); Psmb9
(LMP2). The blue circle indicates gene
repression for H2-Ab1 (MHC-IIβ), which did
not pass the regulation criteria because of
low expression level.
Increased leukocyte trafficking into the liver of chronically stressed mice
Immediately after acute stress, several markers of leukocyte migration were differentially
regulated in hepatic tissue compared with non-stressed controls. This set included lymphocyte
chemoattractive molecules such as Cxcl11 and Cxl12, which were repressed, or neutrophils
attractive Cxcl1, which was highly expressed after acute but also after chronic stress exposure. An
induction of Vcam1 after acute and chronic stress and of Arhgap5 selectively after repeated
stress as well as the repression of Bcar3 in the liver after acute stress exposure indicated that the
extravasation of leukocytes was facilitated. Additionally, mRNA profiling uncovered a repression
of genes whose products are involved in maintaining the endothelial barrier function such as
claudin 1 (Cldn1), selectively after acute stress, or claudin 2 and 3 after both acute and chronic
stress. To investigate whether these changes in gene expression also caused physiological
changes in leukocyte composition of the liver, immunofluorescence staining was performed
which revealed that no changes of immune cell numbers in the liver were detectable
immediately after a single acute stress session by staining leukocytes with anti-CD45 antibodies
(data not shown). However, in the periportal areas of chronically stressed animals there were
increased numbers of CD45-positive cells, which supports the data of the gene expression
profiling that indicated stress-induced leukocyte migration (data not shown).
Increased oxidative stress in the liver after acute psychological stress exposure is counterregulated
after repeated stress
Inflammatory responses are associated with increased generation of reactive oxygen and
nitrogen species. Gene expression profiling gave hints for increased oxidative stress in the liver of
acutely and chronically stressed mice. Alas1 or As3mt, which are important for the regulation of
the oxidative state, were differentially regulated after acute and chronic stress. The redox
71

Maren Depke
Results
Liver Gene Expression Pattern in a Mouse Psychological Stress Model
molecule Nnmt was up-regulated selectively after repeated stress. To verify the biological
relevance of an altered redox homeostasis, protein carbonylation was analyzed as a marker of
oxidative stress. Remarkably, increased protein carbonyl content of plasmatic and hepatic
proteins was detectable already immediately after acute stress, indicating a rapid and significant
protein damage by increased formation of reactive oxygen species (data not shown). After
chronic psychological stress the protein-carbonyl content had returned to the level seen in nonstressed
control mice (data not shown).
Finally, the expression of Glrx (glutaredoxin) and Gsta (glutathione S-transferase) was
significantly increased in the liver of chronically stressed mice indicating that compensatory antioxidative
mechanisms were induced. This may explain why the carbonyl content of proteins in
plasma and liver in repeatedly stressed mice had declined to the normal level observed in nonstressed
control mice (data not shown).
Repeated stress-induced apoptosis in the liver
Several genes that were regulated in the liver of repeatedly stressed mice provided hints for
increased cell death. By comparing the mRNA profile of liver homogenates of acutely and
repeatedly stressed mice a highly similar pattern of altered gene expression of cell cycle and
apoptosis-related genes was observed. In both acutely and chronically stressed animals,
increased mRNA levels compared to non-stressed controls were detected for Tnfrsf1b, Cdkn1a,
Cebpb, Igfbp1, and Gadd45b (Fig. R.1.8 A). Other genes, such as Fos and Jun, which were induced
immediately after acute stress exposure, did not reveal prolonged high mRNA expression
(Fig. R.1.8 A, B), whereas yet another group, including Ccnd1 and Xbp1, were selectively
repressed after the ninth stress session (Fig. R.1.8 B). Using TUNEL techniques, induction of
hepatocyte apoptosis was not detectable immediately after one single acute stress exposure
(Fig. R.1.8 D), but high numbers of dead cells in livers after repeated stress exposure (Fig. R.1.8 E)
compared with control mice (Fig. R.1.8 C) were documented. However, the liver mass after the
ninth stress session was only slightly decreased compared to non-stressed control animals
(Fig. R.1.8 F). This may be caused by induction of repair processes for which heightened hepatic
expression of IL-6 receptor and Ptp4a1 after chronic stress are indications.
Fig. R.1.8. Differential regulation of cell death associated genes in liver of stressed mice.
Genes resulting from a search for “apoptosis AND liver”, “apoptosis AND hepatocyte”, “cell death AND liver” and “cell death AND
hepatocyte” using Ingenuity Pathway Analysis (IPA, Ingenuity Systems, www.ingenuity.com) were collected in a list. All genes of this
list which displayed differential regulation in liver after acute and/or chronic stress exposure were added to a new user-defined
pathway as analysis nodes and interactions between these selected genes were drawn using the “Connect” tool of IPA. Edge lines that
are continuous represent direct interactions, whereas broken lines indicate indirect influences. The pathway was overlaid with
expression data from (A) acute stress and from (B) chronic stress. Red indicates increased and green decreased expression after stress
exposure compared with the control. Intensity of the node color represents the degree of up (red) and down (green) regulation in the
stressed livers. Different shapes of the nodes indicate the functional classes of the gene products (e. g. vertical
ellipse − transmembrane receptor; horizontal ellipse − transcription regulator; vertical rectangle − G-protein coupled receptor;
horizontal rectangle − ligand-dependent nuclear receptor; trapezium − transporter; triangle − phosphatase; inverted triangle − kinase;
vertical rhomb − enzyme; horizontal rhomb − peptidase; quadrat − cytokine; double lined circle − group or complex; single lined
circle − other).
The level of apoptosis TUNEL assay (x400): compared to the control (C), livers of mice exposed to acute stress did no display any
significant increase in the level of apoptosis (D). However, high numbers of apoptotic hepatocytes were detected in repeatedly
stressed mice (E). Each picture is representative for n = 9 mice/group. Increased apoptosis did not alter liver mass (F), even after the
9th stress cycle (black box plots) compared with non-stressed mice (white box plots) n = 9 mice/group; data representative for least 2
independent experiments.
72

liver weight [g]
Maren Depke
Results
Liver Gene Expression Pattern in a Mouse Psychological Stress Model
A
C
D
B
E
F
1.4
1.2
1
0.8
0.6
73

log cfu/g liver
Maren Depke
Results
Liver Gene Expression Pattern in a Mouse Psychological Stress Model
Reduced antibacterial response in the liver after chronic stress exposure [Cornelia Kiank]
Kiank et al. already showed that chronic psychological stress increased the bacterial spreading
after experimental infection with the extracellular pathogen E. coli ATCC 25922 (Kiank et al.
2006). To elucidate whether altered hepatic immune responsiveness of chronically stressed
animals also alters the antibacterial response to intracellular microbes, BALB/c mice were
intraperitoneally challenged with 150 cfu Salmonella typhimurium. Salmonella infection could not
be restricted at 48 and 72 h after infection when mice were exposed to nine stress sessions prior
to the bacterial challenge (Fig. R.1.9). This shows that although immune effector cells infiltrated
the liver of chronically stressed mice, immune suppression dominated, which resulted in an
insufficient protection against infections.
6
**
***
***
***
5
Fig. R.1.9 Bacterial load of livers after infection with Salmonella
typhimurium.
Colony forming units (cfu) in the liver of chronically stressed (black
box plots) and non-stressed mice (white box plots) 48 and 72 h
after intraperitoneal infection with 150 cfu of S. typhimurium wt
12023 are displayed. (n = 9 mice/group, ** p < 0.01, *** p < 0.001
by Mann-Whitney U-test, data representative for two independent
experiments).
4
3
48 h 72 h
74

infection rate
cfu / 10 mg tissue
cfu / 10 mg tissue
Maren Depke
Results
KIDNEY GENE EXPRESSION PATTERN IN AN
IN VIVO INFECTION MODEL
Infection rate in kidney samples
The two Staphylococcus aureus strains RN1HG and RN1HG ΔsigB were used to infect mice
with an almost equal infection dose for both strains. It was known from other genetic
backgrounds that virulence and bacterial load are often similar for sigB deletion mutants and
their parental strain. Nevertheless, individual mice might still display differences. Therefore, the
infection rate of each sample was tested on molecular basis to identify potentially existing outlier
samples of divergent bacterial load.
Infection rates of individual mice kidney samples were comparable in range, mean, and
median in the biological replicates (BR) samples as well used as in for samples 10 arrays_Exp of infection Feb 09with the two different
strains (Fig. R.2.1). The infection rate of samples and infected exp. Apr 2009 with S. aureus RN1HG ranged from
4.3E+05 cfu/10 mg tissue to 2.2E+06 cfu/10 mg tissue (mean: 1.1E+06; median: 9.2E+05) and of
those infected with RN1HG ΔsigB from 4.7E+05 cfu/10 mg tissue to 2.3E+06 cfu/10 mg tissue
Infection rate cfu / 10 mg tissue
(mean: 1.0E+06; median: 8.6E+05) in both biological [experiments replicates. february and april 2009]
3.0E+06 3.010 6
Fig. R.2.1:
Infection rate in kidney samples of
mice infected with S. aureus RN1HG
and RN1HG ΔsigB.
The infection rate of each sample was
determined in a qPCR approach in
comparison to data from a mixture of
non-infected kidney tissue and in vitro
cultivated staphylococcal cells. The
line indicates the mean of infection
rate for each biological replicate.
BR – biological replicate.
2.0E+06 2.010 6
1.0E+06 1.010 6
infection with
RN1HG
first biological
replicate BR1
wt (2 kidneys)_Feb09
NMRI infection infected with with infection RN1HGwith
RN1HG ΔsigB RN1HG
first biological second biological
replicate BR1 replicate BR2
delta sigB (2 kidneys)_Feb09
wt (2 kidneys)_Apr09
infection with
RN1HG ΔsigB
second biological
replicate BR2
delta sigB (2 kidneys)_Feb09
Reproducibility of replicates and clustering of Exp. treatment Feb. 09; group members Exp. Apr. 09;
used for array analysis (4/7; 5/8) used for array analysis (5/8; 5/8)
Principal Component Analysis (PCA) was applied for a first general impression of the array
data set. This method calculates the direction of strongest variation from the multidimensional
array data set, and reduces it to a new value of the parameter called Principle Component (PC).
The remaining variation in the data set is subsequently addressed in the same way until all or a
pre-defined fraction of variation is collapsed into new values. This procedure results in a set of
PCs, each of which accounts for a fraction of the total variance in the data set. Usually, the first 2
or 3 PCs are displayed in a 2- or 3- dimensional plot, respectively. In such a plot, the distance of
the points that represent the individual data sets correlates to the difference between them.
75

Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
In this study, the PCA plot was derived from log-transformed intensity data of 24 arrays,
analyzing 2 biological replicates of mice infected with S. aureus RN1HG and RN1HG ΔsigB and one
additional group of sham infection (Fig. R.2.2). The analysis was performed on sequence level
(probe sets). Control sequences and sequences that are absent on all 24 arrays (p > 0.01 on
intensity profile level) were not included.
The biological reproducibility is visualized in the PCA plot by the arrangement of
corresponding data points close to each other. The PCA clearly distinguished two groups:
infection with S. aureus and sham infection / NaCl control (Fig. R.2.2 A). Furthermore, the PCA
depicted that the data sets of infection with S. aureus RN1HG and infection with RN1HG ΔsigB
were highly similar (Fig. R.2.2 A−C).
A view from front B view from right side C view from above
Fig. R.2.2: Visualization of transcriptome data using Principal Component Analysis (PCA).
Log-transformed data were used to derive a PCA plot containing 24 arrays (analyzing 2 biological replicates of 2 groups with different
infecting strains and one additional group of sham infection). Resulting principal components (PC) were set to cover 95 % of total
variation, while the total number of principal components was not limited. The analysis was restricted to non-control probe sets that
were not absent on all 24 arrays, when absence of expression is defined via a p-value > 0.01 on intensity profile level in Rosetta
Resolver.
All 3 images (A-C) belong to the same PCA shown in the normal view from the front in the first image (A). In the second image the
PCA-graph is turned by 90° to the left around the axis of principle component 2 resulting in a view from the right side into the plot (B).
Finally, in the third image the view from above into the graph is achieved by turning the plot by 90° around the axis of principle
component 1 (C).
Arrays of the first biological replicate (BR1) are represented by rhombi () and arrays of the second biological replicate (BR2) by
dots (•). Coloring distinguishes the three treatment groups of infection with S. aureus RN1HG (red), infection with S. aureus
RN1HG ΔsigB (blue) and sham infection / NaCl control (green).
Comparison of treatment groups reveals the same host reaction to infection with S. aureus
RN1HG and its sigB-mutant
Already the PCA had shown a high similarity between the reaction to the two different
infecting strains S. aureus RN1HG and RN1HG ΔsigB. Based on the PCA results the strategy for
statistical testing consisted of four main types of comparison (Fig. R.2.3):
comparison of the two groups with different infecting strains (infection with RN1HG ΔsigB
vs. infection with RN1HG) after combining both biological replicates
comparison of the two groups with different infecting strains for each biological replicate
separately
comparison of the same experimental groups of different biological replicates (infection with
RN1HG: BR1 vs. BR2; infection with RN1HG ΔsigB: BR1 vs. BR2)
comparison of infection with S. aureus and sham infection in the second biological replicate
(infection with RN1HG vs. sham infection; infection with RN1HG ΔsigB vs. sham infection)
76

Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
infection with
RN1HG
BR1
(4 arrays)
3
infection with
RN1HG
BR2
(5 arrays)
4
2
1
2
sham infection
(NaCl)
BR2
(5 arrays)
infection with
RN1HG ΔsigB
BR1
(5 arrays)
3
infection with
RN1HG ΔsigB
BR2
(5 arrays)
4
Fig. R.2.3: Overview on the comparisons between groups in this study that were addressed with statistical testing and visualized with
scatter plots.
The two groups with different infecting strains (infection with RN1HG ΔsigB vs. infection with RN1HG) were compared after
combining both biological replicates () and for each biological replicate separately (). Same experimental groups of different
biological replicates (infection with RN1HG: BR1 vs. BR2; infection with RN1HG ΔsigB: BR1 vs. BR2) were also checked against each
other (). The last comparison focused on infection and sham infection in the second biological replicate (infection with RN1HG vs.
sham infection; infection with RN1HG ΔsigB vs. sham infection; ).
BR – biological replicate.
The comparisons were visualized using scatter plots (Fig. R.2.4). Comparable to the PCA
results, a striking concordance between expression values of kidney after infection with RN1HG
and infection with RN1HG ΔsigB was observed, especially when both biological replicates were
combined (Fig. R.2.4, panel ), but also when the biological replicates were considered
separately (Fig. R.2.4, panel ). In the comparison of the same experimental groups in both
replicates high similarity was observed (Fig. R.2.4, panel ), although the inter-replicate variation
of the same treatment was higher than the intra-replicate variation of the different treatments
for S. aureus infected samples. The inter-replicate variation might be due to the difference of one
day in the sampling time point (d 4 or d 5). Strong effects of S. aureus infection independent of
the infecting strain emerged in the comparison to the sham infected / NaCl control sample group
(Fig. R.2.4, panel ).
The different experimental groups were compared with statistical testing to obtain lists of
differentially expressed genes. Sequences not expressed and control sequences were not
included in statistical testing.
The first statistical test compared kidney samples of mice infected with S. aureus RN1HG ΔsigB
vs. infection with RN1HG in an approach that combined both biological replicates (for
comparison see Fig. R.2.3; and Fig. R.2.4, panel ). Only one sequence corresponding to one
gene was significantly different in intensity between both groups (Table R.2.1). For all arrays
except one array from a specific animal the signal intensities of this gene were low and their p-
value for expression was high. i. e. the expression was absent (Fig. R.2.5 A). The array data from
the single outlying animal caused statistical significance, but the result is not biologically relevant
because it is obviously due to an unknown, animal-specific factor and not to treatment.
77

BR2: inf. RN1HG
BR2: inf. RN1HG
BR1: inf. RN1HG ΔsigB
BR1+2: inf. RN1HG ΔsigB
BR2: inf. RN1HG ΔsigB
BR2: inf. RN1HG ΔsigB
BR2: inf. RN1HG ΔsigB
Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
BR1+2: inf. RN1HG
BR1: inf. RN1HG
BR2: inf. RN1HG
BR1: inf. RN1HG
BR1: inf. RN1HG ΔsigB
BR2: sham inf. (NaCl)
BR2: sham inf. (NaCl)
Fig. R.2.4: Scatter plots comparing mean signal intensities of treatment groups.
The signals of the three groups “infection with RN1HG”, “infection with RN1HG ΔsigB”, and “sham infection / NaCl control” are plotted
separately for biological replicates (BR1, BR2) or combined for both biological replicates (BR1+2). Control sequences and sequences
that were absent on all of the arrays used for each scatter plot are not shown. (Absence of expression is defined via a
p-value > 0.01 on intensity profile level in Rosetta Resolver). Numbers in the first column refer to the comparison scheme in Fig. R.2.3.
78

Table R.2.1: Genes displaying statistically different expression values in the corresponding comparisons.
a
The GeneChip Mouse Gene 1.0 ST array contains 35557 sequences (probe sets) of which 6613 are controls (e. g. negative, positive, PolyA, and hybridization controls). The control sequences and sequences that
were not expressed on all arrays in the selected group comparison were not included in statistical testing.
Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
b Rosetta Resolver software maps the sequences (probe sets) to 20074 EntrezGene records (genes). This mapping is used to calculate the numbers of genes that are significant in statistical testing and when indicated
group comparison
number
of
arrays
number of
sequences
absent on
all arrays
significant with p*

Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
In the search for specific differences in infection with S. aureus RN1HG ΔsigB vs. infection with
RN1HG, the same comparison was performed for the biological replicates separately (for
comparison see Fig. R.2.3; and Fig. R.2.4, panel ). This analysis resulted in 7 or 6
sequences/genes which were differentially expressed in the first biological replicate BR1 or the
second biological replicate BR2, respectively (Table R.2.1). Both lists did not display any overlap
indicating that the few regulated genes were not reproducible in the biological replicates,
although statistically significant (Fig. R.2.5 B).
Additionally, the statistical significance in the biological replicate BR1 was again caused by one
single sample, i. e. by one specific animal, which provided an outlier value to the data set
(Fig. R.2.5 C). The same sample accounted for the outlier in the values of the single gene
significantly regulated when combining the biological replicates (Fig. R.2.5 A).
The statistical significance of the 6 differentially expressed genes in the second biological
replicate was not caused by outliers (Fig. R.2.5 D), but the fold change in this comparison was
only moderate ranging from to -1.34 to -2.18 (mean: -1.72; median: -1.67).
In conclusion, the study could not provide any hints for statistically significant differences in
the expression pattern in murine kidney upon infection with S. aureus RN1HG or its isogenic sigB
mutant.
When asking for the reproducibility of the same treatment in both biological replicates in the
light of a possible influence on the detection of differential gene expression between the groups
infected with the two different S. aureus strains (for comparison see Fig. R.2.3; and Fig. R.2.4,
panel ), 33 sequences (i. e. 26 genes) were significantly different between the replicates of
infection with RN1HG, and 62 sequences (i. e. 31 genes) significantly differed between the
replicates of infection with RN1HG ΔsigB (Table R.2.1).
Of these genes, only a small fraction of approximately 20 % possessed an absolute fold change
greater than 2. For these few genes a high expression variation within one biological replicate
was observed or the statistical difference even was again due to one outlier array. Additionally,
most genes which differed between the replicates did not display any overlap with the genes
which were significantly different between the groups infected with the two different S. aureus
strains. Only some genes were statistically differentially expressed between the equally treated
replicates as well as between infection with RN1HG and RN1HG ΔsigB in the first biological
replicate. But as the statistical significance between samples infected with the two strains was
only caused by one single outlier value (Fig. R.2.5 C), it was clear that the small difference
between biological replicates did not prevent the identification of differential gene expression in
the comparison of host reaction to the two infecting S. aureus strains.
In summary, the comparison of same treatment groups in different biological replicates
revealed minor, negligible differences that did not influence the detection of differential
regulation when comparing the groups of infection with different S. aureus strains. The identical
reaction to infection with S. aureus RN1HG and S. aureus RN1HG ΔsigB was reliably measured in
the array study and the detection of differences was not prevented by biological variation
between the two independent infection experiments.
80

Maren Depke
BR1 sign DsigB vs wt
BR1 sign DsigB vs wt
A
DsigB
Igk-V28
wt
infection with RN1HG ΔsigB vs. RN1HG;
BR1 and und BR2 BR2 combined DsigB vs wt
Igk-V28_DsigB
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
B
BR1 sign DsigB vs wt
BR1 sign DsigB vs wt
7 0 6
C
Igk-V28_DsigB
1 10 100 1000 10000
Igk-V28_wt
signal intensity
infection with S. aureus RN1HG ΔsigB
infection with S. aureus RN1HG wt
Igk-V28_wt
Cyp11b1_DsigB
D
sequences/genes differentially
expressed between
infection with RN1HG ΔsigB
and infection with RN1HG
in BR1
sequences/genes differentially
expressed between
infection with RN1HG ΔsigB
and infection with RN1HG
in BR2
Cyp11b1_DsigB
infection BR1 with sign RN1HG DsigB ΔsigB vs wt vs. RN1HG; BR1
Cyp11b1_wt
infection with S. aureus RN1HG ΔsigB
infection with S. aureus RN1HG wt
Igk-V28_DsigB
Igk-V28
Cyp11b1_wt
Igk-V28_wt
4921521F21Rik_DsigB
infectionBR2 withsign RN1HG DsigB ΔsigBvs vs. wt RN1HG; BR2
Cyp11b1_DsigB
Cyp11b1
4921521F21Rik_DsigB
Cyp11b1_wt
4921521F21Rik_wt
Gpr84_DsigB
Gpr84
Gpr84_wt
4921521F21Rik_DsigB
4921521F21Rik
4921521F21Rik_wt
4921521F21Rik_wt
Star_DsigB
Cd274_DsigB
Cd274
Cd274_wt
Star_DsigB
Star
Star_wt
Star_DsigB
Star_wt
Cxcl9_DsigB
Cxcl9
Cxcl9_wt
Hsd3b1_DsigB
Cybb_DsigB
Hsd3b1
Hsd3b1_wt
Star_wt
Hsd3b1_DsigB
Cybb
Cybb_wt
Cyp21a1_DsigB
Cyp21a1
Cyp21a1_wt
Hsd3b1_DsigB
Hsd3b1_wt
C3ar1_DsigB
C3ar1
C3ar1_wt
Cyp11a1_DsigB
Cyp11a1
Cyp11a1_wt
Hsd3b1_wt
Cyp21a1_DsigB
Plekho2_DsigB
Plekho2
Plekho2_wt
1 10 100 1000 10000
signal intensity
Cyp21a1_DsigB
Fig. R.2.5: Signal intensities Cyp21a1_wt and list comparison for differentially expressed genes.
A. Comparison of murine kidney expression profiles after infection with RN1HG ΔsigB and infection with RN1HG with statistical testing
results in one differentially expressed gene when combining both available biological replicates. However, the signal intensity is
extremely low (not expressed with p > 0.01 on intensity profile level in Rosetta Resolver software) in all arrays except in one array
(i. e. a kidney sample) Cyp21a1_wt from one mouse infected with RN1HG (BR1) which is visible as outlier and which causes the statistical
significance. Cyp11a1_DsigB
B. Comparison of genes differentially regulated between infection with RN1HG and its isogenic sigB mutant in BR1 and BR2. Both
10 biological replicates result 100in completely different 1000 lists of differentially expressed 10000 genes.
C, D. Signal intensity values of differentially expressed genes in the comparison of infection with RN1HG ΔsigB and infection with
RN1HG in BR1 Cyp11a1_DsigB
(C) and in BR2 (D). In BR1, statistical significance is caused by a single array that includes an outlier value.
Cyp11a1_wt
BR – biological replicate.
10 100 1000 10000
81
1 10 100 1000 10000
signal intensity
Cyp11a1_wt1 10 100 1000 10000
1 10 100 1000 10000

log 2 ratio (inf. RN1HG ΔsigB / sham inf.)
Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
Comparative analysis of S. aureus infected samples with sham infection identifies strong
infection/inflammation reactions of the kidney tissue
As high reproducibility of kidney expression profiles in biological replicates of infection with
S. aureus had been shown, the comparison between S. aureus infected samples and sham
infection / NaCl control was conducted with only one representative biological replicate (for
comparison see Fig. R.2.3; and Fig. R.2.4, panel ).
When comparing sham infection with infection with S. aureus RN1HG, 5264 sequences were
differentially expressed which correspond to 4613 genes (EntrezGene records in Rosetta Resolver
software). Of these genes, 1083 possessed an absolute fold change of at least 2. A similar
difference was observed in the comparison of sham infection with infection with S. aureus
RN1HG ΔsigB: 4826 sequences (4248 genes) differed significantly of which 970 genes exhibited
an absolute fold difference of at least 2 (Table R.2.1).
When calculating ratios of highly similar sample data sets to a common baseline, a result of
conforming values is expected. The log-ratio plot of the two ratios “infection with
RN1HG” / “sham infection” and “infection with RN1HG ΔsigB” / “sham infection” in the second
biological replicate BR2 on EntrezGene level visualizes this similarity of the ratio data: Data points
are mainly arranged near the diagonal of the plot (Fig. R.2.6). Some data points display more
scattering, but the difference in the ratios was not significant when tested statistically and was
mainly due to higher variation in signal intensities of one group, as described above.
6
Fig. R.2.6:
Ratio plot comparing the log 2ratio of “infection with
RN1HG” / “sham infection” with the log 2ratio of
“infection with RN1HG ΔsigB” / “sham infection” in the
second biological replicate BR2 on EntrezGene level.
Genes differentially regulated (p* < 0.01 in errorweighted
one-way ANOVA with Benjamini-Hochberg
False Discovery Rate multiple testing correction in the
Rosetta Resolver software) in both comparisons are
colored in light gray, genes specifically significant in the
comparison “infection with RN1HG” / “sham infection”
are depicted in black while genes specifically significant
in the comparison “infection with RN1HG
ΔsigB” / “sham infection” are shown in dark gray. In
total, 5025 genes are visible.
BR – biological replicate.
4
2
0
-2
-4
-6
-6 -4 -2 0 2 4 6
log 2 ratio (inf. RN1HG / sham inf.)
The genes differentially regulated in S. aureus infected kidney tissue compared to sham
infection were further analyzed using the Ingenuity Pathway Analysis tool. Only genes displaying
an at least twofold difference in expression between sham infected and S. aureus infected
animals (either RN1HG or RN1HG ΔsigB from the second biological replicate BR2) were
considered in this analysis. Using this approach, the list of genes eligible for analysis was
restricted to 1156 genes of which 4 could not be mapped to internal data base entries (Ingenuity
Pathway Knowledge Base, IPKB) by IPA. Since expression values of experiments involving
infection with S. aureus RN1HG or its isogenic sigB mutant did not differ significantly, the average
of both expression values was compared to sham infected animals, which allows intuitive
visualization of the data.
The global functional analysis using IPA offers an overview on the biological functions that are
associated with the analyzed data set. As expected from the experimental setting, influence on
the functional category “Inflammatory Response” was most pronounced with p-values of
2.36E−75 to 1.24E−10 for the sub-categories and with 343 associated genes differentially
expressed between infection and sham infection. Additionally, other infection, immune response
82

Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
and inflammation related categories (e. g. Inflammatory / Immunological / Infectious Disease,
Antigen Presentation) and categories showing the impact of infection on the tissue (e. g. Cellular
Compromise, Cell Death) were significantly affected.
More detailed analysis of the tissue reaction is offered by the so-called Canonical Pathway
analysis in IPA. This tool contains predefined pathways and the associated genes, displayed in
order of cellular location and function, which can be overlaid with gene expression data, e. g. fold
change values. A p-value for enrichment of pathway members in the data set of interest
(compared to a reference set, e. g. the whole array) is calculated. Many pathways scored with a
highly significant p-value because of the large input data set. Additionally, a ratio of the
pathway’s genes included in the data set and the total number of genes in the pathway is given.
When data of such a Canonical Pathway analysis were ordered by p-value, “Acute Phase
Response Signaling” with p = 5.00E−21 and a ratio of 53/178 of genes from the data set included
in the pathway was the most significantly influenced reaction. This pathway includes signaling
chains starting with TNF-α, IL-1 and IL-6, their signal transduction molecules and transcription
factors and finally the genes whose transcription is induced or repressed. While the acute phase
response is located in hepatic tissue the pathway has a high overlap with a localized
infection/inflammation reaction and therefore is covered by the kidney infection vs. sham
infection data set to this high extent.
Directly infection related genes are included in the canonical pathways “Role of Pattern
Recognition Receptors in Recognition of Bacteria and Viruses” (p = 1.16E−13; ratio = 27/80,
Table R.2.2) and “Toll-like Receptor Signaling” (p = 7.72E−08; ratio = 16/54, Fig. R.2.7) which
contribute as part of the innate immune system to the first line of defense against pathogens. In
infected kidney tissue, an induction of different Toll-like receptors (Tlr1, Tlr2, Tlr4, Tlr6, Tlr7, Tlr8),
the bacterial or dsRNA receptor NALP3 (Nlrp3), and the beta-glucan receptor Dectin-1 (Clec7a)
was observed, and a tendency of induction in Tlr9 and the dsRNA receptors RIG-1 (Ddx58) and
Ifih1. Functionally associated receptors CD14 and LBP were also induced and, following infection,
even members of the signal transduction cascade showed an increased expression (Syk, Myd88,
Irak3, Map3k1, Irf7, Pik3cd, Pik3cg, Pik3r5, Casp1, Fos, Jun, Nfkb2, and a tendency of increase in
Irak4, Mapk13, Map4k4 and Nfkb1). Contrarily, the MAP-kinase-pathway component Map2k6
was repressed and the signal transducer Ecsit and transcription factor PPARα (Ppara) were
repressed by trend. The signaling ends in the transcription of pro-inflammatory cytokines, e. g.
TNFα, IL-6, and RANTES (Ccl5) whose induction was recorded in infected tissue.
Fig. R.2.7:
Toll-like Receptor Signaling (modified
from IPA, www.ingenuity.com).
Colors indicate direction and magnitude
of differential regulation: red – induction;
green – repression. More intense shade of
color is used for higher absolute fold
change values. A trend of induction is
shown in yellow, a trend of repression in
light blue.
83

Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
Table R.2.2: Tabular overview of infection-regulated genes which are included in Ingenuity Pathway Analysis (IPA, Ingenuity Systems,
www.ingenuity.com) canonical pathway “Role of Pattern Recognition Receptors in recognition of Bacteria and Viruses” (modified
from IPA, www.ingenuity.com).
IPA
Rosetta Resolver annotation
fold change
infection vs.
sham infection
localization
and functional
grouping a
gene
name
gene
name
description
EntrezGene
ID
RN1HG
RN1HG
ΔsigB
extracellular PRR C1q C1qa complement component 1, q
12259 5.5 4.5
subcomponent, alpha polypeptide
C1qb complement component 1, q
12260 7.4 6.0
subcomponent, beta polypeptide
C1qc complement component 1, q
12262 6.9 5.4
subcomponent, C chain
C3 C3 complement component 3 12266 5.1 5.1
membranebound
C3aR C3ar complement component 3a receptor 1 12267 10.3 6.7
PRR C5aR C5ar complement component 5a receptor 1 12273 7.6 6.1
TLR1 Tlr1 toll-like receptor 1 21897 2.9 2.3
TLR2 Tlr2 toll-like receptor 2 24088 4.4 4.0
TLR4 Tlr4 toll-like receptor 4 21898 2.4 2.3
TLR6 Tlr6 toll-like receptor 6 21899 2.4 2.0
TLR7 Tlr7 toll-like receptor 7 170743 3.0 2.8
TLR8 Tlr8 toll-like receptor 8 170744 4.7 4.3
TLR9 Tlr9 toll-like receptor 9 81897 1.9 1.9
DECTIN-1 Clec7a C-type lectin domain family 7, member a 56644 3.7 2.8
cytoplasmic PRR NALP3 Nlrp3 NLR family, pyrin domain containing 3 216799 3.3 2.7
RIG-1 Ddx58 DEAD (Asp-Glu-Ala-Asp) box polypeptide 58 230073 1.7 1.6
MDA-5 Ifih1 interferon induced with helicase C domain 1 71586 1.5 1.3
signal
PI3K Pik3cd phosphatidylinositol 3-kinase catalytic delta 18707 2.5 2.0
transduction,
effector
Pik3cg
polypeptide
phosphoinositide-3-kinase, catalytic, gamma 30955 3.4 3.1
molecules
(cytoplasm)
Pik3r5
polypeptide
phosphoinositide-3-kinase, regulatory
320207 2.4 2.0
subunit 5, p101
MYD88 Myd88 myeloid differentiation primary response
17874 2.8 2.5
gene 88
SYK Syk spleen tyrosine kinase 20963 3.6 2.8
CASP1 Casp1 caspase 1 12362 3.4 2.2
IL-1β Il1b interleukin 1 beta 16176 29.0 21.5
OAS Oas1g 2'-5' oligoadenylate synthetase 1G 23960 2.2 1.7
Oas2 2'-5' oligoadenylate synthetase 2 246728 1.9 1.5
Oas3 2'-5' oligoadenylate synthetase 3 246727 1.7 1.4
RNaseL Rnasel ribonuclease L (2', 5'-oligoisoadenylate
24014 1.6 1.4
synthetase-dependent)
transcription IRF-7 Irf7 interferon regulatory factor 7 54123 2.7 1.9
factors (nucleus) NFκB Nfkb1 nuclear factor of kappa light polypeptide
18033 2.0 1.9
gene enhancer in B-cells 1, p105
Nfkb2 nuclear factor of kappa light polypeptide
18034 2.7 2.5
gene enhancer in B-cells 2, p49/p100
transcriptionally TNFα Tnf tumor necrosis factor 21926 3.0 2.3
affected genes IL-6 Il6 interleukin 6 16193 9.2 7.8
RANTES Ccl5 chemokine (C-C motif) ligand 5 20304 6.9 4.9
a PRR – pattern recognition receptor
Pattern recognition includes the complement system. The canonical pathway “Complement
System” was significantly affected after infection (p = 6.86E−10; ratio = 16/36, Fig. R.2.8) which
includes all three activation variants of classical, lectin and alternative pathway: The components
C1q (C1qa, C1qb, C1qc), C1r (C1r, C1rb), C3, C4 (C4b), C7, Factor D / Adipsin (Cfd), Factor
P / Properdin (Cfp), the complement-activating protease Masp1, and the receptors for C3a and
C5a (C3ar, C5ar) were induced. A tendency of increased expression was observed for factors C2,
C6, and Factor B (Cfb). But higher expression was also visible for the C1 inhibitor (Serping1), for
the complement-inhibitory factor I (Cfi), and by trend for the inhibitory factor H (Cfh), probably as
a reaction to restrict complement activation to a physiologically tolerable extent as the
84

Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
complement system can cause immense damage also to host cells and tissue. The complement
component C8 consists of three chains (alpha, beta, and gamma) and is the last component in the
C5b-C6-C7-C8 complex that acts as a catalyst in the polymerization of C9 to form the membrane
attack complex MAC. Surprisingly, the gene C8a coding for the alpha chain was repressed in the
infected samples, and also for the gene C8g (gamma chain) a tendency of repression was
recorded. Possibly the opsonization and chemoattractant function of the complement system
was enhanced in infection leading to phagocytosis, while the bacteriolytic function of the
complement system was not potentiated. Otherwise, the components C8 and C9 – produced in
the liver – might be delivered via the blood stream to the site of infection.
Fig. R.2.8:
Complement System
(modified from IPA, www.ingenuity.com).
Colors indicate direction and magnitude of
differential regulation: red – induction; green –
repression. More intense shade of color is used for
higher absolute fold change values. A trend of
induction is shown in yellow. Factor P / Properdin
(CFP) has been inserted manually into the pathway.
As already mentioned, the Toll-like receptor (TLR) signaling cascades lead to the transcription
of different proinflammatory cytokines which themselves activate signaling pathways and the
adequate cellular responses. When using the Canonical Pathways to specifically find infection
relevant signal transduction, the signaling pathways of interferon, IL-6, TREM1, and IL-10
appeared among others with significant p-values.
The pathway “Interferon Signaling” (p = 1.14E−06; ratio = 11/30; Fig. R.2.9) contains the
induced IFN-γ receptor (Ifngr1, Ifngr2) and the IFNα/β-receptor, of which the gene for one of the
two chains (Ifnar2) was induced, too. In the receptor’s signal transduction, the genes for signal
transducers STAT 1 and 2 were induced together with the transcriptionally regulated genes Tap1,
Ifitm1, Irf1, Psmb8, Irf9 and Oas1. Additionally, the inhibitory protein tyrosine phosphatase TC-
PTP (Ptpn2) displayed a trend of induction, probably indicating a counter-regulatory process.
85

Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
Fig. R.2.9:
Interferon Signaling (modified
from IPA, www.ingenuity.com).
Red color indicates increase of
expression. More intense shade
of color is used for higher
absolute fold change values. A
trend of induction is shown in
yellow.
Similarly, IL-6 receptor (Il6ra), IL-1 receptor (Il1r1, Il1r2) and TNF-α receptor (Tnfrsf1a,
Tnfrsf1b) and their ligands (Il6, Il1a, Il1b, Tnf) were expressed at higher levels in infected tissue
than in control tissue as depicted in the pathway “IL-6 Signaling” (p = 6.88E−11; ratio = 27/93;
Fig. R.2.10). Glycoprotein 130 (Il6st), a signal transducer shared by IL-6 and other cytokines,
exhibited at least a trend of induction. Again, genes for signal transduction molecules and
transcription factors were significantly (Ikbke, Nfkbia, Nfkbia, Nfkbiz, Nfkb2, Stat3, Jun, Fos,
Cebpb) or slightly (Nfkbid, Nfkb1, Traf2, Map4k4, Mapk13, Nras) increased. Signal transduction
was activated in infected tissue, which was demonstrated by the induction of transcriptionally
regulated genes like TSG6 (Tnfaip6), Collagen (Col1a1), alpha-2-macroglobulin (A2m), and IL-6
itself.
Fig. R.2.10: IL-6 Signaling (modified from IPA, www.ingenuity.com).
Direction and magnitude of differential expression are indicated by colors: red – induction; green – repression. More intense shade of
color is used for higher absolute fold change values. A trend of induction is shown in yellow.
86

Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
The canonical pathway “TREM1 Signaling” (p = 8.15E−13; ratio = 23/69) slightly overlaps with
the pathway “IL-6 Signaling” as branches of signal transduction also lead to IL-6 and TNFα
transcription via STAT3 and NFκB (all induced after infection). Also TLR4 is included because after
activation, the receptor can associate with TREM1, activate the kinase IRAK1 and finally the
NFκB-mediated transcription. The increased Trem1 itself, a receptor whose ligand is still
unknown and which is expressed by neutrophils, monocytes and macrophages, and the increased
adapter molecule DAP12 (Tyrobp) are the beginning of a proinflammatory reaction chain that
includes AKT/PKB (Rac2, induced), PLCγ and NTAL (Plcg2 and Lat2, induced by trend). TREM1
signaling not only influences transcription of proinflammatory cytokines (MCP-1/Ccl2, TNFα,
MCP-3/Ccl7, IL-6, MIP-1α/Ccl3), but also the translocation of cytokines from cytoplasm to the
extracellular space (MCP-1/Ccl2, TNFα, IL-6, IL-1β). DAP12 positively influences the activation of
IL-1β by CASP1 (both induced) in the pathway of NLR (intracellular NACHT-LRR receptors)
recognizing intracellular bacteria.
Furthermore, DAP12 leads to the transcription of cell adhesion molecules (CD11c/Itgax,
CD49e/Itga5) and the co-stimulatory proteins CD54/Icam1 and CD86 (all induced). Genes coding
for cell surface proteins CD32/Fcgr2b and CD40, which are additionally regulated via DAP12, were
increased by trend.
Opposed to these proinflammatory signaling pathways, the “IL-10 Signaling” pathway
(p = 1.99E−15; ratio = 27/70; Fig. R.2.11) is an example for a mechanism aiming to limit and
terminate the cellular inflammatory response in addition to regulating growth and differentiation
of different immune cells. Here, the IL-10 receptor α-chain (Il10ra) was induced after infection.
Induced Stat3 is part of the signaling cascade like in IL-6 signaling but SOCS3, suppressor of
cytokine signaling, was additionally increased. This gene is itself transcriptionally regulated at the
end of the IL-10 signaling pathway. Furthermore, Ccr1, Ccr5, Arg2, Il4ra, and Hmox1 were
induced genes responding to IL-10. IL-6 signals are also mediated via STAT3, and the cytokine IL-6
was induced contrarily to IL-10, which was not regulated. Therefore, the IL-10 pathway might not
be activated yet, but just prepared for a later phase of anti-inflammatory reactions.
Fig. R.2.11:
IL-10 Signaling (modified from IPA, www.ingenuity.com).
Red color indicates increase of expression. More intense shade of color is
used for higher absolute fold change values.
87

Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
An essential part of the immune response to infection is the presentation of antigens to
immune cells to direct the defense against the specific pathogen in the adaptive part of the
immune system. Antigens belong to two main groups, extracellular and intracellular antigens,
which are processed in separate pathways. Extracellular antigens are taken up by
phagocytosis/endocytosis and digested in vesicles after fusion with lysosomes. Finally, peptides
are bound to MHC-II molecules originating from exocytic Golgi vesicles and presented on the cell
surface. Intracellular antigens are digested by cytosolic proteasome with sequential transport of
peptides into the endoplasmatic reticulum (ER). These peptides are loaded to MHC-I molecules
and transported via the Golgi to the cell surface for presentation. Presentation of extracellular
antigens preferentially occurs by professional antigen presenting cells (APC) like dendritic cells
(DC), macrophages and B cells. The peptide-MHC-II complex is recognized by CD4 + helper T cells,
which leads to initiation of mechanisms fighting extracellular infection or disposing extracellular
antigens. Immune cells and other cells of the body are capable of presenting intracellular
antigens via MHC-I. They present the antigenic peptide to CD8 + cytotoxic T cells and thereby
activate processes to destroy cells and kill pathogen in case of intracellular infection.
For both pathways an increase in transcription of associated genes was observed after
infection (Fig. R.2.12). The immune-proteasome genes Psmb9, Psmb10, and Psmb8 (LMP2,
LMP7), coding for proteins in the 20S-core of the proteasome, were induced leading to an
immune-response specific reorganization of the proteasomes’ catalytic center. For the genes
Psme1 and Psme2, which contribute to the 11S regulatory complex of the proteasome, a trend of
induction was visible. Induced were also the peptide transporters Tap1 and Tap2 and the TAPbinding
protein tapasin/TPN (Tapbp), which attaches the TAP molecules to the MHC-I complex.
The ER aminopeptidase Erap1, which is responsible for N-terminal trimming of the peptide in the
ER, was induced in tendency. Finally, MHC-I molecules (H2-D1, H2-K1, H2-Q6, H2-Q7, H2-Q8) and
Beta-2-microglobulin (MHC-I-β; B2m) were induced, too.
In the pathway for presentation of extracellular antigens MHC-II molecules (H2-Aa, H2-Ab1,
H2-DMa, H2-DMb1, H2-Ea, H2-Eb1) and the invariant chain Cd74 (CLIP) were increased. Induction
of lysosomal enzymes, particularly of cathepsins, was also observed (Chi3l1, Chi3l3, Hpse, Ctsc,
Ctse, Ctss; induced by trend: Gla, Ctsd, Ctsk, Ctsl).
Fig. R.2.12: Antigen Presentation (modified from IPA, www.ingenuity.com). Red color indicates increase of expression.
88

Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
Metabolic effects of infection in kidney tissue
In an approach to recognize changes in gene expression of metabolic enzymes after infection
lists of differentially regulated genes were subjected to a BIOCYC “omics-Viewer” analysis (BIOCYC,
SRI International, CA, USA, http://biocyc.org/expression.html). This tool allows the display of
gene expression data on highly abstracted metabolic pathway schemes and therefore an intuitive
comprehension of processes in the selected experimental setup. Two different lists were
included in the analysis: a) genes significantly different between infection and sham infection
with a minimal absolute fold change of 2 in at least one of the two comparisons “infection with
RN1HG vs. sham infection” and “infection with RN1HG ΔsigB vs. sham infection” and b) genes
significantly different between infection and sham infection with a minimal absolute fold change
of 1.5 in at least one of the two comparisons described before.
When focusing on regulation that exceeds a factor of 2 in at least one comparison, regulation
was discernible for certain groups of metabolic pathways (Fig. R.2.13 A). After infection, a
repression of genes relevant for cholesterol biosynthesis was visible, and also pathways of amino
acid degradation contained repressed genes. Induction of gene expression was detected for
steroid hormone biosynthesis and for purine degradation. This is surprising, as induction was also
visible for steps of the purine/pyrimidine biosynthesis. Similarly, an increase in gene expression
was measured for certain amino acid biosynthesis steps.
Including differentially expressed genes with lower absolute fold change values (1.5 in at least
one of two comparisons) supplemented the data set with many repressed and only few induced
genes, which is visualized by the over-representation of yellow colored reaction steps
(Fig. R.2.13 B). In this view, repression in further steps in the group of lipid biosynthesis and
amino acid degradation was visible. Additional repression occurred in aerobic respiration, TCA
cycle, glycolysis, but also in gluconeogenesis and glycogen biosynthesis. The pattern of increased
genes dominating both purine degradation as well as biosynthesis was maintained after inclusion
of regulation with lower absolute fold change values.
As the gene expression analysis was performed for tissue samples, i. e. of a mixture of
different kidney tissue cell types and in case of infection additionally of a mixture of different
immune cell types invading the inflamed tissue, the resulting pattern is difficult to match for
particular cell types. While one cell type might induce purine degradation e. g. for using it as
catabolic substrate, another cell type might induce purine biosynthesis e. g. as basis for DNA
replication during proliferation. Such situation could lead to contradictory results concerning
metabolic pathways if the number of cells influencing the pattern in a certain direction is high
enough.
The general gene expression pattern showed a metabolic disturbance in several pathways
that cannot be distinguished into anabolic or catabolic shift because it contained reduction of
several catabolic pathways like glycolysis, amino acid degradation, aerobic respiration, and TCA
cycle, but also reduction in anabolic reactions like gluconeogenesis or amino acid biosynthesis.
The reason of this unclear or mixed reaction might be found in the high extent of infection and
illness. Possibly the infection has caused such a strong damage to the kidney after 5 days that
also organ function, nutrition supply and possibly oxygen availability are impaired.
89

Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
A
co-factor biosynthesis
cholesterol,
triacylglycerol,
phospholipid
biosynthesis
aerobic
respiration
amino acid degradation
amino acid
biosynthesis
other
degradation
pathways
purine/
pyrimidine
biosynthesis
sugar
degradation
gluconeogenesis steroid
and glycogen hormone
biosynthesis biosynthesis
glycolysis
prostaglandin
biosynthesis
TCA
purine
degradation
fatty acid oxidation,
triacylglycerol,
phospholipid
degradation
B
co-factor biosynthesis
cholesterol,
triacylglycerol,
phospholipid
biosynthesis
aerobic
respiration
amino acid degradation
amino acid
biosynthesis
other
degradation
pathways
purine/
pyrimidine
biosynthesis
sugar
degradation
gluconeogenesis steroid
and glycogen hormone
biosynthesis biosynthesis
glycolysis
prostaglandin
biosynthesis
TCA
purine
degradation
fatty acid oxidation,
triacylglycerol,
phospholipid
degradation
Fig. R.2.13: The influence of infection on gene expression in murine metabolic pathways (modified from omics-viewer(s) of BIOCYC,
SRI International, CA, USA, http://biocyc.org/expression.html).
Nodes represent metabolites and lines indicate reactions. The metabolic reactions are colored according to the enzymes’ gene
expression regulation in the comparison of infection vs. sham infection. Red marks an increase and yellow a decrease in infected
tissue. The display is limited to genes whose absolute fold change exceeds 2 in at least one of the two comparisons “infection with
RN1HG vs. sham infection” and “infection with RN1HG ΔsigB vs. sham infection” (A) or to those whose absolute fold change exceeds
1.5 in at least one of the two comparisons mentioned before (B).
90

Maren Depke
Results
GENE EXPRESSION PATTERN OF BONE-MARROW DERIVED
MACROPHAGES AFTER INTERFERON-GAMMA TREATMENT
Experimental application of serum-free differentiation and cultivation of mouse BMM
Bone-marrow derived macrophages (BMM) allow experimental analysis of macrophage
reactions and functions without the impact of immunological conditioning that affects already
differentiated macrophages for example from spleen or peritoneum. The standardization of
experiments was markedly improved in 2009 when Eske et al. published a serum-free culture
system for BMM that allows conduction of studies independently from the influence of
undefined immune-active substances and possible variation introduced to experiments by using
fetal calf serum (FCS) as supplement. BMM cultivated in this new serum-free medium maintain
the already known characteristics of macrophages and have proven to lead to highly reproducible
results (Eske et al. 2009).
The experimental setup in this study included the two mouse strains BALB/c and C57BL/6.
After in vitro differentiation, BMM of each strain were divided in two treatment groups: IFN-γ
treated BMM and non-treated medium control BMM.
The approach of transcriptome analysis using the Affymetrix GeneChip Mouse Gene 1.0 ST
allowed monitoring of 35557 probe sets, of which 6613 belong to the group of controls. The
remaining 28944 probe sets represent 20074 EntrezGene records in the annotations of Rosetta
Resolver software (annotation of 06/2009). The LC-MS/MS proteome analysis provided
information on 946 reliably identified proteins (Dinh Hoang Dang Khoa).
High reproducibility of gene expression in experiments using serum-free differentiation and
cultivation of mouse BMM
For a first general impression of the array data set, the method of Principal Component
Analysis (PCA) was applied. This method calculates the direction of strongest variation from the
multidimensional array data set, and reduces it to a new value of the parameter called Principle
Component (PC). The remaining variation in the data set is subsequently addressed in the same
way until all or a pre-defined fraction of variation is collapsed into new values. This procedure
results in a set of PCs, of which each accounts for a fraction of the total variance in the data set.
Usually, the first 2 or 3 PCs are displayed in a 2- or 3-dimensional coordinate system, respectively.
In such a plot, the distance of the points that represent the individual data sets correlates to the
difference between them.
In this study, the PCA plot was derived from log-transformed intensity data of 12 arrays,
analyzing 3 biological replicates of 2 strains and 2 treatment groups (Fig. R.3.1). The PCA clearly
depicted the high reproducibility of biological replicates obtained from the serum-free model of
BMM differentiation, cultivation and IFN-γ treatment: The 3 biological replicates of each group
were arranged together, while the BMM of different strains and treatments were separated.
Differences on the axis for PC 1 were by a factor of approximately 250 smaller than differences
on the axes for PCs 2 and 3. That implied that PCs 2 and 3 covered the main part of variance in
the data set. This led to the conclusion that the data set was mainly influenced by 2 factors. In an
experimental design of the 2 factors strain and treatment together with the observed grouping of
biological replicates obviously the experimental factors were responsible for the observed
variation.
91

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
Fig. R.3.1:
Transcriptome data visualization using Principal
Component Analysis (PCA).
Log-intensity data were used to derive a PCA
plot containing 12 arrays (analyzing 3 biological
replicates of 2 strains and 2 treatment groups).
Resulting principal components were set to
cover 95 % of total variation, while the total
number of principal components was not
limited. The analysis was restricted to noncontrol
probe sets that were not absent on all
12 arrays, when absence of expression is
defined via a p-value > 0.01 on profile level in
Rosetta Resolver. Differences on the axis for
principal component 1 are by a factor of
approximately 250 smaller than differences on
the axis for principal components 2 and 3,
which is shown in the small insert that depicts
the same PCA but is turned by 90° to the right
around the axis of principle component 2.
Arrays of BALB/c BMM samples are
represented by dots (•), arrays of C57BL/6
BMM samples by rhombi (). IFN-γ treated
samples are indicated in gray, and non-treated
control level samples are colored in black.
As transcriptome data demonstrated a high reproducibility of biological replicates in the
model system, selected samples were used for different proteome applications rather than
analyzing different biological replicates with only one proteomics approach (Dinh Hoang Dang
Khoa).
For a general comparison, gelfree proteome analysis and transcriptome analysis are more
comparable than 2-DE proteome analysis and transcriptome analysis. Therefore, further data
examination focused on that.
Overall comparison of transcriptome [Maren Depke] and LC-MS/MS proteome [Dinh Hoang
Dang Khoa] analyses
900 of 946 identified proteins from gel-free LC-MS/MS approach could be mapped to the
20074 EntrezGene records that were specified to be represented on the Affymetrix GeneChip
Mouse Gene 1.0 ST array by the Rosetta Resolver software (Table R.3.1). As expected, a much
bigger number of genes was accessible by transcriptome analysis than proteins by gelfree
proteome analysis. Nevertheless, for almost all identified proteins the gene expression pattern
could be recorded.
Table R.3.1: Overall comparison transcriptome [Maren Depke] and LC-MS/MS proteome [Dinh Hoang Dang Khoa] analyses and
numbers of differentially regulated genes and proteins with differing abundance resulting from LC-MS/MS analysis.
total number of
detected genes or
proteins
in BALB/c BMM
IFN-γ effects a
in C57BL/6 BMM
at non-treated
control level
strain differences a
at IFN-γ
treated level
microarray 20074 442 (2.2 %) 396 (2.0 %) 222 (1.1 %) 230 (1.1 %)
LC-MS/MS 946 45 (4.8 %) 52 (5.5 %) 218 (23.0 %) 308 (32.6 %)
a Percentage values indicate the proportion of regulation in both approaches. Values were calculated by dividing the number of
regulated genes by the total number of genes available on the array or the number of regulated proteins by the total number of
identified proteins.
92

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
Surprisingly, the percentage of proteins with significant difference in abundance between
BMM of BALB/c and C57BL/6 mice compared to the total of all identified proteins was noticeably
higher than the percentage of differentially expressed genes between BMM of BALB/c and
C57BL/6 mice compared to the total of all genes available on the array (Table R.3.1).
IFN-γ influence on gene expression [Maren Depke] and protein abundances [Dinh Hoang Dang
Khoa] in BMM of BALB/c and C57BL/6 mice
The comparison of IFN-γ treated BMM with the non-treated medium control revealed that
IFN-γ mainly led to induction of gene expression or increase in protein abundance in both strains
(Table R.3.2 A). From the total of 442 (BALB/c BMM) and 396 (C57BL/6 BMM) differentially
expressed genes, 388 (BALB/c BMM) and 330 (C57BL/6 BMM) were induced, while only 54
(BALB/c BMM) and 66 (C57BL/6 BMM) were repressed.
IFN-γ treatment increased the abundance of 41 (BALB/c BMM) and 43 (C57BL/6 BMM)
proteins and reduced the abundance of 4 (BALB/c BMM) and 9 (C57BL/6 BMM) from the total of
45 (BALB/c BMM) and 52 (C57BL/6 BMM) regulated proteins.
The overlap of the list of proteins that differed significantly in their abundance between IFN-γ
treated BMM and the non-treated medium control and the list of differentially expressed genes
that were found after IFN-γ treatment amounted to 24 (BALB/c BMM) and 20 (C57BL/6 BMM).
Compared to 43 (BALB/c BMM) and 50 (C57BL/6 BMM) regulated proteins also available by
transcriptome analysis and 52 (BALB/c BMM) and 53 (C57BL/6 BMM) differentially expressed
genes also accessible by LC-MS/MS proteome analysis, the overlap of regulated proteins and
genes added up to 38 % to 56 % (Table R.3.2 A).
Table R.3.2: Comparison of differentially regulated genes resulting from transcriptome analysis [Maren Depke] with differentially
regulated proteins resulting from LC-MS/MS analysis [Dinh Hoang Dang Khoa].
A Direction of regulation by IFN-γ in BALB/c BMM and
in C57BL/6 BMM.
IFN-γ treatment effects
B Direction of strain difference between C57BL/6 BMM and BALB/c
BMM at non-treated medium-control and IFN-γ treated level.
strain differences
LC-
a intersection
a microarray
MS/MS
LC-
a intersection
a microarray
MS/MS
a
BALB/c BMM
control condition BMM
induced 39 (41) 41 (388) 24 higher in C57BL/6 102 (112) 10 (100) 7
repressed 4 (4) 11 (54) 0 higher in BALB/c 101 (106) 10 (122) 4
total 43 (45) 52 (442) 24 total 203 (218) 20 (222) 12
C57BL/6 BMM
treatment condition BMM
induced 42 (43) 41 (330) 20 higher in C57BL/6 +IFN-γ 123 (135) 13 (93) 10
repressed 8 (9) 12 (66) 0 higher in BALB/c +IFN-γ 165 (173) 11 (137) 7
total 50 (52) 53 (396) 20 Total 288 (308) 24 (230) 19
The first number in the table always refers to the proteins or genes that were available for analysis with both approaches,
microarray and LC-MS/MS analysis. In brackets, the number of regulated proteins or genes without restriction to the accessibility by
both methods is given.
The comparison of IFN-γ effects in BMM of the two strains BALB/c and C57BL/6 resulted in a
high overlap of 307 differentially expressed genes (Fig. R.3.2 A) and 29 proteins with significantly
different abundance (Fig. R.3.2 C). Even when the gene or protein was not significantly expressed
in one of the strains a similar expression trend in BMM of both strains was observed (Fig. R.3.2 B,
D).
93

log2(ratio IFN-γ-treated / control) of C57BL/6 BMM
log2(ratio IFN-γ-treated / control) of C57BL/6 BMM
Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
A
B
10
8
6
4
135 307 89
2
0
-2
genes regulated by IFN-γ
in BALB/c BMM (442)
genes regulated by IFN-γ
in C57BL/6 BMM (396)
-4
-6
-8
-10
-10 -8 -6 -4 -2 0 2 4 6 8 10
log 2 (ratio IFN-γ-treated / control) of BALB/c BMM
C
D
10
8
6
4
16 29 23
2
0
-2
proteins regulated by IFN-γ
in BALB/c BMM (45)
proteins regulated by IFN-γ
in C57BL/6 BMM (52)
-4
-6
-8
-10
-10 -8 -6 -4 -2 0 2 4 6 8 10
log 2 (ratio IFN-γ-treated / control) of BALB/c BMM
Fig. R.3.2: Comparison of IFN-γ effects on gene expression pattern and protein abundance in BALB/c BMM and C57BL/6 BMM.
A, C. Overview on numbers of differentially expressed genes in transcriptome analysis (A) and of proteins with significantly different
abundance in LC-MS/MS analyses (C) when comparing IFN-γ effects in BALB/c BMM and C57BL/6 BMM.
B, D. Log 2-transformed ratio data “IFN-γ treated BALB/c BMM” / “medium control BALB/c BMM” were plotted on the x-axis and log 2-
transformed ratio data “IFN-γ treated C57BL/6 BMM” / “medium control C57BL/6 BMM” on the y-axis from transcriptome (B) and LC-
MS/MS (D) analyses.
Coloring distinguishes three different groups of regulated genes or proteins: Genes/proteins with significant differential
expression/abundance caused by IFN-γ only in BALB/c BMM are shown in dark gray, while that with significant differential
expression/abundance caused by IFN-γ only in C57BL/6 BMM are displayed in light gray. Genes/proteins significantly regulated by
IFN-γ in both BALB/c BMM and C57BL/6 BMM are colored in black. The minimal absolute fold change cutoff of 1.5 has been applied to
all result lists.
All transcriptome ratio data were derived from mean intensities of three biological replicates, and ratio data of LC-MS/MS analyses
were derived from intensities of one biological replicate analyzed in technical triplicates of each group.
Strain differences in gene expression [Maren Depke] and protein abundances [Dinh Hoang
Dang Khoa] between BMM of BALB/c and C57BL/6 mice in untreated medium control and in
IFN-γ treated samples
The comparison of C57BL/6 BMM and BALB/c BMM at the non-treated medium control level
and at the IFN-γ treated level resulted in 100 (control level) and 93 (IFN-γ treated level) genes
that featured a higher expression in C57BL/6 BMM and 122 (control level) and 137 (IFN-γ treated
level) genes that exhibited a higher expression in BALB/c BMM.
94

log2(ratio C57BL/6 / BALB/c) at IFN-γ-activated level
log2(ratio C57BL/6 / BALB/c) at IFN-γ-activated level
Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
The same comparison using the LC-MS/MS proteome data yielded 112 (control level) and 135
(IFN-γ treated level) proteins with a higher abundance in C57BL/6 BMM and 106 (control level)
and 173 (IFN-γ treated level) proteins with a higher abundance in BALB/c BMM (Table R.3.2 B).
A
B
10
8
6
94 128 102
4
2
genes differentially
expressed between
BALB/c BMM and C57BL/6
BMM at the non-activated
control level (222)
genes differentially
expressed between
BALB/c BMM and C57BL/6
BMM at the IFN-γ-treated
level (230)
0
-2
-4
-6
-8
-10
-10 -8 -6 -4 -2 0 2 4 6 8 10
log 2 (ratio C57BL/6 / BALB/c) at control level
C
D
10
8
6
35 183 125
4
2
proteins differentially
regulated between BALB/c
BMM and C57BL/6 BMM
at the non-activated
control level (218)
proteins differentially
regulated between BALB/c
BMM and C57BL/6 BMM
at the IFN-γ-treated level
(308)
0
-2
-4
-6
-8
-10
-10 -8 -6 -4 -2 0 2 4 6 8 10
log 2 (ratio C57BL/6 / BALB/c) at control level
Fig. R.3.3: Comparison of strain differences between BALB/c BMM and C57BL/6 BMM at non-treated control and IFN-γ treated level.
A, C. Overview on numbers of differentially expressed genes in transcriptome analysis (A) and of proteins with significantly different
abundance in LC-MS/MS analyses (C) when comparing strain differences between BALB/c BMM and C57BL/6 BMM at the non-treated
control level and at the IFN-γ treated level.
B, D. Log 2-transformed ratio data “medium control C57BL/6 BMM” / “medium control BALB/c BMM” were plotted on the x-axis and
log 2-transformed ratio data “IFN-γ treated C57BL/6 BMM” / “IFN-γ treated BALB/c BMM” on the y-axis from transcriptome (B) and LC-
MS/MS (D) analyses.
Coloring distinguishes three different groups of regulated genes or proteins: Genes/proteins with significant strain difference only at
non-treated control level are shown in dark gray, while that with significant strain difference only at IFN-γ treated level are displayed
in light gray. Genes/proteins significantly different between BALB/c BMM and C57BL/6 BMM on both treatment levels, i. e. at nontreated
control and IFN-γ treated level, are colored in black. The minimal absolute fold change cutoff of 1.5 has been applied to all
result lists.
All transcriptome ratio data were derived from mean intensities of three biological replicates, and ratio data of LC-MS/MS analyses
were derived from intensities of one biological replicate analyzed in technical triplicates of each group.
In the comparison of the two mouse strains only 20 out of 222 (control level) and 24 out of
230 (IFN-γ treated level) differentially expressed genes were also identified in LC-MS/MS analysis.
Inversely, 203 out of 218 (control level) and 288 out of 308 (IFN-γ treated level) proteins with
significantly differing abundance were also detected on the Affymetrix array. Due to the limiting
95

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
number of genes whose corresponding proteins were accessible by LC-MS/MS analysis, the
overlap of differentially abundant proteins and differentially expressed genes was low in absolute
number (12 at control level and 19 at IFN-γ treated level) but it still held 60 % (control level) and
79 % (IFN-γ treated level) of the genes that were both differentially expressed in the array data
set and accessible by proteome analysis (Table R.3.2 B). Vice versa, for only 5.9 % (control level)
and 6.6 % (IFN-γ treated level) of regulated proteins with accessible microarray data a change in
gene expression level was detected. Hence, the majority of protein abundance differences
between the strains were independent from gene expression level.
A refined comparison between proteome and transcriptome analysis which included the
direction of regulation uncovered that the overlap of proteins and genes higher expressed in
C57BL/6 BMM and the overlap of proteins and genes higher expressed in BALB/c BMM (7 and 4
at control level; 10 and 17 at IFN-γ treated level) did not sum up to the total overlap of 12 at
control level and 19 at IFN-γ treated level (Table R.3.2 B). The explanation for the discrepancy is
that one protein/gene (Scp2) at the control level showed a higher expression in BALB/c BMM in
transcriptome but higher abundance in C57BL/6 BMM in proteome analysis. Correspondingly, at
the IFN-γ treated level two proteins/genes did not possess the same direction of regulation: H2-
AB1 and H2-K1 were higher expressed in C57BL/6 BMM in transcriptome but showed higher
abundance in BALB/c BMM in proteome analysis.
The comparison of strain differences which occurred at the non-treated medium control level
and at the IFN-γ treated level demonstrated similar differences between C57BL/6 BMM and
BALB/c BMM at both treatment levels: 128 genes were expressed differentially at both treatment
conditions (Fig. R.3.3 A), and 183 proteins show significantly different abundance (Fig. R.3.3 C).
Equivalent to the IFN-γ effects, a similar expression trend in the strain comparisons was visible at
both treatment levels, even when regulation was significant only in one (Fig. R.3.3 B, D).
Revelation of similar main effects of IFN-γ in BALB/c BMM and C57BL/6 BMM by network
analysis and confirmation of known IFN-γ effects in serum-free BMM cultivation system
Ingenuity Pathway Analysis (IPA, www.ingenuity.com) allows to arrange genes of interest, e. g.
differentially expressed genes, in networks which show the interactions in the selected group of
genes. The two networks with the highest score from the separate analyses for IFN-γ effects in
BALB/c BMM and C57BL/6 BMM exhibited an overlap of 22 out of 35 genes (nodes). Therefore,
they covered similar functional aspects. After merging both networks, the comparison of
differential gene expression after IFN-γ treatment in BALB/c BMM (Fig. R.3.4 A) and in C57BL/6
BMM (Fig. R.3.4 B) showed that many of the genes of this network were subjected to comparable
regulation by IFN-γ in BMM of both strains, sometimes with differences in magnitude.
The network confirms already known IFN-γ effects in the new serum-free BMM cultivation
system. The 26 S proteasome was also known to be changed in response to IFN-γ stimulus. The
proteasome, consisting of a ring-shaped 20 S core complex of several subunits, which exhibits
besides others protease function, and two regulatory complexes of 19 S, is the center of
intracellular ubiquitin-dependent protein degradation which is not only needed for the normal
protein turnover but also for the generation of peptides for the presentation to immune cells via
MHC-I molecules. The change of the normal proteasome into the so-called immunoproteasome is
mediated by the exchange of three β-subunits from the core complex by gene products of
Psmb9, Psmb10, and Psmb8 and is induced by IFN-γ. The exchange of these subunits leads to a
change in the proteolytic activity and specificity (Wang/Maldonado 2006, Strehl et al. 2005). Also
an alternative regulatory complex of 11 S (PA28) exists (Rivett et al. 2001). Here, expression of
Psmb9 (3.6 / 3.4), Psmb10 (3.1 / 3.7), and Psmb8 (2.8 / 3.1), coding for β-subunits of the core
complex, and Psme1 (2.1 / 2.1) and Psme2 (2.7 / 3.1), which encode components of the regulatory
11 S complex, were induced in both BALB/c and C57BL/6 BMM after IFN-γ treatment.
96

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
A B
Fig. R.3.4: Ingenuity Pathway Analysis of IFN-γ effects in BALB/c and C57BL/6 BMM.
EntrezGene ID lists of IFN-γ regulated genes were uploaded from Rosetta Resolver software to Ingenuity Pathway Analysis (IPA, www.ingenuity.com). The two networks with the highest score from the separate
analyses for BALB/c BMM and C57BL/6 BMM exhibited an overlap of 22 genes (nodes). Therefore, they covered similar functional aspects. Both networks were merged using the automatic merge-tool from IPA. The
overlay data tool allows coloring of nodes corresponding to the fold change values from the comparisons “IFN-γ treated BALB/c BMM” vs. “non-treated medium control BALB/c BMM” (A) or “IFN-γ treated C57BL/6
BMM” vs. “non-treated medium control C57BL/6 BMM” (B). Red indicates an increase while green marks a decrease of gene expression. The shade of the color correlates to the magnitude of change: more intense
color shows a higher absolute fold change. Transcriptome expression data included only genes which passed the regulation criteria of ANOVA significance p* < 0.01 and the absolute fold change cutoff of 1.5.
97

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
The peptide transporters Tap1 and Tap2 translocate peptides from cytosolic proteasome into
the endoplasmatic reticulum, where they bind to MHC-I molecules, while the peptide-MHCcomplexes
are lateron transferred to the cell surface. Tap1 and Tap2 are described to be induced
by IFN-γ (Brucet et al. 2004, e. g. human, Ma W et al. 1997, Schiffer et al. 2002). In this study,
Tap1 (5.0 / 5.2) and Tap2 (3.3 / 3.0) were higher expressed after IFN-γ treatment in BMM of both
strains BALB/c and C57BL/6 (Fig. R.3.4 A, B). Erap 1, an endoplasmic reticulum aminopeptidase
that takes part in further processing and N-terminal trimming of the peptides in the ER (Chang et
al. 2005, Jung et al. 2009), was additionally induced after IFN-γ treatment in BALB/c BMM (2.1)
and C57BL/6 BMM (2.2) [data not shown]. Finally, several genes of MHC class I and class II
molecules were induced after IFN-γ treatment in BMM of both strains BALB/c and C57BL/6 [data
not shown].
Comparison of differentially expressed gene lists (Fig. R.3.2 A), comparison of log 2 -ratio data
(Fig. R.3.2 B), and Ingenuity Pathway Analyis (Fig. R.3.4) revealed highly similar gene expression
signatures in BALB/c and C57BL/6 BMM after IFN-γ treatment even for genes which were only
differentially expressed in one of the two comparisons. Therefore, further analysis of biological
effects resulting from IFN-γ treatment was performed using the union list of IFN-γ effects in
BALB/c and C57BL/6 BMM.
First, the list of in total 531 IFN-γ dependent genes was subjected to a global functional
analysis using Ingenuity Pathway Analysis (IPA, www.ingenuity.com), which compared functional
categories within the data set with those of the complete array and assigned a p-value to rate on
significant overrepresentation (Table R.3.3).
Table R.3.3: Global functional analysis of IFN-γ dependent genes in BMM of at least one strain, BALB/c or C57BL/6, using Ingenuity
Pathway Analysis (Ingenuity Systems, www.ingenuity.com).
Diseases and Disorders
category p-value a of
number
genes
Inflammatory
Response
Immunological
Disease
Inflammatory
Disease
Connective
Tissue Disorders
Skeletal and
Muscular
Disorders
Infection
Mechanism
Cancer
Respiratory
Disease
Antimicrobial
Response
Infectious
Disease
2.06E-06 -
3.53E-40
5.28E-06 -
1.39E-23
5.61E-06 -
5.75E-23
8.25E-07 -
5.12E-21
8.25E-07 -
5.12E-21
2.83E-06 -
3.51E-19
7.13E-06 -
4.05E-17
2.20E-06 -
2.20E-16
4.46E-08 -
6.66E-15
2.20E-06 -
1.14E-14
146
150
157
Molecular and Cellular Functions
category p-value a of
number
genes
Cellular Growth and
Proliferation
Cellular
Development
Cell-To-Cell
Signaling and
Interaction
111 Cellular Movement
150 Cell Death
43
141
Cellular Function
and Maintenance
Antigen
Presentation
56 Cell Cycle
29 Gene Expression
87 Cell Signaling
a Ten categories with the smallest p-values are cited.
D. a. F. – Development and Function
6.38E-06 -
4.55E-25
6.28E-06 -
8.22E-25
6.78E-06 -
2.73E-23
1.36E-09 -
5.54E-21
5.67E-06 -
1.21E-18
4.90E-06 -
1.76E-18
6.38E-06 -
3.53E-16
1.04E-06 -
4.82E-10
1.98E-07 -
1.00E-09
9.08E-07 -
3.04E-09
154
117
Physiological System
Development and Function
category p-value a of
number
genes
Hematological
System D. a. F.
Immune Cell
Trafficking
127 Hematopoiesis
101
150
84
56
71
69
62
Tissue
Morphology
Cell-mediated
Immune
Response
Organismal
Survival
Tissue
Development
Humoral
Immune
Response
Cardiovascular
System D. a. F.
Organismal
Development
6.38E-06 -
4.55E-25
4.90E-06 -
6.53E-21
2.06E-06 -
2.65E-20
9.94E-07 -
1.36E-18
4.01E-07 -
2.27E-18
2.50E-07 -
3.64E-15
6.82E-06 -
4.43E-14
9.94E-07 -
2.03E-10
4.05E-06 -
2.39E-09
3.05E-06 -
9.33E-09
138
93
82
70
66
76
71
37
39
41
98

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
This analysis revealed expected categories within the IFN-γ dependently regulated genes in
BMM like “Inflammatory Response”, “Antigen Presentation”, “Immune Cell Trafficking” and other
immune response related functions. Furthermore, growth, proliferation, but also cell cyle and cell
death associated genes were included in the set of differentially expressed genes.
For a more detailed view, IPA’s canonical pathway analysis was applied (Table R.3.4). Here,
the pathway “Interferon Signaling” appeared with the highest ratio value, i. e. the fraction of
genes from the pathway, which was also included in the list of IFN-γ dependent genes. This
observation nicely proved the relevance of the gene expression signature which was recorded
after stimulation of BMM with interferon. Further pathways from the group with the highest
ratio values were among others “Antigen Presentation Pathway”, “Role of Pattern Recognition
Receptors in Recognition of Bacteria and Viruses”, and “Activation of IRF by Cytosolic Pattern
Recognition Receptors” which fits well into the expected characterization of the experimental
system. Again, also cell death associated features appeared with the “Retinoic acid Mediated
Apoptosis Signaling” pathway.
Table R.3.4: Canonical pathway analysis of IFN-γ dependent genes in BMM of at least one strain, BALB/c or C57BL/6, using Ingenuity
Pathway Analysis (Ingenuity Systems, www.ingenuity.com).
Pathway ratio a p-value
Interferon Signaling 11/30 = 0.367 8.34E-12
Antigen Presentation Pathway 11/39 = 0.282 1.61E-09
Role of Pattern Recognition Receptors in Recognition of Bacteria and Viruses 20/86 = 0.233 5.03E-15
Complement System 8/36 = 0.222 4.09E-07
Retinoic acid Mediated Apoptosis Signaling 9/44 = 0.205 2.42E-08
Role of PKR in Interferon Induction and Antiviral Response 9/46 = 0.196 2.95E-07
T Helper Cell Differentiation 8/41 = 0.195 7.42E-06
Activation of IRF by Cytosolic Pattern Recognition Receptors 14/74 = 0.189 1.51E-11
Allograft Rejection Signaling 8/45 = 0.178 2.34E-05
TREM1 Signaling 12/69 = 0.174 3.84E-09
a Ten pathways with the highest ratio values are cited.
Manual data mining revealed different cytokines and cytokine receptors, which were
differentially expressed in BMM upon treatment with IFN-γ. Here, the induction of several
chemotactic chemokines (Ccl5, Ccl12, Ccl22, Cxcl9, Cxcl10, Cxcl11, Cxcl16) was noticeable. Among
the receptors, Ccr5 induction fitted to the induction of one of its ligands, Ccl5. The induced
receptor Ccrl2 is known to be induced after activation and during differentiation from monocytes
to macrophages, but its function is not described yet. Contrarily, the receptor Cxcr4 was
repressed after treatment of BMM with IFN-γ. Two interleukins were included in the
differentially expressed cytokines: IL-15, a regulatory, anti-apoptotic cytokine with structural
similarity to IL-2, and the subunit IL27-p28 of the heterodimeric IL-27, a regulatory cytokine
produced by antigen presenting cells. Several interleukin receptors or receptor subunits were
induced after IFN-γ treatment. The induction of the IL-15 receptor alpha-chain (Il15ra)
corresponded to the already mentioned induction of IL-15. Furthermore, the strongly
interdependent receptor subunits Il4ra, Il13ra1, Il21r, and Il2rg were induced. The alpha-chain of
IL-4 receptor (Il4ra) and the IL-21 receptor (Il21r) employ the IL-2 receptor gamma-chain (Il2rg) as
their second subunit. Similarly, the primary IL-13 binding subunit of the IL-13 receptor (Il13ra1)
builds a receptor complex together with the alpha-chain of IL-4 receptor (Il4ra). Also Il12rb1,
coding for the low-affinity binding subunit of the IL-12 receptor, was induced. The induction of
Il10ra, high-affinity binding subunit of the IL-10 receptor, and of Il18bp, cytokine IL-18 binding
protein, is an example for inhibitory processes in BMM after IFN-γ treatment. Repression of
interleukin receptor subunits was also found after IFN-γ treatment: Il1rl2, receptor for interleukin
1 family member 9, and Il6st alias gp130, signal transducer for several cytokines, e. g. IL-6, were
99

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
found to be repressed. Finally, differential expression was observed for members of the TNF
ligand and receptor superfamilies. The central inflammation mediator Tnf, the apoptosis inducer
TRAIL (Tnfsf10), and the T cell survival modulator Tnfsf18 were induced in parallel with the
receptor Tnfrsf14 (Table R.3.5).
Table R.3.5: IFN-γ influence on gene expression of cytokines and cytokine receptors in BMM of BALB/c and C57BL/6 mice.
Rosetta Resolver Annotation fold change a strain
difference
gene
Entrez
description
alias
name
Gene ID
BALB/c C57BL/6 control IFN-γ
Ccl5 chemokine (C-C motif) ligand 5
SISd, Scya5, RANTES,
TCP228
20304 3.7 5.0
Ccl12 chemokine (C-C motif) ligand 12 MCP-5, Scya12 20293 3.4 2.1
Ccl22 chemokine (C-C motif) ligand 22
MDC, DCBCK, ABCD-1,
Scya22
20299 2.9 3.6
Ccr5 chemokine (C-C motif) receptor 5 AM4-7, CD195, Cmkbr5 12774 1.7 1.4
Ccrl2 chemokine (C-C motif) receptor-like 2
E01, CCR11, L-CCR,
Cmkbr1l2
54199 2.0 1.7
Cxcl9 chemokine (C-X-C motif) ligand 9 CMK, Mig, Scyb9, crg-10 17329 45.0 56.1
Cxcl10 chemokine (C-X-C motif) ligand 10
C7, IP10, CRG-2, INP10,
Ifi10, mob-1, Scyb10, 15945 47.8 21.5
gIP-10
Cxcl11 chemokine (C-X-C motif) ligand 11
IP9, H174, ITAC, CXC11,
SCYB9B, Scyb11, betaR1
56066 8.3 1.1 x
Cxcl16 chemokine (C-X-C motif) ligand 16 SR-PSOX, Zmynd15 66102 2.6 2.9
Cxcr4 chemokine (C-X-C motif) receptor 4
CD184, LESTR, Sdf1r,
Cmkar4, PB-CKR, 12767 -1.8 -2.2 x x
PBSF/SDF-1
Il1rl2 interleukin 1 receptor-like 2 IL-1Rrp2 107527 -1.7 -1.8
Il2rg interleukin 2 receptor, gamma chain CD132, gamma(c) 16186 1.9 1.7
Il4ra interleukin 4 receptor, alpha Il4r, CD124 16190 2.0 1.6 x x
Il6st interleukin 6 signal transducer CD130, gp130 16195 -1.8 -1.8
Il10ra interleukin 10 receptor, alpha Il10r, CDw210, mIL-10R 16154 1.6 1.7
Il12rb1 interleukin 12 receptor, beta 1 CD212, IL-12R[b] 16161 1.8 3.9
Il13ra1 interleukin 13 receptor, alpha 1 NR4, Il13ra, CD213a1 16164 2.1 1.7 x
Il15 interleukin 15 16168 2.8 2.0 x
Il15ra interleukin 15 receptor, alpha chain 16169 4.3 4.0
Il18bp interleukin 18 binding protein MC54L, Igifbp, IL-18BP 16068 5.5 6.2
Il21r interleukin 21 receptor NILR 60504 3.4 2.8 x x
Il27 interleukin 27 p28, Il30, IL-27p28 246779 2.3 2.4
Tnf tumor necrosis factor
DIF, Tnfa, TNFSF2,
Tnfsf1a, TNF-alpha
21926 2.4 2.7
Tnfsf10
tumor necrosis factor (ligand) superfamily,
member 10
TL2, Ly81, Trail, APO-2L 22035 10.6 5.2
Tnfsf18
tumor necrosis factor (ligand) superfamily,
member 18
Gitrl 240873 3.3 1.1 x
Tnfrsf14
tumor necrosis factor receptor superfamily, Atar, HveA, Hvem,
member 14 (herpesvirus entry mediator) Tnfrs14
230979 2.6 2.5
a Fold change values were calculated from expression intensities of IFN-γ treated BMM in comparison to control BMM from the mean
of three biological replicates. Differential expression in statistical testing with p* < 0.01 and a minimal absolute fold change of 1.5 is
indicated in bold.
Furthermore, GTPases / GTP binding proteins of different families, which are known to be
induced by interferon and play a role in antimicrobial defense, were induced, and some even
belonged to the genes with the strongest induction within the data set. Four families are
described in literature (MacMicking 2004), p47 GTPases, p65 guanylate-binding proteins (GBP),
Mx proteins and very large inducible GTPases (VLIG), and of each family examples were included
in the data set of induced genes in BMM after IFN-γ treatment. The p47 family was present with
the members Igtp, Iigp1, Irgm1, and Tgtp. All guanylate binding proteins of the p65 family were
induced (Gbp1, Gbp2, Gbp3, Gbp4, Gbp5, and Gbp6). Both types of Mx genes, coding for the
nuclear (Mx1) and the cytosolic (Mx2) form, exhibited increased expression in BMM after IFN-γ
treatment. Finally, very large interferon inducible GTPase 1, Gvin1, was included in the list of
induced genes as member of the last family (Table R.3.6).
100

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
Table R.3.6: IFN-γ influence on gene expression of GTPase family members in BMM of BALB/c and C57BL/6 mice.
Rosetta Resolver Annotation fold change a strain
difference
gene
Entrez
description
alias
name
Gene ID
BALB/c C57BL/6 control IFN-γ
Igtp interferon gamma induced GTPase Irgm3 16145 17.4 13.5
Iigp1 interferon inducible GTPase 1 Iigp, Irga6 60440 100.0 100.0
Irgm1 immunity-related GTPase family M member 1
Ifi1, Irgm, Iigp3, Iipg3,
LRG-47
15944 6.9 11.2
Tgtp T-cell specific GTPase Tgtp, Gtp2, Mg21, Irgb6 21822 49.0 100.0
Gbp1 guanylate binding protein 1 Mpa1, Gbp-1, Mag-1 14468 43.9 6.0 x x
Gbp2 guanylate binding protein 2 14469 17.2 31.2 x
Gbp3 guanylate binding protein 3 Gbp4 55932 21.7 15.6
Gbp4 guanylate binding protein 4 Mpa2, Mag-2, Mpa-2 17472 60.6 94.3
Gbp5 guanylate binding protein 5 Gbp5a 229898 50.7 66.8
Gbp6 guanylate binding protein 6 Gbp7 229900 9.3 7.2
Mx1 myxovirus (influenza virus) resistance 1 Mx, Mx-1 17857 10.8 6.9 x
Mx2 myxovirus (influenza virus) resistance 2 Mx-2 17858 6.1 3.5 x
Gvin1 GTPase, very large interferon inducible 1 VLIG, Iigs1, VLIG-1 74558 4.6 6.7
a Fold change values were calculated from expression intensities of IFN-γ treated BMM in comparison to control BMM from the mean
of three biological replicates. Differential expression in statistical testing with p* < 0.01 and a minimal absolute fold change of 1.5 is
indicated in bold.
BMM after IFN-γ treatment induced the expression of complement system components. Of
the classical activation pathway, the C1 subcomponents C1qa, C1qb, C1qc, C1r, and C1rb were
induced, and of the alternative pathway factor B (Cfb). Induced component C3 takes part in all
activation pathways and especially in the amplification loop of complement activation
(Table R.3.7).
Table R.3.7: IFN-γ influence on gene expression of complement components in BMM of BALB/c and C57BL/6 mice.
Rosetta Resolver Annotation fold change a strain
difference
gene
Entrez
description
alias
name
Gene ID
BALB/c C57BL/6 control IFN-γ
C1qa
complement component 1, q subcomponent, alpha
polypeptide
C1q 12259 1.3 2.3
C1qb
complement component 1, q subcomponent, beta
polypeptide
12260 2.2 2.8 x x
C1qc complement component 1, q subcomponent, C chain C1qg, Ciqc 12262 1.6 2.8 x
C1r complement component 1, r subcomponent C1ra, C1rb 50909 4.3 4.7
C1rb complement component 1, r subcomponent, b 270510 5.5 6.3
C3 complement component 3 ASP, Plp 12266 4.9 5.6
Cfb complement factor B B, Bf, Fb, H2-Bf 14962 14.1 19.5
a Fold change values were calculated from expression intensities of IFN-γ treated BMM in comparison to control BMM from the mean
of three biological replicates. Differential expression in statistical testing with p* < 0.01 and a minimal absolute fold change of 1.5 is
indicated in bold.
Other immune function related gene expression changes were conspicuous in BMM after
IFN-γ treatment. Inducible nitric oxide synthase 2 (Nos2) was increased as expected in treated
BMM. The enzyme produces NO, which acts as antimicrobial effector and as signal mediator
during immune defense processes. Another induced enzyme gene was kynureninase (Lkynurenine
hydrolase, Kynu) whose product is involved in the degradation of kynurenines, toxic
intermediates with signaling properties. Their production is amplified in inflammation settings by
the induction of indoleamine 2,3-dioxygenase (IDO), which catalyzes the first step of tryptophan
degradation in the kyurenine pathway. Interestingly, the Ido gene was not differentially
expressed in BMM after IFN-γ treatment, expression was even absent in the majority of samples.
The tryptophanyl-tRNA synthetase (Wars) was induced in IFN-γ treated BMM. Prostaglandin-
101

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
endoperoxide synthase 2 (Ptgs2), coding for the enzyme responsible for prostaglandin E 2
biosynthesis, was induced as well as different 2'-5'-oligoadenylate synthetase genes (Oas2, Oas3,
Oasl1, Oasl2). They are part of the innate immune response. Serum amyloid proteins are part of
the acute phase response. In humans, Saa genes are known to be expressed not only in the liver,
but also in activated monocytes/macrophages (Urieli-Shoval et al. 1994). In this study, serum
amyloid A 3 (Saa3) was induced in murine BMM after IFN-γ treatment.
BMM induced plasma membrane receptors after treatment with IFN-γ. Among others, the
induction of IgG Fc receptor with high affinity I and with low affinity IV (Fcgr1 and Fcgr4), of tolllike
receptors 2, 9, and 12 (Tlr2, Tlr9, Tlr12), and of co-stimulatory receptors Cd40 and Cd86 was
observed. Additionally, CD274 (B7-H1, PD-L1), ligand for the immunoinhibitory receptor PD-1 on
activated B and T cells, and the second PD-1 receptor, programmed cell death 1 ligand 2
(Pdcd1lg2; PD-L2) exhibited induction (Table R.3.8).
Table R.3.8: IFN-γ influence on expression of immune function related genes in BMM of BALB/c and C57BL/6 mice.
Rosetta Resolver Annotation fold change a strain
difference
gene
Entrez
description
alias
name
Gene ID
BALB/c C57BL/6 control IFN-γ
Nos2 nitric oxide synthase 2, inducible iNOS,Nos2a 18126 18.6 25.5
Kynu kynureninase (L-kynurenine hydrolase) 70789 4.0 5.2
Wars tryptophanyl-tRNA synthetase WRS 22375 3.7 4.4
Ptgs2 prostaglandin-endoperoxide synthase 2
COX2, PHS-2, Pghs2,
TIS10, PGHS-2
19225 16.2 15.4
Oas2 2'-5' oligoadenylate synthetase 2 Oasl11 246728 2.3 2.4 x
Oas3 2'-5' oligoadenylate synthetase 3 Oasl10 246727 3.0 4.4
Oasl1 2'-5' oligoadenylate synthetase-like 1 oasl9 231655 4.8 2.4 x x
Oasl2 2'-5' oligoadenylate synthetase-like 2
Oasl, M1204, Mmu-
OASL
23962 2.6 4.2
Saa3 serum amyloid A 3 l7R3, Saa-3 20210 3.8 5.7
Fcgr1 Fc receptor, IgG, high affinity I
CD64, IGGHAFC,
FcgammaRI
14129 4.6 4.9 x
Fcgr4 Fc receptor, IgG, low affinity IV
Fcrl3, CD16-2, FcgRIV,
Fcgr3a, FcgammaRIV
246256 6.0 3.9
Tlr2 toll-like receptor 2 Ly105 24088 2.0 1.5
Tlr9 toll-like receptor 9 81897 2.8 2.3
Tlr12 toll-like receptor 12 Gm1365 384059 1.8 2.8
Cd40 CD40 antigen
IGM, p50, Bp50, GP39,
IMD3, TRAP, HIGM1, T- 21939 5.1 2.6 x
BAM, Tnfrsf5
Cd86 CD86 antigen
B7, B70, MB7, B7-2,
B7.2, CLS1, Ly58, ETC-1,
Ly-58, MB7-2, Cd28l2,
12524 6.5 5.4
TS/A-2
Cd274 CD274 antigen
B7-H1, PD-L1, Pdcd1l1,
Pdcd1lg1
60533 4.0 3.0
Pdcd1lg2 programmed cell death 1 ligand 2 Btdc, B7-DC, PD-L2 58205 5.5 8.6
a Fold change values were calculated from expression intensities of IFN-γ treated BMM in comparison to control BMM from the mean
of three biological replicates. Differential expression in statistical testing with p* < 0.01 and a minimal absolute fold change of 1.5 is
indicated in bold.
Finally, the expression of mitochondrial superoxide dismutase 2 (Sod2) and glutaredoxin
(Glrx), which are involved in the anti-oxidant defense, was induced. Contrarily, glutathione
S-transferase mu 1 (Gstm1), which is associated to detoxification processes, was repressed.
Furthermore, the induction of lysosomal enzymes cathepsin C and H (Ctsc, Ctsh) was observed.
BMM induced cell adhesion molecules, integrins, and protocadherin (Icam1, Itgal, Itgb7, Pcdh7,
Vcam1) after IFN-γ treatment, whereas they repressed the beta-integrin subunit Itgb3
(Table R.3.9).
102

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
Table R.3.9: IFN-γ influence on gene expression of anti-oxidant, detoxification, and lysosomal enzymes and of adhesion molecules in
BMM of BALB/c and C57BL/6 mice.
Rosetta Resolver Annotation fold change a strain
difference
gene
Entrez
description
alias
name
Gene ID
BALB/c C57BL/6 control IFN-γ
Sod2 superoxide dismutase 2, mitochondrial MnSOD 20656 2.1 2.2
Glrx glutaredoxin Grx1, Glrx1, Ttase 93692 2.0 2.2
Gstm1 glutathione S-transferase, mu 1 Gstb1, Gstb-1 14862 -1.9 -1.9
Ctsc cathepsin C DPP1, DPPI 13032 4.0 4.5 x
Ctsh cathepsin H Ctsh 13036 2.2 1.6 x x
Icam1 intercellular adhesion molecule 1 CD54, Ly-47, Icam-1, MALA-2 15894 2.3 2.2
Itgal integrin alpha L Cd11a, LFA-1, Ly-15, Ly-21 16408 2.8 2.0 x
Itgb3 integrin beta 3 CD61, GP3A, INGRB3 16416 -2.3 -2.9
Itgb7 integrin beta 7 Ly69 16421 2.5 2.1
Pcdh7 protocadherin 7 54216 1.7 1.7 x
Vcam1 vascular cell adhesion molecule 1 CD106, Vcam-1 22329 3.1 1.7
a Fold change values were calculated from expression intensities of IFN-γ treated BMM in comparison to control BMM from the mean
of three biological replicates. Differential expression in statistical testing with p* < 0.01 and a minimal absolute fold change of 1.5 is
indicated in bold.
Biological context of gene expression differences between BALB/c BMM and C57BL/6 BMM
Comparison of differentially expressed gene lists (Fig. R.3.3 A) and comparison of log 2 -ratio
data (Fig. R.3.3 B) revealed highly similar strain differences between BALB/c and C57BL/6 BMM at
both treatment levels, non-treated medium control and after IFN-γ treatment, even for genes
which were differentially expressed only in one of the two comparisons. Therefore, the biological
context of strain differences was further analyzed using the union set of differences at control
level and after IFN-γ treatment.
First, the list of in total 324 genes, which exhibited differences between the strains in at least
one of the two treatment conditions, was subjected to a global functional analysis using
Ingenuity Pathway Analysis (IPA, www.ingenuity.com), which compared functional categories
within the data set with those of the complete array and assigned a p-value to rate on significant
overrepresentation (Table R.3.10).
This analysis revealed on the one hand categories with immune response related functions
like “Inflammatory Response”, “Antigen Presentation”, and “Immune Cell Trafficking”, which was
not surprising because the gene expression repertoire of BMM independent from the strain of
which the cells were derived is focused at the immune response as the proprietary function of
the BMM. Metabolic functions and the category “Cell Death” additionally appeared in the
analysis of strain differences.
Also for the strain difference aspect of this study, IPA’s canonical pathway analysis was
applied for a more detailed view (Table R.3.11).
Here, the pathway “Complement System” appeared with the highest ratio value. Four genes
differentially expressed between the strains were included in this pathway. Noticeably, the ratio
values for this and further pathways were lower than in the analysis of IFN-γ effects on BMM.
This might reflect the smaller number of genes exhibiting differences between the strains in
comparison to the number of genes which are influenced by IFN-γ, but also the possibility that
the strain differences are more distributed among the pathways and functions.
103

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
Table R.3.10: Global functional analysis of genes exhibiting strain differences in BMM of at least one treatment group, medium control
or after IFN-γ treatment, using Ingenuity Pathway Analysis (Ingenuity Systems, www.ingenuity.com).
Diseases and Disorders
category p-value a of
number
genes
Inflammatory
Response
Cancer
Antimicrobial
Response
Inflammatory
Disease
Genetic Disorder
Connective
Tissue Disorders
Skeletal and
Muscular
Disorders
Gastrointestinal
Disease
Metabolic
Disease
Immunological
Disease
1.51E-02 -
2.19E-14
1.51E-02 -
6.88E-14
5.60E-05 -
6.53E-11
1.51E-02 -
1.09E-10
1.51E-02 -
3.99E-08
1.51E-02 -
4.61E-08
1.51E-02 -
4.61E-08
1.51E-02 -
6.60E-07
1.51E-02 -
8.95E-07
1.51E-02 -
4.22E-06
69
Molecular and Cellular Functions
category p-value a of
number
genes
Cellular Growth and
Proliferation
95 Lipid Metabolism
18
Small Molecule
Biochemistry
88 Cell Death
146 Cellular Movement
58
78
Cell-To-Cell
Signaling and
Interaction
Antigen
Presentation
70 Cell Morphology
a Ten categories with the smallest p-values are cited.
D. a. F. – Development and Function
75
66
Cellular
Development
Carbohydrate
Metabolism
1.36E-02 -
2.43E-08
1.51E-02 -
4.00E-08
1.51E-02 -
4.00E-08
8.68E-03 -
5.49E-07
1.51E-02 -
5.53E-07
1.51E-02 -
1.14E-06
1.19E-02 -
1.08E-05
1.51E-02 -
1.86E-05
1.51E-02 -
1.86E-05
1.51E-02 -
3.81E-05
92
28
42
81
47
64
24
38
44
Physiological System
Development and Function
category p-value a of
number
genes
Hematological
System D. a. F.
Immune Cell
Trafficking
Cardiovascular
System D. a. F.
Organismal
Development
Humoral
Immune
Response
Organismal
Survival
Cell-mediated
Immune
Response
Tissue
Development
Tissue
Morphology
20 Hematopoiesis
1.51E-02 -
7.63E-07
1.37E-02 -
7.63E-07
1.51E-02 -
3.45E-05
9.01E-03 -
3.45E-05
1.51E-02 -
5.60E-05
1.45E-02 -
6.36E-05
1.51E-02 -
7.36E-05
1.51E-02 -
1.43E-04
1.37E-02 -
2.29E-04
1.51E-02 -
7.48E-04
65
38
30
18
14
19
13
29
35
35
Table R.3.11: Canonical pathway analysis of genes exhibiting strain differences in BMM of at least one treatment group, medium
control or after IFN-γ treatment, using Ingenuity Pathway Analysis (Ingenuity Systems, www.ingenuity.com).
Pathway ratio a p-value
Complement System 4/36 = 0,111 1,04E-03
Crosstalk between Dendritic Cells and Natural Killer Cells 8/98 = 0,082 1,45E-05
Role of Pattern Recognition Receptors in Recognition of Bacteria and Viruses 7/86 = 0,081 1,08E-04
Antigen Presentation Pathway 3/39 = 0,077 4,87E-03
Allograft Rejection Signaling 3/45 = 0,067 6,18E-03
Autoimmune Thyroid Diseases Signaling 3/47 = 0,064 7,69E-03
Caveolar-mediated Endocytosis Signaling 5/83 = 0,060 4,44E-03
Role of RIG1-like receptors in Antiviral Innate Immunity 3/52 = 0,058 1,34E-02
IL-10 Signaling 4/70 = 0,057 1,02E-02
Toll-like Receptor Signaling 3/54 = 0,056 3,06E-02
a Ten pathways with the highest ratio values which additionally are over-represented with p < 0.05 are cited.
Manual data mining revealed different cytokines and cytokine receptors, which were
differentially expressed between BMM of the two strain at control level, upon treatment with
IFN-γ, or in both conditions (Table R.3.12).
Certain chemotactic chemokines were higher expressed in BALB/c BMM (Ccl24, Cxcl11,
Cxcl14), but also for the receptor Ccr2, which binds Ccl2, Ccl7, and Ccl13 chemokines, and Cxcr4,
whose ligand is Cxcl12, higher expression was observed in BALB/c BMM. Additionally, IL-15, a
104

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
regulatory, anti-apoptotic cytokine with structural similarity to IL-2, and the immune modulators
Tnfsf8 and Tnfsf18, members of the TNF ligand superfamily, were higher expressed in BALB/c
BMM.
Strain differences were observed for the interleukin receptors Il4ra (alpha-chain of IL-4
receptor), Il13ra1 (primary IL-13 binding subunit of the IL-13 receptor), and Il21r (IL-21 receptor).
Il4ra and Il13ra1 are linked, because Il13ra1 builds a receptor complex together with Il4ra. All
three interleukin receptor chains were higher expressed in C57BL/6 BMM (Table R.3.12).
Table R.3.12: Strain differences in gene expression of cytokines and cytokine receptors between BMM of BALB/c and C57BL/6 mice.
Rosetta Resolver Annotation fold change a IFN-γ effect
gene
Entrez
description
alias
name
Gene ID
control IFN-γ BALB/c C57BL/6
Ccl24 chemokine (C-C motif) ligand 24 CKb-6, MPIF-2, Scya24 56221 -2.7 -2.2
Ccr2 chemokine (C-C motif) receptor 2
Ckr2, Ccr2a, Ccr2b,
Ckr2a, Ckr2b, mJe-r, 12772 -2.8 -4.0
Cmkbr2, Cc-ckr-2
Cxcl11 chemokine (C-X-C motif) ligand 11
IP9, H174, ITAC, b-R1,
CXC11, I-TAC, SCYB9B, 56066 -1.0 -7.6 x
Scyb11, betaR1
Cxcl14 chemokine (C-X-C motif) ligand 14
KS1, Kec, BMAC, BRAK,
NJAC, MIP-2g, Scyb14, 57266 -4.0 -6.1
bolekine, MIP2gamma
Cxcr4 chemokine (C-X-C motif) receptor 4
CD184, LESTR, Sdf1r,
Cmkar4, PB-CKR, 12767 -2.1 -2.5 x
PBSF/SDF-1
Il4ra interleukin 4 receptor, alpha Il4r, CD124 16190 2.2 1.7 x x
Il13ra1 interleukin 13 receptor, alpha 1 NR4, Il13ra, CD213a1 16164 1.7 1.4 x x
Il15 interleukin 15 16168 -1.5 -2.0 x
Il21r interleukin 21 receptor NILR 60504 3.3 2.8 x
Tnfsf18
tumor necrosis factor (ligand) superfamily,
member 18
Gitrl 240873 1.0 -3.0 x
Tnfsf8
tumor necrosis factor (ligand) superfamily,
member 8
CD153, Cd30l, CD30LG 21949 -1.8 -5.2
a Fold change values were calculated for the comparison of C57BL/6 BMM with the baseline of BALB/c BMM of mean values from
three biological replicates per group. Positive values indicate higher expression in C57BL/6 BMM than in BALB/c BMM, whereas
negative values indicate higher expression in BALB/c BMM than in C57BL/6 BMM. Differential expression between BMM of the two
strains at the same treatment level (IFN-γ treated or medium control) using statistical testing in combination with an absolute fold
change cutoff of 1.5 is indicated in bold.
Furthermore, four GTPases / GTP binding proteins of the p65 and the Mx families, which were
already observed to be induced by IFN-γ within the data set of BALB/c and C57BL/6 BMM,
exhibited strain differences. Gbp1, Mx1, and Mx2 were higher expressed in BALB/c BMM,
whereas Gbp2 was higher expressed in C57BL/6 BMM.
Complement system components C1qb and C1qc, the receptor for complement factor
fragment C5a, C5ar1, but also the complement factor C1 inhibitor, Serping1, were higher
expressed in C57BL/6 BMM (Table R.3.13).
Other immune function related gene expression differences between BMM of the two strains
were conspicuous, because the genes were clearly affected by IFN-γ treatment. First, MHC genes
of class I (H28, H2-Q6, H2-Q8, H2-T24) and class II (H2-Ea) were higher expressed in BALB/c BMM.
But also H2-Ab1 (class II), H2-K1, and H2-M2 (class I) exhibited strain differences, however, their
expression was higher in C57BL/6 BMM. Second, a group of interferon-inducible genes possessed
differences in their expression between BMM of both strains: Ifi27l2a, Ifi44, Ifi202, Ifi203, Ifih1,
Ifit3, Ifitm3, Irf7, and Isg20. All of them were higher expressed in BALB/c BMM than in C57BL/6
BMM (Table R.3.14).
105

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
Table R.3.13: Strain differences in gene expression of GTPase family members and complement system components between BMM of
BALB/c and C57BL/6 mice.
Rosetta Resolver Annotation fold change a IFN-γ effect
gene
Entrez
description
alias
name
Gene ID
control IFN-γ BALB/c C57BL/6
Gbp1 guanylate binding protein 1 Mpa1, Gbp-1, Mag-1 14468 -4.1 -30.0 x x
Gbp2 guanylate binding protein 2 14469 1.3 2.3 x x
Mx1 myxovirus (influenza virus) resistance 1 Mx, Mx-1 17857 -2.0 -3.1 x x
Mx2 myxovirus (influenza virus) resistance 2 Mx-2 17858 -3.0 -5.3 x x
C1qb
complement component 1, q subcomponent,
beta polypeptide
12260 6.2 7.8 x
C1qc
complement component 1, q subcomponent,
C chain
C1qg, Ciqc 12262 2.0 3.5 x
C5ar1 complement component 5a receptor 1 C5r1, Cd88, D7Msu1 12273 2.9 2.1
Serping1
serine (or cysteine) peptidase inhibitor, clade G,
C1nh, C1INH
member 1
12258 1.1 2.1 x
a Fold change values were calculated for the comparison of C57BL/6 BMM with the baseline of BALB/c BMM of mean values from
three biological replicates per group. Positive values indicate higher expression in C57BL/6 BMM than in BALB/c BMM, whereas
negative values indicate higher expression in BALB/c BMM than in C57BL/6 BMM. Differential expression between BMM of the two
strains at the same treatment level (IFN-γ treated or medium control) using statistical testing in combination with an absolute fold
change cutoff of 1.5 is indicated in bold.
Table R.3.14: Strain differences in gene expression of MHC molecules and interferon-inducible factors between BMM of BALB/c and
C57BL/6 mice.
Rosetta Resolver Annotation fold change a IFN-γ effect
gene
Entrez
description
alias
name
Gene ID
control IFN-γ BALB/c C57BL/6
H28 histocompatibility 28 H-28, H28-1, NS1178 15061 -21.1 -100.0 x
H2-Ea histocompatibility 2, class II antigen E alpha
Ia3, Ia-3, H-2Ea,
AI323765, E-alpha-f
14968 -64.2 -100.0 x
H2-Q6 histocompatibility 2, Q region locus 6 Qa6, Qa-6, H-2Q6 110557 -2.8 -1.9 x x
H2-Q8 histocompatibility 2, Q region locus 8
Qa8, Qa-2, H-2Q8,
Ms10t, MMS10-T
15019 -4.8 -3.2 x
H2-T24 histocompatibility 2, T region locus 24 H-2T24 15042 -4.6 -2.5 x x
H2-Ab1 histocompatibility 2, class II antigen A, beta 1
IAb, Ia2, Ia-2, Abeta,
H-2Ab, H2-Ab, Rmcs1, 14961 2.1 2.0
I-Abeta
H2-K1 histocompatibility 2, K1, K region H-2K, H2-K, H-2K(d) 14972 1.5 1.6
H2-M2 histocompatibility 2, M region locus 2 H-2M2, Thy19.4 14990 6.2 4.4 x x
Ifi27l2a interferon, alpha-inducible protein 27 like 2A Ifi27, Isg12, Isg12(b1) 76933 -5.4 -4.0
Ifi44 interferon-induced protein 44 p44, MTAP44 99899 -36.1 -5.4 x x
Ifi202 interferon activated gene 202 15949 -20.1 -78.6 x
Ifi203 interferon activated gene 203 15950 -3.9 -2.0 x x
Ifih1 interferon induced with helicase C domain 1
Hlcd, MDA5, MDA-5,
Helicard
71586 -1.5 -1.9 x x
Ifit3
interferon-induced protein with
tetratricopeptide repeats 3
Ifi49 15959 -4.5 -2.5 x x
Ifitm3 interferon induced transmembrane protein 3
Fgls, IP15, Cd225, mil-
1, Cdw217
66141 -5.0 -2.2 x x
Irf7 interferon regulatory factor 7 54123 -7.3 -3.2 x x
Isg20 interferon-stimulated protein DnaQL, HEM45 57444 -1.3 -3.9 x
a Fold change values were calculated for the comparison of C57BL/6 BMM with the baseline of BALB/c BMM of mean values from
three biological replicates per group. Positive values indicate higher expression in C57BL/6 BMM than in BALB/c BMM, whereas
negative values indicate higher expression in BALB/c BMM than in C57BL/6 BMM. Differential expression between BMM of the two
strains at the same treatment level (IFN-γ treated or medium control) using statistical testing in combination with an absolute fold
change cutoff of 1.5 is indicated in bold.
106

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
Table R.3.15: Strain differences in gene expression of plasma membrane receptors between BMM of BALB/c and C57BL/6 mice.
gene
name
description
Rosetta Resolver Annotation fold change a IFN-γ effect
alias
Entrez
Gene ID
control IFN-γ BALB/c C57BL/6
Cadm1 cell adhesion molecule 1
Bl2, ST17, Igsf4, Necl2, Tslc1,
Igsf4a, RA175A, RA175B, RA175C, 54725 -4.7 -4.0
RA175N, SgIGSF, SynCam
L1cam L1 cell adhesion molecule L1, CD171, L1-NCAM, NCAM-L1 16728 -2.2 -2.2
Pcdh7 protocadherin 7 54216 -1.7 -1.7 x
Itgal integrin alpha L Cd11a, LFA-1, Ly-15, Ly-21 16408 2.0 1.4 x x
Cd14 CD14 antigen 12475 -3.2 -2.4
Cd40 CD40 antigen
IGM, p50, Bp50, GP39, IMD3,
TRAP, HIGM1, T-BAM, Tnfrsf5
21939 -1.2 -2.4 x x
Fcgr1 Fc receptor, IgG, high affinity I CD64, IGGHAFC, FcgammaRI 14129 -2.0 -1.9 x x
Tlr1 toll-like receptor 1 21897 -2.0 -1.6
Procr protein C receptor, endothelial Ccca, EPCR 19124 2.3 2.3
a Fold change values were calculated for the comparison of C57BL/6 BMM with the baseline of BALB/c BMM of mean values from
three biological replicates per group. Positive values indicate higher expression in C57BL/6 BMM than in BALB/c BMM, whereas
negative values indicate higher expression in BALB/c BMM than in C57BL/6 BMM. Differential expression between BMM of the two
strains at the same treatment level (IFN-γ treated or medium control) using statistical testing in combination with an absolute fold
change cutoff of 1.5 is indicated in bold.
In the group of plasma membrane receptors, the adhesion molecules Cadm1, L1cam, and
Pcdh7 were higher expressed in BALB/c BMM, whereas expression of integrin Itgal was higher in
C57BL/6 BMM. The expression of CD14 / LPS receptor, CD40 / co-stimulatory receptor, Fcgr1 / high
affinity IgG Fc receptor I, and Tlr1 / toll-like receptor 1, was higher in BALB/c BMM than in
C57BL/6 BMM. Contrarily, Procr / protein C receptor exhibited higher expression in C57BL/6 BMM
(Table R.3.15).
Table R.3.16: Strain differences in gene expression of immune response related proteins, lysosomal, extracellular matrix degrading,
and detoxification enzymes between BMM of BALB/c and C57BL/6 mice.
Rosetta Resolver Annotation fold change a IFN-γ effect
gene
Entrez
description
alias
name
Gene ID
control IFN-γ BALB/c C57BL/6
Lbp lipopolysaccharide binding protein Ly88 16803 1.5 2.3
Lta4h leukotriene A4 hydrolase 16993 1.9 1.7
Oas2 2'-5' oligoadenylate synthetase 2 Oasl11 246728 -2.6 -2.5 x
Oasl1 2'-5' oligoadenylate synthetase-like 1 oasl9 231655 -2.5 -4.9 x
Arg2 arginase type II AII 11847 -5.1 -3.7
Tfpi tissue factor pathway inhibitor EPI, LACI 21788 2.0 2.1
Ctsc cathepsin C DPP1, DPPI 13032 1.5 1.7 x x
Ctse cathepsin E CE, CatE 13034 -4.3 -5.7
Ctsh cathepsin H 13036 2.7 2.0 x
Hyal1 hyaluronoglucosaminidase 1 Hya1, Hyal-1 15586 -2.2 -1.7
Mmp14 matrix metallopeptidase 14 (membrane-inserted)
MT1-MMP, MT-
MMP-1
17387 2.3 1.8
Gstm2 glutathione S-transferase, mu 2 Gstb2, Gstb-2 14863 3.5 3.3
Gstp1 glutathione S-transferase, pi 1 GstpiB 14870 1.9 2.0
a Fold change values were calculated for the comparison of C57BL/6 BMM with the baseline of BALB/c BMM of mean values from
three biological replicates per group. Positive values indicate higher expression in C57BL/6 BMM than in BALB/c BMM, whereas
negative values indicate higher expression in BALB/c BMM than in C57BL/6 BMM. Differential expression between BMM of the two
strains at the same treatment level (IFN-γ treated or medium control) using statistical testing in combination with an absolute fold
change cutoff of 1.5 is indicated in bold.
LBP, LPS binding protein, and Lta4h, whose gene product catalyzes the first reaction step in
the biosynthesis of leukotrienes, i. e. the reaction from leukotriene A4 to leukotriene B4, were
higher expressed in C57BL/6 BMM than in BALB/c BMM. In contrast, 2'-5'-oligoadenylate
synthetase genes Oas2 and Oasl1, which are part of the innate immune response, exhibited
107

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
higher expression in BALB/c BMM. The same direction of difference was observed for arginase
type II (Arg2). Tissue factor pathway inhibitor (Tfpi), coding for a protease inhibitor with anticoagulant
function, was higher expressed in C57BL/6 BMM. Genes of lysosomal, matrix
degrading, and detoxification enzymes were differentially expressed between BMM of the two
mouse strains. Lysosomal enzyme genes showed a mixed pattern with higher expression in
BALB/c BMM for cathepsin E (Ctse) and hyaluronidas (Hyal1) and higher expression in C57BL/6
BMM for cathepsins C and H (Ctsc, Ctsh). Matrix metallopeptidase 14 (Mmp14), which is involved
in degradation of extracellular matrix, and glutathione S-transferases mu 1 and pi 1 (Gstm1,
Gstp1) which are associated with detoxification processes, were detected with higher expression
in C57BL/6 BMM (Table R.3.16).
Strain-specific difference in expression of IFN-γ related genes
BMM of BALB/c and C57BL/6 differ in their ability to eliminate infecting bacteria, especially
after IFN-γ treatment (Breitbach et al. 2006). In this study, the primary reaction of BMM to IFN-γ
in comparison between both strains was analyzed. Therefore, a network centered around the
starting node IFN-γ was created using Ingenuity Pathway Analysis (IPA, www.ingenuity.com). This
network contained 181 genes (Fig. R.3.5 A, B) additionally to the starting node IFN-γ. Genes were
grouped manually into four categories: 1) difference existed between BALB/c BMM and C57BL/6
BMM at both treatment levels, control and after IFN-γ treatment; 2) difference between BMM of
the two strains occurred only at control level; 3) difference between BMM of the two strains
occurred only after IFN-γ treatment; 4) no strain difference was observed, but the gene
expression was influenced at least in BMM of one mouse strain. A second categorization divided
each group exhibiting strain differences into genes higher expressed in BALB/c BMM and into
genes higher expressed in C57BL/6 BMM.
After restricting the network to genes described to be linked to IFN-γ in macrophages or RAW
cells in the Ingenuity Pathway Knowledge Base (IPKB), a list of 132 new macrophage-related
genes out of 181 IFN-γ linked genes could be obtained. Correspondingly, a list of 49 of 181 genes
that are already included in the IPKB as IFN-γ related in macrophages or RAW cells were
exported.
Fig. R.3.5: Ingenuity Pathway Analysis (IPA, www.ingenuity.com) of IFN-γ related genes.
A, B. Network resulting from application of the grow-tool on transcriptome data
To create a network centered around the starting node IFN-γ the grow-tool of IPA‘s pathway function was applied. The addition of
further nodes according to information stored in the so-called Ingenuity Pathway Knowledge Base (IPKB) was restricted to a list of
genes that are differentially expressed in at least 1 of 4 comparisons 1) IFN-γ effects in BALB/c BMM, 2) IFN-γ effects in C57BL/6 BMM,
3) strain difference at non-treated control level and 4) strain difference after IFN-γ treatment. Further restrictions were not included.
Networks are colored according to strain difference at control level (A) or at IFN-γ treated level (B). Green represents a higher
expression in BALB/c BMM than in C57BL/6 BMM whereas red indicates higher expression in C57BL/6 BMM than in BALB/c BMM. The
intensity of color correlates with the magnitude of difference: The shade of color becomes darker with increasing absolute fold
change.
108

Maren Depke
Results
Gene Expression Pattern of Bone-Marrow Derived Macrophages after Interferon-gamma Treatment
A
regulation by IFN-γ at least
in BMM of one strain; no
significant difference in
expression between BMM
of both strains
difference between BMM of
the two mouse strains only
after IFN-γ treatment
B
higher expression
in BALB/c BMM
higher expression
in C57BL/6 BMM
difference between BMM of
the two mouse strains only
at control level
difference between BMM of
the two mouse strains at
both control and IFN-γ
treated level
regulation by IFN-γ at least
in BMM of one strain; no
significant difference in
expression between BMM
of both strains
difference between BMM of
the two mouse strains only
after IFN-γ treatment
higher expression
in BALB/c BMM
higher expression
in C57BL/6 BMM
difference between BMM of
the two mouse strains only
at control level
difference between BMM of
the two mouse strains at
both control and IFN-γ
treated level
109

cfu / S9 cell
cfu / S9 cell
% PI-negative S9 cells
Maren Depke
Results
HOST CELL GENE EXPRESSION PATTERN IN AN
IN VITRO INFECTION MODEL
Staphylococcal cfu per host cell after internalization and host cell vitality [Melanie Gutjahr,
Petra Hildebrandt]
After 2.5 h, an averaged number of 1.3 staphylococcal colony forming units (cfu)/ S9 cell were
found in internalization experiments (data courtesy of Melanie Gutjahr). Although not statistically
significant in the four biological replicates used for this study, a trend of increasing cfu / S9 cell
along with increasing infection time was discernible and led to an internalization mean value of
2.3 cfu / S9 cell 6.5 h after start of infection (Fig. R.4.1 A).
A
S9 infection proteome and transcriptome
4
3
2
1
B
100
80
60
40
20
0
RN1HG GFP RN1HG GFP
infected S9 cells infected S9 cells
2.5 h after start 6.5 h after start
of time infection after start of infection of infection
2.5 h
6.5 h
00
untreated
control
1 2non-sorted3 4 5 sorted 6 7
infected non-infected (GFP – ) infected (GFP + )
2.5 h 6.5 h 2.5 h 6.5 h 2.5 h 6.5 h
S9 cell samples
Fig. R.4.1: Staphylococcal cfu per host cell after internalization and host cell vitality.
A. Staphylococcal viable cell counts per S9 cell in infection experiments. The line indicates mean values per each analyzed time point.
B. Percentages of viable PI – S9 cells in untreated control and after infection with S. aureus RN1HG GFP determined by FACS counting.
Infected samples were analyzed for the two time points in different populations: without sorting or after sorting separately for GFP –
and GFP + S9 cells. Mean values of n = 2 experiments.
Cfu-data: courtesy of Melanie Gutjahr. FACS-data: courtesy of Petra Hildebrandt.
Host cell vitality was determined by propidium iodide (PI) staining and counting in FACS
(n = 2). Although 7.0 % and 12.1 % of cells in the infected, but non-sorted cell suspension were
positive for PI and therefore not viable 2.5 h and 6.5 h after infection, respectively, compared to
3.6 % of PI-positive dead cells in the untreated control, most of the dead, infected cells were
discarded during the sorting procedure (Fig. R.4.1 B). Thus, cell vitality was not influenced in the
subsets of sorted cells 2.5 h after infection (98.7 % in GFP + S9 cells, 99.0 % in GFP – S9 cells). After
6.5 h, a minor drop in viability was recorded for the infected samples (97.3 % in GFP + S9 cells,
98.8 % in GFP – S9 cells; data courtesy of Petra Hildebrandt).
Reproducibility of replicates and clustering of treatment group members
Staphylococcus aureus RN1HG GFP was used to infect S9 cells, a human bronchial epithelial
cell line. The study included samples of two treatments, infected green fluorescence-positive
FACS-sorted S9 cells and control cells that were exposed to medium, and analyzed the two time
points 2.5 h and 6.5 h after start of infection. Hierarchical clustering visualized the grouping of
the array data sets according to their similarity (Fig. R.4.2). Biological replicates of each treatment
111

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
and time point had the highest similarity to each other, and thus, the four treatment/time point
groups defined the three lowest level clusters (Fig. R.4.2, orange dashed line). When further
analyzing the next level of clustering the high similarity between the 2.5 h groups, i. e. infected
samples and control samples, was obvious (Fig. R.4.2, red dashed line). In the following cluster
level, the 6.5 h control samples showed higher similarity to both 2.5 h samples groups than to the
6.5 h infected samples. Summing up, the biological replicates exhibited a high reproducibility. The
highest similarity on the level of treatment groups was observed between 2.5 h infected and
control samples, and not between the control samples of both time points as it might have been
expected. The medium control samples of the 6.5 h timepoint were more similar to the 2.5 h
samples (infection and control) than to the 6.5 h infected samples.
medium control
infected; GFP positive
time point 2.5 h
time point 6.5 h
biological replicate 1
biological replicate 2
biological replicate 3
biological replicate 4
Fig. R.4.2: Hierarchical clustering of 16 array data sets from S9 infection experiment.
The following clustering algorithms were applied on log-transformed data: Agglomerative clustering with average linkage using
euclidian distance weighted by error as similarity measure. Sequences which were not expressed on all 16 arrays (defined by p > 0.01
on intensity profile level in Rosetta Resolver software) and control sequences were excluded from the cluster analysis.
The orange dashed line indicates the three lowest level clusters which led to a segmentation into the the four treatment/time point
groups. The next level of clustering grouped the 6.5 h control samples next to 2.5 h control and infected samples (red dashed line).
Comparison of treatment groups and assessment of differentially regulated genes
For identification of differential gene expression, infected samples were compared to their
time-matched controls. Additionally, same treatments were compared between the two time
points for control purposes. In detail, four comparisons of sample groups were carried out:
2.5 h infected GFP-positive samples vs. 2.5 h medium control
6.5 h infected GFP-positive samples vs. 6.5 h medium control
6.5 h infected GFP-positive samples vs. 2.5 h infected GFP-positive samples
6.5 h medium control vs. 2.5 h medium control
All comparisons are depicted by scatter plots (Fig. R.4.3). A small difference appeared
between the infected and control samples 2.5 h after infection (Fig. R.4.3, panel ), while a
strong reaction of infected cells compared to the time-matched control became visible after 6.5 h
(Fig. R.4.3, panel ). Consequently, the infected samples strongly differed between both time
points (Fig. R.4.3, panel ). But also in the control samples a certain change of signal intensities
depending on incubation time was recorded. This change mainly consisted of lower signal values
after 6.5 h than after 2.5 h (Fig. R.4.3, panel).
112

infected; 2.5 h
infected; 6.5 h
mean intensity
mean intensity
infected; 6.5 h
medium control; 6.5 h
Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
medium control; 2.5 h
infected; 2.5 h
medium control; 6.5 h
medium control; 2.5 h
Fig. R.4.3: Scatter plots comparing mean signal intensities of treatment groups.
The signals of the four groups “medium control sample after 2.5 h”, “GFP-positive sorted infected sample after 2.5 h”, “medium
control sample after 6.5 h”, and “GFP-positive sorted infected sample after 2.5 h” were plotted after combining the four biological
replicates. Control sequences and sequences that were absent on all of the arrays used for each scatter plot are not shown. (Absence
of expression is defined via a p-value > 0.01 on intensity profile level in Rosetta Resolver).
When looking at examples of signal intensity data, gene expression pattern were recognized
that were characterized by similar expression in the three groups of infected GFP-positive sorted
cells after 2.5 h, medium control cells after 2.5 h, and infected GFP-positive sorted cells after
6.5 h. A divergent signal intensity, lower or higher than in the three other groups, was noticeable
for the last group of medium control cells after 6.5 h (Fig. R.4.4). However, the time-dependent
changes in the control samples’ gene expression were not in the focus of this study, but the
infection-dependent changes should be accessed. Therefore, a strategy was defined that allowed
to exclude genes which are mainly influenced by a differing value in the group of 6.5 h medium
control cells from the list of infection-dependent differential gene expression after 6.5 h of
infection.
A
Fkbp4 (EntrezGene ID 2288)
B
Cldn1 (EntrezGene ID 9076)
2500
3000
Fig. R.4.4:
Examples of mean signal
intensities characterized by a
higher (A) and lower value
(B) for the medium control
after 6.5 h compared to the
other three sample groups.
2000
1500
1000
500
0
pMEM control
tryps.
medium ActD/NaN3
control
2.5 h 6.5 h
2500
2000
1500
1000
500
2288 FKBP4
0
pMEM control
tryps.
ActD/NaN3 medium
control
RN1HG infected pMEM control RN1HG infected
RN1HG GFP wt GFP tryps. RN1HG GFP wt GFP
infected,
FACS sorted
GFP positive
GFPpositivpositive
medium ActD/NaN3
control
infected,
FACS sorted
GFP positive
GFP-
2.5 h 6.5 h
sorted
sorted
RN1HG infected pMEM control RN1HG infected
RN1HG GFP wt GFP tryps. RN1HG GFP wt GFP
FACS infected, sorted
GFP positive
GFPpositive
ActD/NaN3 medium
control
FACS infected, sorted
GFP positive
GFP-
6.5 h positive
2.5 h
sorted
sorted
2.5 h 6.5 h
9076 CLDN1
First, differential gene expression was determined with statistical testing for the comparison
of infected GFP-positive sorted cells vs. medium control, each after 6.5 h. Second, the differential
gene expression was calculated for the comparison of infected GFP-positive sorted cells after
6.5 h with medium control cells after 2.5 h. Third, only genes which featured differential gene
expression in both comparisons were included in the list of infection-dependent differential gene
expression 6.5 h after infection. For assessing the infection-dependent differential gene
113

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
expression after 2.5 h, infected GFP-positive sorted cells after 2.5 h were statistically compared
to their time matched medium control cells.
Altogether, 86 genes were defined as differentially expressed in infection compared to control
2.5 h after infection. Of these genes, 40 possessed a minimal absolute fold change of 1.5, and 11
genes held a minimal absolute fold change of 2 (Table R.4.1 A). In comparison, for the 6.5 h time
point differential expression in infection compared to control was made up of 2296 genes, of
which 1196 passed an absolute fold change cutoff of 1.5 and still 495 passed a cutoff of 2
(Table R.4.1 B).
Comparison of infection-dependent differentially expressed genes after 2.5 h and 6.5 h
When comparing the infection-dependent changes in gene expression with a minimal
absolute fold change of 1.5 after 2.5 h and after 6.5 h, an overlap of 26 genes was recorded. This
corresponds to 65 % of the genes differentially expressed 2.5 h after infection (Fig. R.4.5 A). If
genes that did not exhibit differential expression relative to the baseline of 2.5 h medium control
had not been excluded from the list of infection-dependent regulated genes after 6.5 h, the
overlap would have even increased by 6 genes to 80 %.
A
B
2.5 h
repressed
repressed 6.5 h
14 26 1170
induced
5
333
induced
9
2
837
genes differentially expressed
in S9 cells between infection
with RN1HG GFP and medium
control 2.5 h after infection
genes differentially expressed
in S9 cells between infection
with RN1HG GFP and medium
control 6.5 h after infection
1
23
Fig. R.4.5: Comparison of infection-dependent regulated genes in S9 cells 2.5 h and 6.5 h after start of infection.
After applying a minimal absolute fold change cutoff of 1.5, 40 genes were differentially expressed in infection compared to control
2.5 h after start of infection, while the number of differentially expressed genes 6.5 h after start of infection increased to 1196. The
overlap between both lists contained 26 genes (A). This group of 26 genes in the overlap could be subdivided into 23 genes which
were induced 2.5 h as well as 6.5 h after start of infection, 2 genes which were repressed in both time points and 1 genes which was
induced 2.5 h after start of infection and repressed after 6.5 h. In general, a bigger fraction of differentially regulated genes was
induced than repressed in each time point. This summed up to 33 or 860 induced genes 2.5 h or 6.5 h after start of infection,
respectively. Repression occurred for 7 genes at the 2.5 h time point and 336 genes at the 6.5 h time point (B).
In a finer resolution, also the direction of regulation was comparable between the two
analyzed time points. Of the 26 genes, which were differentially expressed after 2.5 h as well as
after 6.5 h, 23 were induced and 2 were repressed in both time points. Only for one gene
induction was observed after 2.5 h while the expression was repressed after 6.5 h (Fig. R.4.5 B).
Concluding from this high degree of conformance in the two time points, the RNA expression
profiling 2.5 h after start of infection seemed to catch the very early beginning of the host cell’s
reaction, which was very strongly aggravated during the following 4 hours until the measurement
6.5 h after start of infection.
In each time point, induction of gene expression was more pronounced than repression.
Induction summed up to 33 or 860 genes 2.5 h or 6.5 h after start of infection, respectively.
Repression covered 7 genes at the 2.5 h time point and 336 genes at the 6.5 h time point
(Fig. R.4.5 B).
114

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
A Genes displaying statistically different expression values in the corresponding comparisons.
testing a significant with p*

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
Differentially abundant proteins in infected samples compared to medium-treated controls
[Melanie Gutjahr]
Samples for transcriptome analysis were generated from the same experiments which yielded
samples for a proteome study (Melanie Gutjahr). Here, a gel-free mass spectrometric analysis
was performed, which allowed the identification of 1049 proteins in all analyzed samples. Of
these, Melanie Gutjahr classified 112 and 224 as different in abundance 2.5 h and 6.5 h after start
of infection, respectively. At the 2.5 h time point, reduced abundance prevailed with 73 proteins
in comparison to 39 proteins with increased abundance. In contrary, 96 reduced and 128
increased proteins were observed 6.5 h after start of infection. The common signature of both
2.5 h and 6.5 h time point included 63 proteins, whereas 49 and 161 proteins were specifically
regulated at the 2.5 h and 6.5 h time point, respectively. Reverse regulation at both time points
did not exist.
UniProt accession numbers of proteins exhibiting differential abundance were mapped to
EntrezGene IDs using the UniProt ID mapping tool (www.uniprot.org). The majority of these
EntrezGene IDs was available as annotation for sequences of the GeneChip Human Gene 1.0 ST
array in Rosetta Resolver software. Finally, 98 regulated proteins were available for comparison
with transcriptome data from samples 2.5 h after infection, and at the 6.5 h time point, 198
regulated proteins could be compared to array results (Table R.4.2).
Table R.4.2: Mapping of UniProt entries to EntrezGene IDs and EntrezGene records in Rosetta Resolver software.
time
point
UniProt accession numbers
of regulated proteins
EntrezGene IDs
after ID mapping
EntrezGene IDs available in Rosetta Resolver software
for the GeneChip Human Gene 1.0 ST array
2.5 h 112 110 98
6.5 h 224 221 198
Comparison of differentially abundant proteins with differentially expressed genes
[in cooperation with Melanie Gutjahr]
At the first time point 2.5 h after start of infection, the small list of differentially expressed
genes did not overlap with the list of regulated proteins.
When comparing transcriptome and proteome results of the 6.5 h time point, an overlap of 11
genes/proteins was discernible. These included 7 (ALCAM, APOL2, ERAP1, IFIT2, IFIT3, OASL,
WARS) and 3 (FASN, KPNA2, PPME1) genes/proteins which exhibited in both analysis methods
up- and down-regulation, respectively. The remaining gene/protein DYNC1H1 was induced at
transcript level and reduced in protein abundance.
The comparison between differentially expressed genes at the earlier 2.5 h time point with
regulated proteins at the later 6.5 h time point revealed that for 2 (IFIT2, IFIT3) increased
proteins the transcripts already were induced 2.5 h after start of infection.
When the comparison was performed in reversion, i. e. when the 2.5 h time point regulated
proteins were compared with the later 6.5 h time point’s gene expression changes, 4 (CNP,
DYNC1H1, IFIT1, ZC3HAV1) proteins were reduced at 2.5 h whereas gene transcripts were
induced at 6.5 h. Furthermore, one protein (ALCAM) was increased after 2.5 h, but the transcript
was induced only at the 6.5 h time point. This gene/protein was already included in a list
mentioned above. Finally, the protein FASN was reduced 2.5 h after start of infection, but the
transcript was repressed after 6.5 h. The FASN gene/protein was already included in the list of
gene and protein repression at the 6.5 h time point (Table R.4.3).
116

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
Table R.4.3: Differentially abundant proteins which exhibit differential gene expression.
gene name
description
UniProt
accession
number
EntrezGene
ID
protein
fold change a
transcript
2.5 h 6.5 h 2.5 h 6.5 h
ALCAM activated leukocyte cell adhesion molecule Q13740 214 4.5 4.3 1.1 1.8
APOL2 apolipoprotein L, 2 Q9BQE5 23780 2.1 4.5 1.2 4.5
ERAP1 endoplasmic reticulum aminopeptidase 1 Q9NZ08 51752 N/A 5.6 -1.0 1.6
IFIT2
interferon-induced protein with
tetratricopeptide repeats 2
P09913 3433 1.1 6.2 2.4 5.2
IFIT3
interferon-induced protein with
tetratricopeptide repeats 3
O14879 3437 1.1 2.6 1.6 3.4
OASL 2'-5'-oligoadenylate synthetase-like Q15646 8638 N/A 16.9 1.1 5.1
WARS tryptophanyl-tRNA synthetase P23381 7453 -1.3 1.6 1.1 4.4
FASN fatty acid synthase P49327 2194 -1.6 -1.6 -1.1 -1.7
KPNA2
karyopherin alpha 2 (RAG cohort 1, importin
alpha 1)
P52292 3838 -1.5 -1.7 -1.3 -1.7
PPME1 protein phosphatase methylesterase 1 Q9Y570 51400 -2.9 -18.0 -1.1 -1.9
CNP 2',3'-cyclic nucleotide 3' phosphodiesterase P09543 1267 -2.4 1.2 -1.1 1.6
IFIT1
interferon-induced protein with
tetratricopeptide repeats 1
P09914 3434 -2.1 1.3 1.6 3.3
ZC3HAV1 zinc finger CCCH-type, antiviral 1 Q7Z2W4 56829 -2.0 1.2 1.4 3.4
a Fold change values were calculated for the comparison of infected GFP + S9 cell with the baseline of medium control samples.
Differential regulation is indicated in bold.
Evaluation of infection-dependent differentially expressed genes using Ingenuity Pathway
Analysis (IPA) for the two time points 2.5 h and 6.5 h after start of infection
For each of the two analyzed time points, 2.5 h and 6.5 h after start of infection, a list with
genes differentially expressed in infected samples compared to controls was uploaded to
Ingenuity Pathway Analysis (IPA, www.ingenuity.com). Adequate fold change value accompanied
the EntrezGene identifiers. A minimal absolute fold change cutoff of 1.5 had been applied.
In the statistical testing performed by the different IPA tools, the size of the input list
influences the significance levels of the results. As the two lists were very different in size (40
genes at the 2.5 h time point vs. 1196 genes at the 6.5 h time point; Table R.4.1), the significance
levels for each analysis were by far not directly comparable. Therefore, the 2.5 h data were
examined separately and additionally served for illustration purposes like overlaying gene
expression data for 2.5 h on 6.5 h data results. The analysis concerning networks and pathways
was focused on the differential gene expression at the 6.5 h time point.
In the samples of the time point 2.5 h after infection, the big fraction of genes with increased
expression was noticeable (33 of 40 genes; Fig. R.4.5 B). Strongest induction (fold change 4.2)
was observed for IL-6, a cytokine occupying a key position in regulation of the immune response.
Another immune-stimulating cytokine, IFN-β, was strongly induced (IFNB1, fold change 2.8;
rank 4 of induced genes), and the response to interferon was seen in the induction of interferoninduced
proteins IFIT2 (fold change 2.4) and IFIT3 (fold change 1.6). IFN-β also influences IL-6
expression (Fig. R.4.6).
Interestingly, prostaglandin-endoperoxide synthase 2 transcription was induced (PTGS2 alias
Cox-2, fold change 2.3; rank 6 of induced genes). This enzyme generates the inflammation
mediators prostaglandin (PG) G 2 and H 2 from arachidonic acid. From PGH 2 , the other
prostaglandins like PGE 2 are formed by different synthases. In the data set, also PGE 2 receptor
PTGER4 was induced (and induction of PGE 2 receptor PTGER2 was seen only at the 6.5 h time
point). Therefore, prostaglandin synthesis as well as prostaglandin receptor were induced on
117

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
transcriptome level at the 2.5 h time point. PTGS2 transcription is positively influenced by
IL-6. All three genes, IL6, PTGS2, and PTGER4, are indirectly induced by endothelin, EDN1, which
also has vasoconstrictor functions and itself was induced 2.5 h after start of infection (fold
change 4.2). The influence on expression is at least in parts mediated by ERK1/2, which was not
regulated itself (Fig. R.4.6).
Fig. R.4.6:
Interactions of IL-6, PTGS2,
PTGER4, EDN1, ERK1/2, IFNB1,
IFIT2, and IFIT3 in a custom
pathway design (modified from
IPA, www.ingenuity.com).
Red color indicates induction of
gene expression in infected
samples compared to control 2.5 h
after start of infection. More
intense shade of color is used for
higher absolute fold change
values. Dashed lines indicate
indirect influence on expression
(E), activation (A), phosphorylation
(P), and secretion (S), the solid line
binding (B).
These genes illustrated a beginning pro-inflammatory response. Another interesting candidate
which is assigned an important role at the interface of neuronal and immune system, the
leukemia inhibitory factor LIF, was increased in infected cells (fold change 1.6) already 2.5 h after
start of infection (and stayed induced at the 6.5 h time point). LIF belongs to the IL-6 family of
cytokines and recruits, like IL-6, the ubiquitous glycoprotein gp130 as second subunit of its
receptor.
Several genes for signal transduction molecules and transcription factors were included in the
small set of 40 differentially expressed genes 2.5 h after start of infection. An increased
expression was observed for NR4A2 (fold change 3.5), NFKBIZ (3.0), KLF4 (2.1), ATF3 (2.1), BCL6
(1.7), MYC (1.7), ZFP36L2 (1.6), and KLF6 (1.5).
Increased expression was recorded for TNFAIP3 (2.3), a gene thought to be critical for limiting
inflammation, and for ZC3H12C (1.5), a member of a family of negative regulators in macrophage
activation. In addition, CD274 (B7-H1), ligand for the immunoinhibitory receptor PD-1 on
activated B and T cells and on monocytes, was higher expressed in infected samples than in
controls. These gene expression changes alluded to a role in counterregulatory processes during
the beginning of infection and inflammation.
Other genes like NEDD9 (2.0), TNC (1.6), and RND3 (1.5) are involved in cell adhesion,
rounding and cytoskeleton organization, indicating the beginning of morphological changes in the
infected cells.
Even more, cell cycle progression seems to be negatively affected by the infection, as seen by
increased transcription of KLF4 (2.1), PPP1R15A (2.1), and MYC (1.7). Strikingly, 7 of 40 genes
were repressed 2.5 h after start of infection and all of them are functionally associated with cell
cycle progression: AURKA (−1.9), PLK1 (−1.8), HIST1H2AB (−1.6), BUB1 (−1.5) C13ORF34/BORA
(−1.5), CDCA2 (−1.5), and CDCA8 (−1.5).
About two thirds of regulated genes in the 2.5 h samples were still regulated 6.5 h after
infection, and the absolute fold change was even increased for almost all of these genes at the
later time point (Fig. R.4.7). Only one exception with reversed direction of regulation occurred,
and PTGS2 had a smaller fold change after 6.5 h than after 2.5 h although its expression still was
increased. To conclude, the changes started in the early time point were enforced in the later
time point.
118

fold change (infected vs. control)
fold change (infected vs. control)
fold change (infected vs. control)
Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
14
12
14
12
14
12
Fig. R.4.7:
2
Comparison of fold change value for differentially expressed genes both after
2.5 h and after 6.5 h.
0
Absolute fold change values for almost all of these genes increased at the later
-2
time point (black lines). PTGS2 had a smaller fold change after 6.5 h than after
2.5 h (gray dashed line) and only one gene, ADAMTS1, showed an reversed -4
direction of regulation in both time points (gray dashed/dotted line).
10
8
6
4
10
8
6
4
2
0
-2
-4
10
8
6
4
2
0
-2
-4
2.5 h 2.5 h 2.5 h 6.5 h 6.5 h 6.5 h
In IPA, a global functional analysis can be applied to the complete data set. Results offer an
overview on the data-associated biological functions with a p-value to judge on overrepresentation
of the corresponding function in the data set. The results can be grouped in three
topics: 1) Diseases and Disorders 2) Molecular and Cellular Functions, and 3) Physiological System
Development and Function. Here, especially the first and second topic were relevant for data
analysis because the data were generated in vitro. To get a first impression, only the first five
categories with smallest p-value were selected from the topics 1) and 2) for the 6.5 h time point.
In the topic “Diseases and Disorders” a very dominant association of the data set with diverse
inflammation and infection related categories was visible (Table R.4.4). Category “Cancer”
obtained the smallest p-value, but as such category collects diverse functions that might also be
immune-system associated, the result will not be interpreted as cancer like reaction of S9 cells to
infection. Indeed, in a comparison of all genes in the list associated with “Cancer” and all genes
linked to the other four immune-related categories in this analysis, approximately one third
appeared in both “Cancer” as well as in the immune-related categories.
The top categories of “Molecular and Cellular Function” mainly dealt with cellular growth,
proliferation, death or related aspects (Table R.4.4). First signs of the influence of infection on
this issue have already been recognized in the 2.5 h samples. In the later analysis time point of
6.5 h this influence was even more manifested and resulted in very low p-values.
Table R.4.4: Global functional analysis (IPA, www.ingenuity.com) of the infection-dependent differentially regulated genes in S9 cells
6.5 h after start of infection. First five categories of the topics “Diseases and Disorders” and “Molecular and Cellular Functions” are
cited with the range of p-values of the associated sub-categories.
rank
Diseases and Disorders
Molecular and Cellular Functions
category p-value range category p-value range
1 Cancer 3.44E-09 – 1.85E-03 Cellular Growth and Proliferation 1.04E-12 – 2.11E-03
2 Infection Mechanism 3.76E-09 – 1.50E-03 Cell Death 1.27E-10 – 2.21E-03
3 Antimicrobial Response 5.46E-07 – 7.96E-04 Cellular Development 4.94E-09 – 2.06E-03
4 Inflammatory Response 5.46E-07 – 1.84E-03 Cellular Function and Maintenance 8.19E-08 – 1.95E-03
5 Immunological Disease 1.12E-06 – 1.76E-03 Cell Cycle 1.22E-07 – 2.21E-03
The Canonical Pathways tool in IPA allows analysis of the data set in the context of predefined
pathways and their associated genes, which are displayed in order of cellular location and
function and which can be overlaid with gene expression data.
The first, most significant result for the infection-dependent gene expression changes in S9
cells 6.5 h after start of infection was the pathway “Interferon Signaling” (p = 5.81E−09;
ratio = 13/30; Fig. R.4.8). Induced gene expression was documented for IFN-β, the signal
transduction molecules JAK2, STAT1, and STAT2, the regulatory suppressor of cytokine signaling
119

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
SOCS1, and finally for the transcriptionally regulated genes PSMB8, TAP1, IRF1, IFI35, IFIT1, IFIT3,
OAS1, and MX1.
For the influence of staphylococcal infection on IFN signaling in in vivo kidney gene expression
please refer to Results / Kidney Gene Expression Pattern in an in vivo Infection Model / Fig. R.2.9,
page 86.
Fig. R.4.8:
Interferon Signaling (modified
from IPA, www.ingenuity.com).
Red color indicates increase of
expression. More intense shade
of color is used for higher
absolute fold change values.
The second result in significantly over-represented canonical pathways was “Role of Pattern
Recognition Receptors in Recognition of Bacteria and Viruses” (p = 3.63E−08; ratio = 20/80). This
pathway includes C3AR1 (fold change 3.6), receptor for complement component C3a, pentraxin
PTX3 (3.7), a soluble pattern recognition receptor (PRR) and inflammation regulator involved in
enhancing complement activation and clearance of apoptotic cells, toll-like receptor TLR3 (5.1)
with its signal transduction molecule MYD88 (2.3), intracellular PRR RIG-1 (DDX58; 3.2) and MDA-
5 (IFIH1; 5.6), which are involved in viral double-stranded RNA recognition, and NOD1 (2.5),
receptor for bacterial diaminopimelic acid with its signal transduction kinase RIPK2 (3.8). NOD1 is
known to activate CASP1 (4.4), a protease cleaving the inactive IL-1β precursor.
Three different 2'-5'-oligoadenylate synthetase genes OAS1/2/3 (2.5/2.4/1.9) were induced.
They are part of the innate immune response, induced by interferon, and their product
oligoadenylate activates RNASEL (1.6), which was additionally induced on mRNA level in this
experiment. Cytokine genes IL6 (11.9), IFNB1 (3.6), IL12A (4.0), CCL5 (4.0; RANTES) are included in
this pathway as end point of signal transduction starting with different TLRs via NFκB. Finally,
three subunits of PI3K, phosphoinositide-3-kinase, were induced: PIK3R1 (1.7), regulatory subunit
1/alpha, PIK3R3 (2.4), regulatory subunit 3/gamma, and PIK3CA (1.7), catalytic subunit alpha
polypeptide.
Significant over-representation was also observed for members of the pathway “Death
Receptor Signaling” (p = 1.21E−03; ratio = 12/64). The death receptor FAS (3.7) and cascade
initiating caspases CASP8 (1.6) and CASP10 (2.2) were induced, but also CFLAR (2.6; FLIP) as
inhibitor of CASP8/10. Ligands TNFSF10 (18.5; APO2L, TRAIL) and TNFSF15 (4.6; TL1) were
strongly induced, but not their receptors (DR). Death receptor’s/TNF receptor’s signal
transduction kinase RIPK1 (2.2) was induced together with MAP kinases MAP3K14 (2.0) and
MAP4K4 (2.7) and IκB kinase (IKK) subunit IKBKE (2.3). Additionally, CASP3 was slightly induced
120

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
(1.6), but CASP9, the link to the intrinsic mitochondrial apoptosis pathway, was repressed (−1.6).
Possibly, S9 cells were prepared to react to external apoptosis signals or give signals to infection
relevant immune cells like neutrophils or macrophages. Of course, such situation does not occur
in the S9 cell in vitro infection setting.
Similarly, antigen presentation will not find any reaction partner in vitro. This canonical
pathway was significantly overrepresented in the infection-dependent genes which were
differentially expressed 6.5 h after start of infection (p = 1.21E−02; ratio = 8/39; Fig. R.4.9).
Antigen presentation is divided into two parts, presentation of intracellular and extracellular
antigens. Intracellular antigens are digested by cytosolic proteasome with sequential transport of
peptides into the endoplasmatic reticulum (ER). These peptides are loaded to MHC-I molecules
and transported via the Golgi to the cell surface for presentation. The immune-proteasome gene
PSMB8 (LMP7), coding for a protein in the 20S-core of the proteasome, was induced (fold change
1.8). PSMB9 (LMP2) with a fold change of 1.499 was just marginally below the cutoff and could
be regarded as induced, too. Both induced genes lead to an immune response-specific
reorganization of the proteasomes’ catalytic center. Induced were the peptide transporters TAP1
(2.4) and TAP2 (2.2), the TAP-binding protein tapasin/TPN (TAPBP; 1.5), which attaches the TAP
molecules to the MHC-I complex, and the ER aminopeptidases ERAP1 (1.6) and ERAP2 (4.4),
which are responsible for N-terminal trimming of the peptide in the ER. At the end of this
pathway, MHC-I molecules HLA-E (1.7) and HLA-F (1.8) and MHC class I-related molecule MR1
(2.0) were induced. MR1, a non-classical MHC-gene, features a strong homology to MHC-I
molecules, is highly conserved in human and mouse, ubiquitously transcribed in different tissue
or cell types, and speculated to present an unknown invariant ligand to certain “innate” mucosaassociated
invariant T cells (MAIT cells) expressing an invariant T cell receptor (TCR) similar as
innate natural killer T cells (iNKT) (Riegert et al. 1998, Hansen TH et al. 2007). Beta-2-
microglobulin (B2M; MHC-I-β) as the second subunit of the MHC-I complex showed a trend of
induction with a fold change of 1.3, but was not significant in statistical testing.
Fig. R.4.9: Antigen Presentation (modified from IPA, www.ingenuity.com). Red color indicates increase of expression.
121

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
Immune cells and other cells of the body are capable of presenting intracellular antigens via
MHC-I. Although a significant increase in expression was observed for MHC-II gene HLA-DOP
(3.8), all other MHC-II genes for both alpha and beta subunits were not differentially expressed.
Additionally, the intensity of MHC-II gene expression is clearly among the lower values on the
whole array. This is not surprising, because presentation of extracellular antigens via MHC-II
preferentially occurs by professional antigen presenting cells (APC) like dendritic cells (DC),
macrophages and B cells, and the human bronchial epithelial cell line S9, which was used in this
study, does not belong to the group of professional APC.
For the influence of staphylococcal infection on antigen presentation in in vivo kidney gene
expression please refer to Results / Kidney Gene Expression Pattern in an in vivo Infection
Model / Fig. R.2.12, page 88.
Persistence and enhancement of S9 reaction to infection with S. aureus RN1HG from the 2.5 h
to the 6.5 h time point
It has already been mentioned that a big fraction of the genes, which were differentially
expressed at the 2.5 h time point, was still regulated at the later 6.5 h time point. This concerned
26 genes (Table R.4.5).
Table R.4.5: Overview on genes which were differentially expressed in S9 cells at both the 2.5 h and the 6.5 h time point in
comparison between infection with S. aureus RN1HG GFP and control treatment.
Rosetta Resolver annotation
fold change a
gene name
description
Entrez
Gene ID
2.5 h 6.5 h
IL6 interleukin 6 (interferon, beta 2) 3569 4.2 11.9
NFKBIZ nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, zeta 64332 3.0 7.9
IFNB1 interferon, beta 1, fibroblast 3456 2.8 3.6
IFIT2 interferon-induced protein with tetratricopeptide repeats 2 3433 2.4 5.2
PTGS2 prostaglandin-endoperoxide synthase 2 (prostaglandin G/H synthase and cyclooxygenase) 5743 2.3 1.6
KLF4 Kruppel-like factor 4 (gut) 9314 2.1 9.9
ATF3 activating transcription factor 3 467 2.1 3.9
PPP1R15A protein phosphatase 1, regulatory (inhibitor) subunit 15A 23645 2.1 2.4
EDN1 endothelin 1 1906 2.0 8.2
NEDD9 neural precursor cell expressed, developmentally down-regulated 9 4739 2.0 5.5
PMAIP1 phorbol-12-myristate-13-acetate-induced protein 1 5366 2.0 4.4
PTGER4 prostaglandin E receptor 4 (subtype EP4) 5734 1.8 5.2
ADAMTS1 ADAM metallopeptidase with thrombospondin type 1 motif, 1 9510 1.8 -1.7
BCL6 B-cell CLL/lymphoma 6 604 1.7 3.5
SAT1 spermidine/spermine N1-acetyltransferase 1 6303 1.6 4.6
ZFP36L2 zinc finger protein 36, C3H type-like 2 678 1.6 4.0
FAM46A family with sequence similarity 46, member A 55603 1.6 3.5
IFIT3 interferon-induced protein with tetratricopeptide repeats 3 3437 1.6 3.4
TNC tenascin C 3371 1.6 2.1
LIF leukemia inhibitory factor (cholinergic differentiation factor) 3976 1.6 2.2
CD274 CD274 molecule 29126 1.6 5.5
KLF6 Kruppel-like factor 6 1316 1.5 2.6
ZC3H12C zinc finger CCCH-type containing 12C 85463 1.5 2.8
C6orf141 chromosome 6 open reading frame 141 135398 1.5 3.7
CDCA8 cell division cycle associated 8 55143 -1.5 -2.1
HIST1H2AB histone cluster 1, H2ab 8335 -1.6 -2.1
a Fold change values were calculated for the comparison of infected GFP + S9 cell with the baseline of medium control samples.
Leukemia inhibitory factor (LIF) was already increased 2.5 h after start of infection (fold
change 1.6). This increase was further elevated after 6.5 h (2.2). Fittingly, increased expression
was also detectable for the LIF receptor (LIFR, 2.2) at this time point.
122

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
CD274 alias programmed cell death 1 ligand 1 (PD-L1) is involved in the regulation of
inflammatory responses (Sharpe et al. 2007). In this study, CD274 gene expression was induced
2.5 h after start of infection and further increased at the 6.5 h time point. Interestingly, at the
later time point, the second PD-1 receptor, programmed cell death 1 ligand 2 (PDCD1LG2; 2.7),
was also induced.
The prostaglandin-endoperoxide synthase 2 (PTGS2) was less induced at the 6.5 h than at the
2.5 h time point (Table R.4.5). Additionally, the intensity values for PTGS2 at 6.5 h were lower in
both control and infected samples (60 and 94) when compared to the 2.5 h control sample (280).
This might hint for an early reaction that was about to stop at the later time point. Induction of
prostaglandin receptor PTGER4 still increased from 2.5 h to 6.5 h, and induction of another
receptor, PTGER2, was only detectable at the 6.5 h time point. The reaction of the genes for
these receptors seemed to lag behind the reaction of the gene coding for the enzyme responsible
for synthesis of their ligand.
Interferon-β (IFNB1) was induced at both analyzed time points. At the 2.5 h time point,
interferon reaction has already been described in reference to the induced genes IFIT2 and IFIT3
(page 117). In the IPA canonical pathway “Interferon Signaling” (page 119), IFI35, IFIT1, IFIT3,
IRF1, and others were included as induced, interferon-regulated genes at the 6.5 h time point,
and IFIH1 has already been introduced in the IPA canonical pathway “Role of Pattern Recognition
Receptors in Recognition of Bacteria and Viruses” (page 120). Manual filtering revealed further
six interferon regulated genes IFI16, IFI44L, IFIT2, IFIT5, IRF2, and ISG20 to be induced 6.5 h after
start of infection.
A histone cluster 1 gene (HIST1H2AB) was repressed significantly with a less than −1.5x fold
change 2.5 h after start of infection. But 6.5 h after infection, a number of 17 other histone
cluster 1 genes together with HIST1H2AB were repressed. Of these 17 additional genes, 4 had
already yielded a significant p-value at the 2.5 h time point, but did not pass the 1.5 fold cutoff.
Furthermore, H1 histone family member 0 (H1F0) and H2A histone family member X (H2AFX)
were repressed at the 6.5 h time point (Table R.4.5, Table R.4.6).
Cyclins CCNB1, CCNB2, and CCNH were repressed, while CCND1 and CCNL1 were induced
6.5 h after start of infection. In this context, induced cyclin-dependent kinases / kinase-like CDK6,
CDKL2, and CDKL5 were conspicuous at the 6.5 h time point, while cyclin-dependent kinase
inhibitors CDKN1B, CDKN2B, and CDKN3 were repressed. Contrarily, the inhibitor CDKN2C was
induced. Similarly, genes coding for cell division cycle associated proteins CDCA3, CDCA4, CDCA8,
and CDC25A, and genes encoding centromer/centrosomal proteins CENPF, CEP55, CEP97,
CEP120, and CEP250 were repressed, while CDC14A and CEP170 were induced. Regulatory genes
CKS2, GAS2L3, PRC1, and SKP2 were repressed, whereas BTG1, CHEK2, and DMTF1 were
observed as induced (Table R.4.6).
In a condensed manner, histones were repressed, some cyclin-dependent kinase (CDK) were
induced, and CDK inhibitors, cyclins, cell division cycle associated proteins, centromer/
centrosomal proteins and regulatory genes were found with both repressed and induced
examples. Nevertheless, repression held a bigger fraction of this subset of genes than induction.
123

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
Table R.4.6: Overview on selected differentially expressed cell cycle and proliferation related genes in S9 cells 6.5 h after start of
infection with S. aureus RN1HG.
Rosetta Resolver annotation
fold change a
gene name
description
Entrez Gene
ID
2.5 h 6.5 h
HIST1H1A histone cluster 1, H1a 3024 -1.4 -3.0
HIST1H1B histone cluster 1, H1b 3009 -1.4 -1.7
HIST1H1C histone cluster 1, H1c 3006 -1.3 -1.9
HIST1H1D histone cluster 1, H1d 3007 -1.4 -2.2
HIST1H2AB histone cluster 1, H2ab 8335 -1.6 -2.1
HIST1H2AC histone cluster 1, H2ac 8334 -1.4 -1.6
HIST1H2AI histone cluster 1, H2ai 8329 -1.3 -1.7
HIST1H2AK histone cluster 1, H2ak 8330 -1.4 -2.2
HIST1H2BH histone cluster 1, H2bh 8345 -1.3 -1.8
HIST1H2BN histone cluster 1, H2bn 8341 -1.3 -1.5
HIST1H3A histone cluster 1, H3a 8350 -1.5 -2.1
HIST1H3B histone cluster 1, H3b 8358 -1.3 -1.5
HIST1H3F histone cluster 1, H3f 8968 -1.3 -1.8
HIST1H3I histone cluster 1, H3i 8354 -1.3 -1.7
HIST1H4B histone cluster 1, H4b 8366 -1.4 -2.0
HIST1H4C histone cluster 1, H4c 8364 -1.4 -1.6
HIST1H4D histone cluster 1, H4d 8360 -1.5 -1.7
HIST1H4F histone cluster 1, H4f 8361 -1.5 -1.8
H1F0 H1 histone family, member 0 3005 -1.0 -1.6
H2AFX H2A histone family, member X 3014 -1.2 -1.5
CCNB1 cyclin B1 891 -1.3 -1.6
CCNB2 cyclin B2 9133 -1.3 -1.6
CCNH cyclin H 902 1.0 -1.7
CCND1 cyclin D1 595 -1.1 1.7
CCNL1 cyclin L1 57018 1.4 2.1
CDK6 cyclin-dependent kinase 6 1021 -1.0 1.6
CDKL2 cyclin-dependent kinase-like 2 (CDC2-related kinase) 8999 1.0 3.2
CDKL5 cyclin-dependent kinase-like 5 6792 1.1 2.6
CDKN1B cyclin-dependent kinase inhibitor 1B (p27, Kip1) 1027 -1.2 -1.6
CDKN2B cyclin-dependent kinase inhibitor 2B (p15, inhibits CDK4) 1030 1.0 -1.9
CDKN3 cyclin-dependent kinase inhibitor 3 1033 -1.3 -1.9
CDKN2C cyclin-dependent kinase inhibitor 2C (p18, inhibits CDK4) 1031 -1.0 1.7
CDCA3 cell division cycle associated 3 83461 -1.4 -2.0
CDCA4 cell division cycle associated 4 55038 -1.0 -1.6
CDCA8 cell division cycle associated 8 55143 -1.5 -2.1
CDC25A cell division cycle 25 homolog A (S. pombe) 993 -1.0 -1.5
CDC14A CDC14 cell division cycle 14 homolog A (S. cerevisiae) 8556 -1.1 2.8
CENPF centromere protein F, 350/400ka (mitosin) 1063 -1.2 -1.9
CEP55 centrosomal protein 55kDa 55165 -1.2 -1.5
CEP97 centrosomal protein 97kDa 79598 -1.1 -1.9
CEP120 centrosomal protein 120kDa 153241 -1.0 -1.6
CEP250 centrosomal protein 250kDa 11190 -1.1 -1.6
CEP170 centrosomal protein 170kDa 9859 -1.1 1.8
CKS2 CDC28 protein kinase regulatory subunit 2 1164 -1.2 -1.7
GAS2L3 growth arrest-specific 2 like 3 283431 -1.4 -2.1
PRC1 protein regulator of cytokinesis 1 9055 -1.2 -1.6
SKP2 S-phase kinase-associated protein 2 (p45) 6502 -1.2 -1.5
BTG1 B-cell translocation gene 1, anti-proliferative 694 1.4 1.8
CHEK2 CHK2 checkpoint homolog (S. pombe) 11200 -1.1 3.2
DMTF1 cyclin D binding myb-like transcription factor 1 9988 1.1 1.9
a Fold change values were calculated for the comparison of infected GFP + S9 cell with the baseline of medium control samples.
Differential regulation is indicated in bold.
124

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
Insights into infection-dependent differential gene expression by selected examples from the
6.5 h time point
Infection of S9 cells with S. aureus RN1HG and internalization of the pathogen affected
metabolism related genes 6.5 h after start of infection.
Interestingly, the second most strongly induced gene was indoleamine 2,3-dioxygenase 1
(IDO1, fold change 23.5), a key molecule in immunomodulation, microbial growth control, and
pathogen immune escape (Zelante et al. 2009). The enzyme catalyzes the reaction of tryptophan
to formylkynurenine. In relation to IDO1, the increased expression of WARS (4.4) whose gene
product is responsible for tryptophan tRNA charging attracted attention.
Lipid metabolism related genes regulated 6.5 h after start of infection in S9 cells featured a
decreased expression in infected samples leading to the assumption of a reduced de novo
lipogenesis (Fig. R.4.10).
This concerned different types of lipids. In the mevalonate pathway and the following
cholesterol biosynthesis steps, five genes were repressed: mevalonate kinase (MVK; −2.2),
farnesyl-diphosphate farnesyltransferase 1 (FDFT1; −1.7), sterol-C4-methyl oxidase-like (SC4MOL;
−2.0), squalene epoxidase (SQLE; −1.8) and NAD(P) dependent steroid dehydrogenase-like
(NSDHL; −1.7). The induced genes of cholesterol 25-hydroxylase (CH25H; 6.0) and oxysterol
binding protein-like 10 (OSBPL10; 1.8) are involved in negative feedback on cholesterol
biosysnthesis. The repression of fatty acid synthase (FASN; −1.7) is linked to two genes indirectly
involved in lipid metabolism. Deiodinase type II (DIO2; −4.1) catalyzes the deiodization of
thyroxine (T 4 ) to the main biologically active form triiodothyronine (T 3 ), which is known to induce
lipogenesis. Although there was probably no hormone T 4 to be converted in the in vitro situation
used for this study, the repression of DIO2 might reflect a reaction that makes sense in vivo for
bronchial epithelial cells leading to less local activation of lipogenesis. In contrast, glucagon (GCG)
is known to suppress lipogenesis (Hillgartner et al. 1995). This hormone was induced (3.6) in
infected S9 cells. Finally, the two genes for mitochondrial glycerol-3-phosphate acyltransferase
(GPAM; −1.6) and 1-acylglycerol-3-phosphate O-acyltransferase 5 (AGPAT5; −1.5), which are
involved in triacylglycerol biosynthesis, were repressed in infected samples.
Fig. R.4.10:
Repression of enzyme genes related to
lipid metabolism.
Lipid metabolism related genes were
chosen with the help of omics-viewer(s)
of BIOCYC (SRI International, CA, USA,
http://biocyc.org) and manually added
to and arranged in the Ingenuity
Pathway Analysis path designer tool
(IPA, www.ingenuity.com). Green color
indicates repression of gene expression
in infected samples at the 6.5 h time
point, red color induction of gene
expression, and more intense color
shows a higher absolute fold change.
Enzymes mevalonate kinase (MVK),
farnesyl-diphosphate farnesyltransferase
1 (FDFT1) in mevalonate
pathway, and sterol-C4-methyl oxidaselike
(SC4MOL), SQLE (squalene
epoxidase), and NAD(P) dependent steroid dehydrogenase-like (NSDHL) in the following steps of cholesterol biosynthesis were
repressed. CH25H (cholesterol 25-hydroxylase) and OSBPL10 (oxysterol binding protein-like 10), which are involved in negative
feedback on cholesterol biosynthesis, were induced. Fatty acid synthase (FASN), the enzyme of fatty acid biosynthesis, was repressed,
and also deiodinase type II (DIO2), an enzyme catalyzing the deiodization of thyroxine (T 4) to the main biologically active form
triiodothyronine (T 3). T 3 induces (A) de novo lipogenesis, while glucagon (GCG) suppresses (I) it. The enzymes mitochondrial glycerol-3-
phosphate acyltransferase (GPAM) and 1-acylglycerol-3-phosphate O-acyltransferase 5 (AGPAT5) are involved in triacylglycerol
biosynthesis. Their expression was reduced 6.5 h after start of infection.
125

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
Enzymes catalyzing synthesis or degradation reactions of lipids are in certain cases also
related to signal transduction and regulatory processes in the cell. In this context, a cumulation of
increased expression values in infected samples 6.5 h after start of infection was noticeable
(Fig. R.4.11).
Phosphatidylinositol-4-kinase and phosphoinositide-3-kinase subunits (PI4K2B, 2.1; PIK3CA,
1.7; PIK3R1, 1.7; PIK3R3, 2.4) belong to the group of genes coding for enzymes involved in lipid
messenger reactions. These enzymes catalyze reactions ending with the production of
phosphatidylinositol-trisphosphate (PIP 3 ), a messenger involved in activation of protein kinase
AKT/PKB. Phospholipases C (PLCB4, 2.9; PLCG2, 3.9; PLCH1, 2.0) generate the messengers
inositol-trisphosphate (IP 3 ) and diacylglycerol (DAG). Cytosolic phospholipase A2 gamma
(PLA2G4C, 2.7) releases among others arachidonic acid, which can serve as messenger with the
function of directly activating ion channels and protein kinase C (Ordway et al. 1991, Khan WA et
al. 1995, Bonventre 1992) or can be processed to eicosanoid mediators like prostaglandins
(Laye/Gill 2003). Interestingly, the acyl-CoA synthetase (ACSL3, 1.7) with preference for
arachidonate, myristate and eicosapentaenoate was increased, too. This might indicate an
inactivation process for the messenger arachidonic acid. Furthermore, PRKD1 (2.0) and PRKD2
(2.5), coding for protein kinase D alias protein kinase C isoform µ, were induced, too (not shown).
Three more genes showed an increase of gene expression 6.5 h after start of infection: Serine
palmitoyltransferase subunit (SPTLC2, 1.8) and ketodihydrosphingosine reductase (KDSR, 1.6),
which are part of ceramide biosynthesis, and alkaline ceramidase (ACER2, 4.1), an enzyme of
ceramide degradation to sphingosine. Noticeably, both final products ceramide and sphingosine
are associated with apoptosis.
3-phosphoinositide
biosynthesis
phospholipases
activation of
fatty acid by CoA
ceramide
biosynthesis
sphingosine
metabolism
preference for
arachidonate,
myristate,
eicosapentaenoate
palmityl-CoA
ceramide
PIP 3
IP 3 + DAG
arachidonic
acid
arachidonyl-
CoA
ceramide
sphingosine
activation of
AKT/PKB
activation of
calcium-channels
and PKC
activity regulation
of ion channels
and PKC
inactivation of
arachidonic acid
apoptosis
Fig. R.4.11: Induction of enzyme genes related to lipid messenger production.
Lipid metabolism related genes were chosen with the help of omics-viewer(s) of BIOCYC (SRI International, CA, USA, http://biocyc.org)
and manually added to and arranged in the Ingenuity Pathway Analysis path designer tool (IPA, www.ingenuity.com). Red color
indicates induction of gene expression in infected samples at the 6.5 h time point, and more intense color shows a higher absolute
fold change. Genes coding for enzymes, which are involved in lipid messenger reactions, were induced: phosphatidylinositol 4-kinase
type 2 beta (PI4K2B), phosphoinositide-3-kinase, catalytic, alpha polypeptide (PIK3CA), phosphoinositide-3-kinase, regulatory subunit
1/alpha (PIK3R1), and phosphoinositide-3-kinase, regulatory subunit 3/gamma (PIK3R3). These enzymes catalyze reactions ending
with the production of phosphatidylinositol-trisphosphate (PIP 3), a messenger involved in activation of protein kinase AKT/PKB.
Phospholipase C, beta 4 (PLCB4), phosphatidylinositol-specific phospholipase C, gamma 2 (PLCG2), and phospholipase C, eta 1 (PLCH1)
generate the messengers inositol-trisphosphate (IP 3) and diacylglycerol (DAG), cytosolic, calcium-independent phospholipase A2,
group IV C (PLA2G4C) generates the messenger arachidonic acid. Acyl-CoA synthetase long-chain family member 3 (ACSL3) was
increased, too. This might indicate in inactivation process for the messenger arachidonic acid.
Serine palmitoyltransferase, long chain base subunit 2 (SPTLC2) and 3-ketodihydrosphingosine reductase (KDSR), which are part of
ceramide biosynthesis, and alkaline ceramidase 2 (ACER2), an enzyme of ceramide degradation to sphingosine, showed a higher
expression in infected samples than in controls at the 6.5 h time point.
126

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
Differential expression of apoptosis associated genes has already been observed in the data
evaluation with Ingenuity Pathway Analysis tools (see pathway “Death Receptor Signaling”,
page 120). Further apoptosis related genes were found when manually filtering the list of
differentially expressed genes at the 6.5 h time point (Table R.4.7). Genes for apoptosis inducing
proteins B-cell CLL/lymphoma 10 (BCL10, 1.7) and apoptosis facilitator BCL2-like 11 (BCL2L11,
2.0) were increased in infected samples. The proteins belong to the group of BH3-only proteins
and achieve their apoptotic effect by binding to proteins of the Bcl-2 family. The mechanism is
discussed to take effect by activating pro-apoptotic members or inactivating pro-survival
members in a de-repression mode (Bouillet/Strasser 2002, Adams 2003, Fletcher/Huang 2006).
Also apolipoprotein L1 (APOL1) is a BH3-only protein whose accumulation leads to autophagy
(Wan et al. 2008, Zhaorigetu et al. 2008) and is postulated to be linked to apoptosis
(Vanhollebeke/Pays 2006). Its expression was increased by a factor of 3.4 in infected samples at
the 6.5 h time point. Interestingly, four other apolipoprotein L genes, which do not possess a
secretion signal peptide like APOL1 and therefore are thought to be localized intracellularly
(Duchateau et al. 1997, Page et al. 2001), showed increased expression: APOL2 (4.6), APOL3 (3.7),
APOL4 (1.9), and APOL6 (4.2). Expression of APOL5 was absent (p > 0.01 on intensity profile level
in Rosetta Resolver analysis) in all 16 infected and control S9 cell samples of this study. These
other APOL proteins also contain BH3-domains and are supposed to be associated with
programmed cell death and immune response (Liu Z et al. 2005). Most interestingly, APOL1 is the
target protein of the antagonistic Trypanosoma brucei rhodesiense protein SRA which helps the
pathogen to evade the immune response. Further pro-apoptotic genes were induced in infected
S9 cells like BH3-like motif containing, cell death inducer (BLID; 2.3), BCL2-antagonist/killer 1
(BAK1; 1.9), caspase 4, apoptosis-related cysteine peptidase (CASP4; 2.8), but also an antiapoptotic
gene, BCL2-associated athanogene (BAG1; 1.8; Table R.4.7).
Table R.4.7: Overview on selected differentially expressed apoptosis related genes in S9 cells 6.5 h after start of infection with
S. aureus RN1HG GFP.
Rosetta Resolver annotation
fold change a
gene
name
description
Entrez
Gene ID
BCL10 B-cell CLL/lymphoma 10 8915 CLAP, mE10, CIPER, c-E10, CARMEN 1.7
BCL2L11 BCL2-like 11 (apoptosis facilitator) 10018
BAM, BIM, BOD, BimL, BimEL, BIMbeta6,
BIM-beta7, BIM-alpha6
2.0
APOL1 apolipoprotein L, 1 8542 APOL, APO-L, APOL-I 3.4
APOL2 apolipoprotein L, 2 23780 APOL3, APOL-II 4.5
APOL3 apolipoprotein L, 3 80833 CG12-1, APOLIII 3.7
APOL4 apolipoprotein L, 4 80832 APOLIV,APOL-IV 1.9
APOL6 apolipoprotein L, 6 80830 APOLVI, APOL-VI 4.2
BLID BH3-like motif containing, cell death inducer 414899 BRCC2 2.3
BAK1 BCL2-antagonist/killer 1 578 CDN1, BCL2L7, BAK-LIKE 1.9
CASP4 caspase 4, apoptosis-related cysteine peptidase 837 TX, ICH-2, Mih1/TX, ICEREL-II 2.8
BAG1 BCL2-associated athanogene 573 RAP46 1.8
a Fold change values were calculated for the comparison of infected GFP + S9 cell with the baseline of medium control samples.
alias
6.5 h
Manual filtering revealed induced cytokines and cytokine receptors 6.5 h after start of
infection, some of which have already been mentioned in the context of IFN signaling or pattern
recognition receptors (Table R.4.8). Only IFNB1 and IL6 were already induced at the 2.5 h time
point. In total, the induced genes revealed a pro-inflammatory response: the acute phase
response and fever mediator IL-6, inducers of MHC-I molecules and antiviral/antibacterial
mechanisms (IFNB1, IL28B), chemoattractants for monocytes, T cells, denditic cells and
granulocytes (CCL2, CCL5, CXCL10, CXCL11, CXCL16), B and T cell activating cytokines (IL-7,
TNFSF13B), macrophage growth factor (CSF1), inducer of other cytokines and immune cell
127

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
differentiation (IL12A), epithelial receptor for the inflammatory cytokine IL-22 (IL22RA1),
scavenger receptor for cytokines (CCRL1), immunomodulatory cytokines (IL28A, IL28B, IL29), and
apoptosis inducers (TNFSF10, TNFSF15), but also anti-apoptotic IL-15 and its receptor IL15RA.
Table R.4.8: Overview on differentially expressed cytokines and receptors in S9 cells 6.5 h after start of infection with S. aureus RN1HG
GFP.
Rosetta Resolver annotation
fold change a
gene
name
description
Entrez
Gene ID
CCL2 chemokine (C-C motif) ligand 2 6347 MCP1 2.8
CCL5 chemokine (C-C motif) ligand 5 6352 RANTES 4.0
CCRL1 chemokine (C-C motif) receptor-like 1 51554 CCR11 2.3
CSF1 colony stimulating factor 1 (macrophage) 1435 MCSF 3.1
CXCL10 chemokine (C-X-C motif) ligand 10 3627 IP-10, INP10, IFI10 28.1
CXCL11 chemokine (C-X-C motif) ligand 11 6373 IP9, I-TAC 7.0
CXCL16 chemokine (C-X-C motif) ligand 16 58191 SR-PSOX 1.8
CXCR4 chemokine (C-X-C motif) receptor 4 7852 CD184, SDF1R, LESTR 1.6
IFNB1 interferon, beta 1, fibroblast 3456 3.6
IL6 interleukin 6 (interferon, beta 2) 3569 HGF,HSF,BSF2,IFNB2 11.9
IL7 interleukin 7 3574 2.0
IL12A interleukin 12A 3592 P35,CLMF,NKSF1 4.0
IL15 interleukin 15 3600 3.4
IL15RA interleukin 15 receptor, alpha 3601 2.1
IL22RA1 interleukin 22 receptor, alpha 1 58985 1.8
IL28A interleukin 28A (interferon, lambda 2) 282616 IFNL2 5.1
IL28B interleukin 28B (interferon, lambda 3) 282617 IFNL3 3.0
IL29 interleukin 29 (interferon, lambda 1) 282618 IFNL1 5.7
TNFSF10 tumor necrosis factor (ligand) superfamily, member 10 8743 TRAIL, TL2, APO2L, CD253 18.5
TNFSF13B tumor necrosis factor (ligand) superfamily, member 13b 10673 DTL, BAFF, BLYS, CD257 2.2
TNFSF15 tumor necrosis factor (ligand) superfamily, member 15 9966 TL1, VEGI 4.6
a Fold change values were calculated for the comparison of infected GFP + S9 cell with the baseline of medium control samples.
alias
6.5 h
Further interesting induced genes included myxovirus (influenza virus) resistance 1 and 2
(interferon-inducible protein p78, MX1; 2.2; MX2; 1.9). These genes code for interferon-induced,
antiviral proteins forming a ring-shaped oligomeric structure, which recognizes and sequesters
viral nucleocapsid proteins (Gao et al. 2010). MX1 and MX2 belong to the same family of
interferon-inducible GTPases as the four guanylate binding proteins 1, 3, 4, 5 (interferoninducible,
67kDa, GBP1; 2.6; GBP2; 2.9; GBP3; 14.9; GBP4; 11.1), which were induced in S9 cells
6.5 h after start of infection with S. aureus RN1HG GFP. These GBPs are involved in immune
defense, probably oligomerize, act antivirally and inhibit cellular proliferation (MacMicking 2004).
The fifth member of the p65 GBP family, GBP6, was not expressed on all 16 arrays in this study
(Table R.4.9).
The endothelial protein C receptor (EPCR; PROCR; 1.6) was also induced in infected S9 cells.
Protein C activation by thrombin/thrombomodulin is enhanced when it is bound to EPRC, and
subsequently, activated protein C can act in an anticoagulant and anti-inflammatory manner. In
inflammation, where coagulation is facilitated by different mechanisms, e. g. via C-reactive
protein CRP, the induction of PROCR might indicate a counterregulatory process (Esmon 2006).
Interestingly, the receptor shedding from the cellular surface is mediated by tumor necrosis
factor-alpha converting enzyme / TACE, alias ADAM metallopeptidase domain 17 / ADAM17 (Qu et
al. 2006), which was also induced in the infected S9 cells (ADAM17; 1.8; Table R.4.9).
Further gene expression changes in infected S9 cells could be assigned to a related
physiological aspect: Tissue factor pathway inhibitor 2 (TFPI2; 2.2) belongs like activated
protein C to the natural anticoagulants (Esmon 2006).
128

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
Plasminogen activator urokinase (PLAU, −2.0) activates plasmin and therefore leads to anticoagulant
activity. But its proteolytic activity also leads to degradation of extracellular matrix
(Dass et al. 2008). The expression of PLAU was decreased in infected S9 cells. Contrarily,
plasminogen activator urokinase receptor (PLAUR; 4.0), receptor for PLAU, was induced. The
receptor promotes and localizes extracellular protease activity and plasmin formation by PLAU,
but also participates in internalization of complexes between PLAU and its inhibitors (Cubellis et
al. 1990, Dass et al. 2008, Blasi/Sidenius 2009). Plasminogen activator inhibitor-1 (PAI-1) is the
main inhibitor of PLAU, and it was induced in infected S9 cells (SERPINE1; 3.3; Table R.4.9)
In infected S9 cells, induction was observed for mitochondrial superoxide dismutase 2 (SOD2;
2.6), a gene known to be inducible in response to oxidative stress and by inflammatory cytokines
like IL-6 (Dougall/Nick 1991, Miao et al. 2009). Furthermore, thioltransferase glutaredoxin (GLRX;
1.9) and thioredoxin interacting protein (TXNIP; 2.2), which are involved in anti-oxidant defense,
were induced (Table R.4.9).
Four components of the complement system were induced in infected S9 cells at the 6.5 h
time point: complement component 1, r subcomponent-like and s subcomponent (C1RL; 1.7;
C1S; 1.7) of the classical pathway, complement factor B (CFB; 3.1) of the alternative pathway and
complement component 3a receptor 1 (C3AR1; 3.6; Table R.4.9).
Lysosomal enzyme genes of cathepsin S (CTSS; 2.6) and legumain (LGMN; 2.3) and
additionally, lysosomal membrane protein genes LAMP3 (5.7) and LAPTM5 (1.9) were induced.
Table R.4.9: Overview on selected differentially expressed immune defense related genes in S9 cells 6.5 h after start of infection with
S. aureus RN1HG GFP.
Rosetta Resolver annotation
fold change a
gene name
description
Entrez
Gene ID
MX1 myxovirus (influenza virus) resistance 1, interferoninducible
4599 MxA, IFI78 2.2
protein p78 (mouse)
MX2 myxovirus (influenza virus) resistance 2 (mouse) 4600 MXB 1.9
GBP1 guanylate binding protein 1, interferon-inducible, 67kDa 2633 2.6
GBP3 guanylate binding protein 3 2635 2.9
GBP4 guanylate binding protein 4 115361 Mpa2 14.9
GBP5 guanylate binding protein 5 115362 11.1
PROCR protein C receptor, endothelial (EPCR) 10544 EPCR, CCD41, CD201 1.6
ADAM17 ADAM metallopeptidase domain 17 6868 CSVP, TACE, CD156B 1.8
TFPI2 tissue factor pathway inhibitor 2 7980 PP5, REF1 2.2
PLAU plasminogen activator, urokinase 5328 ATF, UPA, URK, u-PA -2.0
PLAUR plasminogen activator, urokinase receptor 5329 CD87, UPAR, URKR 4.0
SERPINE1 serpin peptidase inhibitor, clade E (nexin, plasminogen 5054 PAI, PAI-1, PLANH1 3.3
activator inhibitor type 1), member 1
SOD2 superoxide dismutase 2, mitochondrial 6648 IPO-B, Mn-SOD 2.6
GLRX glutaredoxin (thioltransferase) 2745 GRX, GRX1 1.9
TXNIP thioredoxin interacting protein 10628 TXNIP, THIF, VDUP1 2.2
C1RL complement component 1, r subcomponent-like 51279 1.7
C1S complement component 1, s subcomponent 716 1.7
CFB complement factor B 629 3.1
C3AR1 complement component 3a receptor 1 719 3.6
CTSS cathepsin S 1520 2.6
LGMN legumain 5641 AEP, PRSC1 2.3
LAMP3 lysosomal-associated membrane protein 3 27074 CD208, TSC403 5.7
LAPTM5 lysosomal multispanning membrane protein 5 7805 CLAST6 1.9
a Fold change values were calculated for the comparison of infected GFP + S9 cell with the baseline of medium control samples.
alias
6.5 h
Several adhesins were induced in S. aureus RN1HG GFP infected S9 cells, like integrins alpha
and beta (ITGA2; 2.9; ITGA5; 2.0; ITGA11; 1.6; ITGB8; 2.6) and their regulator cytohesin 1 (CYTH1;
2.1), protocadherins (PCDH7; 4.1; PCDH17; 2.1) and cadherin CDH8 (1.6). Another cadherin was
129

Maren Depke
Results
Host Cell Gene Expression Pattern in an in vitro Infection Model
repressed (CDH10, −1.9). Increased expression was also observed for adhesion molecules ALCAM
(1.8) and CEACAM1 (9.9). ALCAM was one of the genes whose protein abundance change was in
accordance with the differential expression (Table R.4.10, Table R.4.3).
Table R.4.10: Overview on selected differentially expressed cell adhesion related genes in S9 cells 6.5 h after start of infection with
S. aureus RN1HG GFP.
gene
name
description
Rosetta Resolver annotation
Entrez
Gene
ID
alias
fold change a
ITGA2 integrin, alpha 2 (CD49B, alpha 2 subunit of VLA-2 receptor) 3673 BR, GPIa, CD49B,VLA-2 2.9
ITGA5 integrin, alpha 5 (fibronectin receptor, alpha polypeptide) 3678 FNRA, CD49e,VLA5A 2.0
ITGA11 integrin, alpha 11 22801 1.6
ITGB8 integrin, beta 8 3696 ITGB8 2.6
ITGB1BP2 integrin beta 1 binding protein (melusin) 2 26548 CHORDC3, ITGB1BP 1.8
CYTH1 cytohesin 1 9267 B2-1,SEC7,PSCD1 2.1
PCDH7 protocadherin 7 5099 BHPCDH 4.1
PCDH17 protocadherin 17 27253 PCH68, PCDH68 2.1
CDH8 cadherin 8, type 2 1006 1.6
CDH10 cadherin 10, type 2 (T2-cadherin) 1008 -1.9
ALCAM activated leukocyte cell adhesion molecule 214 MEMD, CD166 1.8
CEACAM1
carcinoembryonic antigen-related cell adhesion molecule 1
(biliary glycoprotein)
634 BGP, BGP1, BGPI 9.9
a Fold change values were calculated for the comparison of infected GFP + S9 cell with the baseline of medium control samples.
6.5 h
130

Maren Depke
Results
PATHOGEN GENE EXPRESSION PROFILING
Growth Media Comparison Study
Reproducibility of replicates and clustering of experimental condition groups
Cultivation of S. aureus RN1HG in pMEM medium was performed in triplicate. In the tiling
array analysis of the sample points exponential growth, stationary phase t 2 , and stationary
phase t 4 , the replicates of each group were arranged together in a cluster. As expected, for the
stationary t 4 time point, which is defined as 4 h after entry into stationary phase, a higher
similarity to the t 2 samples, which were harvested 2 h earlier, than to the exponential growth
samples was detected. This similarity was especially visible for the t 4 biological replicates 2 and 1,
whereas the replicate 3 held more distance to the other samples (Fig. R.5.1).
exponentialgrowth
stationary phase t 2
stationary phase t 4
biological replicate 1
biological replicate 2
biological replicate 3
Fig. R.5.1: Hierarchical clustering of 9 tiling array data sets from growth in pMEM medium.
The following clustering algorithms were applied on z-score-transformed data: Agglomerative clustering with average linkage using
cosine correlation as similarity measure. All sequences were included in the cluster analysis.
Three major classes according to the sample points during growth were discernible: 1) exponential growth, 2) stationary phase t 2, and
3) stationary phase t 4. As expected for the stationary t 4 time point, a higher similarity to the t 2 samples than to the exponential growth
samples was detected. This similarity was especially visible for the t 4 biological replicates 2 and 1, whereas the replicate 3 held more
distance to the other samples.
Comparison of experimental condition groups and assessment of differentially regulated genes
The consumption of medium components and the reduction of growth rate after beginning of
stationary phase was accompanied by a substantial change of gene expression. This change was
visualized by the strong scattering when comparing stationary phase t 2 and stationary phase t 4
samples to the exponential growth in scatter plots (Fig. R.5.2).
The comparison of stationary phase samples with exponential growth samples by statistical
testing resulted in 2243 and 1399 sequences, which exhibited a significant difference to
exponential growth at the t 2 and the t 4 time point, respectively. Of these, 1423 sequences
exceeded an absolute fold change of 2 in the t 2 samples, and 1086 sequences passed the cutoff in
the 2 h later samples of the t 4 time point (Table R.5.1).
131

Maren Depke
Results
Pathogen Gene Expression Profiling
stationary phase t 2
stationary phase t 4
exponential growth
exponential growth
Fig. R.5.2: Scatter plots comparing mean signal intensities of treatment groups.
The signals of the three groups of “stationary phase t 2”, “stationary phase t 4”, and “exponential growth phase” were plotted after
combining the three biological replicates. All sequences available on the array are shown.
Table R.5.1: Results of group comparisons with statistical testing of stationary phase and exponential growth staphylococcal array
data sets.
number of sequences
group comparison a
significant with p* < 0.05 in textbook oneway
ANOVA with Benjamini-Hochberg False
Discovery Rate multiple testing correction
absolute fold change equal
to or greater than 2 AND
significant with p* < 0.05
stationary phase t 2 vs. exponential growth phase 2243 1423
stationary phase t 4 vs. exponential growth phase 1399 1086
a The 080604_SA_JH_Tiling array contains 3882 sequences of which 2825 are known transcripts. The remaining 1057 sequences were
identified as new transcripts in the tiling array analysis. All sequences were included in statistical testing.
Comparison of differentially regulated genes in t 2 and t 4 of stationary phase in pMEM
The gene expression signatures of t 2 and t 4 stationary phase S. aureus RN1HG, which resulted
from statistical comparison with the baseline of exponential growth including multiple testing
correction and minimal absolute fold change cutoff 2, were compared in a Venn diagram
(Fig. R.5.3 A). For 940 sequences, differential expression was observed in both time points of
stationary phase, while 483 and 146 sequences were specific for the t 2 and t 4 stationary phase
signature, respectively.
This corresponds to a fraction of 66 % of sequences regulated at t 2 , which were also
differentially expressed at t 4 . Vice versa, 87 % of differentially expressed sequences at the t 4 time
point were also found to be regulated at t 2 . When examining the induced and repressed
sequences separately, it was first obvious that sequences regulated at both time points always
possessed the same regulation direction in these two time points (Fig. R.5.3 B). Further, a slightly
higher number of repressed sequences was visible in the differentially expressed sequences, but
induction accounted for almost the same number of sequences as repression.
In the t 2 samples, 696 sequences were induced (225 specifically for that time point), 727 were
repressed (258 specifically for that time point), while in the t 4 samples 540 exhibited induction
(69 specifically for that time point) and 546 exhibited repression (77 specifically for that time
point). Therefore, the 940 sequences differentially expressed in both time points consisted of 471
induced and 469 repressed sequences.
132

Maren Depke
Results
Pathogen Gene Expression Profiling
A
B
483 940 146
repressed
repressed
t 2
t 4
induced
258 77
induced
225
469
69
sequences differentially expressed
between stationary growth phase t 2
and exponential growth phase of
S. aureus RN1HG cultivated in pMEM
sequences differentially expressed
between stationary growth phase t 4
and exponential growth phase of
S. aureus RN1HG cultivated in pMEM
471
Fig. R.5.3: Comparison of the t 2 and t 4 stationary phase signatures of S. aureus RN1HG in pMEM.
Both stationary phase samples were compared to the baseline of exponential growth with statistical testing and multiple testing
correction (p* < 0.05), and a minimal absolute fold change cutoff of 2 was applied. The comparison of differentially expressed
sequences at the t 2 and t 4 time point was performed for all regulated sequences (A) and for induced and repressed sequences
separately (B).
Fractions of known genes and newly identified transcribed sequences included on the tiling
array and in the lists of differentially expressed sequences derived from the comparisons of
different growth phase samples
The 080604_SA_JH_Tiling array contains 3882 sequences of which 2825, corresponding to
73 %, are known transcripts and associated with a LocusTag number (SAOUHSC_*). The remaining
1057 sequences were identified as new transcripts in the tiling array analysis. These new
transcripts are expected to belong to all types of RNAs like mRNA and regulatory RNA, but also
formerly unknown transcribed fragments 5’ and 3’ to known genes might be included.
Furthermore, some artifacts might still be included. All sequences, known and new, were
included in statistical testing.
For the stationary phase-specific signature, 444 and 307 differentially expressed transcripts
belonged to the group of new transcripts at the t 2 and t 4 time point, respectively (Table R.5.2).
Table R.5.2: Fractions of known annotated and newly detected transcripts in the results of group comparisons with statistical testing
of different growth phase array data sets.
group comparison a
total number of
sequences with
absolute fold change
equal to or greater
than 2 AND significant
with p* < 0.05
number of known
transcribed sequences
with absolute fold
change equal to or
greater than 2 AND
significant with p* < 0.05
number of new
transcribed sequences
with absolute fold
change equal to or
greater than 2 AND
significant with p* < 0.05
stationary phase t 2 vs. exponential growth phase 1423 979 444
stationary phase t 4 vs. exponential growth phase 1086 779 307
a The 080604_SA_JH_Tiling array contains 3882 sequences of which 2825 are known transcripts. The remaining 1057 sequences were
identified as new transcripts in the tiling array analysis. All sequences were included in statistical testing.
Physiological aspects of stationary phase response
Stationary phase is characterized by a strongly reduced growth rate and by a reorganization of
the metabolic processes of the bacterial cell. Causative for these alterations is the consumption
of nutrients of the growth medium and the following starvation for the preferred carbon source.
Such changes have been published for S. aureus COL during growth in TSB in a global gel-based
and gel-free proteome study by Kohler et al. in 2005.
133

Maren Depke
Results
Pathogen Gene Expression Profiling
The tiling array approach aimed at the identification of the maximal possible number of
transcriptional units − known and newly identified – and therefore included several different
growth media in a cooperation between several laboratories. On the other hand, the tiling array
analysis could be used to get insights into the stationary phase response in the newly established
pMEM medium on transcriptome level under restriction to the already defined transcriptional
units and annotated genes. For this purpose, differential expression was assumed for genes
regulated in at least one of the two analyzed time points of stationary phase, t 2 and t 4 .
Using this application of the array data set, a very clear picture of induced TCA cycle enzyme
genes (citZ, citB, citC, sucA, sucB, sucD, sucC, sdhA, sdhB, sdhC; all induced) was revealed. In
parallel, reduced expression of glycolysis enzyme genes (pgi, pfkA, pgk, pgm, pykA; all repressed)
and induced expression of gluconeogenetic enzyme genes (pckA, gap2, fbp) was observed. This is
in accordance with published stationary phase response of S. aureus (Kohler et al. 2005).
Furthermore, repression of amino acid biosynthesis pathways was described for the stationary
phase. In pMEM, tyryptophan (trpB, trpC, trpD, trpE, trpG), histidine (hisB, hisD, hisG), and
aspartate/arginine (argG, argH, SAOUHSC_00150) biosynthesis genes were repressed. But
contrarily, induction of lysine (lysC, asd), histidine/glutamate (hutH, hutI, hutU, hutG, rocA), and
citruline/ornithine (argF, arg, arcC) biosynthesis genes was visible in the stationary phase
samples cultivated in pMEM. Other amino acid biosynthesis genes (leu, ilv, dap, thr, and others)
were not significantly differentially expressed. Reduced or even ceased growth diminishes the
need for new protein synthesis. Therefore, associated molecules like ribosomal proteins,
translation elongation factors, and chaperones do not need to be newly synthesized, which was
observed on proteome level for S. aureus COL in TSB (Kohler et al. 2005). This phenomenon is
part of the stringent response e. g. to glucose starvation. Here, using pMEM medium, S. aureus
RN1HG and transcriptome analysis, only two ribosomal protein genes, rpsD and rpsT, and prmA, a
ribosomal protein L11 methyltransferase, were repressed. Only translation elongation factor P
(efp) exhibited a fold change of less than −2 in the t 4 samples, but this repression was not
significant. Chaperones dnaK, grpE, groEL, and groES exhibited slightly decreased fold change
values in the range of −1.4 to −1.9, but the difference was also not significant. Of the Clp
complex, expression of subunit X was repressed and of subunit L was induced. Finally, repression
of tRNA synthetases is known to be part of the stringent response in S. aureus (Anderson KL et al.
2006). In stationary phase in pMEM, repression of tyrS, leuS, alaS, aspS, hisS, valS, thrS, glyS,
proS, ileS, pheS, pheT, gltX, metS/metG, and serS was detected. The other tRNA synthetases
except cysS showed a trend of repression with fold change values almost reaching the cutoff (−2).
Virulence factors differentially expressed in stationary vs. exponential growth phase
In stationary growth phase, differential expression of virulence-associated genes was
observed. Many of these genes were regulated at both analyzed time points of stationary phase,
t 2 and t 4 . Surface proteins A and G (spa, sasG) were repressed. Other membrane-bound adhesins
were induced like clfA, fnbA, and isaB. Secreted adhesins and immunomodulatory molecules efb,
chp, and sbi were repressed, whereas ebh and eap were induced. One of the two staphylococcal
superoxide dismutases, sodM, was found to be induced. The toxins hla, hlb, hlgC, and hlgB
exhibited an increase in expression. In the group of extracellular enzymes, only nuc and htrA
were repressed. Contrarily, for secreted lipases geh and lip and for proteases sspB, splC, splB,
splA, and aur a higher expression was detected in stationary phase than in exponential growth.
The complete cap operon of 15 capsular genes (SAOUHSC_00114 to SAOUHSC_00128) was
induced in t 2 samples and still 10 of these genes were also induced in t 4 samples. The biofilm
repressor icaR was induced in t 2 samples. Fittingly, two genes of the ica operon, icaB and icaC,
were observed to be repressed in both analyzed stationary phase time points. Finally, differential
expression of five staphylococcal accessory regulators was detected: sarX and sarT were
repressed, while sarV, sarR, and sarZ were induced in stationary phase (Table R.5.3).
134

Maren Depke
Results
Pathogen Gene Expression Profiling
Table R.5.3: Expression of virulence associated genes in stationary phase in comparison to exponential growth in pMEM.
LocusTag gene annotation
fold change in
comparison to baseline
exponential growth
stationary
phase t 2 a
stationary
phase t 4 a
SAOUHSC_00069 spa protein A -7.4 -8.8
SAOUHSC_02798 sasG surface protein G -4.2 -3.4
SAOUHSC_00812 clfA clumping factor 2.8 2.1
SAOUHSC_02803 fnbA fibronectin-binding protein precursor, putative 3.1 2.7
SAOUHSC_02972 isaB immunodominant staphylococcal antigen B 3.2 3.4
SAOUHSC_01114 efb fibrinogen-binding protein -2.7 1.4
SAOUHSC_02169 chp chemotaxis inhibitory protein -4.3 -2.6
SAOUHSC_02706 sbi immunoglobulin G-binding protein Sbi, putative -3.9 -1.7
SAOUHSC_01447 ebh extracellular matrix-binding protein ebh 4.7 2.9
SAOUHSC_02161 eap MHC class II analog protein 2.5 3.9
SAOUHSC_00093 sodM superoxide dismutase, putative 2.1 1.6
SAOUHSC_01121 hla alpha-hemolysin precursor 15.1 36.8
SAOUHSC_02163 hlb hypothetical protein SAOUHSC_02163 1.9 7.2
SAOUHSC_02709 hlgC leukocidin s subunit precursor, putative 16.2 43.6
SAOUHSC_02710 hlgB leukocidin f subunit precursor 11.5 29.6
SAOUHSC_00818 nuc thermonuclease precursor -2.4 -1.4
SAOUHSC_00958 htrA serine protease HtrA, putative -5.7 -6.3
SAOUHSC_00300 geh lipase precursor 10.4 11.1
SAOUHSC_03006 lip lipase 70.4 97.8
SAOUHSC_00987 sspB cysteine protease precursor, putative 2.1 1.8
SAOUHSC_01939 splC serine protease SplC 2.8 5.2
SAOUHSC_01941 splB serine protease SplB 3.1 6.3
SAOUHSC_01942 splA serine protease SplA 2.9 6.0
SAOUHSC_02971 aur aureolysin, putative 2.9 5.0
SAOUHSC_00114 capA capsular polysaccharide biosynthesis protein, putative 7.0 6.4
SAOUHSC_00115 capB capsular polysaccharide synthesis enzyme Cap5B 6.5 5.5
SAOUHSC_00116 capC capsular polysaccharide synthesis enzyme Cap8C 6.0 5.2
SAOUHSC_00117 capD capsular polysaccharide biosynthesis protein Cap5D, putative 5.7 4.9
SAOUHSC_00118 capE capsular polysaccharide biosynthesis protein Cap5E, putative 5.3 4.2
SAOUHSC_00119 capF capsular polysaccharide synthesis enzyme Cap8F 4.8 3.8
SAOUHSC_00120 capG UDP-N-acetylglucosamine 2-epimerase 4.1 3.4
SAOUHSC_00121 capH
capsular polysaccharide synthesis enzyme O-acetyl transferase Cap5H,
putative
3.9 3.2
SAOUHSC_00122 capI capsular polysaccharide biosynthesis protein Cap5I, putative 3.7 2.9
SAOUHSC_00123 capJ capsular polysaccharide biosynthesis protein Cap5J, putative 3.2 2.3
SAOUHSC_00124 capK capsular polysaccharide biosynthesis protein Cap5K, putative 3.0 2.1
SAOUHSC_00125 capL cap5L protein/glycosyltransferase, putative 2.8 1.9
SAOUHSC_00126 capM capsular polysaccharide biosynthesis protein Cap8M 2.6 1.8
SAOUHSC_00127 capN cap5N protein/UDP-glucose 4-epimerase, putative 2.7 1.8
SAOUHSC_00128 capO cap5O protein/UDP-N-acetyl-D-mannosaminuronic acid dehydrogenase 2.6 1.7
SAOUHSC_00248 lytM peptidoglycan hydrolase, putative -4.2 -2.9
SAOUHSC_00869 dltA D-alanine-activating enzyme -2.5 -2.0
SAOUHSC_00870 dltB dltB protein, putative -2.6 -2.3
SAOUHSC_00872 dltD extramembranal protein -2.6 -2.3
SAOUHSC_02883 ssaA staphylococcal secretory antigen ssaA -5.5 -3.3
SAOUHSC_03001 icaR ica operon transcriptional regulator IcaR, putative 2.3 1.4
SAOUHSC_03004 icaB icaB protein, putative -3.2 -2.1
SAOUHSC_03005 icaC intercellular adhesion protein C, putative -2.6 -2.2
SAOUHSC_00674 sarX staphylococcal accessory regulator X -4.8 -5.1
SAOUHSC_02799 sarT staphylococcal accessory regulator T -8.8 -8.4
SAOUHSC_02532 sarV staphylococcal accessory regulator V 1.7 2.3
SAOUHSC_02566 sarR staphylococcal accessory regulator R 1.9 2.1
SAOUHSC_02669 sarZ staphylococcal accessory regulator Z 2.0 2.4
a Differentially regulated genes (significant with p* < 0.05 and with a minimal absolute fold change of 2) are indicated in bold.
135

Maren Depke
Results
Pathogen Gene Expression Profiling
In vitro Infection Experiment Study
Reproducibility of replicates and clustering of experimental condition groups
Staphylococcus aureus RN1HG or RN1HG GFP were used to infect S9 cells, a human bronchial
epithelial cell line. The study included samples from the four time points 0 h (start of
experiment), 1 h, 2.5 h and 6.5 h after start of infection, and included besides internalized
samples also control samples from inoculation condition of exponential growth phase,
serum/CO 2 controls, non-adherent staphylococci, and an anaerobic incubation control, although
not every sample condition was represented in every time point (see Material and
Methods / In vitro Infection Experiment Study / “Cell culture infection model” and “Bacterial
control samples”, pages 57 and 58). Hierarchical clustering visualized the grouping of the array
data sets according to their similarity (Fig. R.5.4). In an upper clustering level, grouping into three
major classes clearly revealed the separation of main experimental groups: 1) internalized
staphylococci, 2) controls of the later time points 2.5 h and 6.5 h, and 3) exponential growth
phase samples, which represent the inoculation sample condition, and controls of the early time
point 1 h (Fig. R.5.4, red dashed line). A lower level of clustering separated the individual sample
conditions and time points (Fig. R.5.4, orange dashed line). Here, biological replicates for 2.5 h
internalized, 6.5 h internalized, 2.5 h serum/CO 2 control, 6.5 h serum/CO 2 control, 2.5 h
anaerobic incubation and exponential growth phase samples were arranged together. Samples of
1 h serum/CO 2 control and 1 h non-adherent staphylococci were an exception: Because they
were very similar, the clustering did not lead to a complete segmentation of biological replicates
but additionally grouped 2 samples of the two different conditions (Fig. R.5.4, orange dashed line,
lower part). These two control groups were not only similar to each other, but additionally still
featured a high similarity to the inoculation condition of exponential growth phase.
S. aureus RN1HG sample conditions and time points
internalized
2.5 h internalized
6.5 h internalized
controls of
later time
points 2.5 h
and 6.5 h
2.5 h serum/CO 2 control
6.5 h serum/CO 2 control
2.5 h anaerobic incubation
exponential
growth
phase and
controls of
early time
point 1 h
exponential growth phase
1 h serum/CO 2 control
1 h non-adherent
1 h serum/CO 2 control
1 h non-adherent
Fig. R.5.4: Hierarchical clustering of 23 tiling array data sets from in vitro infection experiment.
The following clustering algorithms were applied on z-score-transformed data: Agglomerative clustering with average linkage using
using cosine correlation as similarity measure. All sequences were included in the cluster analysis. The red dashed line indicates
grouping into three major classes: 1) internalized staphylococci, 2) controls of the later time points 2.5 h and 6.5 h, and 3) exponential
growth phase samples, which represent the inoculation sample condition, and controls of the early time point 1 h. A lower level of
clustering (orange dashed line) separated the individual sample conditions and time points. Here, biological replicates for 2.5 h
internalized, 6.5 h internalized, 2.5 h serum/CO 2 control, 6.5 h serum/CO 2 control, 2.5 h anaerobic incubation and exponential growth
phase samples were arranged together. Samples of 1 h serum/CO 2 control and 1 h non-adherent staphylococci were an exception:
Because they were very similar, the clustering did not lead to a complete segmentation of biological replicates, but additionally
grouped 2 samples of the two different conditions (orange dashed line, lower part).
136

Maren Depke
Results
Pathogen Gene Expression Profiling
Summing up, the biological replicates exhibited a high reproducibility. Most importantly, the
internalized condition had characteristics very distinct from either the starting condition of
exponential growth phase and the similar early time point controls, but also from the later
controls which were incubated in the presence of serum in 5 % CO 2 -atmosphere without agitation
and from the anaerobic control samples.
Choice of adequate baseline samples for comparison of the experimental conditions:
Time-matched controls are not suitable because of their similarity to stationary growth phase /
stringent response
Many bacterial species are capable of inducing a so-called stringent response with the aim to
adapt their metabolism to situations of nutrient limitation, especially to carbon and amino acid
starvation. The stringent response leads to induced stress resistance, decelerated growth and
alleviated metabolism. The main mediators of the stringent response are pyrophosphorylated
GTP or GDP molecules, which are also called alarmones: guanosine tetraphosphate (ppGpp) and
guanosine pentaphosphate (pppGpp), depending on bacterial species. In the course of the
stringent response, up to one third of the transcriptome can be modulated. The stringent
response has been intensively studied in E. coli, but also S. aureus is capable of such response
(Condon et al. 1995, Crosse et al. 2000, Anderson KL et al. 2006, Wolz et al. 2010).
Entry into stationary growth phase is initiated by a beginning nutrient limitation after
consumption in the exponential phase of growth. Therefore, in stationary phase bacteria’s
stringent response is triggered.
When comparing the gene expression signature of S. aureus RN1HG, which was incubated for
2.5 h in serum-containing infection medium under 5 % CO 2 -atmosphere (37°C), with the signature
of stationary phase time point t 2 , which is defined as 2 h after entry into the stationary phase
(each in comparison to exponentially growing bacteria), a high overlap between both signatures
was recognized. Almost 44 % of sequences in the 2.5 h serum/CO 2 control signature were also
found in the stationary phase signature (Fig. R.5.5; for S. aureus RN1HG stationary phase
response refer to chapter “Growth Media Comparison Study”, page 131). This first hint for
similarities between the controls of later time points in the infection experiment and the
bacterial stationary phase samples provoked a more detailed analysis of expression data of genes
already known to be involved in the stringent response.
Fig. R.5.5:
Comparison of 2.5 h serum/CO 2 control signature
with stationary phase signature.
Each sample condition was compared to its
corresponding baseline samples of exponential
growth in log-transformed space using textbook
one-way ANOVA with Benjamini-Hochberg False
Discovery Rate multiple testing correction, and
p* < 0.05 was regarded as significant. An absolute
fold change cutoff of 2 was applied.
sequences differentially expressed in
the comparison of incubation for
2.5 h in serum-containing medium
and 5% CO 2 atmosphere vs.
exponential growth phase in pMEM
487 418 1005
sequences differentially expressed
between stationary growth phase (t 2 )
and exponential growth phase of
S. aureus RN1HG cultivated in pMEM
Ribosomal proteins are repressed during stringent response (Anderson KL et al. 2006). As the
bacterial metabolism adapts to slower growth and limited energy resources, the cell also
diminishes protein synthesis and therefore does not need to produce further ribosomes.
Reduced expression of 55 genes for ribosomal proteins was observed in late serum/CO 2 controls
(2.5 h and 6.5 h) and in anaerobiosis (Fig. R.5.6 A).
137

S. aureus RN1HG sample conditions and time points
mean intensity of 2 or 3 biological replicates
S. aureus RN1HG sample conditions and time points
mean intensity of 2 or 3 biological replicates
*
*
*
*
S. aureus RN1HG sample conditions and time points
mean intensity of 2 or 3 biological replicates
*
*
*
Maren Depke
* p

S. aureus RN1HG sample conditions and time points
mean intensity of 2 or 3 biological replicates
Maren Depke
Results
Pathogen Gene Expression Profiling
Similarly, tRNA synthetase genes are repressed in stringent response (Anderson KL et al.
2006). Here, repression of these genes was detectable in internalized staphylococci (2.5 h), late
serum/CO 2 controls (2.5 h and 6.5 h) and in samples after 2.5 h anaerobic incubation
(Fig. R.5.6 B). An exception is the isoleucin tRNA synthetase (ileS), which is induced in stringent
response (Anderson KL et al. 2006). But in the 2.5 h and 6.5 h serum/CO 2 controls, this gene
shows a trend of repression similar to the other tRNA synthetases with a fold change of −1.5 and
−1.8, respectively.
Furthermore, translation elongation factor gene expression is known to be suppressed in
stringent response (Anderson KL et al. 2006). Such repression physiologically acts in concert with
repressed ribosomal protein and tRNA synthetase genes and leads to diminished protein
synthesis. In this study, 4 genes for translation elongation factors revealed a trend of repressed
expression in late serum/CO 2 controls (2.5 h and 6.5 h). With p = 0.06, statistical significance was
only slightly missed in the comparison between these two controls and exponential growth
samples (Fig. R.5.6 C).
Stringent response is mediated by the regulator RelA whose transcript lateron is increased in
stringent response conditions (Anderson KL et al. 2006). A strong trend of induction for relA was
observed for the 2.5 h and 6.5 h serum/CO 2 controls. Although relA exhibited significantly
different expression (p* < 0.01) in the comparison of all sequences between the 2.5 h serum/CO 2
control and exponential growth phase samples, the 1.8-fold change did not pass the cutoff level
of 2. On the other hand, in the comparison between the later 6.5 h serum/CO 2 control and
exponential growth phase samples, the fold change amounted to 2.1, but in this comparison
statistical testing was not possible because of limited number of replicates relA at the 6.5 h time point
(n = 2). Nevertheless, relA could be regarded as induced in the late serum/CO 2 controls, because
both time points showed the same trend and all six other experimental samples possessed
almost the same, lower gene expression intensity, which proved relA highly reliable measurement of
the relA expression by the tiling array approach (Fig. R.5.7).
Fig. R.5.7:
Comparison of mean expression
intensities for the stringent response
regulatory gene relA.
A strong trend of relA induction was
visible for the late serum/CO 2 controls
(2.5 h and 6.5 h). Although relA exhibited
significantly different expression (one-way
textbook ANOVA with Benjamini-
Hochberg False Discovery Rate multiple
testing correction and p* < 0.01) in the
comparison between the 2.5 h serum/CO 2
control and exponential growth phase
samples, the 1.8-fold change did not pass
the cutoff level of 2. On the other hand, in
the comparison between the later 6.5 h
serum/CO 2 control and exponential
growth phase samples, the fold change
amounted to 2.1, but in this comparison
statistical testing was not possible
because of limited number of replicates at
the 6.5 h time point (n = 2).
exponential growthOD phase 0.4
1 h serum/CO 2 control 1h CO2
1 non-adh. h non-adherent MOI 25
2.5 h internal. internalized 2.5h
6.5 h internal. internalized 6.5h
2.5 h serum/COCO2 2 control 2.5h
6.5 h serum/COCO2 2 control 6.5h
2.5 h anaerobic anaerobic incubation 2.5h
1.0E+04
10000
2.0E+04
20000
mean intensitySamples
of biological replicates
3.0E+04
30000
From repressed or in trend repressed ribosomal protein, tRNA synthetase, and translation
elongation factor genes and from the strong tendency of relA infection induction experiments it was reasoned medium comparison that the
gene expression signature of late serum/CO 2 controls (2.5 h and 6.5 h) in fact resembled the
stationary phase/stringent response. As internalized staphylococci and the early control samples
did not possess this similarity, it was concluded that the 1 h serum/CO 2 control sample, although
not time-matched, was the most appropriate baseline sample for this infection experiment study.
139

Maren Depke
Results
Pathogen Gene Expression Profiling
Comparison of experimental condition groups and assessment of differentially regulated
genes/sequences
The first general data analysis steps had provided insight into the overall similarity of the 23
arrays available after hybridization of samples and controls from the infection experiment.
Biological replicates exhibited good reproducibility. The time-matched serum/CO 2 controls for
the internalized samples built a distinct group of samples. As a more detailed analysis revealed
similarities of these controls with stationary phase/stringent response, it became apparent that
the 1 h serum/CO 2 samples were the most suitable control to use as baseline in this study.
Therefore, the following seven types of comparisons were chosen to elucidate the changes in
gene expression pattern and lateron the physiological reactions of S. aureus RN1HG in the
settings of the in vitro S9 infection model (Fig. R.5.8):
comparison of exponential growth phase samples, which represent the inoculation
condition, with baseline of 1 h serum/CO 2 controls
comparison of non-adherent staphylococci (1 h) with baseline of 1 h serum/CO 2 controls
comparison of 2.5 h internalized samples with baseline of 1 h serum/CO 2 controls
comparison of 6.5 h internalized samples with baseline of 1 h serum/CO 2 controls
comparison of 2.5 h serum/CO 2 controls with baseline of 1 h serum/CO 2 controls
comparison of 6.5 h serum/CO 2 controls with baseline of 1 h serum/CO 2 controls
comparison of 2.5 h anaerobic incubation samples with baseline of 1 h serum/CO 2 controls
inoculation
condition
and early
time point
control
samples
1
1 h
serum/CO 2
control
(3 arrays)
exponential
growth phase
(3 arrays)
2
1h
non-adherent
S. aureus
(3 arrays)
4
3
7
5
6
6.5 h
serum/CO 2
control
(2 arrays)
2.5 h
internalized
S. aureus
(3 arrays)
2.5 h
serum/CO 2
control
(3 arrays)
6.5 h
internalized
S. aureus
(3 arrays)
internalized
samples
2.5 h
anaerobic
incubation
(3 arrays)
late
time point
controls
Fig. R.5.8: Overview on the comparisons between groups in this study that were partly addressed with statistical testing and visualized
with scatter plots.
Baseline for all comparisons was the serum/CO 2 control group of 1 h. The very similar exponential growth phase samples, which
represent the inoculation condition, were compared to this baseline () and the even more similar non-adherent staphylococcal
samples after 1 h (). To elucidate the gene expression signature of internalized bacteria, the 2.5 h () as well as the 6.5 h ()
internalized samples were compared to the samples of 1 h serum/CO 2 control group. For comparison purposes, the late time point
controls were also checked against the early serum/CO 2 control as baseline: 2.5 h serum/CO 2 controls () and samples after 2.5 h of
anaerobic incubation (). The comparison focusing on the difference between 6.5 h and 1 h serum/CO 2 control samples () could
not be performed with statistical testing because of the low number (n = 2) of arrays at the 6.5 h time point.
140

exponential growth phase
2.5 h internalized
2.5 h anaerobic incubation
2.5 h serum/CO 2 control
1 h non-adherent
6.5 h internalized
6.5 h serum/CO 2 control
Maren Depke
Results
Pathogen Gene Expression Profiling
The comparisons were visualized using scatter plots (Fig. R.5.9). Comparable to the clustering
results, a high concordance between expression values of 1 h serum/CO 2 controls and
exponential phase samples, which represent the inoculation condition (Fig. R.5.9, panel ), and
even more striking between 1 h serum/CO 2 controls and non-adherent staphylococcal samples
after 1 h (Fig. R.5.9, panel ) was observed. Strong effects of treatment in comparison to 1 h
serum/CO 2 controls emerged in 2.5 h internalized (Fig. R.5.9, panel ), 6.5 h internalized
(Fig. R.5.9, panel ), 2.5 h serum/CO 2 controls (Fig. R.5.9, panel ), 6.5 h serum/CO 2 controls
(Fig. R.5.9, panel ), and in 2.5 h anaerobically incubated samples (Fig. R.5.9, panel ). Most
variation was detectable in the comparison of 1 h and 6.5 h serum/CO 2 controls (Fig. R.5.9,
panel ), but this can be due to the lower number of only two replicates for the 6.5 h serum/CO 2
control, which led to a lesser compensation for variation in the mean value than in the mean of
three replicates in the other sample condition/time point groups.
1 h serum/CO 2 control
1 h serum/CO 2 control
1 h serum/CO 2 control
1 h serum/CO 2 control
1 h serum/CO 2 control
1 h serum/CO 2 control
1 h serum/CO 2 control
Fig. R.5.9: Scatter plots comparing mean signal intensities
of treatment groups.
The signals of the eight groups of “1 h serum/CO 2
controls”, “exponential growth phase”, “1 h nonadherent
staphylococci”, “2.5 h internalization”, “6.5 h
internalization”, “2.5 h serum/CO 2 controls”, “6.5 h
serum/CO 2 controls”, and “2.5 h anaerobic incubation”
were plotted after combining the three biological
replicates. In case of “6.5 h serum/CO 2 controls” only two
biological replicates were available. All sequences
available on the array are shown. Numbers in the first
column refer to the comparison scheme in Fig. R.5.8.
141

Maren Depke
Results
Pathogen Gene Expression Profiling
The different experimental groups were compared with statistical testing to obtain lists of
differentially expressed sequences (Table R.5.4). It has been already described that the three
groups exponential growth phase as inoculation condition, 1 h serum/CO 2 control, and 1 h nonadherent
staphylococci were similar to each other. Non-surprisingly, when comparing 1 h nonadherent
staphylococci to the 1 h serum/CO 2 control samples, no statistically significant
differential expression was detected. Between exponential growth phase and 1 h serum/CO 2
control, 156 sequences were differentially expressed of which 76 exceeded a minimal absolute
fold change of 2. In the comparison of 2.5 h and 6.5 h internalization vs. 1 h serum/CO 2 control,
1392 and 1205 sequences were regarded as differentially expressed, of which 765 and 627
possessed a minimal absolute fold change of 2, respectively. Additionally, 1749 sequences were
differentially expressed between 2.5 h serum/CO 2 control and 1 h serum/CO 2 control, and 905 of
these passed the minimal absolute fold change cutoff 2. The 2.5 h anaerobic signature in
comparison with the baseline of 1 h serum/CO 2 control included 1339 sequences, and still 680
exhibited a minimal absolute fold change of 2.
Table R.5.4: Results of group comparisons with statistical testing of S9 infection experiment staphylococcal array data sets.
group comparison a
number of sequences
significant with p* < 0.05 in
textbook one-way ANOVA with
Benjamini-Hochberg False Discovery
Rate multiple testing correction
absolute fold change equal
to or greater than 2 AND
significant with p* < 0.05
exponential growth phase vs. 1 h serum/CO 2 control 156 76
1 h non-adherent staphylococci vs. 1 h serum/CO 2 control 0 -
2.5 h internalization vs. 1 h serum/CO 2 control 1392 765
6.5 h internalization vs. 1 h serum/CO 2 control 1205 627
2.5 h serum/CO 2 control vs. 1 h serum/CO 2 control 1749 905
2.5 h anaerobic incubation vs. 1 h serum/CO 2 control 1339 680
a The 080604_SA_JH_Tiling array contains 3882 sequences of which 2825 are known transcripts. The remaining 1057 sequences were
identified as new transcripts in the tiling array analysis. All sequences were included in statistical testing.
Fractions of known genes and newly identified transcribed sequences included on the tiling
array and in the lists of differentially expressed sequences derived from the comparisons of
infection experiment samples
The 080604_SA_JH_Tiling array contains 3882 sequences of which 2825, corresponding to
73 %, are known transcripts and associated with a LocusTag number (SAOUHSC_*). The remaining
1057 sequences were identified as new transcripts in the tiling array analysis. These new
transcripts are expected to belong to all types of RNAs like mRNA and regulatory RNA, but also
formerly unknown transcribed fragments 5’ and 3’ to known genes might be included.
Furthermore, some artifacts might still be included. All sequences, known and new, were
included in statistical testing.
For the internalization-specific signature, 200 and 138 differentially expressed transcripts
belonged to the group of new transcripts at the 2.5 h and 6.5 h time point, respectively
(Table R.5.5). Furthermore, 19 new transcripts were included in the set of differentially expressed
sequences between the inoculation condition of exponential growth and the baseline of 1 h
serum/CO 2 control. In comparison to that baseline, 2.5 h serum/CO 2 control and 2.5 h
anaerobically incubated samples inclosed 250 and 200 differentially expressed new transcripts,
respectively (Table R.5.5).
142

Maren Depke
Results
Pathogen Gene Expression Profiling
Table R.5.5: Fractions of known annotated and newly detected transcripts in the results of group comparisons with statistical testing
of S9 infection experiment staphylococcal array data sets.
group comparison a
total number of
sequences with
absolute fold
change equal to or
greater than 2 AND
significant with
p* < 0.05
number of known
transcribed
sequences with
absolute fold
change equal to or
greater than 2 AND
significant with
p* < 0.05
number of new
transcribed sequences
with absolute fold
change equal to or
greater than 2 AND
significant with
p* < 0.05
exponential growth phase vs. 1 h serum/CO 2 control 76 57 19
2.5 h internalization vs. 1 h serum/CO 2 control 765 565 200
6.5 h internalization vs. 1 h serum/CO 2 control 627 489 138
2.5 h serum/CO 2 control vs. 1 h serum/CO 2 control 905 655 250
2.5 h anaerobic incubation vs. 1 h serum/CO 2 control 680 480 200
a The 080604_SA_JH_Tiling array contains 3882 sequences of which 2825 are known transcripts. The remaining 1057 sequences were
identified as new transcripts in the tiling array analysis. All sequences were included in statistical testing.
Comparison of differentially regulated sequences in 2.5 h and 6.5 h internalized staphylococci
The gene expression signatures of 2.5 h and 6.5 h internalized staphylococci, which resulted
from statistical comparison with the baseline of 1 h serum/CO 2 control including multiple testing
correction and minimal absolute fold change cutoff 2, were compared in a Venn diagram
(Fig. R.5.10 A). For 408 sequences, differential expression was observed in both time points of
internalized samples, while 357 and 219 sequences were specific for the 2.5 h and 6.5 h
internalization signature, respectively. This corresponds to a fraction of 53 % of sequences
regulated after 2.5 h of internalization, which were also differentially expressed after 6.5 h. Vice
versa, 65 % of differentially expressed sequences at the 6.5 h time point were also found to be
regulated after 2.5 h. When examining the induced and repressed sequences separately, it was
first obvious that sequences regulated at both time points always possessed the same regulation
direction in these two time points (Fig. R.5.10 B). Further, induction accounted for a bigger
fraction than reduction in the differentially expressed sequences, although the excess was only
small at the 6.5 h time point.
A
B
357 408 219
2.5 h
repressed
repressed 6.5 h
induced
176 121
induced
181
187
98
sequences differentially
expressed in the comparison
“2.5 h internalization” vs.
“1 h serum/CO 2 control”
sequences differentially
expressed in the comparison
“6.5 h internalization” vs.
“1 h serum/CO 2 control”
221
Fig. R.5.10: Comparison of the 2.5 h and 6.5 h signatures of internalized staphylococci.
Both internalized samples were compared to the baseline of 1 h serum/CO 2 control with statistical testing and multiple testing
correction (p* < 0.05), and a minimal absolute fold change cutoff of 2 was applied. The comparison of differentially expressed
sequences at the 2.5 h and 6.5 h time point was performed for all regulated sequences (A) and for induced and repressed sequences
separately (B).
In the 2.5 h samples 402 sequences were induced (181 specifically for that time point), 363
were repressed (176 specifically for that time point), while in the 6.5 h samples 319 exhibited
induction (98 specifically for that time point) and 308 exhibited repression (121 specifically for
that time point). Therefore, the 408 sequences differentially expressed in both time points
consisted of 221 induced and 187 repressed sequences.
143

Maren Depke
Results
Pathogen Gene Expression Profiling
Comparison of differentially expressed genes [Maren Depke] and proteins with differing
abundance [Sandra Scharf] in internalized staphylococci
In an experimental setup corresponding to the generation of samples for tiling array analysis,
the abundance of proteins in internalized staphylococci was recorded with a combined approach
of stable isotope labeling with amino acids in cell culture (SILAC), FACS cell sorting and mass
spectrometric analysis (Sandra Scharf, Schmidt et al. 2010). The method, which was applied for
this first proteome profiling of in vitro internalized staphylococci, limits the possible protein
identification mainly to cytosolic proteins: Secreted proteins were lost during sample
preparation, and peptides from membrane associated proteins are difficult to obtain by tryptic
digest of whole-cell samples.
In the proteomic study, the time points 1.5 h, 2.5 h, 3.5 h, 4.5 h, 5.5 h, and 6.5 h after start of
infection were included, i. e. the analysis window was divided into smaller sections than in the
tiling array study. In an analysis which included more samples and controls than published in the
pilot study, Sandra Scharf identified 648 proteins and obtained a set of 114 proteins with
differential abundance. Differential abundance was assigned to proteins when the regulation was
observed in at least 2 of 3 biological replicates in at least 1 of 6 analyzed time points, and
proteins with contradictory results were excluded. The regulated proteins were normalized in
reference to the SILAC-method immanent internal control (heavy isotope labeled peptides) and
exhibited an abundance difference when comparing target peptide ratios with the median ratio
for all peptides at the corresponding time point. Regulated proteins which were additionally
different in abundance in a serum/CO 2 control (without presence of host cells) were excluded.
To gain an insight into the comparability of transcriptome and proteome profiling,
differentially expressed sequences were compared to regulated proteins. First, up-regulated
proteins were compared to the sequences which exhibited increased expression 2.5 h or 6.5 h
after start of infection or in both time points (Fig. R.5.11 A). Second, the same comparison was
performed for down-regulated proteins and repressed sequences (Fig. R.5.11 B).
A
B
up-regulated proteins in
internalized staphylococci
down-regulated proteins in
internalized staphylococci
53
34
5 2
14
176 96
207
3 0
6
173 121
181
induced sequences in
“2.5 h internalization” vs.
“1 h serum/CO 2 control”
induced sequences in
“6.5 h internalization” vs.
“1 h serum/CO 2 control”
repressed sequences in
“2.5 h internalization” vs.
“1 h serum/CO 2 control”
repressed sequences in
“6.5 h internalization” vs.
“1 h serum/CO 2 control”
Fig. R.5.11: Comparison of differentially expressed sequences in at least one of the two analyzed time points with proteins exhibiting
different abundance in internalized staphylococci.
A. Up-regulated proteins and induced sequences. B. Down-regulated proteins and repressed sequences.
144

Maren Depke
Results
Pathogen Gene Expression Profiling
In total, the transcripts of 21 up-regulated proteins were increased in at least one of the two
analyzed time points. These consisted of 14 proteins, whose genes were induced at both time
points, and of 5 and 2 proteins, whose gene expression was increased only at the 2.5 h and 6.5 h
time point, respectively. For 6 proteins, whose abundance was reduced, the corresponding gene
expression was repressed at both time points, and for further 3 down-regulated proteins the
gene expression was only repressed at the 2.5 h time point (Table R.5.6). These numbers result in
an overlap of 28 % of up- and 21 % of down-regulated proteins with the corresponding gene
expression changes. As the lists of differentially expressed sequences included newly identified
transcripts from the tiling array, which were not accessed with mass spectrometry identification,
these lists contained many transcriptome-specific results.
Table R.5.6: Differentially expressed sequences in internalization samples which exhibit differences in protein abundance and
regulation in the same direction.
LocusTag gene annotation
difference
in protein
abundance
a
fold change in comparison
to baseline
1 h serum/CO 2 control
2.5 h
internalized b
6.5 h
internalized b
SAOUHSC_01846 acsA acetyl-CoA synthetase, putative + 16.3 4.0
SAOUHSC_01395 asd aspartate-semialdehyde dehydrogenase + 14.1 13.3
SAOUHSC_00086 butA 3-ketoacyl-acyl carrier protein reductase, putative + 2.5 2.2
SAOUHSC_01396 dapA dihydrodipicolinate synthase + 12.5 12.0
SAOUHSC_01398 dapD
2,3,4,5-tetrahydropyridine-2-carboxylate N-
succinyltransferase, putative
+ 9.1 9.8
SAOUHSC_01320 dhoM homoserine dehydrogenase, putative + 2.1 5.4
SAOUHSC_00196 fadB hypothetical protein SAOUHSC_00196 + 84.4 16.6
SAOUHSC_00197 fadD hypothetical protein SAOUHSC_00197 + 100.0 35.8
SAOUHSC_02869 rocA delta-1-pyrroline-5-carboxylate dehydrogenase, putative + 8.0 2.8
SAOUHSC_01418 odhA 2-oxoglutarate dehydrogenase, E1 component + 5.2 2.2
SAOUHSC_01416 odhB
2-oxoglutarate dehydrogenase, E2 component,
dihydrolipoamide succinyltransferase
+ 4.5 2.1
SAOUHSC_00371 hypothetical protein SAOUHSC_00371 + 4.4 2.1
SAOUHSC_01399 hypothetical protein SAOUHSC_01399 + 7.9 8.3
SAOUHSC_02425 hypothetical protein SAOUHSC_02425 + 5.3 2.8
SAOUHSC_01276 glpK glycerol kinase + 2.7 1.7
SAOUHSC_01801 icd, citC isocitrate dehydrogenase, NADP-dependent + 2.1 1.2
SAOUHSC_01972 prsA protein export protein PrsA, putative + 2.0 1.6
SAOUHSC_00767 yfiA hypothetical protein SAOUHSC_00767 + 2.9 1.7
SAOUHSC_00951 hypothetical protein SAOUHSC_00951 + 2.4 1.6
SAOUHSC_01321 thrC threonine synthase + 1.7 5.3
SAOUHSC_00833 hypothetical protein SAOUHSC_00833 + 1.2 2.1
SAOUHSC_01771 hemL glutamate-1-semialdehyde-2,1-aminomutase − -2.2 -2.5
SAOUHSC_01092 pheS phenylalanyl-tRNA synthetase, alpha subunit − -2.5 -3.4
SAOUHSC_01093 pheT phenylalanyl-tRNA synthetase, beta subunit − -2.7 -3.5
SAOUHSC_01010 purC
phosphoribosylaminoimidazole-succinocarboxamide
synthase
− -4.5 -2.2
SAOUHSC_01011 purS
phosphoribosylformylglycinamidine synthase, PurS
protein
− -4.9 -2.3
SAOUHSC_01788 thrS threonyl-tRNA synthetase − -2.1 -2.4
SAOUHSC_01018 purD phosphoribosylamine-glycine ligase − -2.6 -1.1
SAOUHSC_01013 purL phosphoribosylformylglycinamidine synthase II − -3.4 -1.5
SAOUHSC_01012 purQ phosphoribosylformylglycinamidine synthase I − -4.0 -1.9
a Up-regulated protein abundance in internalized staphylococci is indicated by “+”, down-regulated abundance by ”–“.
b Differentially regulated genes (significant with p* < 0.05 and with a minimal absolute fold change of 2) are indicated in bold.
145

Maren Depke
Results
Pathogen Gene Expression Profiling
In a similar comparison, up-regulated proteins were compared with repressed transcripts and
vice versa down-regulated proteins with induced transcripts, which resulted in an overlap of 5
(rocD, ldh2, sodM, fabH, SAOUHSC_02150) and 9 (sufC, sufD, spoVG, dps, SAOUHSC_00533,
SAOUHSC_02443, SAOUHSC_02363, SAOUHSC_01854, SAOUHSC_00845) proteins, respectively.
Such contradictory results could be explained in most cases by the following characteristics:
transcript changes do not necessarily influence the protein abundance at the same time
point; the potential time shift between both effects is not defined and may vary between
the different proteins
effects on protein or expression might be outside the analysis window (between start of
infection and first analysis time point of 1.5 h or 2.5 h; after last analysis time point of
6.5 h; between analysis time points especially in the transcriptome analysis)
special protein effects like temporary new synthesis, increased turnover, and other nontranscriptionally
mediated effects
special protein analysis effects: early changes or changes occurring before the first analysis
time point, which stop in the middle of the observation period, might still lead to an
absolute fold change value exceeding 2 at the later time points in cases of stable proteins
protein analysis challenges like low intensity measurements, high variation between
biological replicates, missing data for one of the biological replicates
Expression of genes containing binding sites for Rex, an anaerobiosis regulator and redox
sensor
During in vitro infection/internalization experiments, staphylococci are subjected to changes
in oxygen availability. First, they are shifted from aerobic shaking cultures into cell culture dishes,
second, they are incubated without agitation in a CO 2 -atmosphere optimal for the host cells, and
finally, internalization into eukaryotic host cells might further influence the oxygen availability. To
answer the question whether oxygen limitation has major impact on internalized staphylococci,
the gene expression of anaerobiosis-related genes was analyzed. Very recently, Pagels et al. have
published transcriptome and proteome data of an anaerobic gene expression regulon (Pagels et
al. 2010). Although indirect effects of Rex have also been described, the central regulator Rex, a
redox sensor, acts as transcriptional repressor. In situations of increasing NADH, DNA-binding of
Rex is inhibited, resulting in de-repression of genes which possess a Rex binding site. Pagels et al.
defined 21 genes with such binding site, 19 of which could be mapped to S. aureus NCTC8325
LocusTags (NCBI’s Entrez Genome Gene Plot; http://www.ncbi.nlm.nih.gov/sutils/geneplot.cgi).
For a first impression, the fraction of these genes which were differentially expressed in
staphylococci after 2.5 h of anaerobic incubation, but also in 2.5 h and 6.5 h internalized
staphylococci was determined. In anaerobically incubated staphylococci, 5 Rex binding site
containing genes were found to be regulated (Fig. R.5.12 A), whereas 10 of these genes were
regulated in internalization at both time points and further 3 genes specifically at the time point
of 6.5 h after start of infection (Fig. R.5.12 B). A more detailed analysis revealed that while
induction of these genes in anaerobically incubated staphylococci was expected, the direction of
regulation in internalized staphylococci was clearly opposite (Fig. R.5.12 C). Therefore,
internalization probably did not result in stimuli which were able to inactivate the Rex repressor.
146

mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
A
14 5 675
B
genes containing Rex binding
sites in S. aureus according to
Pagels et al. 2010
genes containing Rex binding
sites in S. aureus according to
Pagels et al. 2010
sequences differentially
expressed in the comparison
“2.5 h anaerobic incubation”
vs. “1 h serum/CO 2 control”
Fig. R.5.12:
Expression of genes containing Rex binding sites.
A. Comparison of genes with Rex binding sites (according to
Pagels et al. 2010) and S. aureus RN1HG anaerobic gene
expression signature in the infection experiment study.
B. Comparison of genes with Rex binding sites (according to
Pagels et al. 2010) and S. aureus RN1HG S9 internalization gene
expression signature 2.5 h and 6.5 h after start of infection.
C. Overview on mean normalized gene expression intensity for
19 genes containing Rex binding sites. After the global
normalization by inter-chip scaling and detrending, each
individual gene has been normalized to the expression level of
the baseline sample “1 h serum/CO 2 control”. Although not
being continuous data, values were depicted as line graphs
instead of bar charts for better facility of inspection. Black
indicates 8 genes repressed in internalized staphylococci at the
2.5 h or 6.5 h time point, dark gray 5 genes differentially
expressed in at least one internalization time point and after
anaerobic incubation, and light gray marks the remaining 6
genes of which 3 possess a fold change value > 1.5 in the
anaerobic sample compared to baseline.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2
control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h
internalization; 2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. −
6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic
incubation)
C
differentially expressed
sequences in “2.5 h
internalization” vs.
“1 h serum/CO 2 control”
10E+01 10.000
10E+00 1.000
10E-01 0.100
10E-02 0.010
10E-03 0.001
exp.
6
0 3
10
357 216
398
1 h
co.
2.5 h
int.
differentially expressed
sequences in “6.5 h
internalization” vs.
“1 h serum/CO 2 control”
OD 0.4 CO2 (serum) internalized internalized CO2 (serum) CO2 (serum) anaerobic
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
0 h 1 h 2.5 h 6.5 h 2.5 h 6.5 h 2.5 h
S. aureus RN1HG sample conditions and time points
Genes of the SaeRS regulon exhibiting differential expression in internalized staphylococci
The two-component system SaeRS is known to positively regulate different virulence
associated genes, e. g. genes coding for adhesins or toxins. The induction of gene expression of
saeR was observed in internalized staphylococci 6.5 h after start of infection. Additionally, the
gene saeS for the histidine kinase sensor component was significantly different in 6.5 h
internalized staphylococci compared to baseline, but the fold change of 1.7 did not meet the
cutoff criterium of 2 (Fig. R.5.13 A). As the the two-component system SaeRS is known to be
subjected to positive autoinduction, the induction of saeR and trend for saeS provoked a detailed
analysis of genes known to be regulated by this system. In 2006, Rogasch et al. have published
transcriptome and secretome data of the SaeRS regulon in S. aureus strains COL and Newman
(Rogasch et al. 2006). Using a global microarray screening of saeRS mutants and parental strains,
Rogasch et al. defined 44 genes which were induced by the SaeRS two-component system, 40 of
which could be mapped to S. aureus NCTC8325 LocusTags (NCBI’s EntrezGenome Gene Plot;
http://www.ncbi.nlm.nih.gov/sutils/geneplot.cgi). For a first impression, the fraction of these
genes which were differentially expressed in 2.5 h and 6.5 h internalized staphylococci was
determined (Fig. R.5.13 B). At the 2.5 h time point, 4 genes with known regulation by SaeRS were
147

mean normalized intensity
mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
differentially expressed, 10 further genes were found to be regulated specifically after 6.5 h,
whereas another 10 genes were regulated in both internalization time points. Of these 4, 10, and
10 genes, the numbers of 3, 9, and 9 genes were increased in expression. The remaining gene of
each group exhibited repression (Fig. R.5.13 C, Table R.5.7).
A
33
B
genes regulated by SaeRS
according to Rogasch et al. 2006
22
16
11
saeR
saeS
00
OD 0.4 CO2 (serum) internalized internalized CO2 (serum) CO2 (serum) anaerobic
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
0 h 1 h 2.5 h 6.5 h 2.5 h 6.5 h 2.5 h
S. aureus RN1HG sample conditions and time points
differentially expressed
sequences in “2.5 h
internalization” vs.
“1 h serum/CO 2 control”
4 10
10
353 209
398
differentially expressed
sequences in “6.5 h
internalization” vs.
“1 h serum/CO 2 control”
Fig. R.5.13:
Expression of genes known to be regulated by SaeRS.
A. Expression of saeR and saeS in S. aureus RN1HG during
the S9 infection experiment. Induction of both genes was
significant in statistical testing for the 6.5 h time point of
internalization, but did not pass the two-fold cutoff for
saeS.
B. Comparison of genes known to be regulated by SaeRS
(according to Rogasch et al. 2006) and S. aureus RN1HG S9
internalization gene expression signature 2.5 h and 6.5 h
after start of infection.
C. Overview on mean normalized gene expression intensity
for 24 genes known to be regulated by SaeRS. After the
global normalization by inter-chip scaling and detrending,
each individual gene has been normalized to the
expression level of the baseline sample “1 h serum/CO 2
control”. Although not being continuous data, values were
depicted as line graphs instead of bar charts for better
facility of inspection. Black indicates 9 genes induced in
internalized staphylococci at the 2.5 h or 6.5 h time point,
dark gray 9 genes induced at 6.5 h and 3 genes induced at
2.5 h after start of infection, and light gray marks the
remaining 3 genes of which one was repressed at both
time points, another specifically 2.5 h and the third
specifically 6.5 h after start of infection.
C
10E+02 100.000
10E+01 10.000
10E+00 1.000
10E-01 0.100
10E-02 0.010
OD 0.4 CO2 (serum) internalized internalized CO2 (serum) CO2 (serum) anaerobic
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
0 h 1 h 2.5 h 6.5 h 2.5 h 6.5 h 2.5 h
S. aureus RN1HG sample conditions and time points
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2 control;
2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h internalization;
2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2
control; 2.5 h anae. − 2.5 h anaerobic incubation)
The set of 21 SaeRS regulated genes, which exhibited differential expression in at least one of
the two analyzed internalization time points, included certain functional groups of virulence
factors, and some could even be assigned to more than one group. Membrane bound
adhesins/MSCRAMMS were represented by the fibrinogen binding proteins A and B (fnbA, fnbB),
soluble adhesins/SERAMs were covered e. g. by extracellular adherence protein (eap), toxins
were included with the examples of different hemolysins, chemotaxis inhibitory protein (chp)
served as example for immune-evasive proteins, and finally, also secreted enzymes like serine
protease (spl) were listed. Thus, further data mining was performed in consideration of functional
groups of virulence factors.
148

Maren Depke
Results
Pathogen Gene Expression Profiling
Table R.5.7: Expression of genes known to be regulated by SaeRS (according to Rogasch et al. 2006), which exhibited differential
expression in internalized S. aureus RN1HG in S9 cells.
LocusTag gene annotation
fold change in comparison to
baseline 1 h serum/CO 2 control
2.5 h
internalized a
6.5 h
internalized a
SAOUHSC_01942 splA serine protease SplA 3.2 30.4
SAOUHSC_01939 splC serine protease SplC 4.8 43.1
SAOUHSC_02803 fnbA fibronectin-binding protein precursor, putative 6.2 6.4
SAOUHSC_02802 fnbB fibronectin binding protein B, putative 7.0 3.1
SAOUHSC_02709 hlgC leukocidin s subunit precursor, putative 5.3 17.9
SAOUHSC_02708 gamma-hemolysin h-gamma-ii subunit, putative 9.6 19.4
SAOUHSC_00196 fadB hypothetical protein SAOUHSC_00196 84.4 16.6
SAOUHSC_00232 lrgA antiholin-like protein lrgA 4.0 5.3
SAOUHSC_01921 hypothetical protein SAOUHSC_01921 4.8 2.2
SAOUHSC_02710 hlgB leukocidin f subunit precursor 3.4 13.4
SAOUHSC_01121 hla alpha-hemolysin precursor 1.9 3.2
SAOUHSC_02169 chp chemotaxis inhibitory protein 1.7 4.4
SAOUHSC_02161 eap MHC class II analog protein 1.8 11.0
SAOUHSC_01114 efb fibrinogen-binding protein 1.9 2.7
SAOUHSC_00816 emp extracellular matrix and plasma binding protein, putative -1.4 2.9
SAOUHSC_00715 saeR response regulator, putative -1.2 2.1
SAOUHSC_01115 hypothetical protein SAOUHSC_01115 1.8 3.0
SAOUHSC_00354 hypothetical protein SAOUHSC_00354 -1.7 2.9
SAOUHSC_00192 coa coagulase 6.3 1.8
SAOUHSC_00400 hypothetical protein SAOUHSC_00400 9.1 1.7
SAOUHSC_00399 hypothetical protein SAOUHSC_00399 10.4 1.9
SAOUHSC_00818 nuc thermonuclease precursor -4.4 1.2
SAOUHSC_02229 anti repressor 1.4 -2.9
SAOUHSC_00182 hypothetical protein SAOUHSC_00182 -2.6 -2.8
a Differentially regulated genes (significant with p* < 0.05 and with a minimal absolute fold change of 2) are indicated in bold.
Differential expression of virulence associated genes
Very important for colonization of the host, but also for internalization into host cells are the
staphylococcal membrane bound adhesins, MSCRAMMs. Genes coding for fibrinogen binding
protein A and B (fnbA, fnbB) and for clumping factor A and B (clfA, clfB) were induced in
internalized staphylococci 2.5 h after start of infection. Differential expression of fnbA and clfA
was still detectable at the 6.5 h time point whereas that of fnbB and clfB was not significant
anymore, although a trend of induction was visible according to the fold change values. The
control samples exhibited only less consistent expression changes. An exception might be the
expression of fnbA and clfA in the 6.5 h serum/CO 2 control. Here, a trend of induction was visible,
but the difference could not be determined statistically because of the number of replicates
(Fig. R.5.14 A, B). Soluble adhesins bind to host structures. They can mediate bacterial cell
attachment indirectly by involving linker molecules. Some alternatively exist in cell-surface
associated forms. IsaB has an intermediate position as membrane bound and secreted protein,
which was recently discovered to bind extracellular dsDNA and which is implicated in virulence,
but whose exact function is still not identified (Mackey-Lawrence et al. 2009). In staphylococci
internalized by S9 cells, expression of isaB, eap, coa, vwb, emp, and efb was induced in at least
one of the two analyzed time points of internalization. In this group, the repression of coa and
vwb occurred in the 2.5 h serum/CO 2 and 2.5 h anaerobic control samples, and vwb and efb were
significantly less expressed in the exponential growth samples with which inoculation of S9 cell
culture plates took place. Contrarily, the gene eap was two-fold induced in the 2.5 h serum/CO 2
controls, and isaB showed induction in anaerobiosis (Fig. R.5.14 A, B, C).
149

mean normalized intensity
mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
A
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
fnbA fnbB clfA clfB isaB eap coa vwb emp efb
exponential growth phase -1.8 -2.8 1.6 -1.7 1.3 -2.0 -2.2 -5.8 -1.6 -2.8
2.5 h internalized 6.2 7.0 5.3 2.1 2.1 1.8 6.3 4.9 -1.4 1.9
6.5 h internalized 6.4 3.1 2.8 2.1 -1.3 11.0 1.8 2.7 2.9 2.7
2.5 h serum/CO 2 control 1.1 -2.9 2.5 -1.6 1.5 2.0 -3.6 -7.7 -1.9 -1.3
6.5 h serum/CO 2 control 3.4 -2.0 9.2 -3.3 -2.1 4.7 -5.7 -4.2 -1.6 -1.3
2.5 h anaerobic incubation 1.3 -1.2 3.3 -1.4 4.7 3.1 -2.7 -3.3 1.6 1.2
B
C
10.0
10.0
fnbA
eap
fnbB
coa
clfA
vwb
1.0
clfB
isaB
1.0
emp
efb
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
Fig. R.5.14: Adhesins (MSCRAMMs and SERAMs) gene expression.
A. Overview on fold change values relative to the baseline sample “1 h serum/CO 2 control” and on significance in statistical group
comparisons. Differentially expressed genes are marked by gray filling for the corresponding sample condition/time point.
Genes which were significant in statistical testing (p* < 0.05) but did not pass the absolute fold change cutoff 2 are indicated in bold
without filling. The sample “6.5 h serum/CO 2 control” could not be included in statistical testing because of small group size (n = 2).
B, C. Overview on mean normalized gene expression intensity for MSCRAMM (B) and SERAM (C) genes. After the global normalization
by inter-chip scaling and detrending, each individual gene has been normalized to the expression level of the baseline sample “1 h
serum/CO 2 control”. Although not being continuous data, values were depicted as line graphs instead of bar charts for better facility
of inspection.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2 control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h internalization;
2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic incubation)
Toxins, especially lytic toxins, are one of the most outstanding groups of virulence factors
because they directly harm the host cells. In the group of toxins, two bicomponent toxin gene
pairs were observed to be induced in internalized staphylococci: lukD/lukE and hlgB/hlgC.
Furthermore, alpha-hemolysin (hla) was induced. Induction for all genes was significant at the
6.5 h time point except for hlgB, which was not significant although its fold change was the
second highest in this group. The pair hlgB/hlgC was additionally already induced in internalized
staphylococci 2.5 h after start of infection, and also in the 2.5 h serum/CO 2 control. Again, the
6.5 h serum/CO 2 control exhibited an even further increased fold change, but could not be tested
statistically (Fig. R.5.15 A, B).
Further virulence factors, secreted and membrane bound enzymes, were observed with
divergent expression pattern in internalized staphylococci. All six members A, B, C, D, E, F of the
spl operon, coding for serine proteases with similarity to the V8 protease, were found to be
induced at the 6.5 h time point of internalization. Four of them, A, B, C, and E, were additionally
induced 2.5 h after start of infection. For this operon, a similar induction was observed in the
serum/CO 2 control samples. Another exoenzyme, lipase (lip), was induced in internalized
staphylococci, but the peak of induction occurred 2.5 h after start of infection, i. e. earlier than
for the spl operon. In serum/CO 2 control, lip gene expression did not change after 2.5 h, but a
strong induction was detectable after 6.5 h although statistical testing was not possible
(Fig. R.5.16 A, B). A second group of enzymes exhibited repression in internalized staphylococci:
hysA (hyaluronate lyase), htrA (serine protease, a membrane bound enzyme), nuc (nuclease), and
sspA (glutamyl endopeptidase, V8 peptidase).
150

mean normalized intensity
mean normalized intensity
mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
A
B
100.0
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
lukD lukE hlgB hlgC hla
exponential growth phase -1.1 -1.9 -2.9 -5.6 -5.2
2.5 h internalized -1.3 -1.3 3.4 5.3 1.9
6.5 h internalized 6.5 9.0 13.4 17.9 3.2
2.5 h serum/CO 2 control 1.1 1.0 3.1 3.5 1.3
6.5 h serum/CO 2 control 1.4 1.1 8.0 8.9 1.7
2.5 h anaerobic incubation -1.5 -2.0 3.4 5.0 2.4
Fig. R.5.15: Staphylococcal toxin gene expression.
A. Overview on fold change values relative to the baseline sample “1 h serum/CO 2 control” and on significance in statistical group
comparisons. Differentially expressed genes are marked by gray filling for the corresponding sample condition/time point.
Genes which were significant in statistical testing (p* < 0.05) but did not pass the absolute fold change cutoff 2 are indicated in bold
without filling. The sample “6.5 h serum/CO 2 control” could not be included in statistical testing because of small group size (n = 2).
B. Overview on mean normalized gene expression intensity. After the global normalization by inter-chip scaling and detrending, each
individual gene has been normalized to the expression level of the baseline sample “1 h serum/CO 2 control”. Although not being
continuous data, values were depicted as line graphs instead of bar charts for better facility of inspection.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2 control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h internalization;
2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic incubation)
10.0
1.0
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
lukD
lukE
hlgB
hlgC
hla
A
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
splF splE splD splC splB splA lip hysA htrA nuc sspA sspB sspC
exponential growth phase -1.3 -1.7 -1.1 -3.3 -5.4 -4.7 -1.4 -1.5 -1.2 -2.6 -1.4 -1.3 1.0
2.5 h internalized 1.6 3.1 3.2 4.8 3.3 3.2 34.2 -2.5 -2.3 -4.4 -2.5 -1.4 -1.1
6.5 h internalized 14.9 25.9 32.0 43.1 35.5 30.4 10.7 -3.0 -2.0 1.2 -1.2 -1.1 1.3
2.5 h serum/CO 2 control 1.9 3.1 3.1 3.8 2.7 2.9 1.3 -2.6 -2.0 4.0 -1.3 -1.0 1.5
6.5 h serum/CO 2 control 9.6 15.8 17.1 22.2 15.4 10.8 27.7 -3.4 -3.1 -1.0 -1.0 1.4 1.7
2.5 h anaerobic incubation -1.0 1.5 1.5 2.2 1.5 1.7 1.9 -1.7 -1.5 1.0 1.1 -1.0 1.3
B
100.0
C
10.0
splF
hysA
10.0
1.0
splE
splD
splC
splB
splA
1.0
htrA
nuc
sspA
sspB
sspC
lip
0.1
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
Fig. R.5.16: Extracellular and membrane bound enzyme gene expression.
A. Overview on fold change values relative to the baseline sample “1 h serum/CO 2 control” and on significance in statistical group
comparisons. Differentially expressed genes are marked by gray filling for the corresponding sample condition/time point.
Genes which were significant in statistical testing (p* < 0.05) but did not pass the absolute fold change cutoff 2 are indicated in bold
without filling. The sample “6.5 h serum/CO 2 control” could not be included in statistical testing because of small group size (n = 2).
B, C. Overview on mean normalized gene expression intensity for induced extracellular enzyme (B) and repressed extracellular and
membrane bound enzyme (C) genes. After the global normalization by inter-chip scaling and detrending, each individual gene has
been normalized to the expression level of the baseline sample “1 h serum/CO 2 control”. Although not being continuous data, values
were depicted as line graphs instead of bar charts for better facility of inspection.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2 control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h internalization;
2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic incubation)
151

mean normalized intensity
mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
Additionally, repression was observed for hysA, htrA, and nuc in 2.5 h serum/CO 2 control.
While hysA and htrA were repressed in both analyzed time points of internalization, the
repression of sspA was only detected 2.5 h after start of infection. The other two genes of the ssp
operon were not detected as differentially expressed (Fig. R.5.16 A, C).
Staphylococci secrete proteins which help the bacterium to evade the host’s immune
response. In the experimental setting of the in vitro S9 infection and internalization model, some
of them were induced in S. aureus RN1HG after internalization in S9 cells, especially at the 6.5 h
time point: chp (chemotaxis inhibitory protein CHIPS), eap (extracellular adherence protein, also
called MHC class II analog protein Map), and efb (extracellular fibrinogen-binding protein).
Staphylokinase (sak) was repressed in internalized staphylococci at the 2.5 h time point
(Fig. R.5.17 A, B). Staphylococci own two genes coding for superoxide dismutases, sodA and
sodM. These two gene exhibited divergent regulation during internalization: sodA was induced
and sodM was repressed in at least one of the two analyzed time points (Fig. R.5.17 A, C).
A
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
chp eap efb sak sodA sodM
exponential growth phase -7.9 -2.0 -2.8 -2.8 2.4 -1.2
2.5 h internalized 1.7 1.8 1.9 -3.8 2.4 -1.7
6.5 h internalized 4.4 11.0 2.7 -1.9 3.2 -3.0
2.5 h serum/CO 2 control -1.2 2.0 -1.3 2.2 1.9 -1.1
6.5 h serum/CO 2 control -3.1 4.7 -1.3 -2.7 2.7 -2.5
2.5 h anaerobic incubation -4.6 3.1 1.2 1.1 1.3 -1.6
B
100.0
C
10.0
10.0
1.0
chp
eap
efb
sak
1.0
sodA
sodM
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
Fig. R.5.17: Immune evasion gene expression.
A. Overview on fold change values relative to the baseline sample “1 h serum/CO 2 control” and on significance in statistical group
comparisons. Differentially expressed genes are marked by gray filling for the corresponding sample condition/time point.
Genes which were significant in statistical testing (p* < 0.05) but did not pass the absolute fold change cutoff 2 are indicated in bold
without filling. The sample “6.5 h serum/CO 2 control” could not be included in statistical testing because of small group size (n = 2).
B, C. Overview on mean normalized gene expression intensity for different immune evasion (B) and superoxide dismutase (C) genes.
After the global normalization by inter-chip scaling and detrending, each individual gene has been normalized to the expression level
of the baseline sample “1 h serum/CO 2 control”. Although not being continuous data, values were depicted as line graphs instead of
bar charts for better facility of inspection.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2 control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h internalization;
2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic incubation)
Changes in the cell wall allow staphylococci defense against and evasion of the immune
response. The dlt operon for example is responsible for the incorporation of D-alanine into the
teichoic acids, which leads to a reduction of negative charge of the cell wall. Thus, these bacterial
cells are less vulnerable by antimicrobial cationic peptides due to a reduced attraction.
Surprinsingly, this operon was temporarily repressed at the 2.5 h time point of internalization. At
the same time and also 6.5 h after start of infection, the genes of cell wall modulating enzymes
lytM (peptidoglycan hydrolase, endopeptidase) and ssaA (secretory antigen precursor, amidase)
were induced in internalized staphylococci (Fig. R.5.18 A, B).
152

mean normalized intensity
mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
A
B
10.0 10,0
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
dltA dltB dltC dltD lytM ssaA
exponential growth phase 1.0 -1.0 -1.0 -1.2 2.6 1.2
2.5 h internalized -2.5 -2.2 -1.6 -2.2 2.4 3.0
6.5 h internalized -1.5 -1.3 -1.2 -1.3 2.1 2.9
2.5 h serum/CO 2 control 1.0 -1.0 1.1 -1.1 1.4 1.3
6.5 h serum/CO 2 control -2.5 -2.4 -2.5 -2.4 1.5 -1.9
2.5 h anaerobic incubation -2.0 -2.2 -1.9 -2.7 -3.3 -2.5
1.0 1,0
dltA
dltB
dltC
dltD
lytM
ssaA
Fig. R.5.18: Cell wall associated gene expression.
A. Overview on fold change values relative to the baseline sample “1 h serum/CO 2 control” and on significance in statistical group
comparisons. Differentially expressed genes are marked by gray filling for the corresponding sample condition/time point.
Genes which were significant in statistical testing (p* < 0.05) but did not pass the absolute fold change cutoff 2 are indicated in bold
without filling. The sample “6.5 h serum/CO 2 control” could not be included in statistical testing because of small group size (n = 2).
B. Overview on mean normalized gene expression intensity. After the global normalization by inter-chip scaling and detrending, each
individual gene has been normalized to the expression level of the baseline sample “1 h serum/CO 2 control”. Although not being
continuous data, values were depicted as line graphs instead of bar charts for better facility of inspection.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2 control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h internalization;
2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic incubation)
0.1 0,1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
Biofilm formation is another option for S. aureus to protect the bacterial cells from the
immune system and antimicrobials. Here, bacterial cells are surrounded by a polysaccharide
matrix, composed of poly-N-acetylglucosamine (PNAG). The ica operon encodes the proteins
necessary for PNAG biosynthesis. Of this operon, the icaB gene was repressed in internalized and
control samples of the infection experiment. The icaC and icaD genes behaved similarly, but were
significantly repressed only at the 2.5 h time point of internalization. The fourth gene of the
operon, icaA, and the separately encoded repressor gene, icaR, were not differentially expressed
(Fig. R.5.19 A, B). Furthermore, it was obvious that the expression intensity of icaADBC was low.
A
B
10,0 10.0
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
icaR icaA icaD icaB icaC
exponential growth phase -1.0 2.0 1.6 1.7 1.6
2.5 h internalized -1.2 1.4 -2.2 -3.7 -2.8
6.5 h internalized -1.0 1.0 -1.3 -2.2 -1.5
2.5 h serum/CO 2 control -1.4 1.3 1.1 -2.2 1.1
6.5 h serum/CO 2 control -1.2 1.2 1.5 -1.9 -1.4
2.5 h anaerobic incubation -1.3 1.1 -1.6 -2.4 -1.4
Fig. R.5.19: Biofilm associated gene expression.
A. Overview on fold change values relative to the baseline sample “1 h serum/CO 2 control” and on significance in statistical group
comparisons. Differentially expressed genes are marked by gray filling for the corresponding sample condition/time point.
Genes which were significant in statistical testing (p* < 0.05) but did not pass the absolute fold change cutoff 2 are indicated in bold
without filling. The sample “6.5 h serum/CO 2 control” could not be included in statistical testing because of small group size (n = 2).
B. Overview on mean normalized gene expression intensity. After the global normalization by inter-chip scaling and detrending, each
individual gene has been normalized to the expression level of the baseline sample “1 h serum/CO 2 control”. Although not being
continuous data, values were depicted as line graphs instead of bar charts for better facility of inspection.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2 control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h internalization;
2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic incubation)
1,0 1.0
0,1 0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
icaR
icaA
icaD
icaB
icaC
153

mean normalized intensity
mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
Besides the beforementioned two-component system SaeRS, whose gene expression was
increased, further regulators were differentially expressed in the internalization experiment. The
transcription factors SarT, SarU, and Rot belong to the virulence associated Sar family. SarT and
sarU were repressed in internalized staphylococci at the 6.5 h time point, whereas rot was
repressed at the 2.5 h time point (Fig. R.5.20 A, B). Expression intensity of sarU gene was very
low.
A
B
10.00
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
saeR saeS sarT sarU rot
exponential growth phase -1.8 -1.8 2.1 -1.0 -1.6
2.5 h internalized -1.2 -1.2 -1.6 -1.7 -2.4
6.5 h internalized 2.1 1.7 -4.6 -2.2 -1.7
2.5 h serum/CO 2 control 1.7 1.5 -4.5 2.6 1.6
6.5 h serum/CO 2 control 2.5 1.9 -13.3 1.2 -1.2
2.5 h anaerobic incubation 1.9 1.5 -1.0 1.1 1.1
Fig. R.5.20: Expression of regulator genes.
A. Overview on fold change values relative to the baseline sample “1 h serum/CO 2 control” and on significance in statistical group
comparisons. Differentially expressed genes are marked by gray filling for the corresponding sample condition/time point.
Genes which were significant in statistical testing (p* < 0.05) but did not pass the absolute fold change cutoff 2 are indicated in bold
without filling. The sample “6.5 h serum/CO 2 control” could not be included in statistical testing because of small group size (n = 2).
B. Overview on mean normalized gene expression intensity. After the global normalization by inter-chip scaling and detrending, each
individual gene has been normalized to the expression level of the baseline sample “1 h serum/CO 2 control”. Although not being
continuous data, values were depicted as line graphs instead of bar charts for better facility of inspection.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2 control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h internalization;
2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic incubation)
1.00
0.10
0.01
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
saeR
saeS
sarT
sarU
rot
A
B
10.0
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
prsA vraR vraS
exponential growth phase 1.1 -1.1 -1.1
2.5 h internalized 2.0 1.8 1.6
6.5 h internalized 1.6 1.4 1.3
2.5 h serum/CO 2 control -1.2 -1.2 -1.1
6.5 h serum/CO 2 control -1.9 -1.1 -1.0
2.5 h anaerobic incubation 1.9 1.4 1.6
1.0
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
prsA
vraR
vraS
Fig. R.5.21: Gene expression of prsA and its regulatory two-component system’s genes vraR and vraS.
A. Overview on fold change values relative to the baseline sample “1 h serum/CO 2 control” and on significance in statistical group
comparisons. Differentially expressed genes are marked by gray filling for the corresponding sample condition/time point.
Genes which were significant in statistical testing (p* < 0.05) but did not pass the absolute fold change cutoff 2 are indicated in bold
without filling. The sample “6.5 h serum/CO 2 control” could not be included in statistical testing because of small group size (n = 2).
B. Overview on mean normalized gene expression intensity. After the global normalization by inter-chip scaling and detrending, each
individual gene has been normalized to the expression level of the baseline sample “1 h serum/CO 2 control”. Although not being
continuous data, values were depicted as line graphs instead of bar charts for better facility of inspection.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2 control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h internalization;
2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic incubation)
154

Maren Depke
Results
Pathogen Gene Expression Profiling
Proteome analysis of internalized staphylococci, which was performed by Sandra Scharf,
revealed the increased abundance of PrsA. In parallel, the response regulator VraR of the twocomponent
system VraRS was up-regulated. Also transcriptome analysis revealed induction of
prsA in internalized staphylococci at the time point 2.5 h after start of infection. The regulatory
genes vraR and vraS were not differentially expressed. Nevertheless, with a fold change of 1.8
and 1.6 for vraR and vraS, respectively, a tendency for an induction of gene expression was
visible, which fitted well to the proteome data (Fig. R.5.21 A, B).
The VraRS regulon contains 46 genes including vraR and vraS (Kuroda et al. 2003). Of the
published 46 S. aureus N315 LocusTags, 45 could be mapped to S. aureus NCTC8325 LocusTags
(NCBI’s EntrezGenome Gene Plot; http://www.ncbi.nlm.nih.gov/sutils/geneplot.cgi). When asking
for the fraction of the VraRS regulon which was regulated in internalized S. aureus RN1HG in S9
cells, 22 % (2.5 h after start of infection) and 11 % (6.5 h after start of infection) were differentially
expressed upon internalization. In detail, 7 genes were induced 2.5 h after start of infection, of
which 4 were still induced at the later time point of 6.5 h after start of infection. Repressed gene
expression was also observed: Repression was detected for 3 genes at the early 2.5 h time point,
and one of these genes still was repressed at the later 6.5 h time point (Fig. R.5.22 A, B).
A
VraRS regulon according to
Kuroda et al. 2003
B
VraRS regulon according to
Kuroda et al. 2003
35
40
7 3
0
395 360
0
4 1
0
315 307
0
induced sequences in
“2.5 h internalization” vs.
“1 h serum/CO 2 control”
repressed sequences in
“2.5 h internalization” vs.
“1 h serum/CO 2 control”
induced sequences in
“6.5 h internalization” vs.
“1 h serum/CO 2 control”
repressed sequences in
“6.5 h internalization” vs.
“1 h serum/CO 2 control”
Fig. R.5.22: Differential expression of genes belonging to the VraRS regulon in internalized S. aureus RN1HG in S9 cells.
Genes known to be regulated by VraRS (according to Kuroda et al. 2003) were compared to S. aureus RN1HG S9 internalization gene
expression signatures separately for induced and repressed genes.
A. 2.5 h after start of infection.
B. 6.5 h after start of infection.
Differentially regulated metabolic genes
First, BIOCYC “omics-Viewer” pathway mapping (BIOCYC, SRI International, CA, USA,
http://biocyc.org/expression.html) was applied for a comprehensive overview on changes in
gene expression of metabolic enzymes in internalized and control staphylococci. This tool allows
the display of gene expression data on highly abstracted metabolic pathway schemes and
therefore an intuitive comprehension of processes in the selected experimental setup.
In comparison to the baseline of 2.5 h serum/CO 2 control samples, differentially expressed
sequences in a) 2.5 h internalization, b) 6.5 h internalization, c) 2.5 h serum/CO 2 control, and
d) 2.5 h anaerobic incubation were included in the analysis.
Several changes in the expression of metabolic genes became visible, but certain functionally
associated groups of genes accumulated changes in expression (Fig. R.5.23).
155

Maren Depke
Results
Pathogen Gene Expression Profiling
A
phosphonate/
phosphate
transporters
uracil
permease
oligopeptide
transporters
spermidine/
putrescine
transporter
purine and
pyrimidine
nucleotide
biosynthesis
sodium/sulfate, fructose, glycine
betaine, choline, L-lactate, xanthine,
urea transporter/symporter/permease
acetoin
biosynthesis
different
fermentation
pathways
degradation of glycerol, mannitol,
galactose, fructose, nucleosides,
citrulline/ornithine, arg, his, ala
and others
maltose transporter
protease
proline betaine transporter
amino acid transporter
phospho-transferase system
(PTS); glucose-and mannitolspecific
components
glycolysis
fatty acid
biosynthesis
arginine/
ornithine
antiporter
amino acid
biosynthesis
sodium-dependent tr.,
phosphate transporter
sodium/glutamate symporter,
proline uptake protein,
citrate transporter
tRNA charging
pathways
anaerobic
respiration
aerobic
respiration
gluconate
permease
protein modification: oxoacylreductase,
ketoacyl-reductase,
oxoacyl-synthase reactions at
different C positions
B
choline, L-lactate transporter
different
fermentation
pathways
acetoin
biosynthesis
degradation of glycerol, mannitol,
galactose, fructose, nucleosides,
citrulline/ornithine, arg, his, ala
and others
protease
phosponate/
phosphate and
oligopeptide
transporters
spermidine/
putrescine
transporter
proline betaine transporter
amino acid transporter
phospho-transferase system
(PTS); glucose-and mannitolspecific
components
purine and
pyrimidine
nucleotide
biosynthesis
glycolysis
fatty acid
biosynthesis
arginine/
ornithine
antiporter
amino acid
biosynthesis
phosphate
transporter
tRNA charging
sodium/glutamate symporter pathways
proline uptake protein
anaerobic
respiration
aerobic
respiration
gluconate
permease
protein modification: oxoacylreductase,
ketoacyl-reductase,
oxoacyl-synthase reactions at
different C positions
156

Maren Depke
Results
Pathogen Gene Expression Profiling
C
peptidoglycan
biosynthesis
fructose, glycine betaine
permease/transporter
different
fermentation
pathways
acetoin
biosynthesis
degradation of glycerol, mannitol,
galactose, fructose, nucleosides,
citrulline/ornithine, arg, his, ala
and others
ammonium transporter
gluconeogenesis
uracil
permease
oligopeptide
transporters
spermidine/
putrescine
transporter
protease
proline betaine transporter
branched and standard
amino acid transporter
phospho-transferase system
(PTS); glucose-and mannitolspecific
components
purine and
pyrimidine
nucleotide
biosynthesis
glycolysis
fatty acid
biosynthesis
amino acid
biosynthesis
sodium
glutamate
symporter
proline uptake protein
tRNA charging
pathways
anaerobic
respiration
aerobic
respiration
protein modification: oxoacylreductase,
ketoacyl-reductase,
oxoacyl-synthase reactions at
different C positions
D
fructose, glycine betaine, choline,
transporter/permease
Na + /H + antiporter
Mo 2+ transporter
different
fermentation
pathways
acetoin
biosynthesis
degradation of glycerol, mannitol,
galactose, fructose, nucleosides,
citrulline/ornithine, arg, his, ala
and others
S-, L-lactate permease
proline betaine transporter
monovalent
cation/proton
antiporter
oligopeptide
transporter
ferrichrome
transport
permease;
monovalent
cation/proton
antiporter;
spermidine/
putrescine
transporter
fatty acid
biosynthesis
arginine/
ornithine
antiporter
amino acid
biosynthesis
nickel permease,
Na + /H + antiporter
sodium/glutamate symporter
proline uptake protein
tRNA charging
pathways
Na + /H + antiporter
ferrichrome ABC
transporter
branched amino acid
transporter
standard amino acid
transporter
phospho-transferase system
(PTS); glucose-and mannitolspecific
components
protein modification: oxoacylreductase,
ketoacyl-reductase,
oxoacyl-synthase reactions at
different C positions
Fig. R.5.23: The influence of internalization and control treatment on gene expression in staphylococcal NCTC8325 metabolic
pathways (modified from omics-viewer(s) of BIOCYC, SRI International, CA, USA, http://biocyc.org/expression.html).
Nodes represent metabolites, and lines indicate reactions. The metabolic reactions are colored according to the enzymes’ gene
expression regulation in the samples of 2.5 h internalization (A), 6.5 h internalization (B), 2.5 h serum/CO 2 control (C), and 2.5 h
anaerobic incubation; each sample was compared versus the baseline of 1 h serum/CO 2 control. Red marks an increase and yellow a
decrease of expression in the sample. The display is limited to genes which were significant with p* < 0.05 in statistical testing and
whose mean absolute fold change exceeded 2.
157

mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
At the early internalization time point, several genes of purine and pyrimidine nucleotide
biosynthesis pathways were repressed. Amino acid biosynthesis genes exhibited increase of
expression in both analyzed internalization time points. Furthermore, genes coding for
respiration chain components were induced in expression in internalized staphylococci. In
internalized staphylococci, several tRNA synthetase genes and some glycolytic enzyme genes
were observed to be repressed. Finally, transporters and permeases for sugar molecules,
phosphate, oligopeptides, and other substances were differentially expressed (Fig. R.5.23 A, B).
Serum/CO 2 control samples featured in some aspects similar changes in metabolic enzyme’s gene
expression pattern, e. g. amino acid biosynthesis, glycolysis, tRNA synthetases. On the other
hand, aerobic respiration, sugar uptake phosphotransferase system, and some enzymes of
peptidoglycan biosynthesis were repressed in this group. Anaerobic incubation instead did not
provoke changes in expression of glycolytic enzymes or respiration chain components
(Fig. R.5.23 C, D).
Subsequently to the first approach of BIOCYC pathway mapping, selected genes were analyzed
in more detail. Here, the example of purine biosynthesis was chosen. Although not all genes were
significantly differentially expressed, the gene expression changes were highly similar in the
complete pur operon and guaAB, which are separately encoded but are also involved in purine
biosynthesis. Five genes (purA, purC, purE, purK, purS) were differentially expressed in both
analyzed internalization time points, six genes (purD, purF, purL, purQ, guaA, guaB) were
differentially expressed 2.5 h after start of infection. Three genes, purH, purM, and purN,
exhibited fold change values between −2.6 and −2.8, but were not significant in statistical testing.
Finally, purB was significant in the statistical test in both analyzed time points of internalization,
but did not pass the minimal absolute fold change cutoff of 2. In the 2.5 h serum/CO 2 control,
differential expression occurred only for purA and in trend for purB, whereas almost all genes
exhibited fold change values around −3 to −4 in the 6.5 h serum/CO 2 control. Exceptions were
guaA and guaB, whose fold change values were still around 1, and the repressor gene purR,
which did not exhibit expression changes in any of the analyzed experimental conditions
(Fig. R.5.24, Fig. R.5.25).
A
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
purA purB purC purD purE purF purH purK purL purM purN purQ purR purS guaA guaB
exponential growth phase -1.1 -1.0 1.2 1.3 1.2 1.1 1.2 1.2 1.1 1.1 1.2 1.1 -1.2 1.0 -1.0 -1.1
2.5 h internalized -4.8 -1.6 -4.5 -2.6 -6.2 -3.0 -2.6 -5.4 -3.4 -2.8 -2.6 -4.0 1.1 -4.9 -2.0 -2.3
6.5 h internalized -4.8 -1.6 -2.2 -1.1 -3.4 -1.2 -1.2 -2.9 -1.5 -1.2 -1.1 -1.9 1.3 -2.3 -1.3 -1.4
2.5 h serum/CO 2 control -6.4 -1.6 1.0 1.1 1.0 -1.1 -1.0 1.1 -1.0 -1.0 1.0 1.0 1.3 -1.1 1.4 1.3
6.5 h serum/CO 2 control -28.2 -2.7 -3.2 -3.4 -3.8 -4.1 -4.3 -3.1 -4.0 -3.9 -3.6 -4.0 1.1 -3.9 1.1 -1.0
2.5 h anaerobic incubation -2.1 -1.7 1.4 1.1 1.7 1.3 1.2 1.5 1.3 1.3 1.2 1.5 -1.1 1.1 1.0 1.0
Fig. R.5.24: Purine biosynthesis gene expression.
A. Overview on fold change values relative to the baseline
sample “1 h serum/CO 2 control” and on significance in statistical
group comparisons. The sample “6.5 h serum/CO 2 control” could
not be included in statistical testing because of small group size
(n = 2).
B. Overview on mean normalized gene expression intensity.
After the global normalization by inter-chip scaling and
detrending, each individual gene has been normalized to the
expression level of the baseline sample “1 h serum/CO 2 control”
(1 h co.). Although not being continuous data, values were
depicted as line graphs instead of bar charts for better facility of
inspection.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2
control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h
internalization; 2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. −
6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic
incubation)
B
1.00
0.10
0.01
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
purA
purB
purC
purD
purE
purF
purH
purK
purL
purM
purN
purQ
purR
purS
guaA
guaB
158

Maren Depke
Results
Pathogen Gene Expression Profiling
2
0
-2
-4
-6
-8
01014
purF
2
0
-2
-4
-6
-8
01018
purD
2
0
-2
-4
-6
-8
01016
purN
2
0
-2
-4
-6
-8
01013
purL
2
0
-2
-4
-6
-8
01012
purQ
2
0
-2
-4
-6
-8
01011
purS
2
0
-2
-4
-6
-8
01015
purM
2
0
-2
-4
-6
-8
2
0
-2
-4
-6
-8
01009
purK
01008
purE
2
0
-2
-4
-6
-8
01010
purC
2
0
-2
-4
-6
-8
02126
purB
2
0
-2
-4
-6
-8
00467
purR
2
0
-2
-4
-6
-8
01017
purH
2
0
-2
-4
-6
-8
00019
purA
2
0
-2
-4
-6
-8
02126
purB
AMP
PRPP
I-5‘-
GMP
PRPP : 5-phosphoribosyl-1-pyrophosphate
I-5‘- : inosine-5‘-phosphate
AMP : adenosine-5‘-monophosphate
GMP : guanosine-5‘-monophosphate
00374
guaB
2
0
-2
-4
-6
-8
00375
guaA
2
0
-2
-4
-6
-8
Fig. R.5.25: Purine biosynthesis pathway and gene expression.
Pathway reactions and gene LocusTags were extracted from omics-viewer(s) of BIOCYC (BIOCYC, SRI International, CA, USA,
http://biocyc.org/expression.html). Arrows represent reactions and dots substrates/products when these are not specially named.
Numbers next to enzyme reactions will result in the LocusTag of the gene when combined with “SAOUHSC_*” instead of the wildcard
character. Bar charts indicate fold change values of internalized samples in reference to the baseline sample of 1 h serum/CO 2 control.
Values for the 2.5 h internalized samples are depicted in light gray (first column) and those of the 6.5 h internalized samples in dark
gray (second column).
A third analysis of metabolic gene expression involved the KEGG pathway maps (KEGG: Kyoto
Encyclopedia of Genes and Genomes). Already BIOCYC pathway mapping had revealed changes in
amino acid biosynthesis and glycolysis. Corresponding pathways and additionally pathways of
TCA and urea cycle were selected in KEGG specific for S. aureus NCTC8325. The strain’s genes,
which were mapped by KEGG to the pathways, were extracted and checked for differential
expression in at least one time point after internalization in S9 cells. The biosynthesis pathways
for asparagine, histidine, glutamine, serine, tryptophan, leucine, valine, isoleucine, lysine, and
threonine were induced to a large extent (Fig. R.5.26). A fraction of associated genes were not
included in the lists of differentially expressed genes. A further analysis of these genes identified
some of them to be either significant in statistical testing without passing the minimal absolute
fold change cutoff or to pass that cutoff without reaching statistical significance. These
expression changes were rated as trend, which in many cases further confirmed the findings of
expression changes in the metabolic pathways. In addition to induced expression of amino acid
biosynthesis genes, induction of four TCA cycle enzyme genes was observed, which was even
more substantiated by a trend of increase for further four genes. Contrarily, especially genes
encoding enzymes of irreversible glycolysis reactions were repressed. Finally, genes coding for
enzymes of anabolic reactions during gluconeogenesis were induced in internalized
staphylococci. Using energy, their gene products reverse the reactions whose equilibrium is
normally set to the side of glucose degradation products.
Repressed tRNA synthetase gene expression has already been mentioned before. The
synthetase genes for tyrosine, leucine, threonine, phenylalanine, serine, aspartate, and histidine
(tyrS, leuS, thrS, pheS, pheT, serS, aspS, hisS) were repressed in at least one time point of
internalization, many of them even in both (Fig. R.5.27 A). Further genes (ileS, alaS, metS, asnS,
gltX, trpS) exhibited a trend of repression in at least one time point of internalization with
p* < 0.05, but an absolute fold change of less than 2 (Fig. R.5.27 B). Three genes (valS, glyS, cysS)
behaved contrarily and possessed a trend of induction in one time point of internalization
(Fig. R.5.27 C). Finally, the remaining three tRNA synthetase genes (argS, proS, lysS) showed no
change in expression (Fig. R.5.27 D).
159

Maren Depke
Results
Pathogen Gene Expression Profiling
purine
metabolism
glycerol
01276
glpK
lactate
00206 ldh1
02922 ldh2
pyruvate
01818 ald1 01452 ald2
alanine
asparagine
01497
ahsA
aspartate
00899
argG
imidazolacetolphosphate
00733
hisC
02607
hutU
AICAR
03008 hisF
03010 hisG
03009
hisA
alanine aspartate
02916
panD
glycolysis /
gluconeogenesis
glucose
01430 crr (PTSIIA)
00209 (PTSIIBC)
00900 pgi
fru-6-P
02822 fbp
01807 pfkA
fru-1,6-BP
02366 fba
DHAP
00797
tpiA
GA-3-P
01794 gap2
00795 gap1
00796 pgk
01833
serA
00798 pgm
01910 pckA
PEP
pyruvate
01806
pykA
02289
ilvA
01064 pycA
00898
argH
malate
fumarate
01983 fumC
oxalo
acetate
01103 sdhC
01104 sdhA
01105 sdhB
succinate
01216
sucC
01218
sucD
01802
citZ
citrate
succinyl-
CoA
TCA
cycle
03013
hisD
histidine
02607
hutU
03014
hisG
phosphoribosylpyrophosphate
pentose
phosphate
pathway
01394
lysC
01395
asd
01396
dapA
01397
dapB
01398
dapD
01319
thrA
01322 thrB
homoserine
01320 dhoM
00013
acetylhomoserine
01321 thrC
cystathionine
02286
leuB
threonine
02289 ilvA
2-oxobutanoate
cysteine
02287 leuC
02288 leuD
pyruvate
02282
ilvB
acetyl-CoA
indole
acetaldehyde
indole
acetate
00153
ipdC
indole
pyruvate
indole
serine
01371
trpB
tryptophan
01371
trpB
01846
acsA
00132 aldA
02363
(02142 aldA2)
01347
citB
01416
sucB
01418
sucA
01801 citC
00895 gudB
NH 3
02606
hutI
02610
hutG
glutamate
02869
rocA
00148
arcJ
2-oxoglutarate
pyrroline-5-
carboxylate
02244
01401
lysA
lysine
homocysteine
methionine
isoleucine
02284
ilvC
02281
ilvD
valine
02285 leuA
02287 leuC
02288 leuD
02919 panB
02739
panE
pantoate
00286
leuB
spontaneous
01369
trpC
01370
trpF
chorismate
prephenate
01364
tyrA
00113 adhE
00608 adhA
00733 hisC
phenylalanine tyrosine
argininosuccinate
carbymoylphosphate
citruline
02968
arcB
02969
arcA
00898
argH
00149
argC
00150
urea
cycle
00894
proline
02409 arg, rocF
urea
glutamate
semialdehyde
leucine
acetyl-CoA acetate acetaldehyde ethanol
ornithine arginine
Fig. R.5.26: Pathway map of amino acid biosyntheses, glycolysis/gluconeogenesis, TCA cycle and urea cycle.
Pathway reactions were extracted from KEGG (KEGG: Kyoto Encyclopedia of Genes and Genomes, http://www.genome.jp/kegg/) and
arranged according to genes differentially expressed in S9 cell internalized staphylococci in reference to the baseline sample of 1 h
serum/CO 2 control. Arrows represent reactions and dots substrates/products when these are not specially named. Numbers next to
enzyme reactions will result in the LocusTag of the gene when combined with “SAOUHSC_*” instead of the wildcard character.
Colored arrows in combination with colored gene names indicate differential expression, whereas a colored gene name alone (with a
black arrow) indicates a trend of regulation when only one criterium, p* < 0.05 or minimal absolute fold change > 2, is fulfilled.
Induction of gene expression is marked in red and repression in green color.
160

mean normalized intensity
mean normalized intensity
mean normalized intensity
mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
A
fold change in comparison to
baseline 1 h serum/CO 2 control
tyrS leuS thrS pheS pheT serS aspS hisS ileS alaS metS asnS gltX trpS valS glyS cysS argS proS lysS
S. aureus
RN1HG
sample
conditions
and time
points
exponential growth phase -1.1 -1.0 -1.4 -1.2 -1.3 -1.2 -1.3 -1.2 -1.2 -1.2 -1.1 -1.1 -1.1 1.2 -1.2 3.3 -1.0 -1.1 -1.1 -1.0
2.5 h internalized -2.7 -2.0 -2.1 -2.5 -2.7 -2.2 -1.6 -1.8 -1.5 -1.3 -1.6 -2.0 -1.5 -1.4 -1.1 -1.1 1.5 -1.3 -1.0 -1.2
6.5 h internalized -3.3 -2.3 -2.4 -3.4 -3.5 -2.4 -2.1 -2.3 -1.8 -1.5 -1.4 -1.8 -1.2 -1.3 1.6 1.9 -1.1 -1.2 1.1 -1.0
2.5 h serum/CO 2 control -5.1 -3.0 -2.8 -3.8 -3.9 -2.9 -4.6 -4.5 -3.3 -1.8 -1.2 -1.7 -1.2 1.1 -2.4 2.3 1.2 1.2 -1.6 -1.5
6.5 h serum/CO 2 control -11.2 -5.0 -3.6 -6.2 -4.8 -3.2 -7.5 -8.8 -4.4 -1.8 -1.5 -2.1 -1.3 -1.0 -2.3 1.6 2.0 1.1 -2.0 -1.6
2.5 h anaerobic incubation -3.9 -2.0 -1.5 -2.1 -2.3 -1.9 -3.8 -3.5 -2.9 -1.8 -1.2 -1.5 -1.4 1.0 -2.0 2.1 1.1 1.1 -1.9 -1.9
B
C
10.0
10.0
tyrS
ileS
leuS
alaS
thrS
metS
1.0
pheS
pheT
serS
1.0
asnS
gltX
trpS
aspS
hisS
0.1
0.1
D
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
E
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
10.0
10.0
1.0
valS
glyS
cysS
1.0
argS
proS
lysS
0.1
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
Fig. R.5.27: tRNA synthetase gene expression.
A. Overview on fold change values relative to the baseline sample “1 h serum/CO 2 control” and on significance in statistical group
comparisons. The sample “6.5 h serum/CO 2 control” could not be included in statistical testing because of small group size (n = 2).
B-E. Overview on mean normalized gene expression intensity. After the global normalization by inter-chip scaling and detrending,
each individual gene has been normalized to the expression level of the baseline sample “1 h serum/CO 2 control” (1 h co.). Although
not being continuous data, values were depicted as line graphs instead of bar charts for better facility of inspection.
Twenty tRNA synthetase genes were assigned to four groups: genes repressed in at least one time point of internalization (B), genes
with trend of repression in at least one time point of internalization with p* < 0.05, but an absolute fold change of less than 2 (C),
genes with trend of induction in one time point of internalization (D), and genes exhibiting no change in expression (E).
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2 control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h internalization;
2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic incubation)
After having observed the described changes in metabolic gene expression, the possible
differential expression of transporter genes was addressed directly. Here, two groups could be
distinguished. The first group consisted of transporter genes with repressed gene expression in at
least one analyzed time point of internalized staphylococci. The urea transporter
SAOUHSC_02557 belonged to this group. Noticeably, the adjacent urease operon
SAOUHSC_02558 to SAOUHSC_02565 (ureABCEFGD) possessed a similar expression pattern,
although not all genes were detected as differentially expressed (Fig. R.5.28).
161

mean normalized intensity
mean normalized intensity
mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
A
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
SAOUHSC_
02557 ureA ureB ureC ureE ureF ureG ureD
exponential growth phase 1.6 3.0 2.2 1.9 1.5 1.4 1.2 1.2
2.5 h internalized -2.6 -2.9 -3.2 -3.0 -1.9 -2.2 -1.9 -2.0
6.5 h internalized -1.1 -1.7 -1.6 -1.4 -1.3 -1.6 -1.4 -1.3
2.5 h serum/CO 2 control -1.8 -3.3 -3.1 -2.1 -1.2 -1.3 -1.3 -1.4
6.5 h serum/CO 2 control -1.5 -2.3 -2.0 -1.1 -1.1 -1.2 -1.4 -1.2
2.5 h anaerobic incubation -1.3 5.0 4.8 3.1 1.7 1.3 1.1 -1.1
Fig. R.5.28: Urea transporter and urease operon gene
expression.
A. Overview on fold change values relative to the baseline
sample “1 h serum/CO 2 control” and on significance in statistical
group comparisons. The sample “6.5 h serum/CO 2 control”
could not be included in statistical testing because of small
group size (n = 2).
B. Overview on mean normalized gene expression intensity.
After the global normalization by inter-chip scaling and
detrending, each individual gene has been normalized to the
expression level of the baseline sample “1 h serum/CO 2 control”
(1 h co.). Although not being continuous data, values were
depicted as line graphs instead of bar charts for better facility of
inspection. (exp. − exponential growth phase; 1 h co. – 1 h
serum/CO 2 control; 2.5 h int. − 2.5 h internalization; 6.5 h int. −
6.5 h internalization; 2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h
co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic
incubation)
B
10.0
1.0
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
SAOUHSC_02557
ureA
ureB
ureC
ureE
ureF
ureG
ureD
A
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
potA potB potC potD gltS arcD opuD opuCd opuCc opuCb opuCa
exponential growth phase -1.4 -1.4 -1.6 -1.4 -1.1 -1.5 1.0 -1.0 -1.1 -1.1 -1.0
2.5 h internalized -7.1 -8.4 -9.2 -4.9 -2.2 -2.4 -2.1 -3.3 -3.3 -4.6 -4.2
6.5 h internalized -3.9 -4.4 -4.9 -3.7 -3.0 -2.3 -1.7 -1.6 -2.1 -2.7 -3.0
2.5 h serum/CO 2 control -3.3 -3.5 -4.1 -3.1 -4.4 1.4 -4.7 2.1 2.0 1.9 2.1
6.5 h serum/CO 2 control -4.8 -4.3 -5.5 -2.7 -6.0 2.0 -7.0 1.2 1.0 -1.1 -1.0
2.5 h anaerobic incubation -2.9 -3.5 -4.5 -3.7 -4.0 2.1 -3.9 3.2 3.7 3.7 4.6
B
C
10.0
10.0
potA
opuD1
potB
opuCd
1.0
potC
potD
1.0
opuCc
opuCb
gltS
opuCa
arcD
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
Fig. R.5.29: Spermidine/putrescine, sodium/glutamate, arginine/ornithine, glycine betaine and amino acid transporter gene
expression.
A. Overview on fold change values relative to the baseline sample “1 h serum/CO 2 control” and on significance in statistical group
comparisons. The sample “6.5 h serum/CO 2 control” could not be included in statistical testing because of small group size (n = 2).
B, C. Overview on mean normalized gene expression intensity. After the global normalization by inter-chip scaling and detrending,
each individual gene has been normalized to the expression level of the baseline sample “1 h serum/CO 2 control” (1 h co.). Although
not being continuous data, values were depicted as line graphs instead of bar charts for better facility of inspection.
Spermidine/putrescine ABC transporter potABCD, sodium/glutamate symporter gltS, arginine/ornithine antiporter arcD (B) and
glycine betaine transporter opuD1 and amino acid ABC transporter opuCdCcCbCa (C).
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2 control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h internalization;
2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic incubation)
162

mean normalized intensity
mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
Fig. R.5.30:
Choline, citrate, xanthine, fructose, and lactate
transporter gene expression.
A. Overview on fold change values relative to the baseline
sample “1 h serum/CO 2 control” and on significance in
statistical group comparisons. Differentially expressed
genes are marked by gray filling for the corresponding
sample condition/time point.
Genes which were significant in statistical testing
(p* < 0.05) but did not pass the absolute fold change
cutoff 2 are indicated in bold without filling. The sample
“6.5 h serum/CO 2 control” could not be included in
statistical testing because of small group size (n = 2).
B. Overview on mean normalized gene expression
intensity. After the global normalization by inter-chip
scaling and detrending, each individual gene has been
normalized to the expression level of the baseline sample
“1 h serum/CO 2 control”. Although not being continuous
data, values were depicted as line graphs instead of bar
charts for better facility of inspection.
Choline transporter cudT, citrate transporter
SAOUHSC_02943, xanthine permease pbuX, fructose
specific permease fruA, and L-lactate permease lldP2.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2
control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h
internalization; 2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h
co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h
anaerobic incubation)
A
B
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
10.00
1.00
0.10
0.01
exp.
cudT
SAOUHSC_
02943 pbuX fruA lldP2
exponential growth phase 1.5 -1.2 -1.1 -4.3 -8.6
2.5 h internalized -2.5 -3.7 -3.5 -3.1 -14.4
6.5 h internalized -2.1 -1.4 -1.6 -2.0 -13.2
2.5 h serum/CO 2 control -1.6 1.4 1.5 -9.0 1.4
6.5 h serum/CO 2 control -1.6 -1.3 -1.2 -6.8 -2.5
2.5 h anaerobic incubation -2.9 -1.5 -1.1 -4.7 2.4
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
cudT
SAOUHSC_02943
pbuX
fruA
lldP2
Furthermore, expression of spermidine/putrescine ABC transporter potABCD,
sodium/glutamate symporter gltS, arginine/ornithine antiporter arcD, glycine betaine transporter
opuD1, and amino acid ABC transporter opuCdCcCbCa genes was repressed in internalized
staphylococci (Fig. R.5.29). Additional examples of repressed transporter genes were the choline
transporter cudT, the citrate transporter SAOUHSC_02943, the xanthine permease pbuX, the
fructose specific permease fruA, and the L-lactate permease lldP2 (Fig. R.5.30).
A
Fig. R.5.31: Phosphate ABC transporter gene
expression.
A. Overview on fold change values relative to the
baseline sample “1 h serum/CO 2 control” and on
significance in statistical group comparisons.
Differentially expressed genes are marked by gray
filling for the corresponding sample condition/time
point.
Genes which were significant in statistical testing
(p* < 0.05) but did not pass the absolute fold change
cutoff 2 are indicated in bold without filling. The
sample “6.5 h serum/CO 2 control” could not be
included in statistical testing because of small group
size (n = 2).
B. Overview on mean normalized gene expression
intensity. After the global normalization by inter-chip
scaling and detrending, each individual gene has been
normalized to the expression level of the baseline
sample “1 h serum/CO 2 control”. Although not being
continuous data, values were depicted as line graphs
instead of bar charts for better facility of inspection.
(exp. − exponential growth phase; 1 h co. – 1 h
serum/CO 2 control; 2.5 h int. − 2.5 h internalization;
6.5 h int. − 6.5 h internalization; 2.5 h co. − 2.5 h
serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2 control;
2.5 h anae. − 2.5 h anaerobic incubation)
B
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
100.0
10.0
1.0
0.1
exp.
phoU pstB pstA pstC SAOUHSC_
01388
pstS
exponential growth phase 1.3 1.1 -1.5 -1.5 -1.9 -3.2
2.5 h internalized 5.6 10.1 16.8 43.7 47.7 38.8
6.5 h internalized 9.8 16.1 27.4 59.6 61.3 34.0
2.5 h serum/CO 2 control -1.3 -1.4 -1.5 -1.1 -1.0 -2.5
6.5 h serum/CO 2 control -1.7 -1.4 -1.7 -1.3 -1.7 -2.2
2.5 h anaerobic incubation -1.2 -1.6 -1.9 -1.5 -1.8 1.2
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
phoU
pstB
pstA
pstC
SAOUHSC _01388
pstS
163

mean normalized intensity
mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
The second group comprised transporter genes with induced gene expression in at least one
analyzed time point of internalized staphylococci. This applied for example to the phosphate ABC
transporter operon SAOUHSC_01384 to SAOUHSC_01389, pho/pst (Fig. R.5.31).
A
fold change in comparison to
baseline 1 h serum/CO 2 control
phnE
SAOUHSC_
00103 phnC SAOUHSC_
00105 metN1 SAOUHSC_
00424
SAOUHSC_
00426 metN2
S. aureus
RN1HG
sample
conditions
and time
points
exponential growth phase -1.4 -1.5 -1.6 -1.2 3.3 2.8 2.0 1.6
2.5 h internalized -1.1 1.2 2.0 4.8 6.6 5.8 3.6 1.6
6.5 h internalized 2.9 3.6 5.2 7.4 16.6 12.8 7.6 2.4
2.5 h serum/CO 2 control -1.0 1.1 1.3 1.5 48.7 39.8 22.7 3.1
6.5 h serum/CO 2 control -1.4 -1.4 -1.1 -1.4 21.6 20.1 10.5 2.4
2.5 h anaerobic incubation -1.5 -1.1 1.1 1.7 3.9 2.8 1.8 1.4
Fig. R.5.32: Phosphonate transporter and methionine
binding protein gene expression.
A. Overview on fold change values relative to the baseline
sample “1 h serum/CO 2 control” and on significance in
statistical group comparisons. The sample “6.5 h serum/CO 2
control” could not be included in statistical testing because
of small group size (n = 2).
B. Overview on mean normalized gene expression intensity.
After the global normalization by inter-chip scaling and
detrending, each individual gene has been normalized to
the expression level of the baseline sample “1 h serum/CO 2
control” (1 h co.). Although not being continuous data,
values were depicted as line graphs instead of bar charts for
better facility of inspection.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2
control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h
internalization; 2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h
co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic
incubation)
B
100.0
10.0
1.0
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
phnE
SAOUHSC_00103
phnC
SAOUHSC_00105
metN1
SAOUHSC_00424
SAOUHSC_00426
metN2
Induction was detected for phosphonate ABC transporters (operon SAOUHSC_00102 to
SAOUHSC_00104, including phnE and phnC, and the adjacent SAOUHSC_00105), methionine
import ATP-binding protein gene metN1 and transporter/permease SAOUHSC_00424 and
SAOUHSC_00426 from the same operon, and the separately encoded methionine import ATPbinding
protein gene metN2 (Fig. R.5.32).
A
fold change in comparison to
baseline 1 h serum/CO 2 control
oppF
SAOUHSC_
00169
SAOUHSC_
00923
oppC oppD SAOUHSC_
00926
oppA
SAOUHSC_
00928
appD appF SAOUHSC_
00931
appC
S. aureus
RN1HG
sample
conditions
and time
points
exponential growth phase 1.3 1.2 1.4 1.4 1.4 1.3 1.3 1.3 1.3 -1.1 1.5 1.5
2.5 h internalized 1.3 5.2 2.7 2.7 3.0 2.4 2.4 1.2 -1.0 -1.5 -1.2 -1.2
6.5 h internalized 2.2 17.8 4.8 4.9 5.0 4.3 4.4 2.8 1.9 1.2 1.5 1.0
2.5 h serum/CO 2 control 4.9 27.6 8.3 8.2 9.0 7.5 7.6 8.2 3.1 1.7 1.6 1.3
6.5 h serum/CO 2 control 3.4 20.6 4.2 5.2 5.7 5.5 5.6 10.2 4.1 2.4 1.8 1.3
2.5 h anaerobic incubation 1.2 1.8 5.4 5.1 6.0 4.6 4.8 3.4 1.5 -1.0 -1.0 -1.0
Fig. R.5.33: Peptide and oligopeptide transporter gene
expression.
A. Overview on fold change values relative to the baseline
sample “1 h serum/CO 2 control” and on significance in
statistical group comparisons. The sample “6.5 h serum/CO 2
control” could not be included in statistical testing because
of small group size (n = 2).
B. Overview on mean normalized gene expression intensity.
After the global normalization by inter-chip scaling and
detrending, each individual gene has been normalized to
the expression level of the baseline sample “1 h serum/CO 2
control” (1 h co.). Although not being continuous data,
values were depicted as line graphs instead of bar charts for
better facility of inspection.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2
control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h
internalization; 2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h
co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic
incubation)
B
100.0 oppF
10.0
1.0
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
SAOUHSC_00169
SAOUHSC_00923
oppC
oppD
SAOUHSC_00926
oppA
SAOUHSC_00928
appD
appF
SAOUHSC_00931
appC
164

mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
Additionally, a high number of peptide/oligopeptide transporters were induced: peptide
transporters oppF and SAOUHSC_00169, oligopeptide transporters SAOUHSC_00926 and oppA,
and genes SAOUHSC_00923, oppC (SAOUHSC_00924), and oppD (SAOUHSC_00925) from the
same operon (SAOUHSC_00923 to SAOUHSC_00927). Also induced SAOUHSC_00928 gene codes
for an oligopeptide ABC transporter. In the same chromosomal region, a four-gene-operon
(SAOUHSC_00929 to SAOUHSC_00932) encodes oligopeptide ABC transporters appD, appF,
SAOUHSC_00931, and appC, but these were not induced in internalized staphylococci
(Fig. R.5.33).
Further example for induced transporter gene expression was another operon including msmX
(SAOUHSC_00175, multiple sugar-binding transport ATP-binding protein), SAOUHSC_00176
(bacterial extracellular solute-binding protein), SAOUHSC_00177 and SAOUHSC_00178 (both
maltose ABC transporter, permease proteins). Also opuD2 (osmoprotectant transporter), glnQ
(amino acid ABC transporter), SAOUHSC_02729 (amino acid ABC transporter-like protein), and
gntP (gluconate permease) were induced (Fig. R.5.34).
A
fold change in comparison to
baseline 1 h serum/CO 2 control
S. aureus
RN1HG
sample
conditions
and time
points
msmX SAOUHSC_
00176
SAOUHSC_
00177
SAOUHSC_
00178
opuD2 glnQ
SAOUHSC_
02729
exponential growth phase -1.3 -1.3 -1.2 -1.0 2.8 1.5 1.0 -1.0
2.5 h internalized 15.6 7.6 3.5 2.7 3.7 2.1 8.4 2.6
6.5 h internalized 8.2 5.6 2.8 1.8 2.9 -1.0 2.9 2.1
2.5 h serum/CO 2 control 1.4 -1.1 1.1 1.0 2.9 -1.2 -1.2 -1.2
6.5 h serum/CO 2 control 43.7 42.4 34.8 29.5 4.2 1.1 7.2 7.0
2.5 h anaerobic incubation 3.5 2.3 1.7 1.7 1.1 -1.1 1.2 -1.0
gntP
Fig. R.5.34: Sugar/maltose, osmoprotectant, amino acid,
and gluconate transporter or permease gene expression.
A. Overview on fold change values relative to the baseline
sample “1 h serum/CO 2 control” and on significance in
statistical group comparisons. The sample “6.5 h serum/CO 2
control” could not be included in statistical testing because
of small group size (n = 2).
B. Overview on mean normalized gene expression intensity.
After the global normalization by inter-chip scaling and
detrending, each individual gene has been normalized to
the expression level of the baseline sample “1 h serum/CO 2
control” (1 h co.). Although not being continuous data,
values were depicted as line graphs instead of bar charts for
better facility of inspection.
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2
control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h
internalization; 2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h
co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic
incubation)
B
100.0
10.0
1.0
0.1
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
msmX
SAOUHSC_00176
SAOUHSC_00177
SAOUHSC_00178
opuD2
glnQ
SAOUHSC_02729
gntP
Gene expression changes in internalized staphylococci not occurring in the same way in any of
the control samples
Analysis of virulence associated or metabolic gene expression changes revealed a common
aspect of the internalization signature: For many genes, some of the control conditions or time
points exhibited a similar expression change as the internalized expression pattern. To get hints
on real specifically internalization dependent gene expression changes, a set of sequences, which
contained sequences with different expression in internalized staphylococci in comparison to all
other control samples, was defined in two steps. First, subsets of regulated sequences in
internalized staphylococci were constituted which were a) not regulated in 2.5 h serum/CO 2
control and b) not regulated in 2.5 h anaerobically incubated samples (each in reference to the
baseline sample of 1 h serum/CO 2 control with p* < 0.05 and a minimal absolute fold change of
2). In these sets, sequences were still included which possessed differing expression in the 6.5 h
serum/CO 2 control in comparison to the baseline samples.
165

mean normalized intensity
mean normalized intensity
mean normalized intensity
mean normalized intensity
mean normalized intensity
mean normalized intensity
Maren Depke
Results
Pathogen Gene Expression Profiling
A
B
45 37 34
98 75 43
C
internalization-specific
differential expression;
induction 2.5 h after
start of infection
internalization-specific
differential expression;
induction 6.5 h after
start of infection
F
internalization-specific
differential expression;
repression 2.5 h after
start of infection
internalization-specific
differential expression;
repression 6.5 h after
start of infection
10E+02
10E+02
10E+01 10.00
10E+01 10.00
10E+00 1.00
10E+00 1.00
10E-01 0.10
10E-01 0.10
D
10E-02 0.01
1 2 3 4 5 6 7
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
G
10E-02 0.01
1 2 3 4 5 6 7
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
10E+02
10E+02
10E+01 10.00
10E+01 10.00
10E+00 1.00
10E+00 1.00
10E-01 0.10
10E-01 0.10
E
10E-02 0.01
10E-02 0.01
10E+02
1 2 3 4 5 6 7
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
H
10E+02
1 2 3 4 5 6 7
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
10E+01 10.00
10E+01 10.00
10E+00 1.00
10E+00 1.00
10E-01 0.10
10E-01 0.10
10E-02 0.01
10E-02 0.01
1 2 3 4 5 6 7
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
1 2 3 4 5 6 7
exp.
1 h
co.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
2.5 h
anae.
S. aureus RN1HG sample conditions and time points
Fig. R.5.35: Internalization-specific differential gene expression.
A, B. Overview on the number of sequences which were differentially expressed in internalized staphylococci in comparison to all
other control samples and exhibited induction (A) or repression (B).
C-H. Overview on mean normalized gene expression intensity. After the global normalization by inter-chip scaling and detrending,
each individual gene has been normalized to the expression level of the baseline sample “1 h serum/CO 2 control” (1 h co.). Although
not being continuous data, values were depicted as line graphs instead of bar charts for better facility of inspection.
Intensity values are depicted for 45 sequences specifically induced after 2.5 h (C), 37 sequences specifically induced at both time
points (D), 34 sequences specifically induced after 6.5 h (E), 98 sequences specifically repressed after 2.5 h (F), 75 sequences
specifically repressed at both time points (G), and 43 sequences specifically repressed after 6.5 h (H).
(exp. − exponential growth phase; 1 h co. – 1 h serum/CO 2 control; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h internalization;
2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2 control; 2.5 h anae. − 2.5 h anaerobic incubation)
166

Maren Depke
Results
Pathogen Gene Expression Profiling
This difference could not be assessed with statistical testing because only two arrays were
available for the 6.5 h control. Thus, in a second step, all sequences which possessed an absolute
fold change equal or greater than 2 in the comparison between 6.5 h and 1 h serum/CO 2 control
were excluded from the subsets generated in the first step. This procedure was performed for
induced and repressed genes separately and resulted in 45 sequences specifically induced after
2.5 h, 37 sequences specifically induced at both time points, 34 sequences specifically induced
after 6.5 h, 98 sequences specifically repressed after 2.5 h, 75 sequences specifically repressed at
both time points, and 43 sequences specifically repressed after 6.5 h in samples of internalized
staphylococci (Fig. R.5.35). Most interestingly, these lists included besides newly identified
transcripts also genes which were described above in the context of virulence or metabolic
genes, e. g. clfB, coa, phoP, prsA, glpK (2.5 h, induced), fnbB, vwb, ssaA, phoU, pstA, pstB, pstC,
pstS (2.5 h and 6.5 h, induced), chp, lukD, lukE, hla, efb, emp (6.5 h, induced), icaD, icaC, rot, nuc,
sspA, urea transporter SAOUHSC_02557, ureF (2.5 h, repressed), sarU (6.5 h, repressed).
New transcripts identified with the tiling array approach which are regulated in S9 cell
internalized staphylococci
All sequences of the tiling arrays were included in the statistical testing to identify the
internalization specific gene expression signature. Therefore, the resulting lists of differentially
expressed genes contained newly identified transcripts: 200 sequences for the 2.5 h and 138
sequences for the 6.5 h time point which were significant in ANOVA with p* < 0.05 and exhibited
a minimal absolute fold change of 2 (Table R.5.5).
When comparing these lists, 100 transcripts were differentially expressed at both time points
(Fig. R.5.36 A). These consisted of 67 induced and 33 repressed sequences (Fig. R.5.36 B).
A
B
100 100 38
2.5 h
repressed
repressed 6.5 h
induced
35 18
induced
new transcripts differentially
expressed in the comparison
“2.5 h internalization” vs.
“1 h serum/CO 2 control”
new transcripts differentially
expressed in the comparison
“6.5 h internalization” vs.
“1 h serum/CO 2 control”
65
33
67
20
Fig. R.5.36: Comparison of newly identified transcripts in the 2.5 h and 6.5 h signatures of internalized staphylococci.
Both internalized samples were compared to the baseline of 1 h serum/CO 2 control with statistical testing and multiple testing
correction (p* < 0.05), and a minimal absolute fold change cutoff of 2 was applied. The comparison of differentially expressed newly
identified transcripts at the 2.5 h and 6.5 h time point was performed for all regulated transcripts (A) and for induced and repressed
transcripts separately (B).
Although these 200 and 138 newly identified transcripts were known to be differentially
expressed in at least one of the two internalized samples in comparison to the 1 h serum/CO 2
control, they were expected to form further subgroups e. g. dependent on the direction of
regulation. Therefore, a k-means clustering was applied to a ratio data set (experimental
condition normalized to the 1 h serum/CO 2 control) consisting of the five experimental conditions
of non-adherent (1 h), internalized (2.5 h and 6.5 h), and serum/CO 2 control (2.5 h and 6.5 h)
staphylococci.
167

log 2 (ratio)
log 2 (ratio)
log 2 (ratio)
log 2 (ratio)
log 2 (ratio)
log 2 (ratio)
log 2 (ratio)
log 2 (ratio)
log 2 (ratio)
log 2 (ratio)
log 2 (ratio)
log 2 (ratio)
Maren Depke
Results
Pathogen Gene Expression Profiling
For the new transcripts included in the 2.5 h internalization signature, 6 cluster groups were
defined (Fig. R.5.37 A). Cluster group 1 contained 26 transcripts. Expression was increased in the
2.5 h samples and still slightly stronger in the 6.5 h samples. The serum/CO 2 controls at both time
points also exhibited increase of expression. Cluster group 2 harbored 32 transcripts which
featured increase in expression at the 2.5 h and to a lesser extent at the 6.5 h internalization time
points and in the later 6.5 h serum/CO 2 control whereas the 2.5 h serum/CO 2 control exhibited in
average no change. Transcripts with decreased expression in the 2.5 h internalized and less
strongly in the 6.5 h sample belonged to cluster group 3 which included 42 transcripts. The 2.5 h
serum/CO 2 control exhibited no apparent trend of regulation whereas the 6.5 h control exhibited
a decrease in expression. In cluster group 4, 26 transcripts were classified, which exhibited a
decrease of expression in both internalization samples, a slight increase in the 2.5 h serum/CO 2
controls and a decline of expression to the reference level in the 6.5 h serum/CO 2 controls.
Internalization samples in cluster group 5 showed an increase of expression 2.5 h after start of
infection, which again declined in samples after 6.5 h. Contrarily, expression in the serum/CO 2
controls was induced at the 2.5 h time point and further increased at the 6.5 h time point. This
cluster contained 49 sequences. Finally, cluster group 6 contained 25 sequences which were
increased in both internalized samples, and possessed a variable expression pattern in the 2.5 h
serum/CO 2 control with increased, repressed and unchanged values, while the 6.5 h serum/CO 2
control samples exhibited a trend of decreased expression.
A
B
cluster group 1
cluster group 2
cluster group 1
cluster group 2
6
4
2
0
-2
-4
-6
1 h
n. ad.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
6
4
2
0
-2
-4
-6
1 h
n. ad.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
6
4
2
0
-2
-4
-6
1 h
n. ad.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
6
4
2
0
-2
-4
-6
1 h
n. ad.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
cluster group 3
cluster group 4
cluster group 3
cluster group 4
6
4
2
0
-2
-4
-6
1 h
n. ad.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
6
4
2
0
-2
-4
-6
1 h
n. ad.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
6
4
2
0
-2
-4
-6
1 h
n. ad.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
6
4
2
0
-2
-4
-6
1 h
n. ad.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
cluster group 5
cluster group 6
cluster group 5
cluster group 6
6
4
2
0
-2
-4
-6
1 h
n. ad.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
6
4
2
0
-2
-4
-6
1 h
n. ad.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
6
4
2
0
-2
-4
-6
1 h
n. ad.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
6
4
2
0
-2
-4
-6
1 h
n. ad.
2.5 h
int.
6.5 h
int.
2.5 h
co.
6.5 h
co.
Fig. R.5.37: K-means clustering of gene expression data from newly identified transcripts, which exhibited differential expression in
the 2.5 h internalized sample (A) or in the 6.5 h internalized sample (B) in comparison to the 1 h serum/CO 2 control with statistical
testing and multiple testing correction (p* < 0.05) and a minimal absolute fold change cutoff of 2. For k-means clustering, the
following parameters were applied: Transcripts should be assigned to the user-defined number of 6 groups and sorted by deviation
using the metric type cosine correlation. No additional data preparation or statistical cuts were applied.
The ratio data set (experimental condition normalized to the 1 h serum/CO 2 control) employed for clustering consisted of the five
experimental conditions of non-adherent (1 h), internalized (2.5 h and 6.5 h), and serum/CO 2 control (2.5 h and 6.5 h) staphylococci.
Although not being continuous data, values were depicted as line graphs instead of bar charts for better facility of inspection.
(1 h n. ad. – non-adherent staphylococci after 1 h of infection; 2.5 h int. − 2.5 h internalization; 6.5 h int. − 6.5 h internalization;
2.5 h co. − 2.5 h serum/CO 2 control; 6.5 h co. − 6.5 h serum/CO 2 control)
Correspondingly, also for the new transcripts included in the 6.5 h internalization signature,
6 cluster groups were defined (Fig. R.5.37 B). Here, cluster group 1 contained 19 transcripts with
168

Maren Depke
Results
Pathogen Gene Expression Profiling
an increase in expression at both internalization time points and in the later 6.5 h serum/CO 2
control whereas the 2.5 h serum/CO 2 control exhibited in average no change. In cluster group 2,
to which 17 sequences were assigned, transcripts exhibited a decrease of expression in both
internalization samples, and in general no change in the 2.5 h serum/CO 2 controls. Some of the
sequences were slightly decreased and some other slightly increased in the later 6.5 h serum/CO 2
controls. Cluster group 3 included 36 transcripts with increased expression at both internalization
and serum/CO 2 control time points. Transcripts with mainly decreased expression in both
internalized samples belonged to cluster group 4 which included 34 transcripts. The 2.5 h
serum/CO 2 control exhibited a varying trend of regulation with some repressed and some
induced sequences, whereas the 6.5 h control exhibited a decrease in expression. The 16
transcripts in cluster group 5 possessed an increase in the 2.5 h and 6.5 h internalization samples.
Also in the serum/CO 2 controls at both time points a slight tendency of induction of expression
was observed, but less prominent than the internalization samples. The last cluster group 6
comprised the remaining 16 transcripts with similar expression pattern as the transcripts in
cluster group 5 with exception of the expression change in the serum/CO 2 controls, where only
minor changes in the 2.5 h and slight decrease in expression in the 6.5 h samples were visible.
Transcripts in all cluster groups of both 2.5 h and 6.5 h internalization signatures showed only
minor variation from the non-changed value of 1 in the sample of non-adherent staphylococci.
This finding was expected as the similarity between this sample and the 1 h serum/CO 2 control
baseline had been unveiled before.
169

Maren Depke
D I S C U S S I O N A N D C O N C L U S I O N S
LIVER GENE EXPRESSION PATTERN IN A MOUSE
PSYCHOLOGICAL STRESS MODEL
BALB/c mice were subjected to combined acoustic and restraint stress for 2 h twice a day
during a period of 4.5 days, which serves as a murine model of chronic psychological stress.
Experiments using the same model had already revealed that mice suffered from an impaired
antibacterial defense and from depression-like behavior (Kiank et al. 2006, 2007a).
The observation of severe loss of total body mass initiated further analysis of metabolic
processes as well as a hepatic gene expression profiling study, which resulted in evidence for the
development of a hypermetabolic syndrome in chronically stressed mice. The results of liver gene
expression profiling and physiological analyses have been published by Depke et al. in 2008 and
2009.
Already a single acute stress exposure caused profound changes in hepatic gene expression.
Genes important for metabolic pathways regulating the carbohydrate turnover showed stressinduced
alterations of mRNA expression in hepatic tissue. An acute stress response is essential
for energy mobilization to “fight or flight” in a potentially harmful situation and to sustain or
reconstitute allostasis (McEwen 2004). Initially, catecholamines activate glycogenolysis,
gluconeogenesis, and accelerate lipolysis that subsequently is assisted by catabolic
glucocorticoid-induced pathways (Bagby et al. 1992, Leibowitz/Wortley 2004, Lundberg 2005,
McGuinness et al. 1999). Acute psychological stress was associated with activation of the stress
axes, and an induction of the expression of gluconeogenic genes and of transporters for
glucogenic amino acids (Pck1, G6pc, Slc37a4, Slc15a4, Slc25a15, Slc38a2, Slc3a1, Sds, Slc6a6, and
Tat).
Despite the observed gene expression changes the metabolic parameters did not change
significantly when stress was restricted to a singular event. Contrarily, when stress was
repeatedly applied, female BALB/c mice developed severe systemic neuroendocrine and
metabolic alterations.
Several researchers found that repeated restraint stress causes a loss of body weight in
rodents, which was mainly mediated by initially increased energy expenditure and reduced food
intake that normalized or even heightened within a few days after starting repeated stress due to
neuroendocrine adaptations (Harris et al. 2002, 2006). Others showed prolonged lowered food
intake during 4.5 days repeated restraint stress due to prolonged neuroendocrine dysregulation
(Ricart-Jané et al. 2002). In the chronic stress model used in this study no changes of total food
171

Maren Depke
Discussion and Conclusions
consumption during 4.5 days in repeatedly stressed animals compared with nonstressed controls
were observed. Unchanged consumption may result from an initially reduced food intake after
the first stress session and increased food intake in the later phase of repeated stress exposure in
which additionally neuroendocrine and metabolic dysregulation were manifested.
Repeatedly stressed mice suffered from an increase in metabolic rate in excess of the normal
metabolic response. Such a hypermetabolic response leads to a marked increase in energy
demands. Protein inappropriately becomes an energy source, and increased use of protein
rapidly depletes lean body mass. A hypercatabolic response is a typical feature in infection,
cancer, and prolonged critical illness, and goes along with fever, dysregulations of the
cardiovascular system, hyperglycemia, dyslipidemia, accelerated proteolysis, tissue damage/cell
death, perfusion disturbances, and invasion by microorganisms (Alberda et al. 2006, Costelli et al.
1993, Mizock 1995, Morley et al. 2006, Vanhorebeek/Van den Berghe 2004, van Waardenburg et
al. 2006, Wilmore 2000, Wray et al. 2006). In contrast, starvation is connected with diminished
food intake, hyperthyroidism, and reduced protein catabolism (Alberda et al. 2006, Morley et al.
2006). During a hypercatabolic response, as it was detected in the model of repeated stress, food
intake is often normal (Alberda et al. 2006, Morley et al. 2006) and associated with
hypothyroidism (Vanhorebeek/Van den Berghe 2004). Recently, it was shown that prolonged
sleep deprivation also can cause such a hypermetabolic response in rats (Everson/Reed 1995,
Koban/Swinson 2005). In this study, the stress experiments were performed in the recovery
phase of the animals, and sleep deprivation may have affected metabolic functions. Both
repeatedly stressed mice and long-term sleep deprived rats showed lowered total T 3 and T 4 levels
that did not depend on altered TSH concentrations (Koban/Swinson 2005). Koban and Swinson
propose a reciprocal relationship of catecholamines, which progressively increase during sleep
deprivation, and thyroid hormone concentrations that decline continuously in prolonged
reduction of sleep time. However, in contrast to the repeatedly stressed mice that showed
unaltered food intake in this study, sleep-deprived rats were hyperphagic while body mass was
massively consumed (Everson/Reed 1995, Koban/Swinson 2005). Sleep deprived rats did not
exhibit altered glucocorticoid levels, whereas chronic psychological stress characteristically was
associated with persistently high corticosterone concentrations in the plasma.
The activation of the central nervous system can profoundly affect metabolic regulation such
as shown for the thyroid hormone release (Koban/Swinson 2005). The target organ of metabolic
regulatory pathways is predominantly the liver (Dallman et al. 2007, Leibowitz/Wortley 2004,
McEwen 2004). The activation of the HPA axis with increased glucocorticoid levels can stimulate
food intake and activate carbohydrate, fat and protein catabolism. In turn, the brain receives
signals such as actual glucose and lipid concentrations or increased energy demand (Lam et al.
2007, Lundberg 2005, McEwen 2004). In consequence, central nervous system activation induces
regulatory pathways that equilibrate metabolism to supply the needed energy, e. g. increasing
glucose formation in the liver (Lam et al. 2007, Leibowitz/Wortley 2004, McEwen 2004).
Hypercortisolism shifted metabolic functions toward carbohydrate, lipid, and protein
catabolism (Aikawa et al. 1972, Dallman et al. 2007, Harris et al. 2002, Lam et al. 2007, Ricart-
Jané et al. 2002, Souba et al. 1985) to sustain energy supply by replenishing glucose that is the
main energy source of the body. Glucose can be released after glycogenolysis or induction of
gluconeogenesis when the supply with food is insufficient. Primarily, alanine and glutamate are
precursor molecules for gluconeogenesis (Aikawa et al. 1972, Brosnan 2000, Hagopian et al.
2003, Pasini et al. 2004, Souba et al. 1985). Hepatic induction of alanine aminotransferase (Gpt) 2
172

Maren Depke
Discussion and Conclusions
and Got1 may supply the metabolites for glucose synthesis (Aikawa et al. 1972, Pasini et al. 2004,
Souba et al. 1985). Thereafter, alanine and glutamate can be reconstituted by biotransformation,
whereas essential amino acids cannot be replenished by biosynthesis in repeatedly stressed mice
(Brosnan 2000, Hagopian et al. 2003). Deamination of amino acids results in the production of
ammonia, which is detoxified in the liver by the urea cycle. Increased expression of the urea cycle
enzyme Asl in the liver of repeatedly stressed mice along with usage of arginine and citrulline as
intermediate products of the urea cycle indicates heightened stress-induced ammonia
detoxification to provide C-bodies of amino acids for metabolic pathways such as
gluconeogenesis. Lactate, which alternatively can serve as substrate for hepatic gluconeogenesis,
was produced in high amounts in peripheral tissues during hypermetabolism such as during
sepsis and can cause lactic acidosis (Aikawa et al. 1972, Banerjee et al. 2004, Capes et al. 2000,
Mizock 1995). Acidosis which was detectable in repeatedly stressed mice is often paralleled with
insulin resistance and hyperglycemia (Capes et al. 2000, Vanhorebeek/Van den Berghe 2004,
van Waardenburg et al. 2006). In fact, in repeatedly stressed mice concentrations of the
adipokine resistin increased. Resistin is is inducible by glucocorticoids, prolactin, and growth
hormone, and has been identified to lower insulin sensitivity (Hagopian et al. 2003). Prolonged
hyperglycemia, especially in critically ill patients, increases the risk of infectious complications,
neuronal damage, and multiorgan dysfunction syndrome (Banerjee et al. 2004, Harris et al. 2006,
van Waardenburg et al. 2006, Wray et al. 2002). In line with this, repeatedly stressed mice
suffered from a reduced antimicrobial response (Kiank et al. 2006, 2007a). Lam et al. showed in
2007 that increased glucose sensing by the brain and elevated intracerebral concentrations of
lactate reduced the hepatic secretion of VLDL cholesterol and caused a decline of plasma
triglyceride levels in rodents. In addition, Ricart-Jané et al. (2002) found that repeated restraint
stress caused a loss of plasmatic triacylglycerols with VLDL levels, which was accompanied by a
reduced food intake. Heightened triglyceride clearance and increased lipolysis to assemble C3
bodies for gluconeogenesis can result in essential fatty acid deficiency (Akhtar et al. 2005,
Nemoto/Sakurai 1995, Rizki et al. 2006).
Hypotriglyceridemia was detected in repeatedly stressed mice, which went along with
hypercholesteremia and up-regulation of glucocorticoid-sensitive cytochrome genes in the liver
(Akhtar et al. 2005, Li-Hawkins et al. 2000, Nemoto/Sakurai 1995). Cyp4a enzymes are involved in
the removal of fatty acids and can counter-regulate hepatic steatosis, which was a typical
consequence of repeated stress exposure in mice. Increased expression of Cyp17a1 gives hints
for hepatic induction of steroidogenesis (Akhtar et al. 2005) in repeatedly stressed mice, and upregulation
of expression of Cyp39a1 and Cyp2b10 indicates increased removal of steroids by bile
acid production (Li-Hawkins et al. 2000). Importantly, the cholesterol metabolism of rodents and
humans is difficult to compare. In mice, HDL is the main lipoprotein present in the blood and
essentially required for steroidogenesis (Martin et al. 1999, Ricart-Jané et al. 2002), whereas
humans use LDL for steroid synthesis and have lower HDL concentrations (Chen W et al. 2005,
Martin et al. 1999). Furthermore, HDL is able to scavenge endotoxins from the plasma (Barter et
al. 2004). The relevance of stress-induced elevation of plasma HDL levels in mice, e. g. for steroid
synthesis or scavenging the bacterial compound lipopolysaccharide, remains to be elucidated.
Other authors found that lipopolysaccharide or TNF challenges particularly increase the catabolic
rate (Ogimoto et al. 2006, Sugita et al. 2002). Such effects seem to be mediated via TNF-induced
neuroendocrine stimulation, e. g. activation of the sympathetic nervous system that primarily
causes glucose formation (Bagby et al. 1992, Finck et al. 1998, Finck/Johnson 2000,
173

Maren Depke
Discussion and Conclusions
Leibowitz/Wortley 2004, Lundberg 2005, McGuinness et al. 1999, Ogimoto et al. 2006, Sugita et
al. 2002). In addition, release of catecholamines potentially induces the expression of the leptin
gene in adipose and hepatic tissue (Arnalich et al. 1999, Finck/Johnson 2000, Mak et al. 2006).
Leptin then can signal afferently to the brain to sustain inhibitory effects on food intake. In this
relation, one should expect reduced food consumption in the hyperleptinemic repeatedly
stressed mice. However, food intake was not altered, whereas total body mass furthermore was
lost. One possible explanation for the drastic loss of body weight along with hyperleptinemia is
that leptin potentially can increase the energy expenditure such as enhancing the activation of
the respiratory chain and, therefore, can increase energy consumption (Fleury et al. 1997,
Ricquier/Bouillaud 2000). In this study, already acute stress drastically induced changes of
hepatic gene expression that did not significantly disrupt allostatic regulation of metabolism in
mice. In turn, repeated psychological stress in BALB/c mice along with systemic
immunodeficiency that was reported previously (Kiank et al. 2006, 2007b) induced a
hypermetabolic stress syndrome.
It is not clear why some individuals lose weight during prolonged stressful situations, whereas
others gain body mass (Alberda et al. 2006, Harris et al. 1998, Morley et al. 2006,
Vanhorebeek/Van den Berghe 2004, Wilmore 2000, Wray et al. 2002). Because there is an
increased number of patients suffering from metabolic syndrome and clinically relevant
associated illness, many publications show that chronic stress is promoting the development of a
metabolic syndrome that is associated with gain of fat mass (obesity), type 2 diabetes,
hyperlipidemia, and hypertension (Alberda et al. 2006, Harris et al. 1998, Lundberg 2005). In
contrast, the animal experiments with BALB/c mice as a mouse strain with high stress
susceptibility (Kiank et al. 2006) led to the detection of metabolically driven wasting because of a
hypercatabolic stress response. It can be assumed that the genetic predisposition influences the
development of either stress-induced metabolic syndrome or loss of body weight phenotype.
Moreover, it is shown that besides genetic predisposition, environmental factors influence
prenatal and postnatal neuronal and neuroendocrine differentiation, resulting in different coping
styles in the adult (Koolhaas et al. 2007). They showed that proactive/aggressive animals
developed stress-induced hypertension, cardiac arrhythmia, and inflammation, whereas
reactive/passive individuals are more susceptible to anxiety disorders, metabolic syndrome,
depression, and infection. However, the neurobiology and endocrine regulation of these different
coping styles are not well understood yet. A loss of biological reserves as in the model of
repeated stress exposure is as fatal as the development of a metabolic syndrome because of
losing the ability to fight infection and cancer (Alberda et al. 2006, Hang et al. 2003, Hansen MB
et al. 1998, Morley et al. 2006, Souba et al. 1985).
In the clinical setting, it is now clear that the catabolic response will become autodestructive if
not contained. The severity of complications will occur in proportion to lost body protein. In the
model of repeatedly stressed mice, particularly arginine deficiency became evident. Interestingly,
alimentation with arginine and omega-3 fatty acids-enriched enteral feeds decreased hospital
days and infectious complications in critically ill patients (Beale et al. 1999), which commonly
show a loss of about 10 % of lean body mass (Arnalich et al. 1999, Fleury et al. 1997,
Ricquier/Bouillaud 2000). Healthy adults require about 0.8 g protein/kg body weight·d to
maintain homeostasis. Stressful events such as traumatic injury or infection increase the body’s
protein requirement up to 1.5–2 g protein/kg body weight·d or even more. However, humans
cannot metabolize more than 2 g/kg body weight·d. This often results in a fatal negative nitrogen
174

Maren Depke
Discussion and Conclusions
balance in severely ill patients (Demling 2007). Finally, amplified protein breakdown with a loss of
more than 40 % of lean body mass leads to irreversible cell damage (Beale et al. 1999, Demling
2007).
In conclusion, the physiological measurements and hepatic gene expression profiling
demonstrated that highly demanding psychological stress in the absence of injury or infection is
able to induce a severe hypermetabolic syndrome in mice. Such overwhelming wasting condition
can reduce the individual’s resistance to further stressful stimuli such as injury or infection.
The same data set of hepatic gene expression was used to gain deeper insight into hepatic
immune regulatory and cell cycle processes of BALB/c mice, which develop systemic
immunodeficiency with a reduced antibacterial defense after repeated stress exposure (Kiank et
al. 2006, 2007a,b). This second part of the results has been published by Depke et al. in 2009.
Already a single acute stress exposure drastically altered the expression of immune response
and of apoptosis related genes. An activation of immunosuppressive mechanisms in the liver was
measured in both acutely and chronically stressed mice including increased expression of
Tsc22d3 (GILZ, glucocorticoid-induced leucine zipper) and Fkbp5, which both are inducible by
high glucocorticoid levels (Ayroldi et al. 2007, Berrebi et al. 2003, D'Adamio et al. 1997,
Mittelstadt/Ashwell 2001). Tsc22d3 inhibits the expression of CD80 and CD86 co-stimulatory
molecules or toll-like receptor 2 and reduces the production of proinflammatory mediators
(Berrebi et al. 2003, Cohen et al. 2006). Fkbp5, an immunophilin family member, inhibits
calcineurin signaling in lymphocytes and modifies glucocorticoid signaling by binding heat shock
proteins of steroid receptors (Baughman et al. 1995, Sinars et al. 2003). Other B and T
lymphocyte signaling molecules such as protein kinase C ν (Prkcn), or several GTPases showed
altered hepatic mRNA expression in both acutely and chronically stressed mice, respectively.
Importantly, the down-regulation of several IFN-γ inducible genes after acute stress was
amplified by repeated stress exposure. Some of the interferon-inducible mRNA transcripts like
Igtp and Stat1, which were repressed in the liver of chronically stressed mice, encode for host
proteins of well characterized antimicrobial activity (Döffinger et al. 2002, Johnson/Scott 2007,
Martens et al. 2004). Diminished expression of IFN-γ inducible GTPases was shown to be related
to a reduced defense against Listeria monocytogenes and Mycobacterium avium infection
(Martens et al. 2004). Stat1 is important for IFN-γ, IL-6, IL-10, and IL-12 signaling and for
maintaining MHC- and co-stimulatory molecule expression in dendritic cells, which all are
important for the antibacterial response (Döffinger et al. 2002, Johnson/Scott 2007). In line with
this, chronically stressed animals showed a repression of the Stat1 gene and reduced expression
of MHC-molecules along with a heightened susceptibility to bacterial infections (Kiank et al. 2006,
2007b) and an increased spreading of translocated commensals into liver and lung, which went
along with a reduced ex vivo inducibility of IFN-γ (Kiank et al. 2008).
Besides the reduced expression of cellular signaling molecules, genes associated with immune
cell migration and tissue invasion were differentially expressed. These included T cell
chemoattractive peptides such as Cxcl11 and Cxcl12 and neutrophil chemoattractive Cxcl1
(Ghosh et al. 2006, Helbig et al. 2004, Wiekowski et al. 2001) and an increased expression of cell
adhesion molecules such as Vcam-1 in the liver homogenate after both acute and chronic stress.
Arhgap5 was selectively induced after repeated stress exposure (Cook-Mills 2002, Haddad/Harb
2005, Kim et al. 2006, Petri/Bixel 2006) while the expression of several claudins, which are tight
junction proteins that regulate paracellular permeability (Amasheh et al. 2002), was significantly
175

Maren Depke
Discussion and Conclusions
reduced. Thus, data from the murine psychological stress model indicate that leukocyte
migration associated molecules were increasingly transcribed already after acute stress.
However, leukocyte infiltration into hepatic tissue was not measurable at this particular moment
immediately after a single stress exposure but this may be due to the point in time of data
acquisition which was chosen. Other authors showed leukocyte recruitment into the liver 3−6 h
after hepatic ischemia/reperfusion (Zwacka et al. 1997) or 2−4 h after initiation of an acute phase
response (APR) following brain injury (Campbell et al. 2003). Therefore, one can propose that cell
recruitment into the liver may be detectable in the next few hours after termination of acute
stress, which remains to be elucidated in further experiments, and later manifests during
repeated stress exposure as shown by the data of chronically stressed mice.
An invasion of inflammatory cells such as neutrophils and macrophages into tissue is often
associated with increased release of proinflammatory cytokines, which in turn enhance
catabolism and stimulate the hypothalamus-pituitary-adrenal axis (Bartolomucci 2007, Herold et
al. 2006, Lundberg 2005). Although increased expression of cytokine genes in the liver was not
shown, plasma glucocorticoids levels were increased after both acute and chronic psychological
stress exposure causing up-regulation of glucocorticoid-inducible genes, such as acute phase
proteins (Kiank et al. 2006, Depke et al. 2008). In line with these data after chronic stress, Yoo
and Desiderio found increased expression of several APR markers, including C-reactive protein
(CRP), serum amyloid A proteins, lipocalins or orsomucoid at 3 h, 6 h and 12 h after low-dose
bolus application of LPS (Yoo/Desiderio 2003), which can serve as a model of minute bacterial
translocation. Acute phase proteins (APPs) have several immune and metabolic regulatory
functions such as enhancing the antimicrobial response: CRP can opsonize microorganisms and
activate the complement system (Casey et al. 2008, Mold/Du Clos 2006), apolipoproteins, or
lipocalin 2 regulate the cholesterol transport, endotoxin scavenging, and free radical production
(Levels et al. 2003, Roudkenar et al. 2007, Tseng et al. 2004, Wurfel et al. 1994). In contrast, other
APPs such as haptoglobin, which were found to be induced after repeated stress, have strong
anti-inflammatory effects (Tseng et al. 2004) and therewith may contribute to the reduced
antibacterial defense of chronically stressed mice, which even suffered from long-lasting bacterial
infiltration into liver and lung (Kiank et al. 2008). Along with the increased expression of APR
genes, high expression levels of cytochrome P450 enzymes were detected which normally are
suppressed during an APR (Siewert et al. 2000). This can be explained by findings of others who
showed that hepatic levels of P450 enzymes increased during fasting and then became resistant
to the suppression during inflammatory states (Iber et al. 2001, Morgan 2001). Therefore, the
increased cytochrome expression in the chronic stress model may be a part of the hypercatabolic
stress response that overcomes an inflammation-induced repression of P450 enzyme expression.
An activation of cytochromes enhances the generation of reactive oxygen species (ROS) (Morgan
2001, Zangar et al. 2004). Oxidative stress, in turn, can cause lipid peroxidation or increase
protein carbonyl content (Ermak/Davies 2001, Garg/Aggarwal 2002, Ott et al. 2007, Zangar et al.
2004), which was also observed in plasma and liver after acute stress exposure in this study.
Carbonylated proteins are likely non-functional and prone to degradation. Oxidant-induced
damage in hepatic tissue of acutely stressed mice is supported by the finding that several cell
cycle and cell death associated genes such as Cdkn1a, Gadd45b, Gadd45g or Fkbp5 were already
induced immediately after acute stress – a moment when detection of cellular apoptosis is
probably too early – and remained highly expressed after repeated stress when increased
hepatocyte apoptosis was measured.
176

Maren Depke
Discussion and Conclusions
However, tissue protective mechanisms seem to be active in the liver of chronically stressed
mice as well. These include an increased expression of Alas1, which functions as an oxygen
radical scavenger (Kaliman et al. 2005), and of glutaredoxin and glutathione S-transferase alpha
that are ubiquitous proteins with redox-active cysteine-groups (Garg/Aggarwal 2002,
Haddad/Harb 2005, Jurado et al. 2003). Increased expression of these anti-oxidant enzymes may
compensate and protect from further oxidative stress-induced damage. Therefore, long lasting
increased glucocorticoid levels, which are highly potent mediators of inducing apoptotic
processes, are proposed to be the main mediators of hepatic cell death in chronically stressed
animals (Ayroldi et al. 2007, Herold et al. 2006), but this remains to be elucidated.
The liver has an extraordinary ability to regenerate or heal itself by replacing or repairing
injured tissue (Fausto et al. 2006, Klein et al. 2005, Riehle et al. 2008). Hepatic gene expression
profiling after acute and chronic stress included increased hepatic expression of protein tyrosine
phosphatase 4a1, which initiates the liver regeneration response, and of IL-6 receptor whose
ligation was shown to be protective during liver injury (Fausto et al. 2006, Klein et al. 2005, Riehle
et al. 2008). Thus, acute stress-induced up-regulation of genes coding for tissue destructive
mechanisms associated with oxidative stress were counter-regulated by increased expression of
anti-oxidative and tissue regenerative genes which may have protected mice from more severe
damage of hepatic tissue during chronic psychological stress.
In summary, the data show that already acute stress exposure induces hepatic mRNA
expression of several LPS and glucocorticoid-sensitive immune response genes as well as of
apoptosis-related markers. Besides increased oxidative stress, this does not cause measurable
immune alterations or cell death immediately after stress exposure. However, when stress is
repeated and becomes chronic, reduced antigen presentation and altered gene expression of
cellular signaling molecules of immune cells along with an ongoing expression of apoptosisrelated
molecules may contribute to biologically relevant immunosuppression which was
demonstrated by a reduced ability of stressed mice to clear bacterial infections from the liver.
Moreover, the induction of cell protective and liver regenerative genes in hepatic tissue of
chronically stressed mice give strong evidence that counter-regulatory mechanisms are activated
to prevent further hepatocyte damage in these animals.
177

Maren Depke
Discussion and Conclusions
KIDNEY GENE EXPRESSION PATTERN IN AN
IN VIVO INFECTION MODEL
The SigB regulon contains several virulence associated genes, and expression of this regulon in
general promotes adhesion and reduces production of extracellular toxins (Bischoff et al. 2004).
While regulation of SigB activity has already been studied in vitro (Senn et al. 2005), its impact on
in vivo physiology, e. g. in infection models, is still matter of debate. The aim of this study was the
analysis of the relevance of SigB-dependent gene expression regulation for the adaptation of the
host transcriptome in murine infection.
In an i. v. murine infection model the two Staphylococcus aureus strains RN1HG and its
isogenic sigB mutant were used to infect mice with comparable infection doses. In mouse kidney,
the infection resulted in similar infection rates of approximately 1.0E+06 cfu/10 mg of tissue. The
role of SigB has already been analyzed in different in vivo infection models. A general observation
of these studies was that sigB deletion mutants and their parental strain can accomplish
comparable bacterial loads. In the clinical isolate WCUH29 the allelic replacement of sigB did not
lead to differing bacterial load in organs or wounds after testing original strain and isogenic
mutant in different types of infection models (Nicholas et al. 1999). Also Chan et al. (1998)
observed similar numbers of cfu in a skin abcess model, but the results might have been
compromised by the fact that the group used a sigB deletion in the 8325-4 background which
itself harbors an inactivating mutation in rsbU, a gene for a SigB regulatory phosphatase
(Kullik/Giachino 1997). Horsburgh et al. published in 2002 the construction of S. aureus SH1000, a
variant of strain 8325-4, in which the rsbU gene had been repaired. SH1000 and 8325-4 resulted
in the recoverage of similar numbers of cfu in a murine subcutaneous skin abcess model
(Horsburgh et al. 2002b). Contrarily, another study reported a significantly different bacterial
burden in kidney after i. v. infection with S. aureus 8325-4 (RsbU − ), SH1000 (RsbU + ), and a sigB
knockout mutant of SH1000 (SigB − ). In this setting, the repair of rsbU leads to an increased cfu
count 14 days after inoculation (Jonsson et al. 2004). In the study described in this thesis, almost
equal mean values of infection rate were observed for S. aureus RN1HG and its isogenic sigB
mutant, a further confirmation of literature data indicating similar virulence of sigB deficient and
positive S. aureus strains.
In order to gain a much more detailed view of the host reaction to both strains, kidney gene
expression profiles were compared using Affymetrix GeneChip arrays. In a total number of 19
infected samples, highly similar gene expression profiles were recorded 4 or 5 days after infection
with RN1HG or its sigB derivative. Even with this sophisticated analysis no difference in the
expression pattern in murine kidney after infection with the two strains was observed, indicating
that SigB indeed does not influence the host response.
However, the comparison to sham-infected samples revealed a rather strong reaction of the
murine host to infection, bacterial accumulation and proliferation in the kidney. Several immune
response pathways were induced on transcriptional level indicating a vivid activation of the
immune response. This applies first to pathways of the innate immune system: Pattern
recognition receptors (PRR) like toll-like receptors (TLR) and complement system components
178

Maren Depke
Discussion and Conclusions
were induced and strengthened the first line of defense against the pathogen. Second, several
immune response signaling pathways exhibited increased gene expression of their members, e. g.
proinflammatory pathways of IFN, IL-6, and TREM1 signaling, and several proinflammatory
cytokines were also induced, e. g. TNFα, IL-6, RANTES/Ccl5, MCP-1/Ccl2, MCP-3/Ccl7, and MIP-
1α/Ccl3. Members of IL-10 signaling, like the IL-10 receptor α-chain (Il10ra) and the signal
transducer STAT3, were induced, too. This pathway aims at limitation of the cellular
proinflammatory response. But as the ligand IL-10 was not induced and STAT3 is also part of the
IL-6 signaling, it is not clear whether the IL-10 signaling was active at the time point of analysis or
whether it was only prepared to receive signals in a later phase of infection. As third example, the
infected samples exhibited a distinct transcriptional induction of several members of the
presentation pathways for intracellular and extracellular antigens via MHC molecules. The
comparison of infection and sham infection additionally revealed metabolic disturbances shifting
to both anabolic and catabolic direction of metabolism. This might indicate a high extent of
infection and illness that impairs organ function, nutrition supply and possibly oxygen availability.
The comparison of S. aureus infected samples with non-infected controls proved the strong
reaction of the host to infection. But in the model described in this thesis the host reaction does
not differ depending on the strain used for infection.
Until now, the role of sigB in infections has not been completely clarified. There is evidence
for an influence of sigB on biofilm formation. Attachment and microcolony formation are the two
central steps in the early biofilm formation. In an in vitro biofilm formation model (using
polystyrene microtiter plates) it was shown that increased sigB expression leads to increased
attachment and to increased microcolony formation with increased size of staphylococcal
autoaggregates (Bateman et al. 2001).
In detail, biofilm formation is mediated by polysaccharide intercellular adhesin/PIA, an
extracellular adhesin of staphylococci composed of poly-N-acetylglucosamine/PNAG. Although
highly conserved in staphylococcal strains, the ica operon, coding for enzymes responsible for PIA
synthesis, is expressed in vitro only in a few strains. For the mucosal isolate MA12 it was proven
that a sigB insertion mutant (sigB::ermB) lost the biofilm formation ability after osmotic stress
(3 % NaCl) compared to the wild type strain. Additionally, it was demonstrated under osmotic
stress conditions that the mutant exhibited strongly reduced ica expression compared to the wild
type (Rachid et al. 2000).
Even more, a major part of ica expression and biofilm formation was proven to be mediated
by SarA. It is known that SigB induces sarA expression, but it has not been demonstrated yet
whether the influence on ica expression is an indirect effect of SigB or whether SarA acts
independently of SigB in this context. Interestingly, experiments predict the existence of a SigBactivated
factor, which either might degrade PIA or repress its synthesis. This factor in turn is
repressed by SarA (Valle et al. 2003).
Complementary information regarding the influence of sigB knock-out on biofilm formation
and virulence is available from device-associated staphylococcal infection models in mice. When
MA12 and its isogenic sigB mutant were applied to a central venous catheter (CVC) and later to
the blood stream, both strains were able to form multilayered biofilms inside the catheter.
Therefore, sigB is not essential for biofilm formation. On the other hand, there were structural
differences between the biofilms of both strains: The mutant’s biofilm was lacking certain
extracellular substance. SigB is thus proposed to affect the ability of spreading or release from
adhesive sites/biofilms. In the quantitation of total bacterial burden in the inner organs, a
179

Maren Depke
Discussion and Conclusions
significant decrease was apparent after infection with the sigB mutant compared to the parental
strain (Lorenz et al. 2008). Others have published that sigB mutations have no influence on cfu
counts. This emphasizes the important influence of other experimental parameters like infection
dose, infection route, site of infection or its manifestation, analysis time point after infection, the
host factors – at least in parts determined by the host genome – acting on the bacterium, and the
genetic background of the pathogen.
Repair of rsbU or sigB overexpression in the genetic background of BB255, a rsbU mutated
S. aureus strain, leads to an increase in transcription of clfA, mediated via a SigB-promoter in the
clfA gene, and an increase in transcription of fnbA which is indirectly caused by SigB and probably
mediated via a positive effect of SigB on sarA and subsequently of SarA on the fnbA transcription.
The sigB overexpression mutant features an increased attachment to fibrinogen and fibronectin
coated surfaces and to platelet-fibrin clots, which imitate the environment in injured
endothelium of blood vessels. When the same mutant is applied in an infection setup focusing on
adherence by inducing catheter-associated aortic vegetation in rat, a transient effect of higher
bacterial density in early stages of infection was observed. After long-term infection, the
advantage of the early phase was counterbalanced in comparison to sigB deficient strains. The
authors argue that S. aureus is still able to produce sufficient amounts of adhesins leading to
aortic vegetation even without a functional sigB (Entenza et al. 2005).
In a murine sepsis and arthritis model, the SigB positive strain SH1000 (rsbU repaired) caused
more severe infection than strain 8325-4 which is phenotypically SigB − (11 bp rsbU deletion) as
illustrated by higher mortality, more pronounced weight loss and higher serum levels of IL-6 as
indicator of inflammation. Additionally, at day 14 after i. v. infection higher arthritis frequency
and severity were observed after infection with SH1000. In an attempt to elucidate whether the
effect originated from increased elimination of the SigB negative strain or a decreased virulence
in situ the authors infected mice intra-articularly. In this experiment no difference in the ability to
cause inflammation was recognized between infections with the two strains. This led to the
conclusion that SigB-influenced steps before arthritical joint involvement contributed to the
differences between the strains after i. v. infection, possibly including SigB regulated cell surface
adhesins as crucial virulence factors (Jonsson et al. 2004).
In summary, the impact of SigB in vivo is still disputed, and results in literature references are
partly inconsistent. A hypothesis reconfirmed by several authors is that the effect of SigB in
in vivo infection models possibly only manifests in early phases of the infection and only impacts
processes in a transient manner.
Hence, it can be speculated that in the model described in this thesis the analyzed time point
might have been too late for detection of differences in the host reaction.
Furthermore, the effects of SigB are part of a complex and interfering regulation mechanism
including several other regulatory molecules like RNAIII from the agr locus or SarA. SigB
dependent transcription of sar locus was demonstrated (Ziebandt et al. 2004). The SarA protein is
a transcriptional regulator, which represses (e. g. spa) or activates (e. g. fnbA) gene expression
directly. It also activates expression of the agr system and therefore influences indirectly the
expression of genes responding to RNAIII (Chien et al. 1999). SigB itself has a negative effect on
the amount of RNAIII (Horsburgh et al. 2002b). Several genes or their promoter regions possess
binding sites for more than one transcriptional regulator or sigma factor, e. g. the sae locus coded
two-component system also influences SigB-regulated target genes (Goerke et al. 2005).
Therefore, the transcriptional pattern of the regulators overlap, and the mutation of one of them
180

Maren Depke
Discussion and Conclusions
might be compensated at least in parts by others. Such compensation might contribute to the
non-distinguishable host reaction in the model described in this thesis. Additionally, the balance
between the different regulators might be different in in vivo settings than in the well-studied
in vitro conditions. Goerke et al. (2005) observed less SigB activity in vivo in a guinea pig model of
implant infection than in vitro under induced conditions like in the post-exponential growth
phase. But differences were also observed between in vitro and in vivo analyses in the sequential
order of gene expression pattern (Goerke et al. 2005).
Even more complicated, a transposon mutagenesis study identified 13 SigB-independently
transcribed genes whose mutation led to an increase in hla transcription and protease activity.
This effect had a different magnitude depending of the mutation-affected gene and could be
traced back to a gradual reduction of SigB activity after mutation. As these genes are part of
normal cellular metabolic pathways and do not exhibit DNA binding motifs the authors propose
them to be indirectly involved in the control of RsbU activity, preventing full signaling through
RsbU but not fully blocking the phosphatase activity. Such hypothesis assigns these metabolic
genes a function similar to that of the B. subtilis RsbRSTX regulation system sensing
environmental conditions, which is absent in S. aureus (Shaw et al. 2006).
In this context, the question arises whether activation of SigB really takes place in the specific
in vivo infection setting or whether there are differences depending on the model used, the site
of infection or other experimental factors influencing S. aureus during infection. Reduced in vivo
activity of SigB might also result in the similarity of host reaction to infection with S. aureus
RN1HG and its sigB mutant as described in this thesis. The question whether sigB and the SigB
regulon are really expressed in infection models strongly suggests the examination of sigB and
SigB-dependent marker genes by real-time qPCR in infected samples for further investigations on
the relevance of SigB in in vivo settings.
In summary, the results of this study do not provide any hints for differences in the
pathogenesis or pathomechanism of the S. aureus strains RN1HG and ΔsigB in the selected model
of i. v. infection in mice. If really existing, such differences might be transient and only apparent
at earlier time points. Effects of SigB might also be superimposed in in vivo infection by the
interlaced pattern of other regulators. There is also the possibility of missing activity of SigB
in vivo which could explain the similarity of host reaction to infection with S. aureus RN1HG and
its sigB mutant in the model used in this study. SigB might possess only to a lesser extent
characteristics attributed to virulence factors and might act in vivo more like a virulence
modulator and fine tune bacterial reactions. Assuming such function, the missing of detectable
differences in the host’s reaction to S. aureus RN1HG and its isogenic sigB mutant is explainable.
181

Maren Depke
Discussion and Conclusions
GENE EXPRESSION PATTERN OF BONE-MARROW DERIVED
MACROPHAGES AFTER INTERFERON-GAMMA TREATMENT
The phagocytes of the innate immune system, neutrophils, monocytes/macrophages and
dendritic cells, accomplish central functions during the encounter of host and pathogen. They are
the first group of immune cells which recognize and react to an infection and the main effectors
of clearance of pathogens. Beside the different location in the non-reactive situation, when
neutrophils and monocytes circulate in the blood stream and macrophages and dendritic cells are
found in the tissue, macrophages and dendritic cells are long-living in comparison to the shortliving
neutrophils. Macrophages and dendritic cells have important functions in the regulation of
the immune reaction, e. g. by their antigen presentation ability (Gordon S 2007). The central
position of macrophages in the immune defense accounts for the relevance of research on
macrophage related topics.
Accordingly, the study described in this thesis aimed to analyze on a molecular level the
reaction of macrophages to interferon-γ (IFN-γ), which in low doses is known to be a priming
signal for macrophages and prepares the cells for a faster reaction to a second stimulus (Dalton
et al. 1993, Huang et al. 1993, Ma J et al. 2003, Mosser 2003). Here, by utilization of bonemarrow
derived macrophages (BMM), which were differentiated in vitro from bone marrow stem
cells under the influence of granulocyte-macrophage colony stimulating factor (GM-CSF), any
influence of immunological conditioning or in vivo stimuli were avoided. Such influences could
impair the experimental results when using mature macrophages prepared from animal organs.
Furthermore, recently established serum-free culture conditions (Eske et al. 2009) were applied.
This new system circumvented uncontrollable influences on BMM experiments, which would
have been introduced by vendor- and batch-varying, cytokine-, hormone- or endotoxincontaining
serum if traditional cultivation conditions had been applied.
The study included BMM of the two mouse strains BALB/c and C57BL/6, which were chosen
because of the differences observed between these mouse strains in in vivo and in vitro infection
studies (Breitbach et al. 2006, Autenrieth et al. 1994, van Erp et al. 2006). Thus, the data set from
this study was used to compare on the one hand non-stimulated control BMM of both strains,
and on the other hand BMM of the two strains after IFN-γ treatment.
Another feature of this study was the combined approach of transcriptome (Maren Depke)
and proteome (Dinh Hoang Dang Khoa) analysis. In the comparison of transcriptome and
proteome results, it was expected that the microarray as whole genome array covered a much
higher number of genes than the number of proteins covered by the gel-free LC-MS/MS
proteome approach, which was already chosen as best comparable method. Anyway, a defined
part of the microarray results like information on genes encoding membrane or secreted proteins
can practically not match proteome results for lack of accessibility. Hence, it was not surprising to
find a high number of differentially expressed genes, for which protein data were not available.
Vice versa, corresponding gene expression values were recorded for almost all protein data.
A good agreement of the results from both analysis levels was evident for the IFN-γ effects.
Depending on the selected comparison, the overlap of differentially expressed genes and
182

Maren Depke
Discussion and Conclusions
proteins different in abundance amounted to approximately 40 % to 60 % in reference to
genes/proteins which were accessible with both methods. Most recently, a study on LPSactivated
RAW 264.7 macrophages and C57BL/6 BMM was published. Applying isotope coded
affinity tagging (ICAT), multidimensional liquid chromatography, and mass spectrometry, 1064
proteins were identified and quantified of which 36 differed in abundance between LPS-treated
and control cells. Comparison to microarray data revealed approximately 75% concordance
(Swearingen et al. 2010). Although the treatments in both studies were different, the total,
regulated and overlap numbers were roughly comparable.
Nevertheless, a different phenomenon was observed when analyzing differences between
BMM of both strains. Expectedly, a high number of differentially expressed genes was observed,
which were not accessible with the proteome approach. But also a high number of regulated
proteins was recorded, for which transcriptome data were mostly available but did not show
differential expression. Thus, BMM in both strains might differ in post-transcriptional,
translational, and protein stability/turnover processes.
Although the regulated genes/proteins were not always identical for the different
comparisons, similar global results were received by both approaches.
1) IFN-γ treatment mainly led to induction of gene expression or an increase in protein
abundance.
2) IFN-γ induced changes were highly similar in BALB/c BMM and C57BL/6 BMM, since many of
them were observed in BMM of both strains or, when observed only in one of them, showed a
highly similar trend in the other in reference to their fold change values.
3) In the comparison between BMM of the two analyzed mouse strains, about 50 % of regulated
genes/proteins exhibited higher expression/abundance in BALB/c BMM whereas the other
50 % showed higher expression/abundance in C57BL/6 BMM.
4) Strain differences were highly similar in the comparison of control level BMM and in the
comparison of IFN-γ treated BMM.
Apart from the comparison to proteome results, the transcriptome data were analyzed for
their biological content. The confirmation of known IFN-γ effects in the data set proved the
relevance of the results. Very prominent, the induction of immunoproteasome and antigen
presentation genes became visible. These included Psmb8, Psmb9, Psmb10, Psme1, and Psme2,
genes of proteasome subunits, Tap1 and Tap2, peptide transporters from cytosol to the
endoplasmatic reticulum (ER), Erap1, ER aminopeptidase, and MHC class I and class II genes,
which were already described to be induced by interferon (Aki et al. 1994, Van den Eynde/Morel
2001, Brucet et al. 2004, Ma W et al. 1997, Schiffer et al. 2002, Chang et al. 2005, Jung et al.
2009, Benoist/Mathis 1990, Boehm et al. 1997, Gobin/van den Elsen 2000). IFN-γ is also in vivo
responsible for the complete exchange of the constitutive to the immunoproteasome since this
exchange was shown to be reduced to 50 % in the liver of IFN-γ -/- BALB/c mice after infection with
lymphocytic choriomeningitis virus. The authors suppose TNF-α to be responsible for the
remaining fraction of exchange (Khan S et al. 2001). Other in vivo experiments using the fungal
pathogen Histoplasma capsulatum in infections of IFN-γ -/- C57BL/6 mice resulted in a complete
inability to induce the immunoproteasome. In that study, the authors could not find hints that
other cytokines can compensate for the loss of IFN-γ in reference to the immunoproteasome
(Barton et al. 2002). The knockout of inducible proteasome β-subunit LMP-7 (Psmb8) led to a
strongly increased susceptibility of mice during Toxoplasma gondii infections, which was
183

Maren Depke
Discussion and Conclusions
explained by the missing pathogenic antigen processing in dendritic cells and the consequently
missing activation of specific cytotoxic CD8 + T cells (Tu et al. 2009).
The induction of immunoproteasome β i -subunits but also of Tap1 was reported to be
mediated by STAT1 during signal transduction and IRF-1, which acts as transcriptional activator
(White et al. 1996, Chatterjee-Kishore et al. 1998, Foss/Prydz 1999, Brucet et al. 2004, Namiki et
al. 2005). In agreement with literature information, IFN-γ treatment led to a clear induction of
STAT1 and IRF-1 in BMM of both strains in the study of murine BMM described in this thesis (data
not shown).
The gene expression profiling revealed the influence of IFN-γ on several cytokines. BMM
induced several chemokines (Ccl5, Ccl12, Ccl22, Cxcl9, Cxcl10, Cxcl11, Cxcl16), chemokine
receptors (Ccr5, Ccrl2), interleukins (Il15, Il27), interleukin receptors (Il2rg, Il4ra, Il10ra, Il12rb1,
Il13ra1, Il15ra, Il21r) and also TNF and TNF family ligands (Tnf, Tnfsf10, Tnfsf18) and receptor
(Tnfrsf14). These were interpreted as preparation for a potentially following second stimulus and
consequent immune reaction, which includes pro- and anti-inflammatory directions.
Cxcl9, Cxcl10, and Cxcl11 are known to be induced by IFN-γ (Luster et al. 1985, Farber 1990,
Taub et al. 1993, Liao et al. 1995, Cole KE et al. 1998). These chemokines bind the receptor Cxcr3,
which is preferentially found on IL-2 activated Th1 cells (Loetscher M et al. 1996). Of these three
chemokines, Cxcl11 has highest affinity to Cxcr3 and additionally is the most potent agonist
(Cole KE et al. 1998). Because the cytokine exerts its influence only on activated T cells, the
authors additionally infer that the cytokine mainly acts during the immune response and aims to
direct effector T cells to the manifestation site, when T cell activation, IL-2 production, and
proliferation of specific T cells had occurred in lymphoid organs (Cole KE et al. 1998).
The three chemokines Cxcl9, Cxcl10, and Cxcl11 exert their function not only in the
chemotaxis of Th1 cells via Cxcr3. It was reported that they also can bind to Ccr3, the receptor for
several Ccl chemokines (Ccl5, Ccl7, Ccl8, Ccl11, Ccl13, Ccl24), which is found on Th2 cells. Cxcl9,
Cxcl10, and Cxcl11 act antagonistic when binding to Ccr3. Thus, these chemokines do not only
attract Th1 cells but also inhibit chemotaxis of Th2 cells and further polarize the immune
response (Loetscher P et al. 2001). More recently, a similar antagonistic effect of Cxcl11 on the
receptor Ccr5 was reported (Petkovic et al. 2004).
Most interestingly, beside agonist and antagonist functions, Cxcl9, Cxcl10, and Cxcl11 have
been reported to exhibit direct antimicrobial functions against Escherichia coli and Listeria
monocytogenes. The observation was rated as biologically relevant since antimicrobially effective
concentrations of chemokines were determined in vitro and in vivo (Cole AM et al. 2001). The
antimicrobial effect of Cxcl9, Cxcl10, and Cxcl11 also applied to S. aureus (Yang D et al. 2003).
Another study determined constitutive expression of Cxcl9 in the human male reproductive
system and secretion into seminal plasma where antibacterial activity against the urogenital
pathogen Neisseria gonorrhoeae was demonstrated leading to the conclusion that Cxcl9 is
relevant for the host’s local immune defense (Linge et al. 2008). The multifunctional
characteristic of chemokine and antimicrobial function is not a unique feature of Cxcl9, Cxcl10,
and Cxcl11, but was observed for several other chemokines, too. Antimicrobial activity could be
assigned to about two third of all studied chemokines (Yang D et al. 2003).
The cytokine TNF was induced in BMM after IFN-γ treatment. TNF is the central inflammation
mediator. Since long it is known that TNF is the mediator of lethal endotoxin effects (Beutler et
al. 1985), which finding was worth a reprint as part of the “Pillars of Immunology” series in The
Journal of Immunology (Vilcek 2008). In the meantime, knowledge was expanded and
184

Maren Depke
Discussion and Conclusions
superfamilies of TNF ligands and receptors were identified (Hehlgans/Pfeffer 2005). TNF is
associated with several functions linked to pro-inflammatory effects, which involve a
multiprotein signaling complex at the cell membrane and MAP-kinase signaling and which are
finally mediated via NFκB and c-Jun (Chen G/Goeddel 2002, Wajant et al. 2003). Rarely, TNF can
induce apoptosis via caspase activation, depending on the cellular balance of antiapoptosis/proliferation
signals mediated by NFκB (Gaur/Aggarwal 2003, Wajant et al. 2003).
Furthermore, TNF is involved in the regulation of fever with both pyrogenic and antipyretic
effects depending on physiological or experimental conditions (Leon 2002). In macrophages, TNF
has an important function in controlling intracellular mycobacterial growth (Bekker et al. 2001).
TNF deficient mice are susceptible to Listeria monocytogenes infections, exhibit reduced contact
hypersensitivity and impairment of splenic follicular architecture and of maturation of the
humoral immune response (Pasparakis et al. 1996). Knockout of TNF and lymphotoxin-α (TNF-β)
leads to an increased susceptibility to S. aureus infections, but also to a reduced frequency of
arthritis indicating a damaging influence of the cytokine during the inflammatory reaction
(Hultgren et al. 1998). Reflecting the central position of TNF in the regulation of the immune
response to infection, several strategies of pathogens to influence the TNF activity were
reported. The interference can lead to diverse and contrary effects depending on the pathogen,
like inhibition of apoptosis, enhancement of apoptosis, inhibition of TNF production or, in case of
S. aureus, the induction of a TNF-like response by interaction of protein A with the TNF receptor
TNFR1 (Rahman/McFadden 2006).
Beside other functions, TNF was described to further enhance the expression of the
immunoproteasome (Loukissa et al. 2000). In non-professional antigen-presenting cells, the
induction of immunoproteasome, TAP peptide transporters, and MHC class I complexes by TNF
independent of IFN-γ activity was observed and in addition increased stability of the MHC class I
complexes at the cell surfaces (Hallermalm et al. 2001). Thus, induction of TNF in BMM after
IFN-γ treatment might intensify the effect on antigen processing and presentation after a second
stimulus. Also in the serum-free system, the inducibility of TNF in IFN-γ and LPS treated BMM was
described (Eske et al. 2009). IL-6, IL-10, and IL-12, which were also inducible by IFN-γ and LPS
(Eske et al. 2009), were not expressed in IFN-γ treated BMM (this study), and the IFN-γ and LPS
inducible chemokine Ccl2 (MCP-1; Eske et al. 2009) was expressed but not differentially regulated
in IFN-γ treated BMM (this study). This strongly hints for the explanation that the induction only
occurs when a second stimulus like LPS is given.
After IFN-γ treatment, the induction of anti-inflammatory IL-10 receptor (Moore et al. 2001)
as well as the induction of Il18bp coding for an IL-18 binding protein, which is a natural
endogenous inhibitor of the proinflammatory IL-18 (McInnes et al. 2000, Gracie et al. 2003,
Novick D et al. 1999), was observed (this study). Not only Stat1 and Irf1 of the positive IFN-γ
feedback were induced, but also Stat3 and Socs1 which are implicated in negative feedback
(Schroder et al. 2004, Hu et al. 2008). This might indicate the possible reactivity to antiinflammatory
stimuli and a confinement of inflammatory response.
After the priming signal of IFN-γ, the BMM are not fully activated and are not expected to
secrete cytokines, but rather after a second stimulus (Hu et al. 2008). Analysis of cytokine
secretion of BMM beyond the examples analyzed until now will be of interest, either after
stimulation with LPS, or after stimulation with molecules of Gram-positive bacteria or even
infection e. g. with S. aureus, as it will be in focus of research for follow-up experiments to this
study.
185

Maren Depke
Discussion and Conclusions
Three other induced genes, kynreninase (Kynu), inducible nitric oxide synthase 2 (Nos2), and
tryptophanyl-tRNA synthetase (Wars), found attention. These are closely linked to inflammation
and interferon action.
Kynureninase was reported to be induced in immortalized murine macrophages (cell line
MT2) by IFN-γ, which was interpreted as an activation of the kynurenine pathway leading to the
formation of quinolinic acid from tryptophan since the cells also induced indoleamine 2,3-
dioxygenase (IDO), which catalyzes the first step of tryptophan degradation (Alberati-Giani et al.
1996). Interestingly, BMM after IFN-γ treatment (this study) did not differentially express
indoleamine 2,3-dioxygenase (IDO) which is known to be induced by interferons in many cell
types (Carlin et al. 1989, Taylor MW/Feng 1991, Alberati-Giani et al. 1996). Even more,
expression of Ido1 and Ido2 was low and even absent in several samples. In literature references,
a lack of inducible IDO activity by IFN-γ was observed in murine peritoneal macrophages. These
cells were able to kill pathogens, and the antimicrobial effect was not diminished by addition of
tryptophan like it was observed for monocyte-derived macrophages (Murray HW et al. 1989).
It was reported that the immune defense mechanisms of IDO/tryptophan degradation and
nitric oxide synthase (NOS)/arginine degradation influence each other. Nitric oxide synthase
catalyzes the first, rate-limiting step of arginine degradation to NO and citrulline
(Knowles/Moncada 1994). NO decreases IDO activity in human mononuclear phagocytes by
interfering with the heme iron located at the active site of IDO enzyme (Thomas et al. 1993).
Experiments gave hints for species-specific reactions to IFN-γ since human macrophages induced
IDO and murine macrophages NOS. Nevertheless, in presence of inhibitors of NO synthesis also
murine macrophages exhibited IDO activity (Thomas et al. 1993). NOS inhibitors in combination
with IFN-γ further increased the induction of IDO mRNA in murine macrophages MT2, which was
already induced after IFN-γ treatment alone (Alberati-Giani et al. 1997). Influence of NO on IDO
mRNA could not be detected in the human epithelial cell line RT4, but NO led to accelerated
proteosomal degradation of IDO protein. The authors stated that the transcriptional regulation of
IDO might be species or cell type specific and needs to be furthermore analyzed (Hucke et al.
2004). Not only NO and reactive nitrogen species influence IDO, but there are also hints for
influences in the opposite direction. The intermediate 3-hydroxyanthranilic acid from the
kynurenine pathway of tryptophan degradation was shown to inhibit the enzymatic activity of
iNOS and the iNOS mRNA induction by IFN-γ and LPS in mouse macrophage RAW 264.7 cells. The
decrease in iNOS mRNA level resulted from inhibition of NFκB activation (Sekkaï et al. 1997).
Even when there are still open questions concerning the interplay between IDO and iNOS, and
even when details in their regulation might differ between mouse and human and between the
different cell types, many information support the proposition that in each physiological situation
a cellular decision towards one of the two antimicrobial mechanisms has to be taken since each
mechanism compromises the other. Fittingly, during the treatment of BMM with IFN-γ the
induction of Nos2, inducible nitric oxide synthase 2, was observed in parallel to the lacking
induction/expression of IDO (this study).
Wars is known to be regulated by IFN-γ (Fleckner et al. 1991, Buwitt et al. 1992, Bange et al.
1992, Rubin et al. 1991). The biological function of the induction of this metabolic enzyme in
inflammatory situations is discussed in different ways: 1) protection of the host from self-induced
tryptophan starvation: In an environment of tryptophan degradation, the amino acid would be
saved from IDO activity when complexed to tRNA, not accessible for pathogens, but still available
for host protein synthesis (Boasso et al. 2005, Murray MF 2003). 2) compensation for increased
186

Maren Depke
Discussion and Conclusions
tryptophanyl-tRNA need: Interferon inducible proteins with tryptophan-enriched sequences
require above average tryptophanyl-tRNA (Xue/Wong 1995). 3) expanded function of the
tryptophanyl-tRNA synthetase: Splice variants were shown to act anti-angiogenic (Tolstrup et al.
1995, Otani et al. 2002, Wakasugi et al. 2002).
It has been shown that bone marrow-derived myeloid dendritic cells (Hara et al. 2008) and
astrocytes (Speciale et al. 1989) take up exogenous kynurenine, which would be an explanation
for the induction of kynureninase in BMM after treatment with IFN-γ (this study).
It can be speculated that the induction of Kynu and Wars together in BMM after treatment
with IFN-γ might indicate a preparation of the BMM for their presence in inflamed tissue where
they might encounter a tryptophan depleted environment.
Interferons are described to induce a group of GTPases / GTP binding proteins of different
families which play a role in antimicrobial defense. Of each of the four families described in
literature (MacMicking 2004) examples were found in the BMM data set after IFN-γ treatment, in
part exhibiting the highest induction factors of the data set: p47 GTPases (Igtp, Iigp1, Irgm1,
Tgtp), p65 guanylate-binding proteins (Gbp1, Gbp2, Gbp3, Gbp4, Gbp5, Gbp6), Mx proteins (Mx1,
Mx2), and very large inducible GTPases (Gvin1). The observation of induction is entirely in
agreement with literature references and impressive in its entity. This group of GTPases provides
the host with resistance against viral and microbial pathogens by cell-autonomous resistance
mechanisms.
GTPase p47 and p65 cellular knockdown studies and experiments using knockout animals led
to a higher susceptibility during infection. These GTPases were described to be membraneassociated
via myristoylation, isoprenylation, or interaction with e. g. Golgi-proteins. GTPases of
the p47 family were recruited to phagosomes after infection. Thus, they target intracellular,
vacuolarized pathogens by a mechanism proposed to remodel the pathogen-containing
compartment and to increase the fusion with lysosomes. GTPases of the p65 family were
described to be involved in the control of virus infections. More recently, a publication reported a
role for p65 GTPases during defense against bacterial and protozoan infections (Shenoy et al.
2007, Taylor GA et al. 2007, Anderson SL et al. 1999, Carter et al. 2005, Degrandi et al. 2007). Mx
proteins, which intracellularly bind to virus particles, are part of the innate immune defense
against several RNA viruses. In mice, the location of the Mx protein type (nuclear Mx1, cytosolic
Mx2) correlates with the replication site of the viruses against which the Mx proteins provide
resistance (Haller/Kochs 2002, Haller et al. 2007). Finally, very large inducible GTPases (VLIG) are
proteins of approximately 280 kDa. Despite the big size, the protein is encoded in a single exon.
Mice harbor six different, but similar VLIG genes, which are organized in a gene cluster on
chromosome 7 (Klamp et al. 2003). The genomes of primates and carnivores do not include VLIG
sequences, which probably were lost during evolution since other vertebrata own VLIG (Li et al.
2009).
Literature data indicate that C57BL/6 BMM do not synthesize Gbp-1 protein after interferon
induction because they own a different allele of the Gbp-1 gene (Staeheli et al. 1984). Although
this study detected a 6-fold increase in Gbp-1 mRNA of C57BL/6 BMM after IFN-γ treatment,
even the induced mRNA level was still very low. This change in mRNA was not detectable on
protein level: LC-MS/MS data of Dinh Hoang Dang Khoa revealed equal protein abundance in
control and IFN-γ treated C57BL/6 BMM, which was in accordance with the phenotype described
in literature.
187

Maren Depke
Discussion and Conclusions
Only few genomewide transcriptome profiling studies on IFN-γ and macrophages are available
in literature. Kota et al. analyzed the reaction of murine RAW 264.7 cells to IFN-γ exposition for
4 h (Kota et al. 2006). Despite the difference of macrophage cell line to BMM and of treatment
time and IFN-γ dose (4 h and 1000 U/ml vs. 24 h and 300 U/ml) between the published study and
the study described in this thesis, a very good agreement in the differential regulation of several
genes was observed (e. g. Oas, GTPases, Socs1, Socs3, Nos2, Cxcl9, Cxcl10, Irf1, Cxcr4,
immunoproteasome and associated genes, MHC molecules, Il18bp, Il10ra, Ctsc, Ptgs2 and
others). Ehrt et al. analyzed BMM after 72 h of IFN-γ treatment (100 U/ml), infection for 24 h
with Mycobacterium tuberculosis, or both (Ehrt et al. 2001). The authors observed – in contrast
to the results of this study – slightly more repressed than induced genes after IFN-γ treatment
alone. The infection setting resulted in an additional, specific set of differentially expressed
genes. Such additional set of genes is also expected in case of future inclusion of further infected
samples in the settings of the study described in this thesis. The authors defined a set of
approximately 1300 genes to be regulated upon treatment of BMM with IFN-γ (Ehrt et al. 2001).
This high number is not directly comparable with the results of this study, because Ehrt and
colleagues treated each probe set of the array as if it represented a single gene. This
simplification was necessary because the analysis was performed before the complete
sequencing of the mouse genome was finished. Thus, at that time, the partially overlapping EST
and cDNA sequences were difficult to assign to gene information. Furthermore, the authors
explain the result of the high number of regulated genes with the long IFN-γ stimulation time of
72 h. But also between the study by Ehrt et al. and the study described in this thesis accordance
was observed (e. g. MHC molecules, Gbp2 and other GTPases, Irg1, Nos2, Cxcl10, Cxcr4).
Interestingly, the authors monitored a strong influence of Nos2 deficiency on the gene
expression profile after IFN-γ treatment leading to the conclusion that Nos2 directly or indirectly
influences the cellular reaction to IFN-γ stimulus. Zocco and coworkers focused their analyses on
rat hepatic macrophages, i. e. Kupffer cells, stimulated with 1000 U/ml IFN-α or IFN-γ for 8 h
(Zocco et al. 2006). Using an Affymetrix array with about 8800 probe sets, they observed 70
induced and 72 repressed genes by IFN-γ, the relation of which resembles more that observed by
Ehrt et al. than the relation determined in the study described in this thesis. Nevertheless, also
here a certain concordance of cellular reaction became visible (e. g. Mx, Gbp2, Nos2, Irf1, Stat1,
Oas, immunoproteasome subunits, MHC molecules). In the study of Pereira et al. BMM of A/J or
BALB/c mice were treated for 18 h with 50 U/ml of IFN-γ and analyzed on a 1536 feature cDNA
array from a fetal thymus library (Pereira et al. 2004). For BALB/c BMM, the induction of 297
genes and repression of 58 genes was recorded. Here, the relation of induction to repression is
similar to that observed in the study described in this thesis although the comparison of results is
difficult because of the very different arrays, which were used in both studies. The comparison
with similar transcriptome studies from literature leads to the conclusion that the results of this
study are well supported by knowledge on cellular reactions to IFN-γ. In addition, similar to
comparisons of other studies, also in this study specific gene expression changes were observed
which might be explained with experimental differences. Nevertheless, for a comprehensive
knowledge of IFN-γ effects the study of different experimental settings is necessary to which this
study contributes.
During data analysis using the Ingenuity Pathway Analysis tool (IPA, www.ingenuity.com), it
became clear that only a fraction of the IFN-γ related, differentially expressed genes were linked
to macrophages or RAW cells in the IPA database. Of about 180 genes, approximately 130 were
188

Maren Depke
Discussion and Conclusions
associated with IFN-γ in other cells, but not in macrophages/RAW cells. Thus, this study provides
a significant contribution to extend the scientific knowledge of IFN-γ effects in macrophages.
Strain differences in gene expression occurred between BALB/c BMM and C57BL/6 BMM on
both treatment levels. Most differences were similar at control level and after IFN-γ treatment,
and the differences included genes, which were expressed at a higher level in BALB/c BMM or in
C57BL/6 BMM. Some insights into the data indicate that the differences might be immunerelevant.
For example Cxcl11, the Th1 cell attracting chemokine, differed between the strains
after IFN-γ treatment because it was induced from a common baseline level only in BALB/c BMM.
Since it is known that the immune response of C57BL/6 mice is balanced towards Th1 with high
IFN-γ/low IL-4 production in splenocytes after concanavalin A stimulation and that of BALB/c
mice towards Th2 with low IFN-γ/high IL-4 production in splenocytes after concanavalin A
stimulation (Mills et al. 2000), the difference after IFN-γ treatment was unexpected. But as the
IFN-γ stimulus was added externally, the difference might display a compensatory variation of
BALB/c BMM gene expression. C57BL/6 macrophages were reported to produce more NO than
BALB/c macrophages (Mills et al. 2000). The induction of Nos2 by IFN-γ in C57BL/6 BMM was by a
factor of about 1.4 stronger than in BALB/c BMM in this study, but the difference was not
significant although a trend for confirmation of literature data was visible. Furthermore, BALB/c
macrophages were observed to produce more ornithine/urea because of a stronger arginase
activity (Mills et al. 2000). In this study, BALB/c BMM exhibited higher expression of arginase 2
than C57BL/6 at control and at IFN-γ treated level as indicated by literature data.
The phenotypical differences between the reaction of BALB/c and C57BL/6 BMM were often
determined in the presence of IFN-γ and a second stimulus like LPS or infection (Breitbach et al.
2006, Eske et al. 2009). To elucidate molecular reasons for the observed differences in killing of
pathogens or cytokine production, the inclusion of samples subjected to a second stimulus in
addition to IFN-γ is recommended.
The experimental setting of this study was planned with the aim to analyze basic principles of
murine BMM: the reaction to the priming signal IFN-γ on a molecular level and differences of
reaction between the BMM of both mouse strains. To further elucidate reasons for the different
success of the mouse strains to fight and overcome infections in vivo and in vitro, experiments
will need to be expanded to the analysis of IFN-γ treatment in combination with bacterial
infection. This might include studies using Burkholderia infections for which attenuated mutants
exist of whose study interesting results are anticipated. Furthermore, the first proteome analyses
were performed for infection experiments of BMM with S. aureus, of which the first results are
presented in the thesis of Dinh Hoang Dang Khoa. Further complementing results are expected
from the extension of the analysis to transcriptome level.
189

Maren Depke
Discussion and Conclusions
HOST CELL GENE EXPRESSION PATTERN IN AN
IN VITRO INFECTION MODEL
Several different cell types of mammalian hosts are in the risk of encounter with pathogens
depending of the entry and manifestation site of infection. These are immune cells as well as cells
associated with structural and functional aspects of the tissue. In this study, the human bronchial
epithelial cell line S9 (American Type Culture Collection ATCC, Manassas, VA, USA; www.atcc.org;
S9 cell ATCC number CRL-2778) served as a model system to study the reaction of epithelial
cells to infection. With regard to the identification of S. aureus as one of the leading causative
organisms of pneumonia (Goto et al. 2009), this species was employed as model pathogen. More
specific, S. aureus RN1HG, a rsbU + repaired RN1-derivative strain (Herbert et al. 2010) with a
SigB-positive phenotype, was chosen.
The infection setting involved bacterial cultivation in an adapted cell culture medium (Schmidt
et al. 2010), which allowed inoculation of eukaryotic host cell cultures with a fraction of complete
bacterial culture. This experimental setup allowed the study of host cell reactions to the influence
of both bacterial factors, its secreted proteins from supernatant and the membrane-bound
factors, which both interact during infection and contribute to the success of the bacterial cells. It
also avoided potentially disturbing influences of bacterial cell handling like that occurring during
centrifugation and washing.
In a combined approach of transcriptome (Maren Depke) and proteome (Melanie Gutjahr)
analysis, the host reaction to infection and bacterial internalization was recorded. Two time
points, 2.5 h and 6.5 h after start of infection, were selected for sampling. Because only about
50 % of host cells in infected cell culture plates harbor internalized staphylococci after infection,
the fraction of non-infected cells was removed by FACS-sorting, which was feasible because a
GFP-expressing S. aureus RN1HG strain has been used. The remaining infected S9 cells were
compared to medium control cells.
In the comparison of infected S9 cell proteome and transcriptome signatures, a considerable
time shift of cellular reaction between both analysis levels was evident. In samples of the first
time point 2.5 h, approximately half the number of proteins differed in abundance between
infected and control cells in comparison to the number of regulated proteins at the later time
point of 6.5 h. Conversely, for the transcriptome analysis, the number of differentially expressed
genes at the first time point amounted to about 3.5 % of the number of regulated genes at the
second time point. Thus, the reaction on proteome level started to a stronger extent between
inoculation and first analysis time point 2.5 h, whereas the transcriptional reaction most
articulately started between the first (2.5 h) and the second (6.5 h) time point.
The cellular reaction to infection in this experimental setting seems to lead first to protein
abundance changes independent from mRNA changes, which on the one hand probably rely on
the already existing messengers and on the other hand might represent post-transcriptional,
translational and degradation/turnover processes as reduction in protein abundance prevailed at
the early time point. Differentially expressed genes and regulated proteins did not overlap at the
first time point 2.5 h after start of infection. Two genes, IFIT2 and IFIT3, were induced at both the
190

Maren Depke
Discussion and Conclusions
2.5 h and the 6.5 h time points, but the protein abundance increased only at the 6.5 h time point.
This indicates that the first mRNA expression changes were manifested on proteome level
probably after the first analyzed time point (2.5 h) and before or at latest directly at the second
analyzed time point (6.5 h).
A time shift between detectable messenger and protein level changes can explain the
observation that approximately 5 % of protein abundance changes found their correspondence at
transcript level at the 6.5 h time point, and vice versa, the even smaller fraction of 0.8 % of
differential expression was realized on proteome level at the same time point. It has to be
mentioned that a small number of genes/proteins showed reverse effects in transcriptome and
proteome analysis, or changes in protein abundance preceeded mRNA changes. This can be
explained by post-transcriptional, translational and protein turnover mechanisms, by time shift
aspects or by the possibility of counter-regulatory processes. During proteome analysis, different
methods have to be applied to study proteins from different compartments or with different
physical characteristics. In this study, a gel-free proteomics approach was applied, which is
known to yield less information on membrane proteins because these are not accessed during
extract preparation and enzymatic digestion. Thus, a defined part of the microarray results can
practically not match proteome results for lack of accessibility. The results from the comparison
of the proteome and the transcriptome analysis approach underline the importance of
performing both analysis levels in parallel. Information of both approaches complement each
other and supply valuable data to establish a comprehensive view of the experimental system.
Although transcriptome profiling revealed the very small number of 40 differentially
expressed genes in S9 cells 2.5 h after start of infection with S. aureus RN1HG, these gene
expression changes indicated the beginning response of the host cells to infection. Furthermore,
these changes were in very good agreement with the differential gene expression at the second
analysis time point four hours later.
At the 2.5 h time point, a strong induction of a proinflammatory response was observed. This
included cytokines like IL-6 and IFN-β, genes related to inflammation mediators like PTGS2, and
genes which are linked to these aspects because they are part of signal transduction and
transcription processes. In addition to the observation of proinflammatory mechanisms, gene
expression changes which aim to counter-regulate and balance the inflammation were observed
already 2.5 h after start of infection. Nevertheless, the infection influenced the host cells very
strongly, and hints for morphological changes and a limitation of cell cycle progression became
visible.
The study included a second sampling point 6.5 h after start of infection. The cellular state at
this time point was characterized by a stronger loss of viability, but these non-viable cells were
not included in the sample preparation after FACS-sorting. Thus, a viable, GFP-positive infected
cell fraction was utilized for transcriptome profiling. Resulting differentially expressed genes
distinguished a rather strong cellular response as the number of regulated genes increased by a
factor of approximately 30 in comparison to the first time point.
In a general view, the S9 cells started a regulatory gene expression program 6.5 h after start of
infection, which was strongly associated with diverse inflammation and cell death related
functions. These included signal transduction linked genes for the interferon signaling cascade
together with IFN-β itself, which has already been induced at the 2.5 h time point, pattern
recognition receptors, antigen presentation and immunoproteasome, but also death receptor
signaling. The amplification of cellular response was indicated by the emergence of functional
191

Maren Depke
Discussion and Conclusions
associated groups of genes, of which few were regulated at the early and several at the later time
point of analysis. Examples are given with leukemia inhibitory factor LIF (induced at both time
points) and its receptor (induced after 6.5 h), ligand CD274 (PD-L1, induced at both time points)
and PD-L2 (induced after 6.5 h), several interferon regulated genes at both time points or
additionally at the 6.5 h time point, and the histone genes, of which one was repressed at the
2.5 h time point and further 17 were added at the 6.5 h time point.
Furthermore, indoleamine 2,3-dioxygenase 1 (IDO1) belonged to the genes with the highest
induction factors at the 6.5 h time point. The gene codes for a key molecule in
immunomodulation, microbial growth control, and pathogen immune escape (Puccetti 2007,
Zelante et al. 2009). IDO is known to be induced by interferons (Carlin et al. 1989,
Taylor MW/Feng 1991) e. g. during infection and inflammation, but also in conditions like stress,
which result in a transient increase of proinflammatory cytokines (Kiank et al. 2010). The enzyme
catalyzes the reaction of tryptophan to formylkynurenine. This is the first rate-limiting step of the
so-called “kynurenine pathway”, which leads to NAD + biosynthesis or to the complete oxidative
degradation of kynurenine (King/Thomas 2007, Xie et al. 2002). One of the main effects of
induced IDO activity is the depletion of tryptophan, which is discussed to be a mechanism to
inhibit microbial growth (Pfefferkorn 1984, Byrne et al. 1986, Müller et al. 2009). But this
mechanism can be counteracted by bacteria, which might induce their tryptophan biosynthesis
or which developed own counter-regulatory mechanisms. An example is Chlamydophila psittaci
which owns adapted tryptophan biosynthesis genes that enable the bacterium to produce
tryptophan from kynurenine taken from their host’s metabolism (Xie et al. 2002).
During metabolism of tryptophan, several neuro- and immune-active substances are
produced, of which some additionally exhibit toxic features and thus might act antimicrobially.
The intermediates of tryptophan and kynurenine degradation, 3-hydroxykynurenine and α-
picolinic acid, were reported to inhibit growth of S. aureus and other bacterial species in a murine
transplantation model even in the presence of tryptophan (Saito et al. 2008, Narui et al. 2009).
The most prominent characteristic of IDO is the induction of an anti-inflammatory bias after a
proinflammatory stimulus. This effect aims to limit inflammation and the accompanying damage
to the host tissue. It has to be considered that the experimental setting in this study only
included one single cell type, the bronchial epithelial cell line S9. Thus, interpretation of an antiinflammatory
effect assumes that bronchial epithelial cells might also in vivo induce IDO, e. g.
during staphylococcal pneumonia. Tryptophan degradation intermediates were shown to inhibit
proliferation of CD4 + and CD8 + T cells and of NK cells (Frumento et al. 2002). Other experiments
supported a proposed mechanism of inhibition of T cell proliferation, which was triggered by the
low concentration of tryptophan resulting from IDO activity and mediated by GCN2 kinase
(EIF2AK4, eukaryotic translation initiation factor 2 alpha kinase 4), a sensor for uncharged tRNA
(Munn et al. 2005). The contradiction that tryptophan depletion leads to an antimicrobial effect
on the one hand and to an inhibition of T cell mediated immune response on the other hand was
resolved by Müller and coworkers, who determined the concentration limit necessary for the two
effects. They observed that bacterial inhibition (50 % inhibitory concentration of 1.9 µM) took
already place at a higher tryptophan concentration whereas tryptophan had to be depleted even
further (50 % inhibitory concentration of 0.1 µM) to achieve T cell inhibition (Müller et al. 2009).
Besides the described aspects, in an in vivo infection situation dendritic cells (DC) will contribute
to the immune response. Further IDO-mediated immune-modulation mechanisms involve DC
mediated T cell tolerance (Popov/Schultze 2008). The induction of IDO in epithelial cells near
192

Maren Depke
Discussion and Conclusions
sites of infection might help to constrain the immune defense reactions to the actually infected
cells and to protect the non-infected neighboring cells from damage by immune cells like
cytotoxic T cells (King/Thomas 2007). This confinement of immune response evolutionarily
developed in favor of the host. Nevertheless, in certain context it benefits pathogens, which
might cause persistent infections since the immune system switched to an anti-inflammatory
state (Zelante et al. 2009).
In relation to IDO1, the increased expression of tryptophanyl-tRNA synthetase (WARS) was
conspicuous, which was also confirmed on protein level. In contrast to bacterial tRNA
synthetases, the eukaryotic or mammalian enzymes were studied to a lesser extent. The enzyme
was identified and described to be IFN-γ induced by two groups, Fleckner and coworkers as well
as Buwitt, Bange and colleagues almost 20 years ago (Fleckner et al. 1991, Buwitt et al. 1992,
Bange et al. 1992) and separately, the inducibility by IFN-α and IFN-γ was demonstrated (Rubin et
al. 1991). The protein domain structure (reviewed by Kisselev 1993) as well as the gene’s exonintron-organization
and interferon response elements are long-known (Frolova et al. 1993). It
was hypothesized that the induction of tryptophanyl-tRNA synthetase in parallel to the
tryptophan catabolizing enzyme IDO aims to protect the host from the tryptophan starvation
which is induced in response to proinflammatory stimuli. Tryptophan would be safe from
degradation when complexed to tRNA and still available for host protein synthesis (Boasso et al.
2005, Murray MF 2003). Furthermore, the induction of tryptophanyl-tRNA synthetase by
interferon was described to find its biological rationale in the enrichment of tryptophan in
immune response related and equally interferon inducible proteins like MHC molecules
(especially in the Ig-domains). Hence, induction of tryptophanyl-tRNA synthetase was suggested
to reflect the heightened need of tryptophanyl-tRNA to sustain the ability to synthesize these
proteins (Xue/Wong 1995). More recent literature data assign tRNA synthetases expanded
functions (Yang XL et al. 2004). Proteolytic tryptophanyl-tRNA synthetase fragments (Favorova et
al. 1989), splice variants (Tolstrup et al. 1995), and transcription starting from a newly identified
IFN-γ sensitive promoter (Liu J et al. 2004) were reported. Smaller tryptophanyl-tRNA synthetase
variants were shown to act anti-angiogenic (Otani et al. 2002, Wakasugi et al. 2002).
In infected S9 cells 6.5 h after start of infection, expression changes of cytokine genes,
GTPase/GTP binding proteins, genes involved in (anti)coagulation, anti-oxidant defense, and
complement system, lysosomal proteins and adhesins were obvious. This time point also
included repression of different branches of de novo lipogenesis like cholesterol, unsaturated
fatty acid, and storage lipid biosynthesis. Fatty acid synthase FASN was repressed in
transcriptome analysis and was equally reduced in abundance in proteome analysis. Contrarily,
several genes which are involved in lipid messenger generation were induced. The hints for
induced ceramide/sphingosine biosynthesis link the induction of lipid messenger generating
genes to the observation of induced expression of genes associated with death receptor signaling
and apoptosis.
The transcriptome profiling of infected S9 cells revealed that in the death receptor signaling
cascade, the receptor FAS and some initiating caspases were induced in parallel with the ligands
TNFSF10 and TNFSF15 and signal transduction genes. But also a caspase inhibitor (CFLAR) and an
anti-apoptotic gene (BAG1) were induced, and CASP9, link to the mitochondrial apoptosis
pathway, was repressed. Among further pro-apoptotic genes, a set of five apolipoprotein L genes
(APOL1, APOL2, APOL3, APOL4, APOL6) was detected with induced expression, of which one,
APOL2, was also increased in protein abundance. The APOL genes were of special interest. The
193

Maren Depke
Discussion and Conclusions
induced APOL1 belongs to the group of BH3-only proteins which achieve their apoptotic effect by
binding to proteins of the Bcl-2 family. It is discussed that the apoptotic effect is mediated by a
mechanism of activating pro-apoptotic members of this family or inactivating pro-survival
members in a de-repression mode (Bouillet/Strasser 2002, Adams 2003, Fletcher/Huang 2006).
Accumulation of apolipoprotein L1 leads to autophagy (Wan et al. 2008, Zhaorigetu et al. 2008)
and is postulated to be linked to apoptosis (Vanhollebeke/Pays 2006). APOL genes were reported
to be induced in proinflammatory environment (Smith/Malik 2009). Unlike the secreted APOL1,
the other four induced apolipoprotein L genes APOL2, APOL3, APOL4, and APOL6 do not possess
a secretion signal peptide and therefore are thought to be localized intracellularly (Duchateau et
al. 1997, Page et al. 2001). Like APOL1, they also contain BH3-domains and are supposed to be
associated with programmed cell death and immune response (Liu Z et al. 2005).
Most interestingly, a second function of APOL1 as human serum factor which lyses serumsensitive
Trypanosoma species was identified (Vanhamme et al. 2003). After uptake, APOL1
forms pores in the trypanosomal lysosomes, and the subsequent lysosomal disruption leads to
killing of the parasite (Pérez-Morga et al. 2005). The species Trypanosoma brucei rhodesiense is
serum-resistant. Its antagonistic protein SRA targets APOL1, inhibits the lytic function, and thus
helps the pathogen to evade the immune response (Xong et al. 1998, De Greef et al. 1989,
Lecordier et al. 2009).
The APOL1 domain which is responsible for APOL1-SRA binding is called SID, and the other
APOL proteins exhibit homologous regions to this domain. The region was subjected to rapid
evolution in all six APOL proteins, and another region – called MAD – evolved rapidly only in
APOL6. The authors Smith and Malik predict from these results the existence of further unknown
antagonists to these regions, especially in intracellular pathogens, which might take advantage
from inhibition of host cell death (Smith/Malik 2009).
At the 6.5 h post-infection time point, S9 cells induced the expression of several chemokines
and cytokines. This set included CCL2 and CCL5, CXCL10, IFNB1, IL6, IL12A and others, and was
interpreted as a pro-inflammatory response. Literature references also report the induction of
chemokines/cytokines by epithelial cells after a challenge with S. aureus with the aim to recruit
immune cells and activate innate and adaptive immune responses (Moreilhon et al. 2005,
Peterson ML et al. 2005). Also endothelial cells exhibit a similar response to infection with
S. aureus (Matussek et al. 2005). The induced chemokines/cytokines from this study and the
published references overlap partly (e. g. IL6, IL15), but still contain specifically regulated genes.
The experiments were performed with different S. aureus strains. They probably influence the
host gene expression to a different extent depending on their differing repertoire of virulence
factors as it was shown for different S. aureus strains infecting endothelial cells (Grundmeier et
al. 2010), and – more distantly related – also in plasma cytokine profiles of patients suffering
from Gram-positive or Gram-negative sepsis and in microarray data sets of LPS or heat-killed
S. aureus Cowan ex vivo stimulated whole blood samples (Freezor et al. 2003). Possibly also the
different host cell lines have different specificities for the induction of pro-inflammatory
cytokines. Despite the difference, in conformity between the different studies the epithelial cells
were recognition sites for infection and mediators of this information to the immune system.
Since the effect was very explicit they were even termed to be components of the innate immune
system (Peterson ML et al. 2005).
Infected S9 cells effectuate a clearly immune defense associated transcriptomic response to
infection. Interestingly, more than 100 genes of the infection-regulated set in S9 cells were also
194

Maren Depke
Discussion and Conclusions
found in the IFN-γ dependently regulated genes in BALB/c or C57BL/6 macrophages, when the
gene list was translated to murine homologues in Rosetta Resolver software with the help of the
EntrezGene HomoloGene database (http://www.ncbi.nlm.nih.gov/homologene). Some genes
even fit together in their regulation in both studies: S9 cells induced cytokine IL12A and
indoleamine 2,3-dioxygenase 1 (IDO1) upon infection, BMM induced the receptor Il12rb1 and
kynureninase (Kynu) after activation by IFN-γ.
An important mediator of inflammation is prostaglandin E 2 (PGE 2 ) whose production is mainly
mediated by prostaglandin-endoperoxide synthase (PTGS), which generates the inflammation
mediators prostaglandin (PG) G 2 and H 2 from arachidonic acid. From PGH 2 , the other
prostaglandins like PGE 2 are formed by different synthases (Spencer et al. 1998, Steer/Corbett
2003, Park JY et al. 2006). In this study, PTGS2 and PGE 2 receptors PTGER4 and PTGER2 were
induced at least at one analyzed time point. The mechanism of PTGS2 induction involves among
others NFκB (Newton et al. 1997, Lin et al. 2000, Tsatsanis et al. 2006). More specifically, it was
published that lipoteichoic acid (LTA) from S. aureus induced PTGS2 protein in human pulmonary
epithelial A549 cells and that this also led to PGE 2 production to which phospholipase A 2 (PLA2)
contributed (Lin et al. 2001). The authors demonstrate an induction of PTGS2 via a mechanism in
which LTA first activates phosphatidylcholine-phospholipase C or D (PC-PLC, PC-PLD), whose
product diacylglycerol (DAG) subsequently activates protein kinase C (PKC), which finally leads to
the activation of NFκB and NFκB-dependent PTGS2 (Lin et al. 2001).
In S. aureus RN1HG infected S9 cells, the induction of phospholipase A 2 (PLA2G4C),
phospholipase C (PLCB4, PLCG2, PLCH1), protein kinase C µ (PRKD1, PRKD2), and PTGS2 was
observed, which fits very well to the literature data. As it is known that NFκB activation results in
autoregulation and induction of its inhibitors (Sun et al. 1993, Le Bail et al. 1993, Liptay et al.
1994, Eto et al. 2003, Trinh et al. 2008), the induction of NFKBIZ can be regarded as indirect hint
for NFκB activity in infected S9 cells.
Airway epithelial cells have been used to study S. aureus in vitro infection before. In a farreaching
microarray and RT-PCR study, Moreilhon and coworkers analyzed the reaction of the
human airway glandular cell line MM-39 to the S. aureus 8325-4 strain in two settings: First,
diluted bacterial supernatants were added to the host cell culture. In a second experiment, PBSwashed
bacterial cells were used to infect the eukaryotic cell culture (Moreilhon et al. 2005).
Bacteria were cultivated in TSB, and when a concentration of 5E+08 cfu/ml was reached, cells or
supernatants were used for the infection experiments. This concentration corresponds to a time
point during exponential growth. Inoculation was performed with 10 % supernatant for a time
span from 1 h to 24 h or with a MOI of 50 for infection with viable staphylococci for 3 h. Thus,
additional to a different host cell line and S. aureus strain – whose known SigB-negative
phenotype (Kullik/Giachino 1997) probably is the main difference to the phenotypically SigBpositive
strain RN1HG – the experimental setting differed from the one used in this study. This
necessarily has impact on the comparability of the results. Furthermore, the strain 8325-4 was
observed to produce a lipase and protease sensitive lipoprotein inhibitor of internalization into
endothelial cells. Reduced internalization consequently led to a reduction of the endothelial cells’
cytokine production (Yao et al. 2000).
Moreilhon et al. observed distinct gene expression profiles of supernatant and viable
staphylococci treated host cells in which bacterial supernatant led to stronger alterations (higher
number as well as stronger magnitude) than washed viable staphylococci.
195

Maren Depke
Discussion and Conclusions
The comparison between the published bacterial supernatant effects and the infection effects
recorded in this study revealed similar functional aspects, although these functions were not
always represented by the same examples. Nevertheless, both studies confirmed each other: 1)
NFκB action (e. g. IL6, JUNB, NR4A2), 2) inflammation (e. g. LIF), 3) inflammatory mediator-related
genes (e. g. PTGS2 alias COX-2). Inflammation-related signal transduction processes were present
with different members of the same families. Moreilhon et al. found induction of JAK1, STAT1,
STAT3 and SOCS2, whereas this study detected induction of JAK2, STAT2 and SOCS1. Similarly,
Moreilhon et al. described induction of pro-inflammatory chemokines/interleukins CXCL1, CXCL2,
CXCL3, CCL20, IL-1α, IL-1β, IL-8, IL-20 and IL-24, and this study found a different set, namely CCL2,
CCL5, CSF1, CXCL10, CXCL11, CXCL16, IL-7, IL-12, IL-15, IL-28, IL-29, and in addition IFNB. From the
group of toll-like receptors, TLR2 was induced in the experiments of Moreilhon et al. while TLR3
was detected with increased expression in the study described in this thesis.
Also cell cycle/apoptosis related genes were found in both studies, but with different
examples. Moreilhon et al. could determine a necrotic phenotype with a transient apoptosis
phase 8 h to 10 h after host-pathogen interaction. In this study, the final fate of the infected cells
was difficult to judge on, because the functional/physiological measurements have not been
performed yet. Here, only an arrest of growth could be postulated from the transcriptome data.
Nevertheless, this is in agreement with proteome result interpretation of Melanie Gutjahr.
Also another study using the cell line MM-39 and washed S. aureus 8325-4 cells from
stationary growth phase (da Silva et al. 2004) describes the epithelial cell reaction as necrotic cell
death after a phase of apoptosis, where a higher concentration of bacterial cells accelerated the
begin of necrosis. Finally, after 24 h a very strong damage of host cells was observed.
Nevertheless, the authors observed that MM-39 host cells were able to defend themselves in the
early phase of infection by the production of SLPI, secretory leukocyte peptidase inhibitor, which
was exhausted in later phases of infection. It has to be mentioned that extracellular bacteria
were not killed in the experiments of da Silva and coworkers and that the strong replication of
non-internalized and host-cell escaped bacteria in the medium during the infection assay and the
parallel production of toxic bacterial products might have influenced the host cells’ reactions
(da Silva et al. 2004). However, transcriptome data of S9 cells during the analysis window of the
study described in this thesis do not allow confirmation of an active defense strategy using SLPI:
SLPI mRNA was expressed, but at a low level, and differential expression was not observed after
infection. Possibly the bronchial epithelial cell line S9 has a different gene expression repertoire
from that of the airway glandular cell line MM-39.
In summary, the in vitro infection study of human bronchial epithelial S9 cells and S. aureus
RN1HG resulted in a picture of a fast, strong proinflammatory reaction of host cells which
revealed central immune response related processes. In agreement with literature reports, the
data led to the conclusion that epithelial cells act as recognition sites for infection and pass this
information to immune cells. Nevertheless, the comparison with other studies revealed that
especially each host cell – bacterial strain combination, but also the additional experimental
settings provoke a specific aspect of host response which supports and necessitates the
extension of research to further of these combinations. Furthermore, the comparison of
proteome and transcriptome results emphasized that the application of both complementing
approaches is recommended and strongly helps to achieve a comprehensive view of hostpathogen
interactions on molecular level.
196

Maren Depke
Discussion and Conclusions
PATHOGEN GENE EXPRESSION PROFILING
Beside its characteristics as commensal colonizer as well as origin of a variety of infectious
diseases, S. aureus was identified as one of the leading causative organisms of pneumonia
besides Streptococcus pneumonia and Haemophilus influenzae (Goto et al. 2009). The pathogen
is easily transferred to the lung, e. g. by aspiration or medical devices, where it encounters
immune cells as well as cells associated with structural and functional aspects of the lung like
epithelial cells.
In this study, the human bronchial epithelial cell line S9 was applied to study the interactions
of epithelial cells with S. aureus RN1HG. Here, the analysis of the pathogen expression profile
complements the analysis of host cell expression patterns, which has been described before.
Similar as in the other studies described in this thesis, also the internalized staphylococci were
monitored in a combined approach of transcriptome (Maren Depke) and proteome (Sandra
Scharf) analysis. Internalized staphylococci were extracted from their S9 host cells and the
bacterial RNA profile was recorded using a tiling array approach. Bacterial intracellular proteins
were monitored and quantified after stable isotope labeling with amino acids in cell culture
(SILAC), FACS-sorting and mass spectrometric analysis.
The advantage of the experimental settings in this study was the application of complete
bacterial culture with both secreted proteins from supernatant and whole bacterial cells to the
host S9 cell culture. Such experimental design was improved after establishment of an adapted
cell culture medium named pMEM (Schmidt et al. 2010) which allows reproducible bacterial
growth, but avoids the inoculation of host cell culture with highly artificial established bacterial
culture media like TBS. Hence, the study of host-pathogen interactions in the context of all
bacterial factors, membrane-bound and secreted, and additionally without any influence on
bacterial physiology by prolonged handling, centrifugation and washing of bacteria was
performed using exponential growth phase cultures of the rsbU + repaired RN1-derivative strain
S. aureus RN1HG (Herbert et al. 2010).
Additional information on gene expression pattern of S. aureus RN1HG during stationary
growth phase in the pMEM medium was available from a second study, which dealt with the
comparison of gene expression pattern in different growth media. That study was part of an
international co-operation in the settings of the EU-IP-FP6-project BaSysBio (LSHG-CT2006-
037469) consortium. Here, proteome analysis was not included.
Both gene expression studies, of medium comparison and of the infection experiment, applied
custom tiling arrays produced and processed by NimbleGen.
First, the knowledge on the gene expression pattern of stationary growth phase was used to
determine the most suitable control sample group for the infection experiment study.
In stationary growth phase, nutrients become limited after consumption in the exponential
phase of growth, which is the trigger for the so-called stringent response. The stringent response
comprises an adaptation of the bacteria’s metabolism to situations of nutrient limitation or
starvation, which includes induction of stress resistance, decelerated growth and alleviated
metabolism. Many of the necessary changes are mediated by transcriptional variation (Condon et
197

Maren Depke
Discussion and Conclusions
al. 1995, Crosse et al. 2000, Anderson KL et al. 2006, Wolz et al. 2010). In a comparison between
the stationary phase gene expression profile and that of the late serum/CO 2 control (each in
relation to exponential phase growth) a high overlap was recognized. This included reduced
expression of ribosomal protein genes, tRNA synthetase genes, translation elongation factor
genes and a clear trend of induction of the stringent response regulator relA as it was reported in
literature (Anderson KL et al. 2006). Thus, the gene expression in the late serum/CO 2 controls
resembled the stationary phase/stringent response. A corresponding similarity was not observed
for the internalized staphylococci as well as for the early control samples. In conclusion from
these observations, it was decided that the 1 h serum/CO 2 control sample, although not timematched,
was the most appropriate baseline sample for this infection experiment study.
Since the shift of bacterial cells from aerobic agitated culture to infection settings in cell
culture plates and 5 % CO 2 atmosphere probably resulted in a change of oxygen availability, genes
harboring binding sites for Rex, the central anaerobiosis regulator and transcriptional repressor
(Pagels et al. 2010), were analyzed to elucidate whether the changes of oxygen availability lead
to an anaerobiosis gene expression signature (this study). De-repression of Rex-binding site
containing genes was visible during anaerobic incubation, but despite the shift from agitated
culture to cell culture plate, internalized samples did not show the same pattern of anaerobic
gene expression. Contrarily, many of these Rex binding sites containing genes, which were in
trend or significantly induced in anaerobically incubated samples (indicating de-repression after
Rex lost its DNA affinity in anaerobic conditions), were in trend or significantly repressed in
internalized staphylococci in S9 cells (indicating sustained DNA binding activity of Rex). Thus, it
was concluded that internalized staphylococci did not suffer from reduced oxygen availability.
One of the first observations during the analysis of the internalized staphylococcal gene
expression pattern was that the response regulator gene saeR of the two-component system
SaeRS was induced in internalized staphylococci 6.5 h after start of infection and that the sensor
component gene saeS was significantly differentially expressed but only did not exceed a fold
change of 2. This observation led to a more detailed analysis of the SaeRS regulon in internalized
staphylococci.
In 2006, Rogasch and coworkers published a transcriptomic and proteomic characterization
study of the SaeRS regulon (Rogasch et al. 2006). Of the genes defined to be regulated in that
publication, 21 were found to be regulated (mostly increased) in at least one of the two analyzed
time points of internalized staphylococci in the study described in this thesis. These included the
auto-induced saeR, adhesins, serine protease, toxins, immune-evasive genes and others. Thus,
most probably the SaeRS two-component system was activated in internalized staphylococci.
The SaeRS system was furthermore studied by DNA array analysis using a saeS gene
replacement mutant (Sa371ko) and its parental strain, the clinical isolate S. aureus WCUH29
(Liang et al. 2006). Less than 20 genes (positive influence on coa, fnbB, fnb, efb, hla, hlb, hlgC,
saeS, SA1000, which corresponds to SAOUHSC_01110, and others; negative influence on agrA
and a gene coding for a hypothetical protein) were observed to be SaeRS dependently regulated.
In vivo, the saeS mutant strain resulted in less bacterial load in the kidney compared to the wild
type strain in a hematogenous pyelonephritis model of i. v. infected mice (Liang et al. 2006).
Interestingly, S. aureus deficient for SaeS exhibited less adhesion to and less invasion in human
lung epithelial A549 cells and resulted in a reduced rate of host cell apoptosis which was possibly
mediated in parts by reduced expression of hla (Liang et al. 2006). Furthermore, negative
influence on adhesion and internalization was documented for efb and SA1000 replacement
198

Maren Depke
Discussion and Conclusions
mutants. Thus, especially saeRS as regulatory system, but also efb and SA1000 (SAOUHSC_01110)
contribute to adhesion and invasion (Liang et al. 2006). Strikingly, these genes were observed to
be induced (saeR, efb) or in trend induced (saeS; SAOUHSC_01110; data not shown for
SAOUHSC_01110) in internalized staphylococci in the study described in this thesis. Furthermore,
also coa, fnbB, hla, and hlgC, determined as SaeRS-dependent by Liang and colleagues, were
induced in internalized staphylococci in this study. The final conclusion of Liang et al. stated that
“activation of the SaeRS system is required for S. aureus to adhere to and invade epithelial cells”.
Supportingly, such activation was also concluded in this study.
Nevertheless, the activation SaeRS and induction of saeRS might depend on further
experimental aspects which are not defined yet. Infection and gene expression studies by Garzoni
and coworkers did not reveal induction but repression of saeR and saeS in S. aureus 6850 upon
internalization in human lung epithelial A549 cells (Garzoni et al. 2007). Contrarily, upon
phagocytosis by human polymorphonuclear leukocytes the induction of saeR and saeS was
observed in different S. aureus strains (Voyich et al. 2005).
While adherence of sae mutant and wild type to endothelial cells did not differ, the mutant
was less invasive than the wild type (Steinhuber et al. 2003). Further affirmation of the
participation of saeS in infection models was derived from mutagenesis studies in combination
with murine systemic infection. Here, non-functional saeS led to attenuated virulence (Benton et
al. 2004). A different bacterial strain background led to similar observations of reduced virulence
in mice (Rampone et al. 1996). Also Goerke et al. identified SaeRS as important regulator which is
active in vivo, although the deletion mutant did not yield different bacterial densities compared
to the wild type in a device-related infection model (Goerke et al. 2005).
Besides the already mentioned saeR and saeS, three other regulators were observed with
differential expression in internalized staphylococci (sarT, sarU, rot). Since sarT and sarU are less
well characterized and since all these three genes were repressed, the effect of this differential
expression cannot easily be rated in the setting of the in vitro infection model.
Fittingly to the concluded activity of the SaeRS system in internalized staphylococci,
membrane-bound adhesins, which are partly known to be regulated by SaeRS (Liang et al. 2006),
were induced (fnbA, fnbB, clfA, clfB).
Fibronectin binding proteins (FnBPs) are – together with other surface proteins – important
mediators of bacterial adhesion to host cells. They exhibit an even more central position for
internalization processes. FnBPs were partly required for adhesion and mainly required for
internalization into the bovine mammary gland epithelial cell line, MAC-T (Dziewanowska et al.
1999). The mechanism includes binding of FnBP to β 1 integrins via a bridge formed by fibronectin,
but also direct binding of FnBP to host membrane-located Hsp60 (Dziewanowska et al. 2000).
Anyway, another study demonstrated variability of fnb genes between different clinical isolates
as well as variation in the ability to adhere to fibronectin by the different isolates. Strains which
were derived from orthopaedic implant-associated infection showed higher adherence than the
other isolates. Community-acquired invasive disease isolates harbored more often two fnb genes
than carriage isolates, but the number of fnb genes correlated only to a low extent with the
adhesion capacity (Peacock et al. 2000).
Fibronectin-binding protein deficiency reduced the colonization of heart tissue in a rat
endocarditis model indicating a role of FnBP in heart valve infection. Other tissue samples did not
show the same effect (Kuypers/Proctor 1989). In order to avoid compensation of a deleted gene
by other, functionally redundant staphylococcal genes, an overexpression study of clfA and fnbA
199

Maren Depke
Discussion and Conclusions
in Lactococcus lactis, a bacterium with low pathogenic potential, was performed (Que et al.
2001). Possession of ClfA or FnbA rendered L. lactis as adherent to fibrinogen or fibronectin as
the gene-donor S. aureus strain. Furthermore, overexpressing L. lactis strains were able to cause
endocarditis in a rat model with a similar minimal infection dose as S. aureus. The authors stated
that this overexpression experiment proved the involvement of clfA and fnbA in endocarditis with
higher confidence than deletion experiments (Que et al. 2001).
Thus, fnb genes have proven to be relevant for adhesion, invasion and in vivo infection
situations. Surprisingly, the induction of adhesins was observed after internalization of S. aureus
RN1HG into S9 cells (this study). It can be speculated that this induction took place in advance to
react to a possible re-liberation from the host cell e. g. after host cell death.
In internalized staphylococci, also soluble adhesins were induced (eap, coa, vwb, emp, efb).
These are also associated with other functions like immune-evasion. Eap, whose expression is
essentially mediated by sae (Harraghy et al. 2005), is one of the examples for a staphylococcal
protein which is related to a variety of functions (Harraghy et al. 2003). Eap was reported for
example to mediate adhesion to cultured fibroblasts (Hussain et al. 2002), to enhance
internalization into human fetal lung fibroblasts and HACAT keratinocytes (Haggar et al. 2003), to
impair wound healing via inhibition of angiogenesis (Athanasopoulos et al. 2006), to block Ras
activation and signal transduction cascade which leads to anti-angiogenesis (Sobke et al. 2006),
to induce pro-inflammatory IL-6 and TNF-α in human peripheral blood mononuclear cells (Scriba
et al. 2008), but also to act anti-inflammatorily by inhibition of leukocyte recruitment via
blockade of ICAM1 and hence lead to immune evasion (Chavakis et al. 2002, Haggar et al. 2004).
Thus, it was not surprising to find induced expression also after internalization into S9 cells (this
study). Further immune-evasion related genes were increased in S9-internalized S. aureus
RN1HG, e. g. chp, whose gene product CHIPS, chemotaxis inhibitory protein of staphylococci, is
known to block the receptor for the chemotactic complement fragment C5a and for bacterialderived
formylated peptides (de Haas et al. 2004, Murdoch/Finn 2000). Also induced efb, coding
for extracellular fibrinogen-binding protein Efb, interferes with complement action, i. e.
opsonization, by inhibition of C3b-binding to bacterial surfaces (Lee et al. 2004a, 2004b). Thus,
internalized staphylococci induce certain genes which would benefit their survival in an in vivo
situation.
In aerobically living cells, reactive oxygen intermediates like hydroxyl radicals (OH•), hydroxide
(OH – ), superoxide anion (O – 2 ), or hydrogen peroxide (H 2 O 2 ) occur in normal metabolism and may
damage cellular structures like DNA (Fridovich 1978, Imlay/Linn 1988). Phagocytes even produce
high amounts of reactive oxygen intermediates during respiratory burst, e. g. superoxide anion by
their NADPH-oxidase, as part of their defense and killing strategy (Robinson 2009). S. aureus
owns two genes coding for superoxide dismutases: sodA and sodM (Clements et al. 1999,
Valderas/Hart 2001). The enzyme participates in the detoxification of reactive oxygen
intermediates by catalyzing the reaction of two superoxide anion molecules to oxygen and
hydrogen peroxide, of which two molecules are consequently decomposed to oxygen and water
by catalase. The complete enzyme constitutes of 2 homo- or heterodimeric subunits. The gene
sodM is characteristic for S. aureus since coagulase-negative staphylococci like S. epidermidis do
not own this gene or a homologue (Valderas et al. 2002).
In the presence of manganese, the internal (methyl viologen) and external (xanthine/xanthine
oxidase) generation of O –
2 in S. aureus led to increased expression of sodA and sodM,
respectively, but not without the addition of manganese (Karavolos et al. 2003). The observation
200

Maren Depke
Discussion and Conclusions
of increased sodA and repressed sodM in internalized staphylococci (this study) fits to an
independent regulation of both genes. Anyway, the regulation of staphylococcal sod genes has
not been intensively studied until now. The analysis of mutants in the regulator genes sigB, agr,
or sarA did not influence the sodA expression in experiments of Clements et al. in 1999. Use of a
sigB deficient S. aureus strain gave hints for a negative influence of σ B on the expression of sodM
(Karavolos et al. 2003). Contradictory to the results of Clements and colleagues, SarA was
reported as negative regulator of especially sodM but also sodA expression via its activity as
transcriptional repressor since gel-shift analysis and consensus sequence search revealed SarAbinding
to the sod promoter regions (Ballal/Manna 2009). The observations of regulatory
influence do not completely explain the expression pattern in this study, and further attempts of
explanation would be too speculative because of only little existing knowledge on the
internalized gene expression regulation.
Detoxification of reactive oxygen intermediates and correlated gene expression might have
impact on virulence since this detoxification, i. e. evasion of immune defense, can give an
advantage to the bacterial cell. Studies from literature give contradictory evidence. Superoxide
dismutase did not correlate with lethality in mouse infection but catalase was proposed to
enhance virulence using clinical isolates and Wood-46 strain of S. aureus (Mandell 1975).
S. aureus 8325-4 sodA mutant did not possess reduced virulence in comparison to wild type
strains in a mouse abscess model (Clements et al. 1999). S. aureus SH1000 sodA, sodM, and
sodA sodM single and double mutants exhibited reduced virulence in a murine model of
subcutaneous infection (Karavolos et al. 2003). The SigB – phenotype of 8325-4, which has been
reversed in SH1000 (Kullik/Giachino 1997, Horsburgh et al. 2002b), might have contributed to
these observations. The superoxide dismutase gene sodA was reported to be linked to the
bacterial acid-adaptive response (Clements et al. 1999). Such function might provide a link to the
expression pattern in internalized staphylococci, which probably encounter reduced pH in the
phagosome/endo-lysosome.
Internalized staphylococci induced two pairs of bicomponent toxins, lukD/lukE, hlgB/hlgC, and
alpha-hemolysin, hla. The subunits of the bicomponent toxins were described to be
interchangeable and to result in different toxins with distinct properties, which led to the
activation of granulocytes (König et al. 1997). Leukocidins target the membrane and lead to
disruption of host cells for example of human polymorphonuclear neutrophils by PVL (Colin et al.
1994) or of erythrocytes by a LukF/Hlg combination (Nguyen et al. 2003) and cause severe
inflammatory damage in a model of toxin injection into rabbit eyes (Siqueira et al. 1997).
LukD/LukE were first characterized by Gravet et al. in 1998. From a set of 429 S. aureus isolates,
the prevalence of lukD/lukE was determined as 82 % in blood and 60.5 % in nasal isolates,
whereas hlg was detected in almost 100 % of isolates (von Eiff et al 2004). The LukD/LukE
combination was shown to result in dermonecrosis after infection of rabbit skin. The
dermonecrosis inducing potency of LukD or LukE was reduced when combinations with other
toxin subunits were applied. Especially the combinations with HlgB led to reduced activity.
HlgB/HlgC possessed lytic activity for erythrocytes, but the LukD/LukE combination did not lead
to erythrocyte lysis. Erythrocyte lysis could not be achieved by combining LukD or LukE with
other toxin subunits. Finally, human polymorphonuclear leukocytes (granulocytes) could be
activated by HlgB/HlgC, LukE/HlgB, LukD/HlgC, and LukD/LukE which decreased potency in the
given order (Gravet et al. 1998). Thus, the different toxins have a distinct functional bias as well
as potency. The specific effects on S9 cells or in vivo, e. g. in the lung, still need to be determined.
201

Maren Depke
Discussion and Conclusions
Similarly, the expression of extracellular and membrane-bound enzymes, which contribute to
virulence or spreading, was difficult to interpret. Expression of serine proteases (splABCDEF) and
lipase (lip) was increased, whereas the expression of hyaluronate lyase (hysA), a membrane
bound serine protease (htrA), nuclease (nuc), and glutamyl endopeptidase alias V8 peptidase
(sspA) was repressed in internalized staphylococci. Some spl genes were found to be SaeRSregulated
(Rogasch et al. 2006). Thus, the induction of the spl operon fits to the proposed activity
of the SaeRS two-component system.
Gene expression changes which might have influence on cell wall remodeling were also
observed in internalized staphylococci. Surprisingly, this included repression of the dlt operon.
The proteins encoded by dlt participate in incorporation and esterification of D-alanine residues
into the cell wall teichoic acids. In this way, the negative charge of the cell wall structures is
reduced and thus is less attractant for positively charged antimicrobial molecules (Peschel et al.
1999). With the aim to evade the immune response, an induction of dlt expression had been
expected. In parallel, the induction of peptidoglycan hydrolase lytM and staphylococcal secretory
antigen ssaA was recorded (this study), which are also involved in cell wall remodeling (Dubrac et
al. 2007). Repression of dlt and induction of lytM were also observed by Garzoni and coworkers
(Garzoni et al. 2007). Furthermore, the induction of prsA, coding for a foldase involved in
translocation and folding of secreted proteins (Sarvas et al. 2004), was observed in internalized
staphylococci 2.5 h after infection of S9 cells (this study). This regulation was in accordance with
protein abundance data in internalized staphylococci which were determined by Sandra Scharf
(Schmidt et al. 2010). She also found an increased abundance of VraR, regulatory component of
the VraRS two-component system and known to be positively auto-regulated (Kuroda et al.
2003), which led to the examination of vraRS gene expression: vraR (1.8) and vraS (1.6) were not
significantly different in internalized staphylococci (2.5 h) in comparison to control, but the fold
change values indicated a slight trend for induction. Hence, the observed differential expression
of cell wall remodeling related genes was in agreement with the proteome data of Sandra Scharf.
Additionally, genes described to belong to the VraRS regulon (Kuroda et al. 2003) were examined.
Internalized staphylococci induced about 20 % of the published VraRS regulon (Kuroda et al.
2003) at the same time point when a trend of vraRS induction was detected. Since the VraRS
two-component system was suggested to be activated upon cell wall damage or inhibition of cell
wall biosynthesis (Kuroda et al. 2000, Kuroda et al. 2003), the observation of cell wall remodeling
associated gene expression changes might indicate a reaction to stressful influences which
possibly operate on the bacterial cell wall after internalization into the host cell. Another possible
explanation proposed by Garzoni and colleagues (2007) is that the gene induction (e. g. lytM) is
associated with cell division processes.
The gene expression analysis of internalized staphylococci in S9 cells led to the recognition of
changes in the expression of metabolic genes. Purine biosynthesis, tRNA synthetases, and
glycolysis were repressed, whereas gluconeogenesis, TCA cycle enzyme genes, and especially
genes related to amino acid biosynthesis were induced. The induction of amino acid biosynthesis
and TCA cycle genes as well as the repression of purine biosynthesis and tRNA synthetase genes
was clearly confirmed in the proteome analysis of Sandra Scharf. Additionally, transporter genes
were differentially expressed. Induction was observed especially for phosphate transporters, but
also for transporters of phosphonate, methionine, peptide/oligopeptide, sugar/maltose,
osmoprotectants, amino acids, and gluconate. Repression was detected for urea,
202

Maren Depke
Discussion and Conclusions
spermidine/putrescine, sodium/glutamate, arginine/ornithine, glycine betaine, amino acid,
choline, citrate, xanthine, fructose, and lactate transporters.
A good metabolic performance, especially the unaffected ability for amino acid biosynthesis
and for the uptake of compounds, is important for the survival of pathogens and has been
reported to be linked not only to auxotrophies, but also to virulence. In a mutagenesis study,
which was combined with a model of murine systemic infection, 24 genes were identified whose
mutation resulted in attenuation of S. aureus virulence. Of these, 9 genes (38 %) were part of
purine or amino acid biosynthesis, and 6 genes (25 %) coded for membrane transporters (Benton
et al. 2004). Of these virulence associated genes, four genes related to amino acid biosynthesis
(lysC, dapB, trpF, asd) and pstS, phosphate transporter, were induced in internalized
staphylococci at least in one of the two analyzed time points (this study). Additionally, amino acid
biosynthesis genes tyrA and pycA were induced in trend with a significant p-value but a fold
change value of less than the cutoff.
Few gene expression studies of internalized staphylococci are published until now. One of
them has already been cited above in selected details. Garzoni and coworkers analyzed
internalized staphylococci in epithelial cells using a microarray approach (Garzoni et al. 2007).
Although the general setting resembles that applied in this study, important differences in the
experimental procedures as well as in the results can be found between both studies. Most
importantly, staphylococcal strain (S. aureus 6850 vs. RN1HG) as well as host cell line (A549 vs.
S9) differed. Garzoni et al. mention exactly this topic and conclude that “certain combinations of
host cells and bacteria can result in different biological outcomes”, which is clearly supported in
the comparison of their and this study. Furthermore, infection took place in a cell culture
medium without (Garzoni et al. 2007) or with serum supplement (this study), with washed
(Garzoni et al. 2007) or non-washed, exoproteins containing (this study) bacterial suspensions, in
a multiplicity of infection (MOI) of 100 for 30 min (Garzoni et al. 2007) or in a MOI of 25 for 1 h
(this study). Also RNA preparation and sample processing methods differed: depletion of
eukaryotic RNA and amplification step (Garzoni et al. 2007) or utilization of RNA without
depletion and amplification since pure bacterial RNA of high quality was available in sufficient
amounts (this study). Finally, different arrays were used. Both studies approximately match in the
analyzed time points of 2 h/2.5 h and 6 h/6.5 h after infection.
The total numbers of differentially expressed genes were higher in the literature reference
with 1042 (Garzoni et al. 2007) and 565 (this study, annotated genes) for the early time point and
766 (Garzoni et al. 2007) and 489 (this study, annotated genes) for the later time point, and
Garzoni et al. detected a higher fraction of repressed genes whereas this study resulted in a
higher number of induced genes. Similar in both studies, comparisons between non-adherent
and mock infection samples (staphylococci in infection medium) did not result in the detection of
differential gene expression. In a very rough comparison, at least 100 genes were differentially
expressed between internalized and control staphylococci in both studies. Garzoni and
colleagues describe differential expression of genes or functional groups of genes, which were
recognized also in this study, e. g. repression of tRNA synthetase genes, induction of fnbAB,
induction of hlgB and lukE, induction of sodA, repression of sspA, induction of lytM, and induction
of transporters. Different results were revealed, e. g. clfB repressed (Garzoni et al. 2007) or
induced (this study), hla not differentially expressed (Garzoni et al. 2007) or induced (this study),
cap operon repressed (Garzoni et al. 2007) or not differentially expressed (this study), TCA cycle
203

Maren Depke
Discussion and Conclusions
genes repressed (Garzoni et al. 2007) or induced (this study), saeRS repressed (Garzoni et al.
2007) or induced (this study).
Clearly, the comparability is limited due to experimental differences, which underlines the
importance of performing global molecular studies (transcriptome and proteome profiling) in
different models to finally achieve a complete picture of the characteristics of host-pathogen
interactions.
Such different models include e. g. other epithelial cell lines, like the S9 cells used in this study,
but can also be extended to immune cells. Such a study was performed by Voyich and colleagues
in 2005. The authors analyzed the gene expression profile of different S. aureus strains upon
phagocytosis by human neutrophils (Voyich et al. 2005). Again, several experimental parameters
differed between the studies of Voyich and colleagues (2005), Garzoni and colleagues (2007), and
this study, partly resulting from the very different host cell type employed by Voyich and
coworkers. The authors described immune evasion mechanisms of S. aureus, which are initiated
by transcriptional changes. First, the authors observed induction of stress response genes, e. g.
superoxide dismutases sodA and sodM and several other genes involved in counteracting the
influence of reactive oxygen intermediates. An as strong response was not observed in this study,
which can be explained by the innate function of neutrophils to kill pathogens, which is not given
for epithelial cells. Thus, staphylococci internalized in S9 cells probably encounter less harmful
molecules than staphylococci after uptake by neutrophils. The induction of TCA cycle genes after
phagocytosis by neutrophils (Voyich et al. 2005) corresponds with results from this study using S9
cells. Voyich and coworkers observed repression of genes involved in cell envelope synthesis, cell
division, and replication, which was not as obvious in S9 cell internalized staphylococci.
Nevertheless, cell wall remodeling processes were concluded from gene expression data (this
study). Furthermore, staphylococci in neutrophils induced several toxins and adhesins (Voyich et
al. 2005), of which some were also observed after S9 cell internalization, e. g. hlgB, hlgC, lukD,
lukE; fnbB, coa, clfA (this study). The toxin gene induction appeared to be stronger in the
neutrophil than in the S9 study, although hla was induced after internalization in S9 cells and not
after phagocytosis by neutrophils. Anyway, the different staphylococcal strains, which were used
in the studies, contribute considerably to the characteristics of the gene expression signature
(Voyich et al. 2005). In the neutrophil phagocytosis dependent gene expression pattern,
differential expression of regulatory genes was reported (Voyich et al. 2005). These included
increased expression of vraRS, saeRS, and sarA. While induction or a trend of induction of saeRS
and vraRS was also visible in this study, sarA was not differentially expressed when staphylococci
were internalized in S9 cells. Furthermore, repression of agr or sigB was not observed in this
study, but in the study of Voyich and coworkers.
In total, after comparison of the results of Voyich and colleagues (2005), Garzoni and
colleagues (2007), and this study, it becomes clear that similarities in certain aspects do not
necessarily lead to completely consistent results. This probably depends to a large extent on the
combination of host cell and bacterial strain, but also on further experimental conditions which
varied strongly between the studies. Nevertheless, parts of the results were also found in other
studies. Since an admittedly slightly imprecise attempt to specify real internalization specific gene
expression that did not occur in control samples revealed virulence associated genes in the study
described in this thesis, the recorded gene expression pattern is probably physiologically
relevant. Therefore, the gene expression profiling turned out to be an important basic study
which will entail follow-up experiments like studies of bacterial mutants, comparison of different
204

Maren Depke
Discussion and Conclusions
host cell lines, but also confirmation and validation experiments by application of different
techniques like quantitative real-time RT-PCR. Of course, the broadening of experiments to
in vivo studies would be most interesting. Here, the most prominent problems are difficulties in
recovery of bacterial cells from tissue and the resulting low amount of sample or the sample
contamination with host material. In this setting, application of quantitative real-time RT-PCR
probably will be the method of choice, which has already proven to successfully help resolving
colonization expression patterns (Burian et al. 2010a, 2010b).
Until now, the tiling array data have been used as expression data set under restriction to
already annotated genomic content of S. aureus. The analysis is not yet finished. Although new
transcribed units have been defined for which interesting regulation patterns in the infection
experiments were visible, the sequences will have to be categorized and validated.
Internalization dependently regulated sequences later will be targets for further characterization
experiments. It will be an interesting task to compare the resulting information on new genes
with that generated with the competing approach of high-throughput transcriptome sequencing
(Beaume et al 2010a, 2010b).
205

Maren Depke
R E F E R E N C E S
Adams JM.
Ways of dying: multiple pathways to apoptosis. Genes Dev. 2003 Oct 15;17(20):2481-95.
Aikawa T, Matsutaka H, Takezawa K, Ishikawa E.
Gluconeogenesis and amino acid metabolism. I. Comparison of various precursors for hepatic gluconeogenesis
in vivo. Biochim Biophys Acta. 1972 Sep 15;279(2):234-44.
Akhtar MK, Kelly SL, Kaderbhai MA.
Cytochrome b 5 modulation of 17α hydroxylase and 17-20 lyase (CYP17) activities in steroidogenesis. J Endocrinol.
2005 Nov;187(2):267-74.
Aki M, Shimbara N, Takashina M, Akiyama K, Kagawa S, Tamura T, Tanahashi N, Yoshimura T, Tanaka K, Ichihara A.
Interferon-γ induces different subunit organizations and functional diversity of proteasomes. J Biochem. 1994 Feb;
115(2):257-69.
Alberati-Giani D, Ricciardi-Castagnoli P, Köhler C, Cesura AM.
Regulation of the kynurenine metabolic pathway by interferon-γ in murine cloned macrophages and microglial cells.
J Neurochem. 1996 Mar;66(3):996-1004.
Alberati-Giani D, Malherbe P, Ricciardi-Castagnoli P, Köhler C, Denis-Donini S, Cesura AM.
Differential regulation of indoleamine 2,3-dioxygenase expression by nitric oxide and inflammatory mediators in
IFN-γ-activated murine macrophages and microglial cells. J Immunol. 1997 Jul 1;159(1):419-26.
Alberda C, Graf A, McCargar L.
Malnutrition: etiology, consequences, and assessment of a patient at risk. Best Pract Res Clin Gastroenterol.
2006;20(3):419-39.
Alberti KG, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato KA, Fruchart JC, James WP, Loria CM, Smith SC Jr;
International Diabetes Federation Task Force on Epidemiology and Prevention; Hational Heart, Lung, and Blood
Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; International
Association for the Study of Obesity.
Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force
on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World
Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity.
Circulation. 2009 Oct 20;120(16):1640-5. Epub 2009 Oct 5.
Almeida RA, Matthews KR, Cifrian E, Guidry AJ, Oliver SP.
Staphylococcus aureus invasion of bovine mammary epithelial cells. J Dairy Sci. 1996 Jun;79(6):1021-6.
Amasheh S, Meiri N, Gitter AH, Schöneberg T, Mankertz J, Schulzke JD, Fromm M.
Claudin-2 expression induces cation-selective channels in tight junctions of epithelial cells. J Cell Sci. 2002 Dec 15;
115(Pt 24):4969-76.
Andersen MH, Schrama D, thor Straten P, Becker JC.
Cytotoxic T cells. J Invest Dermatol. 2006 Jan;126(1):32-41.
Anderson KL, Roberts C, Disz T, Vonstein V, Hwang K, Overbeek R, Olson PD, Projan SJ, Dunman PM.
Characterization of the Staphylococcus aureus heat shock, cold shock, stringent, and SOS responses and their effects
on log-phase mRNA turnover. J Bacteriol. 2006 Oct;188(19):6739-56.
Anderson SL, Carton JM, Lou J, Xing L, Rubin BY.
Interferon-induced guanylate binding protein-1 (GBP-1) mediates an antiviral effect against vesicular stomatitis virus
and encephalomyocarditis virus. Virology. 1999 Mar 30;256(1):8-14.
Arnalich F, López J, Codoceo R, Jim nez M, Madero R, Montiel C.
Relationship of plasma leptin to plasma cytokines and human survivalin sepsis and septic shock. J Infect Dis. 1999
Sep;180(3):908-11.
207

Maren Depke
References
Athanasopoulos AN, Economopoulou M, Orlova VV, Sobke A, Schneider D, Weber H, Augustin HG, Eming SA,
Schubert U, Linn T, Nawroth PP, Hussain M, Hammes HP, Herrmann M, Preissner KT, Chavakis T.
The extracellular adherence protein (Eap) of Staphylococcus aureus inhibits wound healing by interfering with host
defense and repair mechanisms. Blood. 2006 Apr 1;107(7):2720-7. Epub 2005 Nov 29.
Autenrieth IB, Beer M, Bohn E, Kaufmann SH, Heesemann J.
Immune responses to Yersinia enterocolitica in susceptible BALB/c and resistant C57BL/6 mice: an essential role for
gamma interferon. Infect Immun. 1994 Jun;62(6):2590-9.
Ayroldi E, Zollo O, Bastianelli A, Marchetti C, Agostini M, Di Virgilio R, Riccardi C.
GILZ mediates the antiproliferative activity of glucocorticoids by negative regulation of Ras signaling. J Clin Invest.
2007 Jun;117(6):1605-15. Epub 2007 May 10.
Bagby GJ, Lang CH, Skrepnik N, Spitzer JJ.
Attenuation of glucose metabolic changes resulting from TNF-α administration by adrenergic blockade. Am J
Physiol. 1992 Apr;262(4 Pt 2):R628-35.
Ballal A, Manna AC.
Regulation of superoxide dismutase (sod) genes by SarA in Staphylococcus aureus. J Bacteriol. 2009 May;191(10):
3301-10. Epub 2009 Mar 13.
Banerjee RR, Rangwala SM, Shapiro JS, Rich AS, Rhoades B, Qi Y, Wang J, Rajala MW, Pocai A, Scherer PE, Steppan CM,
Ahima RS, Obici S, Rossetti L, Lazar MA.
Regulation of fasted blood glucose by resistin. Science. 2004 Feb 20;303(5661):1195-8.
Bange FC, Flohr T, Buwitt U, Böttger EC.
An interferon-induced protein with release factor activity is a tryptophanyl-tRNA synthetase. FEBS Lett. 1992 Mar
30;300(2):162-6.
Barter PJ, Nicholls S, Rye KA, Anantharamaiah GM, Navab M, Fogelman AM.
Antiinflammatory properties of HDL. Circ Res. 2004 Oct 15;95(8):764-72.
Bartolomucci A.
Social stress, immune functions and disease in rodents. Front Neuroendocrinol. 2007 Apr;28(1):28-49. Epub 2007
Feb 16.
Barton LF, Cruz M, Rangwala R, Deepe GS Jr, Monaco JJ.
Regulation of immunoproteasome subunit expression in vivo following pathogenic fungal infection. J Immunol. 2002
Sep 15;169(6):3046-52.
Bateman BT, Donegan NP, Jarry TM, Palma M, Cheung AL.
Evaluation of a tetracycline-inducible promoter in Staphylococcus aureus in vitro and in vivo and its application in
demonstrating the role of sigB in microcolony formation. Infect Immun. 2001 Dec;69(12):7851-7.
Baughman G, Wiederrecht GJ, Campbell NF, Martin MM, Bourgeois S.
FKBP51, a novel T-cell-specific immunophilin capable of calcineurin inhibition. Mol Cell Biol. 1995 Aug;15(8):4395-
402.
Beale RJ, Bryg DJ, Bihari DJ.
Immunonutrition in the critically ill: a systematic review of clinical outcome. Crit Care Med. 1999 Dec;27(12):2799-
805.
Beaume M, Hernandez D, Francois P, Schrenzel J.
New approaches for functional genomic studies in staphylococci. Int J Med Microbiol. 2010 Feb;300(2-3):88-97.
Epub 2009 Dec 14.
Beaume M, Hernandez D, Farinelli L, Deluen C, Linder P, Gaspin C, Romby P, Schrenzel J, Francois P.
Cartography of methicillin-resistant S. aureus transcripts: detection, orientation and temporal expression during
growth phase and stress conditions. PLoS One. 2010 May 20;5(5):e10725.
Bekker LG, Freeman S, Murray PJ, Ryffel B, Kaplan G.
TNF-α controls intracellular mycobacterial growth by both inducible nitric oxide synthase-dependent and inducible
nitric oxide synthase-independent pathways. J Immunol. 2001 Jun 1;166(11):6728-34.
Benoist C, Mathis D.
Regulation of major histocompatibility complex class-II genes: X, Y and other letters of the alphabet. Annu Rev
Immunol. 1990;8:681-715.
Benton BM, Zhang JP, Bond S, Pope C, Christian T, Lee L, Winterberg KM, Schmid MB, Buysse JM.
Large-scale identification of genes required for full virulence of Staphylococcus aureus. J Bacteriol. 2004 Dec;
186(24):8478-89.
208

Maren Depke
References
Bera A, Herbert S, Jakob A, Vollmer W, Götz F.
Why are pathogenic staphylococci so lysozyme resistant? The peptidoglycan O-acetyltransferase OatA is the major
determinant for lysozyme resistance of Staphylococcus aureus. Mol Microbiol. 2005 Feb;55(3):778-87.
Berrebi D, Bruscoli S, Cohen N, Foussat A, Migliorati G, Bouchet-Delbos L, Maillot MC, Portier A, Couderc J,
Galanaud P, Peuchmaur M, Riccardi C, Emilie D.
Synthesis of glucocorticoid-induced leucine zipper (GILZ) by macrophages: an anti-inflammatory and
immunosuppressive mechanism shared by glucocorticoids and IL-10. Blood. 2003 Jan 15;101(2):729-38. Epub 2002
Sep 12.
Beutler B, Milsark IW, Cerami AC.
Passive immunization against cachectin/tumor necrosis factor protects mice from lethal effect of endotoxin.
Science. 1985 Aug 30;229(4716):869-71.
reprint in: Classical article. J Immunol. 2008 Jul 1;181(1):7-9.
Bischoff M, Dunman P, Kormanec J, Macapagal D, Murphy E, Mounts W, Berger-Bächi B, Projan S.
Microarray-based analysis of the Staphylococcus aureus σ B regulon. J Bacteriol. 2004 Jul;186(13):4085-99.
Bjerketorp J, Nilsson M, Ljungh A, Flock JI, Jacobsson K, Frykberg L.
A novel von Willebrand factor binding protein expressed by Staphylococcus aureus. Microbiology. 2002 Jul;
148(Pt 7):2037-44.
Blasi F, Sidenius N.
The urokinase receptor: focused cell surface proteolysis, cell adhesion and signaling. FEBS Lett. 2010 May 3;584(9):
1923-30. Epub 2009 Dec 27.
Boasso A, Herbeuval JP, Hardy AW, Winkler C, Shearer GM.
Regulation of indoleamine 2,3-dioxygenase and tryptophanyl-tRNA-synthetase by CTLA-4-Fc in human CD4 + T cells.
Blood. 2005 Feb 15;105(4):1574-81. Epub 2004 Oct 5.
Bodén MK, Flock JI.
Fibrinogen-binding protein/clumping factor from Staphylococcus aureus. Infect Immun. 1989 Aug;57(8):2358-63.
Boehm U, Klamp T, Groot M, Howard JC.
Cellular responses to interferon-γ. Annu Rev Immunol. 1997;15:749-95.
Bokarewa M, Tarkowski A.
Human alpha-defensins neutralize fibrinolytic activity exerted by staphylokinase. Thromb Haemost. 2004 May;
91(5):991-9.
Bonventre JV.
Phospholipase A2 and signal transduction. J Am Soc Nephrol. 1992 Aug;3(2):128-50.
Bouillet P, Strasser A.
BH3-only proteins - evolutionarily conserved proapoptotic Bcl-2 family members essential for initiating programmed
cell death. J Cell Sci. 2002 Apr 15;115(Pt 8):1567-74.
Breitbach K, Klocke S, Tschernig T, van Rooijen N, Baumann U, Steinmetz I.
Role of inducible nitric oxide synthase and NADPH oxidase in early control of Burkholderia pseudomallei infection in
mice. Infect Immun. 2006 Nov;74(11):6300-9. Epub 2006 Sep 25.
Brosnan JT.
Glutamate, at the interface between amino acid and carbohydrate metabolism. J Nutr. 2000 Apr;130(4S
Suppl):988S-90S.
Brucet M, Marqués L, Sebastián C, Lloberas J, Celada A.
Regulation of murine Tap1 and Lmp2 genes in macrophages by interferon gamma is mediated by STAT1 and IRF-1.
Genes Immun. 2004 Jan;5(1):26-35.
Burian M, Wolz C, Goerke C.
Regulatory adaptation of Staphylococcus aureus during nasal colonization of humans. PLoS One. 2010 Apr 6;
5(4):e10040.
Burian M, Rautenberg M, Kohler T, Fritz M, Krismer B, Unger C, Hoffmann WH, Peschel A, Wolz C, Goerke C.
Temporal expression of adhesion factors and activity of global regulators during establishment of Staphylococcus
aureus nasal colonization. J Infect Dis. 2010 May 1;201(9):1414-21.
Buwitt U, Flohr T, Böttger EC.
Molecular cloning and characterization of an interferon induced human cDNA with sequence homology to a
mammalian peptide chain release factor. EMBO J. 1992 Feb;11(2):489-96.
209

Maren Depke
References
Byrne GI, Lehmann LK, Landry GJ.
Induction of tryptophan catabolism is the mechanism for gamma-interferon-mediated inhibition of intracellular
Chlamydia psittaci replication in T24 cells. Infect Immun. 1986 Aug;53(2):347-51.
Campbell SJ, Hughes PM, Iredale JP, Wilcockson DC, Waters S, Docagne F, Perry VH, Anthony DC.
CINC-1 is an acute-phase protein induced by focal brain injury causing leukocyte mobilization and liver injury.
FASEB J. 2003 Jun;17(9):1168-70. Epub 2003 Apr 22.
Capes SE, Hunt D, Malmberg K, Gerstein HC.
Stress hyperglycaemia and increased risk of death after myocardial infarction in patients with and without diabetes:
a systematic overview. Lancet. 2000 Mar 4;355(9206):773-8.
Carlin JM, Borden EC, Sondel PM, Byrne GI.
Interferon-induced indoleamine 2,3-dioxygenase activity in human mononuclear phagocytes. J Leukoc Biol. 1989
Jan;45(1):29-34.
Carter CC, Gorbacheva VY, Vestal DJ.
Inhibition of VSV and EMCV replication by the interferon-induced GTPase, mGBP-2: differential requirement for
wild-type GTP binding domain. Arch Virol. 2005 Jun;150(6):1213-20. Epub 2005 Feb 18.
Casey R, Newcombe J, McFadden J, Bodman-Smith KB.
The acute-phase reactant C-reactive protein binds to phosphorylcholine-expressing Neisseria meningitidis and
increases uptake by human phagocytes. Infect Immun. 2008 Mar;76(3):1298-304. Epub 2008 Jan 14.
Chambers HF.
Methicillin resistance in staphylococci: molecular and biochemical basis and clinical implications. Clin Microbiol Rev.
1997 Oct;10(4):781-91.
Chan PF, Foster SJ, Ingham E, Clements MO.
The Staphylococcus aureus alternative sigma factor σ B controls the environmental stress response but not starvation
survival or pathogenicity in a mouse abscess model. J Bacteriol. 1998 Dec;180(23):6082-9.
Chang SC, Momburg F, Bhutani N, Goldberg AL.
The ER aminopeptidase, ERAP1, trims precursors to lengths of MHC class I peptides by a "molecular ruler"
mechanism. Proc Natl Acad Sci U S A. 2005 Nov 22;102(47):17107-12. Epub 2005 Nov 14.
Chatterjee-Kishore M, Kishore R, Hicklin DJ, Marincola FM, Ferrone S.
Different requirements for signal transducer and activator of transcription 1α and interferon regulatory factor 1 in
the regulation of low molecular mass polypeptide 2 and transporter associated with antigen processing 1 gene
expression. J Biol Chem. 1998 Jun 26;273(26):16177-83.
Chavakis T, Hussain M, Kanse SM, Peters G, Bretzel RG, Flock JI, Herrmann M, Preissner KT.
Staphylococcus aureus extracellular adherence protein serves as anti-inflammatory factor by inhibiting the
recruitment of host leukocytes. Nat Med. 2002 Jul;8(7):687-93. Epub 2002 Jun 24.
Chen G, Goeddel DV.
TNF-R1 signaling: a beautiful pathway. Science. 2002 May 31;296(5573):1634-5.
Chen W, Suruga K, Nishimura N, Gouda T, Lam VN, Yokogoshi H.
Comparative regulation of major enzymes in the bile acid biosynthesis pathway by cholesterol, cholate and taurine
in mice and rats. Life Sci. 2005 Jul 1;77(7):746-57. Epub 2005 Mar 19.
Cheung AL, Bayer AS, Zhang G, Gresham H, Xiong YQ.
Regulation of virulence determinants in vitro and in vivo in Staphylococcus aureus. FEMS Immunol Med Microbiol.
2004 Jan 15;40(1):1-9.
Chida Y, Sudo N, Kubo C.
Does stress exacerbate liver diseases? J Gastroenterol Hepatol. 2006 Jan;21(1 Pt 2):202-8.
Chien Y, Manna AC, Projan SJ, Cheung AL.
SarA, a global regulator of virulence determinants in Staphylococcus aureus, binds to a conserved motif essential for
sar-dependent gene regulation. J Biol Chem. 1999 Dec 24;274(52):37169-76.
Clements MO, Watson SP, Foster SJ.
Characterization of the major superoxide dismutase of Staphylococcus aureus and its role in starvation survival,
stress resistance, and pathogenicity. J Bacteriol. 1999 Jul;181(13):3898-903.
Cohen N, Mouly E, Hamdi H, Maillot MC, Pallardy M, Godot V, Capel F, Balian A, Naveau S, Galanaud P, Lemoine FM,
Emilie D.
GILZ expression in human dendritic cells redirects their maturation and prevents antigen-specific T lymphocyte
response. Blood. 2006 Mar 1;107(5):2037-44. Epub 2005 Nov 17.
210

Maren Depke
References
Cole AM, Ganz T, Liese AM, Burdick MD, Liu L, Strieter RM.
Cutting edge: IFN-inducible ELR – CXC chemokines display defensin-like antimicrobial activity. J Immunol. 2001 Jul 15;
167(2):623-7.
Cole KE, Strick CA, Paradis TJ, Ogborne KT, Loetscher M, Gladue RP, Lin W, Boyd JG, Moser B, Wood DE, Sahagan BG,
Neote K.
Interferon-inducible T cell alpha chemoattractant (I-TAC): a novel non-ELR CXC chemokine with potent activity on
activated T cells through selective high affinity binding to CXCR3. J Exp Med. 1998 Jun 15;187(12):2009-21.
Colin DA, Mazurier I, Sire S, Finck-Barbançon V.
Interaction of the two components of leukocidin from Staphylococcus aureus with human polymorphonuclear
leukocyte membranes: sequential binding and subsequent activation. Infect Immun. 1994 Aug;62(8):3184-8.
Condon C, Squires C, Squires CL.
Control of rRNA transcription in Escherichia coli. Microbiol Rev. 1995 Dec;59(4):623-45.
Cook-Mills JM.
VCAM-1 signals during lymphocyte migration: role of reactive oxygen species. Mol Immunol. 2002 Dec;39(9):499-
508.
Costelli P, Carbó N, Tessitore L, Bagby GJ, Lopez-Soriano FJ, Argilés JM, Baccino FM.
Tumor necrosis factor-α mediates changes in tissue protein turnover in a rat cancer cachexia model. J Clin Invest.
1993 Dec;92(6):2783-9.
Crosse AM, Greenway DL, England RR.
Accumulation of ppGpp and ppGp in Staphylococcus aureus 8325-4 following nutrient starvation. Lett Appl
Microbiol. 2000 Oct;31(4):332-7.
Cubellis MV, Wun TC, Blasi F.
Receptor-mediated internalization and degradation of urokinase is caused by its specific inhibitor PAI-1. EMBO J.
1990 Apr;9(4):1079-85.
da Silva MC, Zahm JM, Gras D, Bajolet O, Abely M, Hinnrasky J, Milliot M, de Assis MC, Hologne C, Bonnet N,
Merten M, Plotkowski MC, Puchelle E.
Dynamic interaction between airway epithelial cells and Staphylococcus aureus. Am J Physiol Lung Cell Mol Physiol.
2004 Sep;287(3):L543-51. Epub 2004 May 14.
D'Adamio F, Zollo O, Moraca R, Ayroldi E, Bruscoli S, Bartoli A, Cannarile L, Migliorati G, Riccardi C.
A new dexamethasone-induced gene of the leucine zipper family protects T lymphocytes from TCR/CD3-activated
cell death. Immunity. 1997 Dec;7(6):803-12.
Dallman MF.
Modulation of stress responses: how we cope with excess glucocorticoids. Exp Neurol. 2007 Aug;206(2):179-82.
Epub 2007 Jun 18.
Dallman MF, Warne JP, Foster MT, Pecoraro NC.
Glucocorticoids and insulin both modulate caloric intake through actions on the brain. J Physiol. 2007 Sep 1;
583(Pt2):431-6. Epub 2007 Jun 7.
Dalton DK, Pitts-Meek S, Keshav S, Figari IS, Bradley A, Stewart TA.
Multiple defects of immune cell function in mice with disrupted interferon-gamma genes. Science. 1993 Mar 19;
259(5102):1739-42.
Dass K, Ahmad A, Azmi AS, Sarkar SH, Sarkar FH.
Evolving role of uPA/uPAR system in human cancers. Cancer Treat Rev. 2008 Apr;34(2):122-36. Epub 2007 Dec 26.
De Greef C, Imberechts H, Matthyssens G, Van Meirvenne N, Hamers R.
A gene expressed only in serum-resistant variants of Trypanosoma brucei rhodesiense. Mol Biochem Parasitol. 1989
Sep;36(2):169-76.
de Haas CJ, Veldkamp KE, Peschel A, Weerkamp F, Van Wamel WJ, Heezius EC, Poppelier MJ, Van Kessel KP,
van Strijp JA.
Chemotaxis inhibitory protein of Staphylococcus aureus, a bacterial antiinflammatory agent. J Exp Med. 2004 Mar 1;
199(5):687-95.
Degrandi D, Konermann C, Beuter-Gunia C, Kresse A, Würthner J, Kurig S, Beer S, Pfeffer K.
Extensive characterization of IFN-induced GTPases mGBP1 to mGBP10 involved in host defense. J Immunol. 2007
Dec 1;179(11):7729-40.
DeLeo FR, Allen LA, Apicella M, Nauseef WM.
NADPH oxidase activation and assembly during phagocytosis. J Immunol. 1999 Dec 15;163(12):6732-40.
211

Maren Depke
References
Demling R.
The use of anabolic agents in catabolic states. J Burns Wounds. 2007 Feb 12;6:e2.
Deora R, Misra TK.
Purification and characterization of DNA dependent RNA polymerase from Staphylococcus aureus. Biochem Biophys
Res Commun. 1995 Mar 17;208(2):610-6.
Depke M, Fusch G, Domanska G, Geffers R, Völker U, Schuett C, Kiank C.
Hypermetabolic syndrome as a consequence of repeated psychological stress in mice. Endocrinology. 2008 Jun;
149(6):2714-23. Epub 2008 Mar 6.
Depke M, Steil L, Domanska G, Völker U, Schütt C, Kiank C.
Altered hepatic mRNA expression of immune response and apoptosis-associated genes after acute and chronic
psychological stress in mice. Mol Immunol. 2009 Sep;46(15):3018-28. Epub 2009 Jul 9.
Deurenberg RH, Stobberingh EE.
The evolution of Staphylococcus aureus. Infect Genet Evol. 2008 Dec;8(6):747-63. Epub 2008 Jul 29.
Diekema DJ, Pfaller MA, Schmitz FJ, Smayevsky J, Bell J, Jones RN, Beach M; SENTRY Partcipants Group.
Survey of infections due to Staphylococcus species: frequency of occurrence and antimicrobial susceptibility of
isolates collected in the United States, Canada, Latin America, Europe, and the Western Pacific region for the
SENTRY Antimicrobial Surveillance Program, 1997-1999. Clin Infect Dis. 2001 May 15;32 Suppl 2:S114-32.
Dietrich CG, Geier A, Oude Elferink RP.
ABC of oral bioavailability: transporters as gatekeepers in the gut. Gut. 2003 Dec;52(12):1788-95.
Döffinger R, Dupuis S, Picard C, Fieschi C, Feinberg J, Barcenas-Morales G, Casanova JL.
Inherited disorders of IL-12- and IFNγ-mediated immunity: a molecular genetics update. Mol Immunol. 2002 May;
38(12-13):903-9.
Dougall WC, Nick HS.
Manganese superoxide dismutase: a hepatic acute phase protein regulated by interleukin-6 and glucocorticoids.
Endocrinology. 1991 Nov;129(5):2376-84.
Dubrac S, Boneca IG, Poupel O, Msadek T.
New insights into the WalK/WalR (YycG/YycF) essential signal transduction pathway reveal a major role in
controlling cell wall metabolism and biofilm formation in Staphylococcus aureus. J Bacteriol. 2007 Nov;189(22):
8257-69. Epub 2007 Sep 7.
Duchateau PN, Pullinger CR, Orellana RE, Kunitake ST, Naya-Vigne J, O'Connor PM, Malloy MJ, Kane JP.
Apolipoprotein L, a new human high density lipoprotein apolipoprotein expressed by the pancreas. Identification,
cloning, characterization, and plasma distribution of apolipoprotein L. J Biol Chem. 1997 Oct 10;272(41):25576-82.
Duffield JS.
The inflammatory macrophage: a story of Jekyll and Hyde. Clin Sci (Lond). 2003 Jan;104(1):27-38.
Dunkelberger JR, Song WC.
Complement and its role in innate and adaptive immune responses. Cell Res. 2010 Jan;20(1):34-50. Epub 2009 Dec
15.
Dziewanowska K, Patti JM, Deobald CF, Bayles KW, Trumble WR, Bohach GA.
Fibronectin binding protein and host cell tyrosine kinase are required for internalization of Staphylococcus aureus by
epithelial cells. Infect Immun. 1999 Sep;67(9):4673-8.
Dziewanowska K, Carson AR, Patti JM, Deobald CF, Bayles KW, Bohach GA.
Staphylococcal fibronectin binding protein interacts with heat shock protein 60 and integrins: role in internalization
by epithelial cells. Infect Immun. 2000 Nov;68(11):6321-8.
Ehrt S, Schnappinger D, Bekiranov S, Drenkow J, Shi S, Gingeras TR, Gaasterland T, Schoolnik G, Nathan C.
Reprogramming of the macrophage transcriptome in response to interferon-γ and Mycobacterium tuberculosis:
signaling roles of nitric oxide synthase-2 and phagocyte oxidase. J Exp Med. 2001 Oct 15;194(8):1123-40.
Elenkov IJ.
Glucocorticoids and the Th1/Th2 balance. Ann N Y Acad Sci. 2004 Jun;1024:138-46.
Emonts M, de Jongh CE, Houwing-Duistermaat JJ, van Leeuwen WB, de Groot R, Verbrugh HA, Hermans PW,
van Belkum A.
Association between nasal carriage of Staphylococcus aureus and the human complement cascade activator serine
protease C1 inhibitor (C1INH) valine vs. methionine polymorphism at amino acid position 480. FEMS Immunol Med
Microbiol. 2007 Aug;50(3):330-2. Epub 2007 May 10.
212

Maren Depke
References
Eriksen NH, Espersen F, Rosdahl VT, Jensen K.
Carriage of Staphylococcus aureus among 104 healthy persons during a 19-month period. Epidemiol Infect. 1995
Aug;115(1):51-60.
Ermak G, Davies KJ.
Calcium and oxidative stress: from cell signaling to cell death. Mol Immunol. 2002 Feb;38(10):713-21.
Erwig LP, Kluth DC, Walsh GM, Rees AJ.
Initial cytokine exposure determines function of macrophages and renders them unresponsive to other cytokines.
J Immunol. 1998 Aug 15;161(4):1983-8.
Eske K, Breitbach K, Köhler J, Wongprompitak P, Steinmetz I.
Generation of murine bone marrow derived macrophages in a standardised serum-free cell culture system.
J Immunol Methods. 2009 Mar 15;342(1-2):13-9. Epub 2009 Jan 6.
Esmon CT.
Inflammation and the activated protein C anticoagulant pathway. Semin Thromb Hemost. 2006 Apr;32 Suppl 1:49-
60.
Eto A, Muta T, Yamazaki S, Takeshige K.
Essential roles for NF-κB and a Toll/IL-1 receptor domain-specific signal(s) in the induction of IκB-ζ. Biochem Biophys
Res Commun. 2003 Feb 7;301(2):495-501.
Everson CA, Reed HL.
Pituitary and peripheral thyroid hormone responses to thyrotropin-releasing hormone during sustained sleep
deprivation in freely moving rats. Endocrinology. 1995 Apr;136(4):1426-34.
Farber JM.
A macrophage mRNA selectively induced by γ-interferon encodes a member of the platelet factor 4 family of
cytokines. Proc Natl Acad Sci U S A. 1990 Jul;87(14):5238-42.
Fausto N, Campbell JS, Riehle KJ.
Liver regeneration. Hepatology. 2006 Feb;43(2 Suppl 1):S45-53.
Favorova OO, Zargarova TA, Rukosuyev VS, Beresten SF, Kisselev LL.
Molecular and cellular studies of tryptophanyl-tRNA synthetases using monoclonal antibodies. Remarkable
variations in the content of tryptophanyl-tRNA synthetase in the pancreas of different mammals. Eur J Biochem.
1989 Oct 1;184(3):583-8.
Feezor RJ, Oberholzer C, Baker HV, Novick D, Rubinstein M, Moldawer LL, Pribble J, Souza S, Dinarello CA, Ertel W,
Oberholzer A.
Molecular characterization of the acute inflammatory response to infections with Gram-negative versus Grampositive
bacteria. Infect Immun. 2003 Oct;71(10):5803-13.
Finck BN, Kelley KW, Dantzer R, Johnson RW.
In vivo and in vitro evidence for the involvement of tumor necrosis factor-α in the induction of leptin by
lipopolysaccharide. Endocrinology. 1998 May;139(5):2278-83.
Finck BN, Johnson RW.
Tumor necrosis factor (TNF)-α induces leptin production through the p55 TNF receptor. Am J Physiol Regul Integr
Comp Physiol. 2000 Feb;278(2):R537-43.
Finlay BB, Cossart P.
Exploitation of mammalian host cell functions by bacterial pathogens. Science. 1997 May 2;276(5313):718-25.
Fitzpatrick F, Humphreys H, O'Gara JP.
The genetics of staphylococcal biofilm formation – will a greater understanding of pathogenesis lead to better
management of device-related infection? Clin Microbiol Infect. 2005 Dec;11(12):967-73.
Fleckner J, Rasmussen HH, Justesen J.
Human interferon gamma potently induces the synthesis of a 55-kDa protein (gamma 2) highly homologous to
rabbit peptide chain release factor and bovine tryptophanyl-tRNA synthetase. Proc Natl Acad Sci U S A. 1991 Dec 15;
88(24):11520-4.
Fletcher JI, Huang DC.
BH3-only proteins: orchestrating cell death. Cell Death Differ. 2006 Aug;13(8):1268-71. Epub 2006 Jun 9.
Fleury C, Neverova M, Collins S, Raimbault S, Champigny O, Levi-Meyrueis C, Bouillaud F, Seldin MF, Surwit RS,
Ricquier D, Warden CH.
Uncoupling protein-2: a novel gene linked to obesity and hyperinsulinemia. Nat Genet. 1997 Mar;15(3):269-72.
213

Maren Depke
References
Foss GS, Prydz H.
Interferon regulatory factor 1 mediates the interferon-γ induction of the human immunoproteasome subunit
multicatalytic endopeptidase complex-like 1. J Biol Chem. 1999 Dec 3;274(49):35196-202.
Foster TJ.
Immune evasion by staphylococci. Nat Rev Microbiol. 2005 Dec;3(12):948-58.
Foster TJ.
Colonization and infection of the human host by staphylococci: adhesion, survival and immune evasion. Vet
Dermatol. 2009 Oct;20(5-6):456-70.
Foster TJ, Höök M.
Surface protein adhesins of Staphylococcus aureus. Trends Microbiol. 1998 Dec;6(12):484-8.
Fraser JD, Proft T.
The bacterial superantigen and superantigen-like proteins. Immunol Rev. 2008 Oct;225:226-43.
Fridovich I.
The biology of oxygen radicals. Science. 1978 Sep 8;201(4359):875-80.
Frolova LY, Grigorieva AY, Sudomoina MA, Kisselev LL.
The human gene encoding tryptophanyl-tRNA synthetase: interferon-response elements and exon-intron
organization. Gene. 1993 Jun 30;128(2):237-45.
Frumento G, Rotondo R, Tonetti M, Damonte G, Benatti U, Ferrara GB.
Tryptophan-derived catabolites are responsible for inhibition of T and natural killer cell proliferation induced by
indoleamine 2,3-dioxygenase. J Exp Med. 2002 Aug 19;196(4):459-68.
Fujita T, Matsushita M, Endo Y.
The lectin-complement pathway – its role in innate immunity and evolution. Immunol Rev. 2004 Apr;198:185-202.
Gao S, von der Malsburg A, Paeschke S, Behlke J, Haller O, Kochs G, Daumke O.
Structural basis of oligomerization in the stalk region of dynamin-like MxA. Nature. 2010 May 27;465(7297):502-6.
Epub 2010 Apr 28.
Garg AK, Aggarwal BB.
Reactive oxygen intermediates in TNF signaling. Mol Immunol. 2002 Dec;39(9):509-17.
Garzoni C, Francois P, Huyghe A, Couzinet S, Tapparel C, Charbonnier Y, Renzoni A, Lucchini S, Lew DP, Vaudaux P,
Kelley WL, Schrenzel J.
A global view of Staphylococcus aureus whole genome expression upon internalization in human epithelial cells.
BMC Genomics. 2007 Jun 14;8:171.
Gaur U, Aggarwal BB.
Regulation of proliferation, survival and apoptosis by members of the TNF superfamily. Biochem Pharmacol. 2003
Oct 15;66(8):1403-8.
Ghosh S, Preet A, Groopman JE, Ganju RK.
Cannabinoid receptor CB 2 modulates the CXCL12/CXCR4-mediated chemotaxis of T lymphocytes. Mol Immunol.
2006 Jul;43(14):2169-79. Epub 2006 Feb 28.
Giachino P, Engelmann S, Bischoff M.
σ B activity depends on RsbU in Staphylococcus aureus. J Bacteriol. 2001 Mar;183(6):1843-52.
Gibbs DF, Warner RL, Weiss SJ, Johnson KJ, Varani J.
Characterization of matrix metalloproteinases produced by rat alveolar macrophages. Am J Respir Cell Mol Biol.
1999 Jun;20(6):1136-44.
Gibbs DF, Shanley TP, Warner RL, Murphy HS, Varani J, Johnson KJ.
Role of matrix metalloproteinases in models of macrophage-dependent acute lung injury. Evidence for alveolar
macrophage as source of proteinases. Am J Respir Cell Mol Biol. 1999 Jun;20(6):1145-54.
Gill SR, Fouts DE, Archer GL, Mongodin EF, Deboy RT, Ravel J, Paulsen IT, Kolonay JF, Brinkac L, Beanan M, Dodson RJ,
Daugherty SC, Madupu R, Angiuoli SV, Durkin AS, Haft DH, Vamathevan J, Khouri H, Utterback T, Lee C, Dimitrov G,
Jiang L, Qin H, Weidman J, Tran K, Kang K, Hance IR, Nelson KE, Fraser CM.
Insights on evolution of virulence and resistance from the complete genome analysis of an early methicillin-resistant
Staphylococcus aureus strain and a biofilm-producing methicillin-resistant Staphylococcus epidermidis strain.
J Bacteriol. 2005 Apr;187(7):2426-38.
214

Maren Depke
References
Glaser R, Friedman SB, Smyth J, Ader R, Bijur P, Brunell P, Cohen N, Krilov LR, Lifrak ST, Stone A, Toffler P.
The differential impact of training stress and final examination stress on herpesvirus latency at the United States
Military Academy at West Point. Brain Behav Immun. 1999 Sep;13(3):240-51.
Gleeson M, McDonald WA, Cripps AW, Pyne DB, Clancy RL, Fricker PA.
The effect on immunity of long-term intensive training in elite swimmers. Clin Exp Immunol. 1995 Oct;102(1):210-6.
Gobin SJ, van den Elsen PJ.
Transcriptional regulation of the MHC class Ib genes HLA-E, HLA-F, and HLA-G. Hum Immunol. 2000 Nov;
61(11):1102-7.
Goerke C, Fluckiger U, Steinhuber A, Bisanzio V, Ulrich M, Bischoff M, Patti JM, Wolz C.
Role of Staphylococcus aureus global regulators sae and σ B in virulence gene expression during device-related
infection. Infect Immun. 2005 Jun;73(6):3415-21.
Goodman MN.
Tumor necrosis factor induces skeletal muscle protein breakdown in rats. Am J Physiol. 1991 May;260(5 Pt 1):
E727-30.
Gordon RJ, Lowy FD.
Pathogenesis of methicillin-resistant Staphylococcus aureus infection. Clin Infect Dis. 2008 Jun 1;46 Suppl 5:S350-9.
Gordon S.
The macrophage: past, present and future. Eur J Immunol. 2007 Nov;37 Suppl 1:S9-17.
Goto H, Shimada K, Ikemoto H, Oguri T; Study Group on Antimicrobial Susceptibility of Pathogens Isolated from
Respiratory Infections.
Antimicrobial susceptibility of pathogens isolated from more than 10,000 patients with infectious respiratory
diseases: a 25-year longitudinal study. J Infect Chemother. 2009 Dec;15(6):347-60.
Gracie JA, Robertson SE, McInnes IB.
Interleukin-18. J Leukoc Biol. 2003 Feb;73(2):213-24.
Gravet A, Colin DA, Keller D, Girardot R, Monteil H, Prévost G.
Characterization of a novel structural member, LukE-LukD, of the bi-component staphylococcal leucotoxins family.
FEBS Lett. 1998 Oct 2;436(2):202-8.
Greenberg S, Grinstein S.
Phagocytosis and innate immunity. Curr Opin Immunol. 2002 Feb;14(1):136-45.
Grinholc M, Wegrzyn G, Kurlenda J.
Evaluation of biofilm production and prevalence of the icaD gene in methicillin-resistant and methicillin-susceptible
Staphylococcus aureus strains isolated from patients with nosocomial infections and carriers. FEMS Immunol Med
Microbiol. 2007 Aug;50(3):375-9. Epub 2007 May 30.
Gross M, Cramton SE, Götz F, Peschel A.
Key role of teichoic acid net charge in Staphylococcus aureus colonization of artificial surfaces. Infect Immun. 2001
May;69(5):3423-6.
Grundmeier M, Tuchscherr L, Brück M, Viemann D, Roth J, Willscher E, Becker K, Peters G, Löffler B.
Staphylococcal strains vary greatly in their ability to induce an inflammatory response in endothelial cells. J Infect
Dis. 2010 Mar 15;201(6):871-80.
Haddad JJ, Harb HL.
L-γ-Glutamyl-L-cysteinyl-glycine (glutathione; GSH) and GSH-related enzymes in the regulation of pro- and antiinflammatory
cytokines: a signaling transcriptional scenario for redox(y) immunologic sensor(s)? Mol Immunol. 2005
May;42(9):987-1014. Epub 2004 Nov 23.
Haggar A, Ehrnfelt C, Holgersson J, Flock JI.
The extracellular adherence protein from Staphylococcus aureus inhibits neutrophil binding to endothelial cells.
Infect Immun. 2004 Oct;72(10):6164-7.
Haggar A, Hussain M, Lönnies H, Herrmann M, Norrby-Teglund A, Flock JI.
Extracellular adherence protein from Staphylococcus aureus enhances internalization into eukaryotic cells. Infect
Immun. 2003 May;71(5):2310-7.
Haggar A, Shannon O, Norrby-Teglund A, Flock JI.
Dual effects of extracellular adherence protein from Staphylococcus aureus on peripheral blood mononuclear cells.
J Infect Dis. 2005 Jul 15;192(2):210-7. Epub 2005 Jun 13.
215

Maren Depke
References
Hagopian K, Ramsey JJ, Weindruch R.
Caloric restriction increases gluconeogenic and transaminase enzyme activities in mouse liver. Exp Gerontol. 2003
Mar;38(3):267-78.
Haller O, Kochs G.
Interferon-induced Mx proteins: dynamin-like GTPases with antiviral activity. Traffic. 2002 Oct;3(10):710-7.
Haller O, Staeheli P, Kochs G.
Interferon-induced Mx proteins in antiviral host defense. Biochimie. 2007 Jun-Jul;89(6-7):812-8. Epub 2007 May 8.
Hallermalm K, Seki K, Wei C, Castelli C, Rivoltini L, Kiessling R, Levitskaya J.
Tumor necrosis factor-α induces coordinated changes in major histocompatibility class I presentation pathway,
resulting in increased stability of class I complexes at the cell surface. Blood. 2001 Aug 15;98(4):1108-15.
Hammel M, Sfyroera G, Pyrpassopoulos S, Ricklin D, Ramyar KX, Pop M, Jin Z, Lambris JD, Geisbrecht BV.
Characterization of Ehp, a secreted complement inhibitory protein from Staphylococcus aureus. J Biol Chem. 2007
Oct 12;282(41):30051-61. Epub 2007 Aug 15.
Hang CH, Shi JX, Li JS, Wu W, Yin HX.
Alterations of intestinal mucosa structure and barrier function following traumatic brain injury in rats. World J
Gastroenterol. 2003 Dec;9(12):2776-81.
Hansen MB, Dresner LS, Wait RB.
Profile of neurohumoral agents on mesenteric and intestinal blood flow in health and disease. Physiol Res. 1998;
47(5):307-27.
Hansen TH, Huang S, Arnold PL, Fremont DH.
Patterns of nonclassical MHC antigen presentation. Nat Immunol. 2007 Jun;8(6):563-8.
Hara T, Ogasawara N, Akimoto H, Takikawa O, Hiramatsu R, Kawabe T, Isobe K, Nagase F.
High-affinity uptake of kynurenine and nitric oxide-mediated inhibition of indoleamine 2,3-dioxygenase in bone
marrow-derived myeloid dendritic cells. Immunol Lett. 2008 Feb 15;116(1):95-102. Epub 2007 Dec 26.
Harraghy N, Hussain M, Haggar A, Chavakis T, Sinha B, Herrmann M, Flock JI.
The adhesive and immunomodulating properties of the multifunctional Staphylococcus aureus protein Eap.
Microbiology. 2003 Oct;149(Pt 10):2701-7.
Harraghy N, Kormanec J, Wolz C, Homerova D, Goerke C, Ohlsen K, Qazi S, Hill P, Herrmann M.
sae is essential for expression of the staphylococcal adhesins Eap and Emp. Microbiology. 2005 Jun;151(Pt 6):1789-
800.
Harris RB, Zhou J, Youngblood BD, Rybkin II, Smagin GN, Ryan DH.
Effect of repeated stress on body weight and body composition of rats fed low- and high-fat diets. Am J Physiol.
1998 Dec;275(6 Pt 2):R1928-38.
Harris RB, Mitchell TD, Simpson J, Redmann SM Jr, Youngblood BD, Ryan DH.
Weight loss in rats exposed to repeated acute restraint stress is independent of energy or leptin status. Am J Physiol
Regul Integr Comp Physiol. 2002 Jan;282(1):R77-88.
Harris RB, Palmondon J, Leshin S, Flatt WP, Richard D.
Chronic disruption of body weight but not of stress peptides or receptors in rats exposed to repeated restraint
stress. Horm Behav. 2006 May;49(5):615-25. Epub 2006 Jan 19.
Hartleib J, Köhler N, Dickinson RB, Chhatwal GS, Sixma JJ, Hartford OM, Foster TJ, Peters G, Kehrel BE, Herrmann M.
Protein A is the von Willebrand factor binding protein on Staphylococcus aureus. Blood. 2000 Sep 15;96(6):2149-56.
Hauck CR, Ohlsen K.
Sticky connections: extracellular matrix protein recognition and integrin-mediated cellular invasion by
Staphylococcus aureus. Curr Opin Microbiol. 2006 Feb;9(1):5-11. Epub 2006 Jan 6.
Hehlgans T, Pfeffer K.
The intriguing biology of the tumour necrosis factor/tumour necrosis factor receptor superfamily: players, rules and
the games. Immunology. 2005 May;115(1):1-20.
Helbig KJ, Ruszkiewicz A, Semendric L, Harley HA, McColl SR, Beard MR.
Expression of the CXCR3 ligand I-TAC by hepatocytes in chronic hepatitis C and its correlation with hepatic
inflammation. Hepatology. 2004 May;39(5):1220-9.
Herbert S, Ziebandt AK, Ohlsen K, Schäfer T, Hecker M, Albrecht D, Novick R, Götz F.
Repair of global regulators in Staphylococcus aureus 8325 and comparative analysis with other clinical isolates.
Infect Immun. 2010 Jun;78(6):2877-89. Epub 2010 Mar 8.
216

Maren Depke
References
Herman A, Kappler JW, Marrack P, Pullen AM.
Superantigens: mechanism of T-cell stimulation and role in immune responses. Annu Rev Immunol. 1991;9:745-72.
Herold MJ, McPherson KG, Reichardt HM.
Glucocorticoids in T cell apoptosis and function. Cell Mol Life Sci. 2006 Jan;63(1):60-72.
Hillgartner FB, Salati LM, Goodridge AG.
Physiological and molecular mechanisms involved in nutritional regulation of fatty acid synthesis. Physiol Rev. 1995
Jan;75(1):47-76.
Hiramatsu K, Watanabe S, Takeuchi F, Ito T, Baba T.
Genetic characterization of methicillin-resistant Staphylococcus aureus. Vaccine. 2004 Dec 6;22 Suppl 1:S5-8.
Horsburgh MJ, Wharton SJ, Cox AG, Ingham E, Peacock S, Foster SJ.
MntR modulates expression of the PerR regulon and superoxide resistance in Staphylococcus aureus through
control of manganese uptake. Mol Microbiol. 2002 Jun;44(5):1269-86.
Horsburgh MJ, Aish JL, White IJ, Shaw L, Lithgow JK, Foster SJ.
σ B modulates virulence determinant expression and stress resistance: characterization of a functional rsbU strain
derived from Staphylococcus aureus 8325-4. J Bacteriol. 2002 Oct;184(19):5457-67.
Hu X, Chakravarty SD, Ivashkiv LB.
Regulation of interferon and Toll-like receptor signaling during macrophage activation by opposing feedforward and
feedback inhibition mechanisms. Immunol Rev. 2008 Dec;226:41-56.
Huang S, Hendriks W, Althage A, Hemmi S, Bluethmann H, Kamijo R, Vilcek J, Zinkernagel RM, Aguet M.
Immune response in mice that lack the interferon-gamma receptor. Science. 1993 Mar 19;259(5102):1742-5.
Hucke C, MacKenzie CR, Adjogble KD, Takikawa O, Däubener W.
Nitric oxide-mediated regulation of gamma interferon-induced bacteriostasis: inhibition and degradation of human
indoleamine 2,3-dioxygenase. Infect Immun. 2004 May;72(5):2723-30.
Hudson MC, Ramp WK, Nicholson NC, Williams AS, Nousiainen MT.
Internalization of Staphylococcus aureus by cultured osteoblasts. Microb Pathog. 1995 Dec;19(6):409-19.
Hultgren O, Eugster HP, Sedgwick JD, Körner H, Tarkowski A.
TNF/lymphotoxin-α double-mutant mice resist septic arthritis but display increased mortality in response to
Staphylococcus aureus. J Immunol. 1998 Dec 1;161(11):5937-42.
Hussain M, Haggar A, Heilmann C, Peters G, Flock JI, Herrmann M.
Insertional inactivation of eap in Staphylococcus aureus strain Newman confers reduced staphylococcal binding to
fibroblasts. Infect Immun. 2002 Jun;70(6):2933-40.
Hussain M, Schäfer D, Juuti KM, Peters G, Haslinger-Löffler B, Kuusela PI, Sinha B.
Expression of Pls (plasmin sensitive) in Staphylococcus aureus negative for pls reduces adherence and cellular
invasion and acts by steric hindrance. J Infect Dis. 2009 Jul 1;200(1):107-17.
Iandolo JJ, Worrell V, Groicher KH, Qian Y, Tian R, Kenton S, Dorman A, Ji H, Lin S, Loh P, Qi S, Zhu H, Roe BA.
Comparative analysis of the genomes of the temperate bacteriophages phi 11, phi 12 and phi 13 of Staphylococcus
aureus 8325. Gene. 2002 May 1;289(1-2):109-18.
Iber H, Li-Masters T, Chen Q, Yu S, Morgan ET.
Regulation of hepatic cytochrome P450 2C11 via cAMP: implications for down-regulation in diabetes, fasting, and
inflammation. J Pharmacol Exp Ther. 2001 Apr;297(1):174-80.
Iles KE, Forman HJ.
Macrophage signaling and respiratory burst. Immunol Res. 2002;26(1-3):95-105.
Imlay JA, Linn S.
DNA damage and oxygen radical toxicity. Science. 1988 Jun 3;240(4857):1302-9.
Janeway CA Jr, Medzhitov R.
Innate immune recognition. Annu Rev Immunol. 2002;20:197-216. Epub 2001 Oct 4.
Janeway CA Jr., Travers P, Walport M, Shlomchik MJ.
Immunologie. 5. Auflage, 2002, Spektrum Akademischer Verlag GmbH Heidelberg Berlin. ISBN 3-8274-1079-7.
Jensen PE.
Mechanisms of antigen presentation. Clin Chem Lab Med. 1999 Mar;37(3):179-86.
217

Maren Depke
References
Jin T, Bokarewa M, Foster T, Mitchell J, Higgins J, Tarkowski A.
Staphylococcus aureus resists human defensins by production of staphylokinase, a novel bacterial evasion
mechanism. J Immunol. 2004 Jan 15;172(2):1169-76.
Johnson LM, Scott P.
STAT1 expression in dendritic cells, but not T cells, is required for immunity to Leishmania major. J Immunol. 2007
Jun 1;178(11):7259-66.
Jonsson IM, Arvidson S, Foster S, Tarkowski A.
Sigma factor B and RsbU are required for virulence in Staphylococcus aureus-induced arthritis and sepsis. Infect
Immun. 2004 Oct;72(10):6106-11.
Jönsson K, McDevitt D, McGavin MH, Patti JM, Höök M.
Staphylococcus aureus expresses a major histocompatibility complex class II analog. J Biol Chem. 1995 Sep 15;
270(37):21457-60.
Jung T, Catalgol B, Grune T.
The proteasomal system. Mol Aspects Med. 2009 Aug;30(4):191-296. Epub 2009 Apr 14.
Jurado J, Prieto-Alamo MJ, Madrid-Rísquez J, Pueyo C.
Absolute gene expression patterns of thioredoxin and glutaredoxin redox systems in mouse. J Biol Chem. 2003 Nov
14;278(46):45546-54. Epub 2003 Sep 3.
Kaliman PA, Barannik T, Strel'chenko E, Inshina N, Sokol O.
Intracellular redistribution of heme in rat liver under oxidative stress: the role of heme synthesis. Cell Biol Int. 2005
Jan;29(1):9-14. Epub 2005 Jan 26.
Kaneko J, Kamio Y.
Bacterial two-component and hetero-heptameric pore-forming cytolytic toxins: structures, pore-forming
mechanism, and organization of the genes. Biosci Biotechnol Biochem. 2004 May;68(5):981-1003.
Kantari C, Pederzoli-Ribeil M, Witko-Sarsat V.
The role of neutrophils and monocytes in innate immunity. Contrib Microbiol. 2008;15:118-46.
Karavolos MH, Horsburgh MJ, Ingham E, Foster SJ.
Role and regulation of the superoxide dismutases of Staphylococcus aureus. Microbiology. 2003 Oct;149(Pt 10):
2749-58.
Kawabata S, Iwanaga S.
Structure and function of staphylothrombin. Semin Thromb Hemost. 1994;20(4):345-50.
Kawabata S, Morita T, Iwanaga S, Igarashi H.
Enzymatic properties of staphylothrombin, an active molecular complex formed between staphylocoagulase and
human prothrombin. J Biochem. 1985 Dec;98(6):1603-14.
Khan S, van den Broek M, Schwarz K, de Giuli R, Diener PA, Groettrup M.
Immunoproteasomes largely replace constitutive proteasomes during an antiviral and antibacterial immune
response in the liver. J Immunol. 2001 Dec 15;167(12):6859-68.
Khan WA, Blobe GC, Hannun YA.
Arachidonic acid and free fatty acids as second messengers and the role of protein kinase C. Cell Signal. 1995
Mar;7(3):171-84.
Kiank C, Holtfreter B, Starke A, Mundt A, Wilke C, Schütt C.
Stress susceptibility predicts the severity of immune depression and the failure to combat bacterial infections in
chronically stressed mice. Brain Behav Immun. 2006 Jul;20(4):359-68. Epub 2005 Dec 2.
Kiank C, Entleutner M, Fürll B, Westerholt A, Heidecke CD, Schütt C.
Stress-induced immune conditioning affects the course of experimental peritonitis. Shock. 2007 Mar;27(3):305-11.
Kiank C, Koerner P, Kessler W, Traeger T, Maier S, Heidecke CD, Schuett C.
Seasonal variations in inflammatory responses to sepsis and stress in mice. Crit Care Med. 2007 Oct;35(10):2352-8.
Kiank C, Daeschlein G, Schuett C.
Pneumonia as a long-term consequence of chronic psychological stress in BALB/c mice. Brain Behav Immun. 2008
Nov;22(8):1173-7. Epub 2008 Jun 20.
Kiank C, Zeden JP, Drude S, Domanska G, Fusch G, Otten W, Schuett C.
Psychological stress-induced, IDO1-dependent tryptophan catabolism: implications on immunosuppression in mice
and humans. PLoS One. 2010 Jul 28;5(7):e11825.
218

Maren Depke
References
Kim JS, Kim JG, Moon MY, Jeon CY, Won HY, Kim HJ, Jeon YJ, Seo JY, Kim JI, Kim J, Lee JY, Kim PH, Park JB.
Transforming growth factor-β1 regulates macrophage migration via RhoA. Blood. 2006 Sep 15;108(6):1821-9. Epub
2006 May 16.
King NJ, Thomas SR.
Molecules in focus: indoleamine 2,3-dioxygenase. Int J Biochem Cell Biol. 2007;39(12):2167-72. Epub 2007 Jan 20.
Kisselev LL.
Mammalian tryptophanyl-tRNA synthetases. Biochimie. 1993;75(12):1027-39.
Klamp T, Boehm U, Schenk D, Pfeffer K, Howard JC.
A giant GTPase, very large inducible GTPase-1, is inducible by IFNs. J Immunol. 2003 Aug 1;171(3):1255-65.
Klein C, Wüstefeld T, Assmus U, Roskams T, Rose-John S, Müller M, Manns MP, Ernst M, Trautwein C.
The IL-6-gp130-STAT3 pathway in hepatocytes triggers liver protection in T cell-mediated liver injury. J Clin Invest.
2005 Apr;115(4):860-9. Epub 2005 Mar 3.
Kluytmans J, van Belkum A, Verbrugh H.
Nasal carriage of Staphylococcus aureus: epidemiology, underlying mechanisms, and associated risks. Clin Microbiol
Rev. 1997 Jul;10(3):505-20. Review.
Knowles RG, Moncada S.
Nitric oxide synthases in mammals. Biochem J. 1994 Mar 1;298 ( Pt 2):249-58.
Koban M, Swinson KL.
Chronic REM-sleep deprivation of rats elevates metabolic rate and increases UCP1 gene expression in brown
adipose tissue. Am J Physiol Endocrinol Metab. 2005 Jul;289(1):E68-74. Epub 2005 Feb 22.
Kohler C, Wolff S, Albrecht D, Fuchs S, Becher D, Büttner K, Engelmann S, Hecker M.
Proteome analyses of Staphylococcus aureus in growing and non-growing cells: a physiological approach. Int J Med
Microbiol. 2005 Dec;295(8):547-65. Epub 2005 Oct 25.
Kokai-Kun JF, Walsh SM, Chanturiya T, Mond JJ.
Lysostaphin cream eradicates Staphylococcus aureus nasal colonization in a cotton rat model. Antimicrob Agents
Chemother. 2003 May;47(5):1589-97.
König B, Prévost G, König W.
Composition of staphylococcal bi-component toxins determines pathophysiological reactions. J Med Microbiol. 1997
Jun;46(6):479-85.
Koolhaas JM, de Boer SF, Buwalda B, van Reenen K.
Individual variation in coping with stress: a multidimensional approach of ultimate and proximate mechanisms.
Brain Behav Evol. 2007;70(4):218-26. Epub 2007 Sep 18.
Kota RS, Rutledge JC, Gohil K, Kumar A, Enelow RI, Ramana CV.
Regulation of gene expression in RAW 264.7 macrophage cell line by interferon-γ. Biochem Biophys Res Commun.
2006 Apr 21;342(4):1137-46. Epub 2006 Feb 24.
Kullik I, Giachino P.
The alternative sigma factor σ B in Staphylococcus aureus: regulation of the sigB operon in response to growth phase
and heat shock. Arch Microbiol. 1997 Mar 7;167(2/3):151-9.
Kullik I, Giachino P, Fuchs T.
Deletion of the alternative sigma factor σ B in Staphylococcus aureus reveals its function as a global regulator of
virulence genes. J Bacteriol. 1998 Sep;180(18):4814-20.
Kuo LE, Abe K, Zukowska Z.
Stress, NPY and vascular remodeling: Implications for stress-related diseases. Peptides. 2007 Feb;28(2):435-40. Epub
2007 Jan 22.
Kuroda M, Kuwahara-Arai K, Hiramatsu K.
Identification of the up- and down-regulated genes in vancomycin-resistant Staphylococcus aureus strains Mu3 and
Mu50 by cDNA differential hybridization method. Biochem Biophys Res Commun. 2000 Mar 16;269(2):485-90.
Kuroda M, Kuroda H, Oshima T, Takeuchi F, Mori H, Hiramatsu K.
Two-component system VraSR positively modulates the regulation of cell-wall biosynthesis pathway in
Staphylococcus aureus. Mol Microbiol. 2003 Aug;49(3):807-21.
Kuypers JM, Proctor RA.
Reduced adherence to traumatized rat heart valves by a low-fibronectin-binding mutant of Staphylococcus aureus.
Infect Immun. 1989 Aug;57(8):2306-12.
219

Maren Depke
References
Kwan T, Liu J, DuBow M, Gros P, Pelletier J.
The complete genomes and proteomes of 27 Staphylococcus aureus bacteriophages. Proc Natl Acad Sci U S A. 2005
Apr 5;102(14):5174-9. Epub 2005 Mar 23.
Lam TK, Gutierrez-Juarez R, Pocai A, Bhanot S, Tso P, Schwartz GJ, Rossetti L.
Brain glucose metabolism controls the hepatic secretion of triglyceride-rich lipoproteins. Nat Med. 2007 Feb;
13(2):171-80. Epub 2007 Feb 4.
Lambert GW, Straznicky NE, Lambert EA, Dixon JB, Schlaich MP.
Sympathetic nervous activation in obesity and the metabolic syndrome – Causes, consequences and therapeutic
implications. Pharmacol Ther. 2010 May;126(2):159-72. Epub 2010 Feb 19.
Lang CH, Bagby GJ, Spitzer JJ.
Carbohydrate dynamics in the hypermetabolic septic rat. Metabolism. 1984 Oct;33(10):959-63.
Laye JP, Gill JH.
Phospholipase A2 expression in tumours: a target for therapeutic intervention? Drug Discov Today. 2003 Aug1;
8(15):710-6.
Le Bail O, Schmidt-Ullrich R, Israël A.
Promoter analysis of the gene encoding the IκB-α/MAD3 inhibitor of NF-κB: positive regulation by members of the
rel/NF-κB family. EMBO J. 1993 Dec 15;12(13):5043-9.
Lecordier L, Vanhollebeke B, Poelvoorde P, Tebabi P, Paturiaux-Hanocq F, Andris F, Lins L, Pays E.
C-terminal mutants of apolipoprotein L-I efficiently kill both Trypanosoma brucei brucei and Trypanosoma brucei
rhodesiense. PLoS Pathog. 2009 Dec;5(12):e1000685. Epub 2009 Dec 4.
Lee LY, Miyamoto YJ, McIntyre BW, Höök M, McCrea KW, McDevitt D, Brown EL.
The Staphylococcus aureus Map protein is an immunomodulator that interferes with T cell-mediated responses.
J Clin Invest. 2002 Nov;110(10):1461-71.
Lee LY, Höök M, Haviland D, Wetsel RA, Yonter EO, Syribeys P, Vernachio J, Brown EL.
Inhibition of complement activation by a secreted Staphylococcus aureus protein. J Infect Dis. 2004 Aug 1;
190(3):571-9. Epub 2004 Jul 6.
Lee LY, Liang X, Höök M, Brown EL.
Identification and characterization of the C3 binding domain of the Staphylococcus aureus extracellular fibrinogenbinding
protein (Efb). J Biol Chem. 2004 Dec 3;279(49):50710-6. Epub 2004 Aug 26.
Leibowitz SF, Wortley KE.
Hypothalamic control of energy balance: different peptides, different functions. Peptides. 2004 Mar;25(3):473-504.
Leon LR.
Invited review: cytokine regulation of fever: studies using gene knockout mice. J Appl Physiol. 2002 Jun;92(6):2648-
55.
Leonard BE.
HPA and immune axes in stress: involvement of the serotonergic system. Neuroimmunomodulation. 2006;13(5-6):
268-76. Epub 2007 Aug 6.
Levels JH, Lemaire LC, van den Ende AE, van Deventer SJ, van Lanschot JJ.
Lipid composition and lipopolysaccharide binding capacity of lipoproteins in plasma and lymph of patients with
systemic inflammatory response syndrome and multiple organ failure. Crit Care Med. 2003 Jun;31(6):1647-53.
Li G, Zhang J, Sun Y, Wang H, Wang Y.
The evolutionarily dynamic IFN-inducible GTPase proteins play conserved immune functions in vertebrates and
cephalochordates. Mol Biol Evol. 2009 Jul;26(7):1619-30. Epub 2009 Apr 15.
Liang X, Yu C, Sun J, Liu H, Landwehr C, Holmes D, Ji Y.
Inactivation of a two-component signal transduction system, SaeRS, eliminates adherence and attenuates virulence
of Staphylococcus aureus. Infect Immun. 2006 Aug;74(8):4655-65.
Liao F, Rabin RL, Yannelli JR, Koniaris LG, Vanguri P, Farber JM.
Human Mig chemokine: biochemical and functional characterization. J Exp Med. 1995 Nov 1;182(5):1301-14.
Li-Hawkins J, Lund EG, Bronson AD, Russell DW.
Expression cloning of an oxysterol 7α-hydroxylase selective for 24-hydroxycholesterol. J Biol Chem. 2000 Jun 2;
275(22):16543-9.
220

Maren Depke
References
Limmer A, Ohl J, Kurts C, Ljunggren HG, Reiss Y, Groettrup M, Momburg F, Arnold B, Knolle PA.
Efficient presentation of exogenous antigen by liver endothelial cells to CD8+ T cells results in antigen-specific T-cell
tolerance. Nat Med. 2000 Dec;6(12):1348-54.
Lin CH, Sheu SY, Lee HM, Ho YS, Lee WS, Ko WC, Sheu JR.
Involvement of protein kinase C-γ in IL-1β-induced cyclooxygenase-2 expression in human pulmonary epithelial
cells. Mol Pharmacol. 2000 Jan;57(1):36-43.
Lin CH, Kuan IH, Lee HM, Lee WS, Sheu JR, Ho YS, Wang CH, Kuo HP.
Induction of cyclooxygenase-2 protein by lipoteichoic acid from Staphylococcus aureus in human pulmonary
epithelial cells: involvement of a nuclear factor-κB-dependent pathway. Br J Pharmacol. 2001 Oct;134(3):543-52.
Lindsay JA, Holden MT.
Staphylococcus aureus: superbug, super genome? Trends Microbiol. 2004 Aug;12(8):378-85.
Lindsay JA, Ruzin A, Ross HF, Kurepina N, Novick RP.
The gene for toxic shock toxin is carried by a family of mobile pathogenicity islands in Staphylococcus aureus. Mol
Microbiol. 1998 Jul;29(2):527-43.
Linge HM, Collin M, Giwercman A, Malm J, Bjartell A, Egesten A.
The antibacterial chemokine MIG/CXCL9 is constitutively expressed in epithelial cells of the male urogenital tract
and is present in seminal plasma. J Interferon Cytokine Res. 2008 Mar;28(3):191-6.
Liptay S, Schmid RM, Nabel EG, Nabel GJ.
Transcriptional regulation of NF-κB2: evidence for κB-mediated positive and negative autoregulation. Mol Cell Biol.
1994 Dec;14(12):7695-703.
Liu GY, Essex A, Buchanan JT, Datta V, Hoffman HM, Bastian JF, Fierer J, Nizet V.
Staphylococcus aureus golden pigment impairs neutrophil killing and promotes virulence through its antioxidant
activity. J Exp Med. 2005 Jul 18;202(2):209-15. Epub 2005 Jul 11.
Liu J, Shue E, Ewalt KL, Schimmel P.
A new gamma-interferon-inducible promoter and splice variants of an anti-angiogenic human tRNA synthetase.
Nucleic Acids Res. 2004 Feb 2;32(2):719-27. Print 2004.
Liu Z, Lu H, Jiang Z, Pastuszyn A, Hu CA.
Apolipoprotein l6, a novel proapoptotic Bcl-2 homology 3-only protein, induces mitochondria-mediated apoptosis in
cancer cells. Mol Cancer Res. 2005 Jan;3(1):21-31.
Livermore DM.
Antibiotic resistance in staphylococci. Int J Antimicrob Agents. 2000 Nov;16 Suppl 1:S3-10.
Livermore DM.
Has the era of untreatable infections arrived? J Antimicrob Chemother. 2009 Sep;64 Suppl 1:i29-36.
Loetscher M, Gerber B, Loetscher P, Jones SA, Piali L, Clark-Lewis I, Baggiolini M, Moser B.
Chemokine receptor specific for IP10 and Mig: structure, function, and expression in activated T-lymphocytes. J Exp
Med. 1996 Sep 1;184(3):963-9.
Loetscher P, Pellegrino A, Gong JH, Mattioli I, Loetscher M, Bardi G, Baggiolini M, Clark-Lewis I.
The ligands of CXC chemokine receptor 3, I-TAC, Mig, and IP10, are natural antagonists for CCR3. J Biol Chem. 2001
Feb 2;276(5):2986-91. Epub 2000 Nov 10.
Lorenz U, Hüttinger C, Schäfer T, Ziebuhr W, Thiede A, Hacker J, Engelmann S, Hecker M, Ohlsen K.
The alternative sigma factor sigma B of Staphylococcus aureus modulates virulence in experimental central venous
catheter-related infections. Microbes Infect. 2008 Mar;10(3):217-23. Epub 2007 Nov 28.
Loukissa A, Cardozo C, Altschuller-Felberg C, Nelson JE.
Control of LMP7 expression in human endothelial cells by cytokines regulating cellular and humoral immunity.
Cytokine. 2000 Sep;12(9):1326-30.
Lowy FD.
Staphylococcus aureus infections. N Engl J Med. 1998 Aug 20;339(8):520-32.
Lowy FD.
Is Staphylococcus aureus an intracellular pathogen? Trends Microbiol. 2000 Aug;8(8):341-3.
Lundberg U.
Stress hormones in health and illness: the roles of work and gender. Psychoneuroendocrinology. 2005 Nov;
30(10):1017-21.
221

Maren Depke
References
Luster AD, Unkeless JC, Ravetch JV.
Gamma-interferon transcriptionally regulates an early-response gene containing homology to platelet proteins.
Nature. 1985 Jun 20-26;315(6021):672-6.
Luster AD.
The role of chemokines in linking innate and adaptive immunity. Curr Opin Immunol. 2002 Feb;14(1):129-35.
Ma J, Chen T, Mandelin J, Ceponis A, Miller NE, Hukkanen M, Ma GF, Konttinen YT.
Regulation of macrophage activation. Cell Mol Life Sci. 2003 Nov;60(11):2334-46.
Ma W, Lehner PJ, Cresswell P, Pober JS, Johnson DR.
Interferon-gamma rapidly increases peptide transporter (TAP) subunit expression and peptide transport capacity in
endothelial cells. J Biol Chem. 1997 Jun 27;272(26):16585-90.
Mackey-Lawrence NM, Potter DE, Cerca N, Jefferson KK.
Staphylococcus aureus immunodominant surface antigen B is a cell-surface associated nucleic acid binding protein.
BMC Microbiol. 2009 Mar 26;9:61.
MacMicking J, Xie QW, Nathan C.
Nitric oxide and macrophage function. Annu Rev Immunol. 1997;15:323-50.
MacMicking JD.
IFN-inducible GTPases and immunity to intracellular pathogens. Trends Immunol. 2004 Nov;25(11):601-9.
Macpherson AJ, Martinic MM, Harris N.
The functions of mucosal T cells in containing the indigenous commensal flora of the intestine. Cell Mol Life Sci.
2002 Dec;59(12):2088-96.
Maemura K, Zheng Q, Wada T, Ozaki M, Takao S, Aikou T, Bulkley GB, Klein AS, Sun Z.
Reactive oxygen species are essential mediators in antigen presentation by Kupffer cells. Immunol Cell Biol. 2005
Aug;83(4):336-43.
Mak RH, Cheung W, Cone RD, Marks DL.
Leptin and inflammation-associated cachexia in chronic kidney disease. Kidney Int. 2006 Mar;69(5):794-7.
Mandell GL.
Catalase, superoxide dismutase, and virulence of Staphylococcus aureus. In vitro and in vivo studies with emphasis
on staphylococcal--leukocyte interaction. J Clin Invest. 1975 Mar;55(3):561-6.
Mantovani A, Sica A, Sozzani S, Allavena P, Vecchi A, Locati M.
The chemokine system in diverse forms of macrophage activation and polarization. Trends Immunol. 2004 Dec;
25(12):677-86.
Maresso AW, Schneewind O.
Iron acquisition and transport in Staphylococcus aureus. Biometals. 2006 Apr;19(2):193-203.
Marshall JH, Wilmoth GJ.
Pigments of Staphylococcus aureus, a series of triterpenoid carotenoids. J Bacteriol. 1981 Sep;147(3):900-13.
Martens S, Sabel K, Lange R, Uthaiah R, Wolf E, Howard JC.
Mechanisms regulating the positioning of mouse p47 resistance GTPases LRG-47 and IIGP1 on cellular membranes:
retargeting to plasma membrane induced by phagocytosis. J Immunol. 2004 Aug 15;173(4):2594-606.
Martin G, Pilon A, Albert C, Vallé M, Hum DW, Fruchart JC, Najib J, Clavey V, Staels B.
Comparison of expression and regulation of the high-density lipoprotein receptor SR-BI and the low-density
lipoprotein receptor in human adrenocortical carcinoma NCI-H295 cells. Eur J Biochem. 1999 Apr;261(2):481-91.
Matussek A, Strindhall J, Stark L, Rohde M, Geffers R, Buer J, Kihlström E, Lindgren PE, Löfgren S.
Infection of human endothelial cells with Staphylococcus aureus induces transcription of genes encoding an innate
immunity response. Scand J Immunol. 2005 Jun;61(6):536-44.
McEwen BS.
Protection and damage from acute and chronic stress: allostasis and allostatic overload and relevance to the
pathophysiology of psychiatric disorders. Ann N Y Acad Sci. 2004 Dec;1032:1-7.
McGuinness OP, Jacobs J, Moran C, Lacy B.
Impact of infection on hepatic disposal of a peripheral glucose infusion in the conscious dog. Am J Physiol. 1995 Aug;
269(2 Pt 1):E199-207.
McGuinness OP, Snowden RT, Moran C, Neal DW, Fujiwara T, Cherrington AD.
Impact of acute epinephrine removal on hepatic glucose metabolism during stress hormone infusion. Metabolism.
1999 Jul;48(7):910-4.
222

Maren Depke
References
McInnes IB, Gracie JA, Leung BP, Wei XQ, Liew FY.
Interleukin 18: a pleiotropic participant in chronic inflammation. Immunol Today. 2000 Jul;21(7):312-5.
McNamara PJ, Proctor RA.
Staphylococcus aureus small colony variants, electron transport and persistent infections. Int J Antimicrob Agents.
2000 Mar;14(2):117-22.
Menestrina G, Serra MD, Prévost G.
Mode of action of beta-barrel pore-forming toxins of the staphylococcal alpha-hemolysin family. Toxicon. 2001 Nov;
39(11):1661-72.
Miao L, St Clair DK.
Regulation of superoxide dismutase genes: implications in disease. Free Radic Biol Med. 2009 Aug 15;47(4):344-56.
Epub 2009 May 25.
Mills CD, Kincaid K, Alt JM, Heilman MJ, Hill AM.
M-1/M-2 macrophages and the Th1/Th2 paradigm. J Immunol. 2000 Jun 15;164(12):6166-73.
Mittelstadt PR, Ashwell JD.
Inhibition of AP-1 by the glucocorticoid-inducible protein GILZ. J Biol Chem. 2001 Aug 3;276(31):29603-10. Epub
2001 Jun 7.
Mizock BA.
Alterations in carbohydrate metabolism during stress: a review of the literature. Am J Med. 1995 Jan;98(1):75-84.
Mold C, Du Clos TW.
C-reactive protein increases cytokine responses to Streptococcus pneumoniae through interactions with Fcγ
receptors. J Immunol. 2006 Jun 15;176(12):7598-604.
Moore KW, de Waal Malefyt R, Coffman RL, O'Garra A.
Interleukin-10 and the interleukin-10 receptor. Annu Rev Immunol. 2001;19:683-765.
Moreilhon C, Gras D, Hologne C, Bajolet O, Cottrez F, Magnone V, Merten M, Groux H, Puchelle E, Barbry P.
Live Staphylococcus aureus and bacterial soluble factors induce different transcriptional responses in human airway
cells. Physiol Genomics. 2005 Feb 10;20(3):244-55. Epub 2004 Dec 14.
Morgan ET.
Regulation of cytochrome p450 by inflammatory mediators: why and how? Drug Metab Dispos. 2001 Mar;
29(3):207-12.
Mori M, Gotoh T.
Arginine metabolic enzymes, nitric oxide and infection. J Nutr. 2004 Oct;134(10 Suppl):2820S-2825.
Morikawa K, Inose Y, Okamura H, Maruyama A, Hayashi H, Takeyasu K, Ohta T.
A new staphylococcal sigma factor in the conserved gene cassette: functional significance and implication for the
evolutionary processes. Genes Cells. 2003 Aug;8(8):699-712.
Morley JE, Thomas DR, Wilson MM.
Cachexia: pathophysiology and clinical relevance. Am J Clin Nutr. 2006 Apr;83(4):735-43.
Mosser DM.
The many faces of macrophage activation. J Leukoc Biol. 2003 Feb;73(2):209-12.
Müller A, Heseler K, Schmidt SK, Spekker K, Mackenzie CR, Däubener W.
The missing link between indoleamine 2,3-dioxygenase mediated antibacterial and immunoregulatory effects. J Cell
Mol Med. 2009 Jun;13(6):1125-35. Epub 2009 Oct 13.
Munn DH, Sharma MD, Baban B, Harding HP, Zhang Y, Ron D, Mellor AL.
GCN2 kinase in T cells mediates proliferative arrest and anergy induction in response to indoleamine 2,3-
dioxygenase. Immunity. 2005 May;22(5):633-42.
Murdoch C, Finn A.
Chemokine receptors and their role in inflammation and infectious diseases. Blood. 2000 May 15;95(10):3032-43.
Murray HW, Szuro-Sudol A, Wellner D, Oca MJ, Granger AM, Libby DM, Rothermel CD, Rubin BY.
Role of tryptophan degradation in respiratory burst-independent antimicrobial activity of gamma interferonstimulated
human macrophages. Infect Immun. 1989 Mar;57(3):845-9.
Murray MF.
Tryptophan depletion and HIV infection: a metabolic link to pathogenesis. Lancet Infect Dis. 2003 Oct;3(10):644-52.
223

Maren Depke
References
Namiki S, Nakamura T, Oshima S, Yamazaki M, Sekine Y, Tsuchiya K, Okamoto R, Kanai T, Watanabe M.
IRF-1 mediates upregulation of LMP7 by IFN-γ and concerted expression of immunosubunits of the proteasome.
FEBS Lett. 2005 May 23;579(13):2781-7. Epub 2005 Apr 20.
Narui K, Noguchi N, Saito A, Kakimi K, Motomura N, Kubo K, Takamoto S, Sasatsu M.
Anti-infectious activity of tryptophan metabolites in the L-tryptophan-L-kynurenine pathway. Biol Pharm Bull. 2009
Jan;32(1):41-4.
Nemoto N, Sakurai J.
Glucocorticoid and sex hormones as activating or modulating factors for expression of Cyp2b-9 and Cyp2b-10 in the
mouse liver and hepatocytes. Arch Biochem Biophys. 1995 May 10;319(1):286-92.
Newton R, Kuitert LM, Bergmann M, Adcock IM, Barnes PJ.
Evidence for involvement of NF-κB in the transcriptional control of COX-2 gene expression by IL-1β. Biochem
Biophys Res Commun. 1997 Aug 8;237(1):28-32.
Nguyen VT, Kamio Y, Higuchi H.
Single-molecule imaging of cooperative assembly of γ-hemolysin on erythrocyte membranes. EMBO J. 2003 Oct 1;
22(19):4968-79.
Nicholas RO, Li T, McDevitt D, Marra A, Sucoloski S, Demarsh PL, Gentry DR.
Isolation and characterization of a sigB deletion mutant of Staphylococcus aureus. Infect Immun. 1999 Jul;
67(7):3667-9.
Nicolas P, Leduc A, Robin S, Rasmussen S, Jarmer H, Bessières P.
Transcriptional landscape estimation from tiling array data using a model of signal shift and drift. Bioinformatics.
2009 Sep 15;25(18):2341-7. Epub 2009 Jun 26.
Nilsson IM, Lee JC, Bremell T, Rydén C, Tarkowski A.
The role of staphylococcal polysaccharide microcapsule expression in septicemia and septic arthritis. Infect Immun.
1997 Oct;65(10):4216-21.
Novick D, Kim SH, Fantuzzi G, Reznikov LL, Dinarello CA, Rubinstein M.
Interleukin-18 binding protein: a novel modulator of the Th1 cytokine response. Immunity. 1999 Jan;10(1):127-36.
Novick RP.
Mobile genetic elements and bacterial toxinoses: the superantigen-encoding pathogenicity islands of
Staphylococcus aureus. Plasmid. 2003 Mar;49(2):93-105.
Ogimoto K, Harris MK Jr, Wisse BE.
MyD88 is a key mediator of anorexia, but not weight loss, induced by lipopolysaccharide and interleukin-1β.
Endocrinology. 2006 Sep;147(9):4445-53. Epub 2006 Jun 15.
Ordway RW, Singer JJ, Walsh JV Jr.
Direct regulation of ion channels by fatty acids. Trends Neurosci. 1991 Mar;14(3):96-100.
Otani A, Slike BM, Dorrell MI, Hood J, Kinder K, Ewalt KL, Cheresh D, Schimmel P, Friedlander M.
A fragment of human TrpRS as a potent antagonist of ocular angiogenesis. Proc Natl Acad Sci U S A. 2002 Jan 8;
99(1):178-83. Epub 2002 Jan 2.
Ott M, Gogvadze V, Orrenius S, Zhivotovsky B.
Mitochondria, oxidative stress and cell death. Apoptosis. 2007 May;12(5):913-22.
Otter JA, French GL.
Molecular epidemiology of community-associated meticillin-resistant Staphylococcus aureus in Europe. Lancet
Infect Dis. 2010 Apr;10(4):227-39.
Page NM, Butlin DJ, Lomthaisong K, Lowry PJ.
The human apolipoprotein L gene cluster: identification, classification, and sites of distribution. Genomics. 2001
May 15;74(1):71-8.
Pagels M, Fuchs S, Pané-Farré J, Kohler C, Menschner L, Hecker M, McNamarra PJ, Bauer MC, von Wachenfeldt C,
Liebeke M, Lalk M, Sander G, von Eiff C, Proctor RA, Engelmann S.
Redox sensing by a Rex-family repressor is involved in the regulation of anaerobic gene expression in Staphylococcus
aureus. Mol Microbiol. 2010 Jun 1;76(5):1142-61. Epub 2010 Mar 30.
Park JB.
Phagocytosis induces superoxide formation and apoptosis in macrophages. Exp Mol Med. 2003 Oct 31;35(5):325-35.
224

Maren Depke
References
Park JY, Pillinger MH, Abramson SB.
Prostaglandin E 2 synthesis and secretion: the role of PGE 2 synthases. Clin Immunol. 2006 Jun;119(3):229-40. Epub
2006 Mar 15.
Pasini E, Aquilani R, Dioguardi FS.
Amino acids: chemistry and metabolism in normal and hypercatabolic states. Am J Cardiol. 2004 Apr 22;93(8A):3A-
5A.
Pasparakis M, Alexopoulou L, Episkopou V, Kollias G.
Immune and inflammatory responses in TNFα-deficient mice: a critical requirement for TNFα in the formation of
primary B cell follicles, follicular dendritic cell networks and germinal centers, and in the maturation of the humoral
immune response. J Exp Med. 1996 Oct 1;184(4):1397-411.
Patti JM, Allen BL, McGavin MJ, Höök M.
MSCRAMM-mediated adherence of microorganisms to host tissues. Annu Rev Microbiol. 1994;48:585-617.
Peacock SJ, Foster TJ, Cameron BJ, Berendt AR.
Bacterial fibronectin-binding proteins and endothelial cell surface fibronectin mediate adherence of Staphylococcus
aureus to resting human endothelial cells. Microbiology. 1999 Dec;145 (Pt 12):3477-86.
Peacock SJ, Day NP, Thomas MG, Berendt AR, Foster TJ.
Clinical isolates of Staphylococcus aureus exhibit diversity in fnb genes and adhesion to human fibronectin. J Infect.
2000 Jul;41(1):23-31.
Pereira CA, Modolell M, Frey JR, Lefkovits I.
Gene expression in IFN-gamma-activated murine macrophages. Braz J Med Biol Res. 2004 Dec;37(12):1795-809.
Epub 2004 Nov 17.
Pérez-Morga D, Vanhollebeke B, Paturiaux-Hanocq F, Nolan DP, Lins L, Homblé F, Vanhamme L, Tebabi P, Pays A,
Poelvoorde P, Jacquet A, Brasseur R, Pays E.
Apolipoprotein L-I promotes trypanosome lysis by forming pores in lysosomal membranes. Science. 2005 Jul 15;
309(5733):469-72.
Périchon B, Courvalin P.
VanA-type vancomycin-resistant Staphylococcus aureus. Antimicrob Agents Chemother. 2009 Nov;53(11):4580-7.
Epub 2009 Jun 8.
Peschel A, Otto M, Jack RW, Kalbacher H, Jung G, Götz F.
Inactivation of the dlt operon in Staphylococcus aureus confers sensitivity to defensins, protegrins, and other
antimicrobial peptides. J Biol Chem. 1999 Mar 26;274(13):8405-10.
Peschel A, Jack RW, Otto M, Collins LV, Staubitz P, Nicholson G, Kalbacher H, Nieuwenhuizen WF, Jung G, Tarkowski A,
van Kessel KP, van Strijp JA.
Staphylococcus aureus resistance to human defensins and evasion of neutrophil killing via the novel virulence factor
MprF is based on modification of membrane lipids with l-lysine. J Exp Med. 2001 May 7;193(9):1067-76.
Peterson ML, Ault K, Kremer MJ, Klingelhutz AJ, Davis CC, Squier CA, Schlievert PM.
The innate immune system is activated by stimulation of vaginal epithelial cells with Staphylococcus aureus and
toxic shock syndrome toxin 1. Infect Immun. 2005 Apr;73(4):2164-74.
Peterson PK, Chao CC, Molitor T, Murtaugh M, Strgar F, Sharp BM.
Stress and pathogenesis of infectious disease. Rev Infect Dis. 1991 Jul-Aug;13(4):710-20.
Petkovic V, Moghini C, Paoletti S, Uguccioni M, Gerber B.
I-TAC/CXCL11 is a natural antagonist for CCR5. J Leukoc Biol. 2004 Sep;76(3):701-8. Epub 2004 Jun 3.
Petri B, Bixel MG.
Molecular events during leukocyte diapedesis. FEBS J. 2006 Oct;273(19):4399-407. Epub 2006 Sep 11.
Pfefferkorn ER.
Interferon γ blocks the growth of Toxoplasma gondii in human fibroblasts by inducing the host cells to degrade
tryptophan. Proc Natl Acad Sci U S A. 1984 Feb;81(3):908-12.
Plata K, Rosato AE, Wegrzyn G.
Staphylococcus aureus as an infectious agent: overview of biochemistry and molecular genetics of its pathogenicity.
Acta Biochim Pol. 2009;56(4):597-612. Epub 2009 Dec 11.
Popov A, Schultze JL.
IDO-expressing regulatory dendritic cells in cancer and chronic infection. J Mol Med. 2008 Feb;86(2):145-60. Epub
2007 Sep 18.
225

Maren Depke
References
Puccetti P.
On watching the watchers: IDO and type I/II IFN. Eur J Immunol. 2007 Apr;37(4):876-9.
Qu D, Wang Y, Esmon NL, Esmon CT.
Regulated endothelial protein C receptor shedding is mediated by tumor necrosis factor-alpha converting
enzyme/ADAM17. J Thromb Haemost. 2007 Feb;5(2):395-402. Epub 2006 Dec 7.
Que YA, François P, Haefliger JA, Entenza JM, Vaudaux P, Moreillon P.
Reassessing the role of Staphylococcus aureus clumping factor and fibronectin-binding protein by expression in
Lactococcus lactis. Infect Immun. 2001 Oct;69(10):6296-302.
Rachid S, Ohlsen K, Wallner U, Hacker J, Hecker M, Ziebuhr W.
Alternative transcription factor σ B is involved in regulation of biofilm expression in a Staphylococcus aureus mucosal
isolate. J Bacteriol. 2000 Dec;182(23):6824-6.
Rahman MM, McFadden G.
Modulation of tumor necrosis factor by microbial pathogens. PLoS Pathog. 2006 Feb;2(2):e4.
Rampone H, Martínez GL, Giraudo AT, Calzolari A, Nagel R.
In vivo expression of exoprotein synthesis with a Sae mutant of Staphylococcus aureus. Can J Vet Res. 1996 Jul;
60(3):237-40.
Rasmussen S, Nielsen HB, Jarmer H.
The transcriptionally active regions in the genome of Bacillus subtilis. Mol Microbiol. 2009 Sep;73(6):1043-57. Epub
2009 Aug 4.
Recsei PA, Gruss AD, Novick RP.
Cloning, sequence, and expression of the lysostaphin gene from Staphylococcus simulans. Proc Natl Acad Sci U S A.
1987 Mar;84(5):1127-31.
Ricart-Jané D, Rodríguez-Sureda V, Benavides A, Peinado-Onsurbe J, López-Tejero MD, Llobera M.
Immobilization stress alters intermediate metabolism and circulating lipoproteins in the rat. Metabolism. 2002 Jul;
51(7):925-31.
Ricquier D, Bouillaud F.
Mitochondrial uncoupling proteins: from mitochondria to the regulation of energy balance. J Physiol. 2000 Nov 15;
529 Pt 1:3-10.
Riegert P, Wanner V, Bahram S.
Genomics, isoforms, expression, and phylogeny of the MHC class I-related MR1 gene. J Immunol. 1998 Oct 15;
161(8):4066-77.
Riehle KJ, Campbell JS, McMahan RS, Johnson MM, Beyer RP, Bammler TK, Fausto N.
Regulation of liver regeneration and hepatocarcinogenesis by suppressor of cytokine signaling 3. J Exp Med. 2008
Jan 21;205(1):91-103. Epub 2007 Dec 24.
Rivett AJ, Bose S, Brooks P, Broadfoot KI.
Regulation of proteasome complexes by gamma-interferon and phosphorylation. Biochimie. 2001 Mar-Apr;83(3-4):
363-6.
Rizki G, Arnaboldi L, Gabrielli B, Yan J, Lee GS, Ng RK, Turner SM, Badger TM, Pitas RE, Maher JJ.
Mice fed a lipogenic methionine-choline-deficient diet develop hypermetabolism coincident with hepatic
suppression of SCD-1. J Lipid Res. 2006 Oct;47(10):2280-90. Epub 2006 Jul 8.
Robinson JM.
Phagocytic leukocytes and reactive oxygen species. Histochem Cell Biol. 2009 Apr;131(4):465-9. Epub 2009 Feb 18.
Rogasch K, Rühmling V, Pané-Farré J, Höper D, Weinberg C, Fuchs S, Schmudde M, Bröker BM, Wolz C, Hecker M,
Engelmann S.
Influence of the two-component system SaeRS on global gene expression in two different Staphylococcus aureus
strains. J Bacteriol. 2006 Nov;188(22):7742-58.
Rooijakkers SH, van Wamel WJ, Ruyken M, van Kessel KP, van Strijp JA.
Anti-opsonic properties of staphylokinase. Microbes Infect. 2005 Mar;7(3):476-84. Epub 2005 Feb 26.
Rooijakkers SH, Ruyken M, Roos A, Daha MR, Presanis JS, Sim RB, van Wamel WJ, van Kessel KP, van Strijp JA.
Immune evasion by a staphylococcal complement inhibitor that acts on C3 convertases. Nat Immunol. 2005 Sep;
6(9):920-7. Epub 2005 Aug 7.
226

Maren Depke
References
Roudkenar MH, Kuwahara Y, Baba T, Roushandeh AM, Ebishima S, Abe S, Ohkubo Y, Fukumoto M.
Oxidative stress induced lipocalin 2 gene expression: addressing its expression under the harmful conditions.
J Radiat Res (Tokyo). 2007 Jan;48(1):39-44. Epub 2007 Jan 16.
Rubin BY, Anderson SL, Xing L, Powell RJ, Tate WP.
Interferon induces tryptophanyl-tRNA synthetase expression in human fibroblasts. J Biol Chem. 1991 Dec 25;
266(36):24245-8.
Saito A, Motomura N, Kakimi K, Narui K, Noguchi N, Sasatsu M, Kubo K, Koezuka Y, Takai D, Ueha S, Takamoto S.
Vascular allografts are resistant to methicillin-resistant Staphylococcus aureus through indoleamine 2,3-dioxygenase
in a murine model. J Thorac Cardiovasc Surg. 2008 Jul;136(1):159-67. Epub 2008 May 22.
Sarvas M, Harwood CR, Bron S, van Dijl JM.
Post-translocational folding of secretory proteins in Gram-positive bacteria. Biochim Biophys Acta. 2004 Nov 11;
1694(1-3):311-27.
Schiffer R, Baron J, Dagtekin G, Jahnen-Dechent W, Zwadlo-Klarwasser G.
Differential regulation of the expression of transporters associated with antigen processing, TAP1 and TAP2, by
cytokines and lipopolysaccharide in primary human macrophages. Inflamm Res. 2002 Aug;51(8):403-8.
Schindler CA, Schuhardt VT.
Lysostaphin: A new bacteriolytic agent for the Staphylococcus. Proc Natl Acad Sci U S A. 1964 Mar;51:414-21.
Schmidt F, Scharf SS, Hildebrandt P, Burian M, Bernhardt J, Dhople V, Kalinka J, Gutjahr M, Hammer E, Völker U.
Time resolved quantitative proteome profiling of host-pathogen interactions: The response of Staphylococcus
aureus RN1HG to internalisation by human airway epithelial cells. Proteomics. 2010 Aug;10(15):2801-11.
Schröder JM.
Epithelial antimicrobial peptides: innate local host response elements. Cell Mol Life Sci. 1999 Oct 1;56(1-2):32-46.
Schroder K, Hertzog PJ, Ravasi T, Hume DA.
Interferon-γ: an overview of signals, mechanisms and functions. J Leukoc Biol. 2004 Feb;75(2):163-89. Epub 2003
Oct 2.
Schwarz-Linek U, Höök M, Potts JR.
The molecular basis of fibronectin-mediated bacterial adherence to host cells. Mol Microbiol. 2004 May;52(3):
631-41.
Scriba TJ, Sierro S, Brown EL, Phillips RE, Sewell AK, Massey RC.
The Staphyloccous aureus Eap protein activates expression of proinflammatory cytokines. Infect Immun. 2008 May;
76(5):2164-8. Epub 2008 Mar 10.
Sekkaï D, Guittet O, Lemaire G, Tenu JP, Lepoivre M.
Inhibition of nitric oxide synthase expression and activity in macrophages by 3-hydroxyanthranilic acid, a tryptophan
metabolite. Arch Biochem Biophys. 1997 Apr 1;340(1):117-23.
Senn MM, Giachino P, Homerova D, Steinhuber A, Strassner J, Kormanec J, Flückiger U, Berger-Bächi B, Bischoff M.
Molecular analysis and organization of the sigma B operon in Staphylococcus aureus. J Bacteriol. 2005 Dec;
187(23):8006-19.
Sharpe AH, Wherry EJ, Ahmed R, Freeman GJ.
The function of programmed cell death 1 and its ligands in regulating autoimmunity and infection. Nat Immunol.
2007 Mar;8(3):239-45.
Shaw LN, Aish J, Davenport JE, Brown MC, Lithgow JK, Simmonite K, Crossley H, Travis J, Potempa J, Foster SJ.
Investigations into σ B -modulated regulatory pathways governing extracellular virulence determinant production in
Staphylococcus aureus. J Bacteriol. 2006 Sep;188(17):6070-80.
Shaw LN, Lindholm C, Prajsnar TK, Miller HK, Brown MC, Golonka E, Stewart GC, Tarkowski A, Potempa J.
Identification and characterization of σ S , a novel component of the Staphylococcus aureus stress and virulence
responses. PLoS One. 2008;3(12):e3844. Epub 2008 Dec 3.
Shenoy AR, Kim BH, Choi HP, Matsuzawa T, Tiwari S, MacMicking JD.
Emerging themes in IFN-γ-induced macrophage immunity by the p47 and p65 GTPase families. Immunobiology.
2007;212(9-10):771-84. Epub 2007 Nov 28.
Sieprawska-Lupa M, Mydel P, Krawczyk K, Wójcik K, Puklo M, Lupa B, Suder P, Silberring J, Reed M, Pohl J, Shafer W,
McAleese F, Foster T, Travis J, Potempa J.
Degradation of human antimicrobial peptide LL-37 by Staphylococcus aureus-derived proteinases. Antimicrob
Agents Chemother. 2004 Dec;48(12):4673-9.
227

Maren Depke
References
Siewert E, Bort R, Kluge R, Heinrich PC, Castell J, Jover R.
Hepatic cytochrome P450 down-regulation during aseptic inflammation in the mouse is interleukin 6 dependent.
Hepatology. 2000 Jul;32(1):49-55.
Silence K, Hartmann M, Gührs KH, Gase A, Schlott B, Collen D, Lijnen HR.
Structure-function relationships in staphylokinase as revealed by "clustered charge to alanine" mutagenesis. J Biol
Chem. 1995 Nov 10;270(45):27192-8.
Sinars CR, Cheung-Flynn J, Rimerman RA, Scammell JG, Smith DF, Clardy J.
Structure of the large FK506-binding protein FKBP51, an Hsp90-binding protein and a component of steroid
receptor complexes. Proc Natl Acad Sci U S A. 2003 Feb 4;100(3):868-73. Epub 2003 Jan 21.
Singh VK, Moskovitz J.
Multiple methionine sulfoxide reductase genes in Staphylococcus aureus: expression of activity and roles in
tolerance of oxidative stress. Microbiology. 2003 Oct;149(Pt 10):2739-47.
Sinha B, François PP, Nüsse O, Foti M, Hartford OM, Vaudaux P, Foster TJ, Lew DP, Herrmann M, Krause KH.
Fibronectin-binding protein acts as Staphylococcus aureus invasin via fibronectin bridging to integrin alpha5beta1.
Cell Microbiol. 1999 Sep;1(2):101-17.
Sinha B, Herrmann M, Krause KH.
Is Staphylococcus aureus an intracellular pathogen? Trends Microbiol. 2000 Aug;8(8):343-4. Response to Lowy, 2000.
Siqueira JA, Speeg-Schatz C, Freitas FI, Sahel J, Monteil H, Prévost G.
Channel-forming leucotoxins from Staphylococcus aureus cause severe inflammatory reactions in a rabbit eye
model. J Med Microbiol. 1997 Jun;46(6):486-94.
Smith EE, Malik HS.
The apolipoprotein L family of programmed cell death and immunity genes rapidly evolved in primates at discrete
sites of host-pathogen interactions. Genome Res. 2009 May;19(5):850-8. Epub 2009 Mar 19.
Sobke AC, Selimovic D, Orlova V, Hassan M, Chavakis T, Athanasopoulos AN, Schubert U, Hussain M, Thiel G,
Preissner KT, Herrmann M.
The extracellular adherence protein from Staphylococcus aureus abrogates angiogenic responses of endothelial cells
by blocking Ras activation. FASEB J. 2006 Dec;20(14):2621-3. Epub 2006 Oct 31.
Souba WW, Smith RJ, Wilmore DW.
Effects of glucocorticoids on glutamine metabolism in visceral organs. Metabolism. 1985 May;34(5):450-6.
Speciale C, Hares K, Schwarcz R, Brookes N.
High-affinity uptake of L-kynurenine by a Na + -independent transporter of neutral amino acids in astrocytes.
J Neurosci. 1989 Jun;9(6):2066-72.
Spencer AG, Woods JW, Arakawa T, Singer II, Smith WL.
Subcellular localization of prostaglandin endoperoxide H synthases-1 and -2 by immunoelectron microscopy. J Biol
Chem. 1998 Apr 17;273(16):9886-93.
Staeheli P, Prochazka M, Steigmeier PA, Haller O.
Genetic control of interferon action: mouse strain distribution and inheritance of an induced protein with
guanylate-binding property. Virology. 1984 Aug;137(1):135-42.
Stapleton PD, Taylor PW.
Methicillin resistance in Staphylococcus aureus: mechanisms and modulation. Sci Prog. 2002;85(Pt 1):57-72.
Steer SA, Corbett JA.
The role and regulation of COX-2 during viral infection. Viral Immunol. 2003;16(4):447-60.
Steinhuber A, Goerke C, Bayer MG, Döring G, Wolz C.
Molecular architecture of the regulatory Locus sae of Staphylococcus aureus and its impact on expression of
virulence factors. J Bacteriol. 2003 Nov;185(21):6278-86.
Strehl B, Seifert U, Krüger E, Heink S, Kuckelkorn U, Kloetzel PM.
Interferon-gamma, the functional plasticity of the ubiquitin-proteasome system, and MHC class I antigen
processing. Immunol Rev. 2005 Oct;207:19-30.
Sugita H, Kaneki M, Tokunaga E, Sugita M, Koike C, Yasuhara S, Tompkins RG, Martyn JA.
Inducible nitric oxide synthase plays a role in LPS-induced hyperglycemia and insulin resistance. Am J Physiol
Endocrinol Metab. 2002 Feb;282(2):E386-94.
228

Maren Depke
References
Sun SC, Ganchi PA, Ballard DW, Greene WC.
NF-κB controls expression of inhibitor IκBα: evidence for an inducible autoregulatory pathway. Science. 1993 Mar
26;259(5103):1912-5.
Swain MG.
I. Stress and hepatic inflammation. Am J Physiol Gastrointest Liver Physiol. 2000 Dec;279(6):G1135-8.
Swearingen KE, Loomis WP, Zheng M, Cookson BT, Dovichi NJ.
Proteomic profiling of lipopolysaccharide-activated macrophages by isotope coded affinity tagging. J Proteome Res.
2010 May 7;9(5):2412-21.
Taub DD, Lloyd AR, Conlon K, Wang JM, Ortaldo JR, Harada A, Matsushima K, Kelvin DJ, Oppenheim JJ.
Recombinant human interferon-inducible protein 10 is a chemoattractant for human monocytes and T lymphocytes
and promotes T cell adhesion to endothelial cells. J Exp Med. 1993 Jun 1;177(6):1809-14.
Taylor GA, Feng CG, Sher A.
Control of IFN-γ-mediated host resistance to intracellular pathogens by immunity-related GTPases (p47 GTPases).
Microbes Infect. 2007 Nov-Dec;9(14-15):1644-51. Epub 2007 Sep 14.
Taylor MW, Feng GS.
Relationship between interferon-γ, indoleamine 2,3-dioxygenase, and tryptophan catabolism. FASEB J. 1991 Aug;
5(11):2516-22.
Thakker M, Park JS, Carey V, Lee JC.
Staphylococcus aureus serotype 5 capsular polysaccharide is antiphagocytic and enhances bacterial virulence in a
murine bacteremia model. Infect Immun. 1998 Nov;66(11):5183-9.
Thomas SR, Mohr D, Stocker R.
Nitric oxide inhibits indoleamine 2,3-dioxygenase activity in interferon-γ primed mononuclear phagocytes. J Biol
Chem. 1994 May 20;269(20):14457-64.
Tolstrup AB, Bejder A, Fleckner J, Justesen J.
Transcriptional regulation of the interferon-gamma-inducible tryptophanyl-tRNA synthetase includes alternative
splicing. J Biol Chem. 1995 Jan 6;270(1):397-403.
Trinh DV, Zhu N, Farhang G, Kim BJ, Huxford T.
The nuclear IκB protein IκBζ specifically binds NF-κB p50 homodimers and forms a ternary complex on κB DNA. J Mol
Biol. 2008 May 23;379(1):122-35. Epub 2008 Apr 3.
Tsatsanis C, Androulidaki A, Venihaki M, Margioris AN.
Signalling networks regulating cyclooxygenase-2. Int J Biochem Cell Biol. 2006;38(10):1654-61. Epub 2006 Apr 25.
Tseng CF, Lin CC, Huang HY, Liu HC, Mao SJ.
Antioxidant role of human haptoglobin. Proteomics. 2004 Aug;4(8):2221-8.
Tu L, Moriya C, Imai T, Ishida H, Tetsutani K, Duan X, Murata S, Tanaka K, Shimokawa C, Hisaeda H, Himeno K.
Critical role for the immunoproteasome subunit LMP7 in the resistance of mice to Toxoplasma gondii infection.
Eur J Immunol. 2009 Dec;39(12):3385-94.
Uhlén M, Guss B, Nilsson B, Gatenbeck S, Philipson L, Lindberg M.
Complete sequence of the staphylococcal gene encoding protein A. A gene evolved through multiple duplications.
J Biol Chem. 1984 Feb 10;259(3):1695-702.
Urieli-Shoval S, Meek RL, Hanson RH, Eriksen N, Benditt EP.
Human serum amyloid A genes are expressed in monocyte/macrophage cell lines. Am J Pathol. 1994 Sep;
145(3):650-60.
Valderas MW, Hart ME.
Identification and characterization of a second superoxide dismutase gene (sodM) from Staphylococcus aureus.
J Bacteriol. 2001 Jun;183(11):3399-407.
Valderas MW, Gatson JW, Wreyford N, Hart ME.
The superoxide dismutase gene sodM is unique to Staphylococcus aureus: absence of sodM in coagulase-negative
staphylococci. J Bacteriol. 2002 May;184(9):2465-72.
Valle J, Toledo-Arana A, Berasain C, Ghigo JM, Amorena B, Penadés JR, Lasa I.
SarA and not σ B is essential for biofilm development by Staphylococcus aureus. Mol Microbiol. 2003 May;
48(4):1075-87.
van Belkum A, Verkaik NJ, de Vogel CP, Boelens HA, Verveer J, Nouwen JL, Verbrugh HA, Wertheim HF.
Reclassification of Staphylococcus aureus nasal carriage types. J Infect Dis. 2009 Jun 15;199(12):1820-6.
229

Maren Depke
References
van den Akker EL, Nouwen JL, Melles DC, van Rossum EF, Koper JW, Uitterlinden AG, Hofman A, Verbrugh HA,
Pols HA, Lamberts SW, van Belkum A.
Staphylococcus aureus nasal carriage is associated with glucocorticoid receptor gene polymorphisms. J Infect Dis.
2006 Sep 15;194(6):814-8. Epub 2006 Aug 8.
Van den Eynde BJ, Morel S.
Differential processing of class-I-restricted epitopes by the standard proteasome and the immunoproteasome. Curr
Opin Immunol. 2001 Apr;13(2):147-53.
van Erp K, Dach K, Koch I, Heesemann J, Hoffmann R.
Role of strain differences on host resistance and the transcriptional response of macrophages to infection with
Yersinia enterocolitica. Physiol Genomics. 2006 Mar 13;25(1):75-84. Epub 2005 Dec 13.
van Oosten M, van Amersfoort ES, van Berkel TJ, Kuiper J.
Scavenger receptor-like receptors for the binding of lipopolysaccharide and lipoteichoic acid to liver endothelial and
Kupffer cells. J Endotoxin Res. 2001;7(5):381-4.
van Waardenburg DA, Jansen TC, Vos GD, Buurman WA.
Hyperglycemia in children with meningococcal sepsis and septic shock: the relation between plasma levels of insulin
and inflammatory mediators. J Clin Endocrinol Metab. 2006 Oct;91(10):3916-21. Epub 2006 May 30.
Vanhamme L, Paturiaux-Hanocq F, Poelvoorde P, Nolan DP, Lins L, Van Den Abbeele J, Pays A, Tebabi P, Van Xong H,
Jacquet A, Moguilevsky N, Dieu M, Kane JP, De Baetselier P, Brasseur R, Pays E.
Apolipoprotein L-I is the trypanosome lytic factor of human serum. Nature. 2003 Mar 6;422(6927):83-7.
Vanhollebeke B, Pays E.
The function of apolipoproteins L. Cell Mol Life Sci. 2006 Sep;63(17):1937-44.
Vanhorebeek I, Van den Berghe G.
Hormonal and metabolic strategies to attenuate catabolism in critically ill patients. Curr Opin Pharmacol. 2004
Dec;4(6):621-8.
Vilcek J.
First demonstration of the role of TNF in the pathogenesis of disease. J Immunol. 2008 Jul 1;181(1):5-6.
Viswanathan K, Dhabhar FS.
Stress-induced enhancement of leukocyte trafficking into sites of surgery or immune activation. Proc Natl Acad Sci
U S A. 2005 Apr 19;102(16):5808-13. Epub 2005 Apr 7.
von Eiff C, Becker K, Machka K, Stammer H, Peters G.
Nasal carriage as a source of Staphylococcus aureus bacteremia. Study Group. N Engl J Med. 2001 Jan 4;344(1):11-6.
von Eiff C, Friedrich AW, Peters G, Becker K.
Prevalence of genes encoding for members of the staphylococcal leukotoxin family among clinical isolates of
Staphylococcus aureus. Diagn Microbiol Infect Dis. 2004 Jul;49(3):157-62.
von Eiff C, Peters G, Becker K.
The small colony variant (SCV) concept – the role of staphylococcal SCVs in persistent infections. Injury. 2006
May;37 Suppl 2:S26-33.
Voyich JM, Braughton KR, Sturdevant DE, Whitney AR, Saïd-Salim B, Porcella SF, Long RD, Dorward DW, Gardner DJ,
Kreiswirth BN, Musser JM, DeLeo FR.
Insights into mechanisms used by Staphylococcus aureus to avoid destruction by human neutrophils. J Immunol.
2005 Sep 15;175(6):3907-19.
Wajant H, Pfizenmaier K, Scheurich P.
Tumor necrosis factor signaling. Cell Death Differ. 2003 Jan;10(1):45-65.
Wakasugi K, Slike BM, Hood J, Otani A, Ewalt KL, Friedlander M, Cheresh DA, Schimmel P.
A human aminoacyl-tRNA synthetase as a regulator of angiogenesis. Proc Natl Acad Sci U S A. 2002 Jan 8;99(1):173-
7. Epub 2002 Jan 2.
Wan G, Zhaorigetu S, Liu Z, Kaini R, Jiang Z, Hu CA.
Apolipoprotein L1, a novel Bcl-2 homology domain 3-only lipid-binding protein, induces autophagic cell death. J Biol
Chem. 2008 Aug 1;283(31):21540-9. Epub 2008 May 26.
Wang J, Maldonado MA.
The ubiquitin-proteasome system and its role in inflammatory and autoimmune diseases. Cell Mol Immunol. 2006
Aug;3(4):255-61.
230

Maren Depke
References
Weekers F, Van Herck E, Coopmans W, Michalaki M, Bowers CY, Veldhuis JD, Van den Berghe G.
A novel in vivo rabbit model of hypercatabolic critical illness reveals a biphasic neuroendocrine stress response.
Endocrinology. 2002 Mar;143(3):764-74.
Wertheim HF, Vos MC, Ott A, van Belkum A, Voss A, Kluytmans JA, van Keulen PH, Vandenbroucke-Grauls CM,
Meester MH, Verbrugh HA.
Risk and outcome of nosocomial Staphylococcus aureus bacteraemia in nasal carriers versus non-carriers. Lancet.
2004 Aug 21-27;364(9435):703-5.
Wertheim HF, Melles DC, Vos MC, van Leeuwen W, van Belkum A, Verbrugh HA, Nouwen JL.
The role of nasal carriage in Staphylococcus aureus infections. Lancet Infect Dis. 2005 Dec;5(12):751-62.
West NP, Pyne DB, Renshaw G, Cripps AW.
Antimicrobial peptides and proteins, exercise and innate mucosal immunity. FEMS Immunol Med Microbiol. 2006
Dec;48(3):293-304.
White LC, Wright KL, Felix NJ, Ruffner H, Reis LF, Pine R, Ting JP.
Regulation of LMP2 and TAP1 genes by IRF-1 explains the paucity of CD8 + T cells in IRF-1 -/- mice. Immunity. 1996
Oct;5(4):365-76.
Wiegard C, Frenzel C, Herkel J, Kallen KJ, Schmitt E, Lohse AW.
Murine liver antigen presenting cells control suppressor activity of CD4+CD25+ regulatory T cells. Hepatology. 2005
Jul;42(1):193-9.
Wiekowski MT, Chen SC, Zalamea P, Wilburn BP, Kinsley DJ, Sharif WW, Jensen KK, Hedrick JA, Manfra D, Lira SA.
Disruption of neutrophil migration in a conditional transgenic model: evidence for CXCR2 desensitization in vivo.
J Immunol. 2001 Dec 15;167(12):7102-10.
Wilmore DW.
Metabolic response to severe surgical illness: overview. World J Surg. 2000 Jun;24(6):705-11.
Wolz C, Geiger T, Goerke C.
The synthesis and function of the alarmone (p)ppGpp in firmicutes. Int J Med Microbiol. 2010 Feb;300(2-3):142-7.
Epub 2009 Sep 24.
Wray CJ, Mammen JM, Hasselgren PO.
Catabolic response to stress and potential benefits of nutrition support. Nutrition. 2002 Nov-Dec;18(11-12):971-7.
Wrona D.
Neural-immune interactions: an integrative view of the bidirectional relationship between the brain and immune
systems. J Neuroimmunol. 2006 Mar;172(1-2):38-58. Epub 2006 Jan 10.
Wu S, de Lencastre H, Tomasz A.
Sigma-B, a putative operon encoding alternate sigma factor of Staphylococcus aureus RNA polymerase: molecular
cloning and DNA sequencing. J Bacteriol. 1996 Oct;178(20):6036-42.
Wu SW, de Lencastre H, Tomasz A.
Recruitment of the mecA gene homologue of Staphylococcus sciuri into a resistance determinant and expression of
the resistant phenotype in Staphylococcus aureus. J Bacteriol. 2001 Apr;183(8):2417-24.
Wurfel MM, Kunitake ST, Lichenstein H, Kane JP, Wright SD.
Lipopolysaccharide (LPS)-binding protein is carried on lipoproteins and acts as a cofactor in the neutralization of LPS.
J Exp Med. 1994 Sep 1;180(3):1025-35.
Xie G, Bonner CA, Jensen RA.
Dynamic diversity of the tryptophan pathway in chlamydiae: reductive evolution and a novel operon for tryptophan
recapture. Genome Biol. 2002 Aug 29;3(9):research0051. Epub 2002 Aug 29.
Xong HV, Vanhamme L, Chamekh M, Chimfwembe CE, Van Den Abbeele J, Pays A, Van Meirvenne N, Hamers R, De
Baetselier P, Pays E.
A VSG expression site-associated gene confers resistance to human serum in Trypanosoma rhodesiense. Cell. 1998
Dec 11;95(6):839-46.
Xue H, Wong JT.
Interferon induction of human tryptophanyl-tRNA synthetase safeguards the synthesis of tryptophan-rich immunesystem
proteins: a hypothesis. Gene. 1995 Nov 20;165(2):335-9.
Yang D, Chen Q, Hoover DM, Staley P, Tucker KD, Lubkowski J, Oppenheim JJ.
Many chemokines including CCL20/MIP-3alpha display antimicrobial activity. J Leukoc Biol. 2003 Sep;74(3):448-55.
231

Maren Depke
References
Yang XL, Schimmel P, Ewalt KL.
Relationship of two human tRNA synthetases used in cell signaling. Trends Biochem Sci. 2004 May;29(5):250-6.
Yao L, Bengualid V, Berman JW, Lowy FD.
Prevention of endothelial cell cytokine induction by a Staphylococcus aureus lipoprotein. FEMS Immunol Med
Microbiol. 2000 Aug;28(4):301-5.
Yoo JY, Desiderio S.
Innate and acquired immunity intersect in a global view of the acute-phase response. Proc Natl Acad Sci U S A. 2003
Feb 4;100(3):1157-62. Epub 2003 Jan 22.
Zangar RC, Davydov DR, Verma S.
Mechanisms that regulate production of reactive oxygen species by cytochrome P450. Toxicol Appl Pharmacol. 2004
Sep 15;199(3):316-31.
Zeitlin PL, Lu L, Rhim J, Cutting G, Stetten G, Kieffer KA, Craig R, Guggino WB.
A cystic fibrosis bronchial epithelial cell line: immortalization by adeno-12-SV40 infection. Am J Respir Cell Mol Biol.
1991 Apr;4(4):313-9.
Zelante T, Fallarino F, Bistoni F, Puccetti P, Romani L.
Indoleamine 2,3-dioxygenase in infection: the paradox of an evasive strategy that benefits the host. Microbes Infect.
2009 Jan;11(1):133-41. Epub 2008 Oct 25.
Zhang K, McClure JA, Elsayed S, Conly JM.
Novel staphylococcal cassette chromosome mec type, tentatively designated type VIII, harboring class A mec and
type 4 ccr gene complexes in a Canadian epidemic strain of methicillin-resistant Staphylococcus aureus. Antimicrob
Agents Chemother. 2009 Feb;53(2):531-40. Epub 2008 Dec 8.
Zhaorigetu S, Wan G, Kaini R, Jiang Z, Hu CA.
ApoL1, a BH3-only lipid-binding protein, induces autophagic cell death. Autophagy. 2008 Nov 16;4(8):1079-82. Epub
2008 Nov 23.
Ziebandt AK, Becher D, Ohlsen K, Hacker J, Hecker M, Engelmann S.
The influence of agr and σ B in growth phase dependent regulation of virulence factors in Staphylococcus aureus.
Proteomics. 2004 Oct;4(10):3034-47.
Zocco MA, Carloni E, Pescatori M, Saulnier N, Lupascu A, Nista EC, Novi M, Candelli M, Cimica V, Mihm S, Gasbarrini G,
Ramadori G, Gasbarrini A.
Characterization of gene expression profile in rat Kupffer cells stimulated with IFN-α or IFN-γ. Dig Liver Dis. 2006
Aug;38(8):563-77. Epub 2006 Jun 27.
Zwacka RM, Zhang Y, Halldorson J, Schlossberg H, Dudus L, Engelhardt JF.
CD4 + T-lymphocytes mediate ischemia/reperfusion-induced inflammatory responses in mouse liver. J Clin Invest.
1997 Jul 15;100(2):279-89.
232

Maren Depke
P U B L I C A T I O N S
Peer-review articles
Depke M, Fusch G, Domanska G, Geffers R, Völker U, Schuett C, Kiank C.
Hypermetabolic syndrome as a consequence of repeated psychological stress in mice.
Endocrinology. 2008 Jun;149(6):2714-23. Epub 2008 Mar 6. PMID: 18325986
Depke M, Steil L, Domanska G, Völker U, Schütt C, Kiank C.
Altered hepatic mRNA expression of immune response and apoptosis-associated genes after acute and
chronic psychological stress in mice.
Mol Immunol. 2009 Sep;46(15):3018-28. Epub 2009 Jul 9. PMID: 19592098
Grabarczyk P, Przybylski GK, Depke M, Völker U, Bahr J, Assmus K, Bröker BM, Walther R, Schmidt CA.
Inhibition of BCL11B expression leads to apoptosis of malignant but not normal mature T cells.
Oncogene. 2007 May 31;26(26):3797-810. Epub 2006 Dec 18. PMID: 17173069
Grabarczyk P, Nähse V, Delin M, Przybylski G, Depke M, Hildebrandt P, Völker U, Schmidt CA.
Increased expression of Bcl11b leads to chemoresistance accompanied by G1 accumulation.
PLoS One. 2010 Sep 2;5(9). pii: e12532.
Depke M, Dinh Hoang Dang K, Breitbach K, Brinkmann L, Gesell Salazar M, Hammer E, Bast A, Steil L,
Schmidt F, Steinmetz I, Völker U.
Transcriptomic and proteomic characterization of mouse BMM after IFN-γ activation in a serum-free
system.
In preparation.
Poster/Talk
Analysis of the response of airway epithelial cells S9 to bacterial culture supernatants of Staphylococcus
aureus RN1HG – impact of SigB
M. Gutjahr, M. Depke, P. Hildebrandt, D. Hoppe, M. Degner, A. Kühn, E. Hammer, J. Pané-Farré, L. Steil,
M. Hecker, U. Völker
Meeting of Transregional Collaborative Research Center 34 “Pathophysiology of Staphylococci”
Kloster Banz/Bad Staffelstein, Germany, 28.-31.10.2008 (Poster)
Measuring expression signatures in host-pathogen interactions
M. Depke, U. Völker
Summer School “Pathophysiologie von Staphylokokken in der Post-Genom-Ära”
Vilm, Germany, 26.-29.09.2007 (Talk)
Chronic stress-induced metabolic disturbances in female BALB/c mice
M. Depke, C. Kiank, R. Geffers, C. Schütt, U. Völker
2 nd International Alfried Krupp Kolleg Symposium “stress, behaviour, immune response“
Greifswald, Germany, 9.-11.11.2006 (Poster)
233

Maren Depke
A F F I D A V I T / E R K L Ä R U N G
Hiermit erkläre ich, daß diese Arbeit bisher von mir weder an der Mathematisch-
Naturwissenschaftlichen Fakultät der Ernst-Moritz-Arndt-Universität Greifswald noch einer
anderen wissenschaftlichen Einrichtung zum Zwecke der Promotion eingereicht wurde.
Ferner erkläre ich, daß ich diese Arbeit selbständig verfaßt und keine anderen als die darin
angegebenen Hilfsmittel benutzt habe.
Greifswald, den 24. 09. 2010
………………………………………………………………………………………………………
Maren Depke
235

Maren Depke
A C K N O W L E D G M E N T S
Many thanks go to all supervisors for the chance to work on this topic, for the mentoring and
discussions, and notably for the possibility to use the excellent technical equipment in the
Department of Functional Genomics. I like to thank all colleagues for their contribution to cooperation
projects, and in particular for the nice working atmosphere, for the support and help in
challenging times during this work. And last but not least, sincere thanks are given to all friends
and especially to my family for the encouragement and care they gave to me.
Interfaculty Institute for Genetics and Functional Genomics,
Department of Functional Genomics, Ernst-Moritz-Arndt University of Greifswald, Germany
Sabine Ameling
Marc Burian
Dinh Hoang Dang Khoa
Melanie Gutjahr
… and all other colleagues …
Elke Hammer
Petra Hildebrandt
Ulrike Mäder
Marc Schaffer
Sandra Scharf
Frank Schmidt
Leif Steil
Uwe Völker
Liver gene expression pattern in a mouse psychological stress model
Gene expression profiling was performed on mouse samples from a psychological stress
model which was established and performed by Cornelia Kiank in the setting of the DFG-
Graduiertenkolleg GK840 „Wechselwirkungen zwischen Erreger und Wirt bei generalisierten
bakteriellen Infektionen“. After Affymetrix expression profiling method training by Tanja Töpfer
and Robert Geffers, array hybridizations were conducted in the lab of the array facility at the
Helmholtz-Zentrum für Infektionsforschung. Markus Grube provided the protocol for real-time
PCR mastermix. The work was supervised by Christine Schütt, Barbara Bröker, and Uwe Völker.
Cornelia Kiank, Christine Schütt, Barbara Bröker (Institute for Immunology and Transfusion Medicine,
Department of Immunology, Ernst-Moritz-Arndt University of Greifswald, Germany; C. K. now: David
Geffen School of Medicine, UCLA; Digestive Diseases Research Center and Center for Neurobiology of
Stress; Los Angeles, CA, USA)
Tanja Töpfer, Robert Geffers, Jan Buer, Jürgen Wehland (Helmholtz Centre for Infection Research,
Braunschweig, Germany)
Markus Grube, Heyo Krömer (Institute of Pharmacology, Ernst-Moritz-Arndt University of Greifswald,
Germany)
Uwe Völker (Interfaculty Institute for Genetics and Functional Genomics, Department of Functional
Genomics, Ernst-Moritz-Arndt University of Greifswald, Germany)
237

Maren Depke
Acknowledgments
Kidney gene expression pattern in an in vivo infection model
The mouse infection model was established and performed in the setting of the DFG
Sonderforschungsbereich SFB TR 34 “Pathophysiologie von Staphylokokken in der Post-Genom
Ära” by Tina Schäfer and Knut Ohlsen, who also contributed to study conception and design.
Marc Burian co-operated in the wet-lab-work for the second biological replicate of infected
samples and in the final data analysis and interpretation. The strain S. aureus RN1HG was
generated in the lab of Friedrich Götz and its isogenic sigB mutant by Jan Pané-Farré in the lab of
Michael Hecker. The work was supervised by Uwe Völker.
Tina Schäfer, Knut Ohlsen (Institute of Molecular Infection Biology, Julius-Maximilians University of
Würzburg, Germany)
Jan Pané-Farré, Susanne Engelmann, Michael Hecker (Institute of Microbiology, Ernst-Moritz-Arndt
University of Greifswald, Germany)
Friedrich Götz (Department of Microbial Genetics, Eberhards-Karls University of Tübingen, Germany)
Marc Burian, Uwe Völker (Interfaculty Institute for Genetics and Functional Genomics, Department of
Functional Genomics, Ernst-Moritz-Arndt University of Greifswald, Germany)
Gene expression pattern of bone-marrow derived macrophages after interferon-γ activation
Bone-marrow derived macrophages were generated and stimulated with IFN-γ by Kathrin
Breitbach. Proteome analysis was conducted by Dinh Hoang Dang Khoa. The study was
performed in the setting of the DFG Sonderforschungsbereich SFB TR 34 “Pathophysiologie von
Staphylokokken in der Post-Genom Ära”. The work was supervised by Uwe Völker.
Kathrin Breitbach, Antje Bast, Ivo Steinmetz (Institute of Medical Microbiology, Ernst-Moritz-Arndt
University of Greifswald, Germany)
Dinh Hoang Dang Khoa, Lars Brinkmann, Manuela Gesell Salazar, Elke Hammer, Leif Steil, Frank
Schmidt, Uwe Völker (Interfaculty Institute for Genetics and Functional Genomics, Department of
Functional Genomics, Ernst-Moritz-Arndt University of Greifswald, Germany)
Host cell gene expression pattern in an in vitro infection model
Infection experiments were performed in close collaboration with Melanie Gutjahr, who
worked on host proteome analysis and additionally accomplished control measurements and cell
culture. Petra Hildebrandt conducted FACS cell sorting. The strain S. aureus RN1HG was
generated in the lab of Friedrich Götz and its GFP expressing variant RN1HG pMV158GFP by
Leonard Menschner in the lab of Michael Hecker. The work was supervised by Uwe Völker. This
study was performed in the setting of the DFG Sonderforschungsbereich SFB TR 34
“Pathophysiologie von Staphylokokken in der Post-Genom Ära”.
Leonhard Menschner, Susanne Engelmann, Michael Hecker (Institute of Microbiology, Ernst-Moritz-
Arndt University of Greifswald, Germany)
Melanie Gutjahr, Petra Hildebrandt, Elke Hammer, Uwe Völker (Interfaculty Institute for Genetics and
Functional Genomics, Department of Functional Genomics, Ernst-Moritz-Arndt University of Greifswald,
Germany)
238

Maren Depke
Acknowledgments
Pathogen gene expression profiling – growth media and in vitro infection experiment studies
This study was performed in the setting of the DFG Sonderforschungsbereich SFB TR 34
“Pathophysiologie von Staphylokokken in der Post-Genom Ära” and the international
cooperation EU-IP-FP6-project BaSysBio (LSHG-CT2006-037469).
Infection experiments were performed in close collaboration with Melanie Gutjahr, who
additionally accomplished control measurements, and Petra Hildebrandt, who conducted
eukaryotic cell culture. Many thanks are given to Marc Schaffer for his valuable help during
preparation of internalized staphylococci, to Ulrike Mäder for her help concerning all aspects of
tiling array data, and to Marc Burian for the discussions on staphylococcal gene expression. The
work was supervised by Uwe Völker.
Jan Pané-Farré, Susanne Engelmann, Michael Hecker (Institute of Microbiology, Ernst-Moritz-Arndt
University of Greifswald, Germany)
Magda M. van der Kooi-Pol, Annette Dreisbach, Jan Maarten van Dijl (Department of Medical
Microbiology, University Medical Center Groningen / UMCG, Groningen, The Netherlands; A. D.
formerly: Department of Functional Genomics, Ernst-Moritz-Arndt University of Greifswald, Germany)
Hanne Jarmer (Center for Biological Sequence Analysis, Department of Systems Biology, Technical
University of Denmark, Lyngby, Denmark)
Pierre Nicolas, Aurélie Leduc, Philippe Bessières (INRA, Mathématique Informatique et Génome, Jouyen-Josas,
France)
Melanie Gutjahr, Petra Hildebrandt, Marc Schaffer, Marc Burian, Ulrike Mäder, Uwe Völker (Interfaculty
Institute for Genetics and Functional Genomics, Department of Functional Genomics, Ernst-Moritz-
Arndt University of Greifswald, Germany)
FURTHER COOPERATION PARTNERS
Piotr Grabarczyk, Christian A. Schmidt (Molecular Hematology, Department of Hematology and
Oncology, Ernst-Moritz-Arndt University of Greifswald, Germany)
FINANCIAL SUPPORT
DFG-Graduiertenkolleg / Research Training Group GK840
„Wechselwirkungen zwischen Erreger und Wirt bei generalisierten bakteriellen Infektionen“/
“Host-pathogen interactions in generalized bacterial infection“
DFG Sonderforschungsbereich / Transregional Collaborative Research Center SFB TR 34
“Pathophysiologie von Staphylokokken in der Post-Genom Ära”/
„Pathophysiology of Staphylococci in the Post-Genomic Era“
239