Single-Stranded DNA Phages

Modern Microbial Genetics, Second Edition. Edited by Uldis N. Streips, Ronald E. YasbinCopyright # 2002 Wiley-Liss, Inc.ISBNs: 0-471-38665-0 Hardback); 0-471-22197-X Electronic)6Single-Stranded DNA PhagesJ. EUGENE LECLERCMolecular Biology Division, Center for Food Safety and Applied Nutrition, US Food and DrugAdministration, Washington, District of Columbia 20204I. Introduction ................................... 145II. History and Classification. . ...................... 146III. The Virions .................................... 147IV. Genetic Organization............................ 148A. Isometric Phages ............................ 149B. Filamentous Phages.......................... 152V. Phage Life Cycles . .............................. 154A. Adsorption and Penetration................... 1541. Isometric Phages.......................... 1542. Filamentous Phages . ...................... 155B. DNAReplication............................ 1561. SS!RF ................................. 1572. RF!RF................................. 1593. RF!SS ................................. 160C. Phage Assembly and Release . . . ............... 1611. Isometric Phages.......................... 1612. Filamentous Phages . ...................... 162VI. RNABacteriophages ............................ 164A. Double-Stranded RNA Bacteriophages . ........ 164B. Single-Stranded RNABacteriophages . . ........ 165VII. Uses in Biotechnology ........................... 165A. Cloning and Sequencing Vectors ............... 165B. Hybridization Probes . . ...................... 168C. Site-Directed Mutagenesis .................... 168D. Plasmids with Phage Origins of Replication ..... 169E. Phage Display Technology Using FilamentousBacteriophages .............................. 170VIII. Concluding Remarks ............................ 170I. INTRODUCTIONAlthough the discovery of the smallest bacteriophagescan be traced to 1927, interest inthem intensified only after R. L. Sinsheimerdemonstrated in 1959 that the particles containa single-stranded DNAgenome. Theresearch spurred by these featuresÐthe smalland single-stranded DNAgenomeÐled overthe next 20 years to revelations of fundamen-

146 LECLERCtal information on the frugal uses of nucleotidesequence, of viral and host mechanismsfor DNAreplication, and of host functionsadapted for viral reproduction. The fX174phage that Sinsheimer studied was an importantparticipant in the first era of molecularbiology, largely phage biology, whichculminated in the total in vitro synthesis ofinfectious fX174 DNA. During the secondgreat era, fX174 DNAwas the first genometo be entirely sequenced and engineered derivativesof M13 and f1 genomes were themost commonly used vectors for sequencingcloned DNA.In this chapter we describe the life cycles,genetics, and biochemistry of the two generalclasses of single stranded DNAphages: thespherical or isometric phages, includingfX174 commonly called fX), S13, and G4,and the rod-shaped or filamentous phagesM13, fl, and fd. Two other classes of bacteriophages,the isometric single-stranded anddouble-stranded RNAphages, will only bebriefly discussed. Finally, we review some ofthe current and novel uses of the singlestrandedDNAphages, which continue tohave a significant place in genetics and biochemistryresearch.II. HISTORY ANDCLASSIFICATIONThe initial characterization of the smallestphages of enteric bacteria came from measuringthe sizes of bacterial viruses by filtrationand sedimentation analyses see reviewby Hoffman-Berling et al., 1966). S13 wasshown to have a particle diameter of 25mm, less than one-fourth the size of Tphages. Subsequent electron micrographs offX174 showed a similar size; the particleswere polyhedral and contained a knob orspike at each of 12 axes of symmetry Hallet al., 1959). Several fX-like phages havesince been identified, differing in serologicalproperties and preferences for hosts amongstrains of Escherichia, Shigella, and Salmonella.Taking into account their morphologyand nucleic acid content, the phages are currentlyclassified together as isometric DNAphages. Members of the isometric group thathave been studied to varying degrees includefX174, S13, G4, St-1, fK, and a; they seemto have conserved from a common ancestora similar morphology, genome organization,and protein functions, while differing significantlyin nucleotide sequence. The phages ofthe isometric group are virulent; that is, theykill and lyse their host bacteria at the end ofthe phage life cycle.The initial impetus for studying the smallphages was the limited size of their genomes;the potential existed for completely definingthe genetic content and organization ofhomogeneous DNAand for understandingall viral functions upon infection of host cellssee Sinsheimer, 1966, 1991). The realizationof these goals started with the early studiesof fX174 by Sinsheimer and on S13 by 1.and E.S. Tessman. Characterization of thephage DNAs showed unusual features thatcontrasted sharply with the known propertiesof double-stranded DNASinsheimer,1959a,b; Tessman, 1959): fX DNAreactedwith formaldehyde and was precipitatedwith lead ions, indicating that the aminogroups of the purine and pyrimidine baseswere accessible and not involved in basepairing; ultraviolet absorption of fX DNAwas dependent on temperature over a widerange, unlike double-stranded DNA, whichshows a sharp transition upon denaturation;and the density of the DNA, its light-scatteringproperties, and degradation by nucleasewere more like denatured than native DNA.An extensive series of radiobiological experimentsshowed that the inactivation constantsof S13 and fX phages for incorporated 32 Pwas near unity, that is, every disintegrationbreaks a single strand of DNAand causeslethality Tessman et al., 1957; Tessman,1959). Finally, determination of the basecomposition of fX DNArevealed ratios ofA:Tˆ 0.75 and G : C ˆ 1.3, unlike A T wand G C in double-stranded DNA; inwretrospect, the genome of the single-strandedDNAphage was the exception that provedChargaff's rule for complementary basepairing in double-stranded DNA. Since ex-

SINGLE-STRANDED DNA PHAGES 147periments with both intact phage and purifiedDNAexhibited unusual properties forgenomic DNA, it was concluded that eachphage particle contains one DNAmoleculein single-stranded form.The description of single-stranded DNAin fX and S13 phages led quickly to similarfindings for another class of small phages,the filamentous group, specific for E. colithat contain the male fertility factor F ‡ seePorter, this volume). One rationale for thesearch that yielded male-specific phages wasthat sensitivity to phage infection, amongother properties, might distinguish F andF ‡ or Hfr bacteria Loeb, 1960). In 1963 thedescriptions of three new single-strandedDNAphages were reported: f1 Zinder etal., 1963), fd Marvin and Hoffmann-Berling,1963), and M13 Hofschneider, 1963),isolated respectively from sewers in NewYork, Heidelberg, and Munich. It was soonshown that the isolates were closely relatedand, in fact, they may be considered mutantsof the same phage Salivar et al., 1964). Besidestheir specificity for male hosts, explainedby the requirement for F pili duringadsorption, other properties of these phagesdiffer significantly from the isometric group:no cell lysis occurs during their life cycles,and the phage particles are long, thin, andflexible, without heads or spikes. Indeed, itwas initially difficult to distinguish the phageparticles from the pili of host bacteria. Thesesmall phages, measuring about a micron inlength, are classified as filamentous DNAphages, or FV, for filamentous viruses;more specifically, M13, fl, and fd are classifiedas Ff phages, denoting filamentousphages that require the pili encoded by Fconjugative plasmids for infection seereview by Marvin and Hohn, 1969). Phagesof the filamentous group are ubiquitous andinclude If and IKe phages, which require, forinfection, pili specified by conjugative plasmidsof the I and N incompatibility groups,respectively, and phages that infect Pseudomonasand Xanthomonas.After infection by Ff phage, progenyphage are extruded into the medium as hostcells continue to grow, albeit at a slower rate;phage plaques are visible only because of theslower growth rate of infected cells. Amongthe samples from which the f1 turbid plaqueformerswere discovered Loeb, 1960), clearplaques were also evident and were separatelychosen for study; designated f2, theyturned out to be the first RNAphages describedLoeb and Zinder, 1961). IsometricRNAphages, now isolated on every continent,form a group about as diverse as theisometric DNAphages. They are obviouslyfascinating for study of their mode of replication,particularly exploited in the cases ofQb and f2. In addition, f2, MS2, R17, andQb became the workhorses for studies on themechanism of translation and the ideal substratesfor developing RNAcomposition andsequence methodology. For comprehensivediscussions on all aspects of these phages,the reader is referred to RNA Phages, editedby N.D. Zinder 1975), and reviews by Fiers1979) and Van Duin 1988).III. THE VIRIONSThe single-stranded DNAof the isometricphages is enclosed in a capsid of icosahedralsymmetry, its 20 faces requiring 60 identicalsubunits arranged on a sphere Fig. 1). It hasbeen noted that such an arrangement is commonlyfound for virus constructionÐicosahedralsymmetry allows the subunits toenclose a larger volume than other types ofsymmetry see review by Denhardt, 1977).Sixty molecules of gene F protein fulfill thisrole as the major capsid protein. At each of 12vertices of the icosahedron is a spike or knobthat may be considered a short, primitive tail;the five gene G proteins and one gene Hprotein that compose each spike are likelyinvolved in host recognition and attachmentto the cell surface. Within the icosahedralshell of 25 nm to 36 nm including the spikes),the DNAforms a densely packed core, itsphosphate groups neutralized by polyaminesand the positive charges of virion proteins. Inaddition, the core contains a minor virionprotein, from gene J, which may be involvedin condensation of the viral DNA.

148 LECLERCFig. 1. Schematic representation of the fX174icosahedron. A spike of gene H and gene G proteins,represented by the filled area, is located at the apexof each fivefold axis of symmetry. It is surrounded bygene F coat proteins, represented by stippled areas.Reproduced from Hayashi, 1978, with permission ofthe publisher.)In sharp contrast to the fixed fX spheres,the size of filamentous phage particles is determinedby the length of DNAcontainedwithin them, suggesting a fundamentally differentmechanism for phage morphogenesis.Approximately 1% of Ff phage preparationsis composed of ``miniphage'' particles Fig. 2),with 0.2 to 0.5 times the length of normalparticles and DNAGriffith and Kornberg,1974; Enea and Zinder, 1975; Hewitt, 1975),and ``polyphages,'' which contain multiplegenomes in multiple-length particles Salivaret al., 1967; Scott and Zinder, 1967). Anormal particle of 900 nm length and 6 to 9nm width contains a circular strand of DNAin a protein coat of five proteins reviewed byMarvin, 1998). The particle is a protein tubesheathing the DNA, made up of about 2700molecules of the a-helical gene VIII protein,its subunits overlapping and likened to scaleson a fish Marvin et al., 1974). At one end ofthe tube are four or five copies each of geneIII and gene VI proteins, involved in adsorptionof phage at the tip of an F pilus, and atthe other end are similar numbers of geneVII and gene IX proteins. The DNAis embeddedin the 2 nm core of the particle, itscircular single strand lying like a stretchedoutloop in the tube. Afascinating result ofmorphogenesis is that the loop of DNAisalways oriented the same way, gene III sequencesat the gene III protein-bearing endof the particle and, at the other end, anintergenic region encoding the initiation sitesfor DNAreplication Webster et al., 1981).IV. GENETIC ORGANIZATIONThe inability to digest ends of fX174 DNAby exonuclease treatment led Sinsheimer tosurmise that the DNAis circular. This assumptionwas confirmed by physical studiesFiers and Sinsheimer, 1962) and electronmicroscopy Freifelder et al., 1964). GeneticFig. 2. Electron micrograph of an MI 3 bacteriophage filament and two miniphage particles. Courtesy ofJ. Griffith, University of North Carolina.)

SINGLE-STRANDED DNA PHAGES 149recombination tests using phage mutants alsoestablished that the genetic map of S13 iscircular Baker and Tessman, 1967). Indeed,determining the number and organizationof genes on the circle largely depended onthe analysis of phage mutants, which affectedhost range or plaque morphology orwere conditionally lethal. The conditionallethal mutants, either temperature sensitiveor suppressible, were the most useful becausethey identified the gene products essential forphage development; hundreds of mutants forboth the isometric and filamentous phageshave been collected. In order to enumeratethe phage genes, genetic complementationtests were used, wherein mutations areassigned to different genes if two conditionallethal mutants, infecting cells together undernonpermissive conditions high temperatureor in suppressor-free hosts), produced progenyphage. In this way seven genes in bothfX174 and S13, and eight genes in the Ffphages, were identified. Although thesenumbers are now revised with the availabilityof nucleotide sequence information, it shouldbe noted that the bulk of our knowledge onthe viral genes and their protein functionscomes from the work on these extensive setsof phage mutants. See Pratt, 1969, for areview of the early genetic work.)The new era for analysis of genome organizationstarted with the complete nucleotidesequence determination of fX174 DNASanger et al., 1977). Entire sequences arenow known for phages fX174 revised inSanger et al., 1978), G4 Godson et al.,1978) fd Beck et al., 1978), M13 van Wezenbeeket al., 1980), fl Beck and Zink, 1981),Hill and Peterson, 1982), and IKe Peeters, etal., 1985). Although much remains to belearned about the functions encoded inphage genes, the goal of a complete knowledgeof their organization is nearly realized.In addition the sequences provide a richsource of information for evolutionary description.In the isometric group, fX174with 5386 nucleotides and G4 with 5577 nucleotidesshow considerable variability; inaddition to the different lengths of thegenomes, the coding regions show 33% nucleotidechanges. Differences are particularlyevident in the region of the genome involvedin the regulation of DNAreplication; as discussedlater, these phages indeed have differentmechanisms for initiating replication.There is no significant homology betweenthe phage genomes of the isometric groupand the filamentous group. As anticipated,the nucleotide sequences of the Ff phagesM13 6407 nucleotides), fd 6408 nucleotides),and f1 6407 nucleotides) are nearlyidentical, showing 97±99% homology, andmost of the base substitutions do not resultin amino acid changes. Although the geneorganization for the Ff phages and the N-specific phage IKe is identical, overall homologyis only about 55%. Particular divergencehas occurred for the IKe proteinrequired for binding the infecting phage tohost pili, providing one molecular basis forthe host ranges of F-specific as opposed to.N-specific phages.The genetic organization for representativephages is depicted in Figures 3 and 4, givingfX174 for the isometric group and M13 forthe filamentous group. The maps are drawnto show clockwise transcription of fX DNAand counterclockwise transcription of M13DNA. The polarities of the mRNAs of bothphages in fact correspond to that of the viralDNAs, making the single-stranded DNAphages plus-strand viruses; that is, all transcriptionfor gene expression occurs on thecomplementary, minus strand of replicatedphage DNA. The gene-encoded functionsare also summarized in the figures; overallsimilarities between the phages may be noted,particularly in the clustering the initiationsites for complementary strand synthesis ofcommon functions for DNAreplication andphage morphogenesis. For the following discussion,however, it is best to diverge hereand treat the groups separately.A. Isometric PhagesThe fX genome contains 11 genes Fig. 3),roughly grouped corresponding to the functionsof phage DNAreplication and phage

150 LECLERCFig. 3. Genetic organization and gene products of fX174 DNA. The numbers in the inner circle indicatethe first nucleotide of the initiation codon for the respective protein, numbered from a unique Pst I restrictionendonuclease cleavage site position 5386/1). The protein functions if known) and their molecular weightsderived from the DNA sequence) are given outside the circle. The direction of transcription for the majortranscripts on fX DNA and the approximate positions of the initiation sites for complementary strandsynthesis n 0 recognition site) and viral strand synthesis are indicated by arrows. IR, intergenic region.Reproduced from Baas, 1985, with permission of the publisher.)morphogenesis. Transcription starts at threepromoters preceding clusters of genes whoseproducts are used at different stages, or indifferent amounts, during the life cycle:genes Aand A*, for proteins controllingthe early functions of DNAreplication andshutting off host DNAsynthesis; genes B, C,and K, for proteins involved in early steps ofcapsid morphogenesis and DNAmaturationgene K function is unknown); and genes D,E, J, F, G, and H, whose products are usedin phage morphogenesis and host cell lysis. Amain mRNAterminator is located betweengenes H and A; other termination signalshave been mapped, but since readthroughoccurs, the potential exists for more transcriptionof the morphogenesis genes whoseproducts are needed in greater supply seeFujimura and Hayashi, 1978). In additionto the coding sequences, untranslated intergenicregions IR) ranging from 8 to 110nucleotides are found at the borders of genesJ, F, G, H, and A. Although their completefunctions are probably not known, they arecertainly used efficiently: all contain a ribosomebinding site for the proximal gene; theH/Aspace has the gene Apromoter; andthe IR between F and G contains a recognitionsite for proteins that initiate DNAreplication.

SINGLE-STRANDED DNA PHAGES 151Fig. 4. Genetic organization and gene products of Ff phage DNA, represented by the M13 genome.Proteins and sequence positions are noted as in Figure 3; numbering begins at a unique HindII cleavage site6407/1). The direction of transcription is indicated by the arrow and the location of promoters is shown bybars outside the circle. Promoters indicated by open bars are not active in vivo; see text.) IG, the mainintergenic region; terminator, the main termination site for transcription. Reproduced from Baas, 1985, withpermission of the publisher.)The most astonishing result revealed bythe fX nucleotide sequence, combined withprotein and mutant analyses, was the discoveryof overlapping genes, or one stretch ofDNAcoding for more than one proteinBarrell et al., 1976; Smith et al., 1977; Weisbeeket al., 1977). As shown in Figure 3, geneB lies totally within gene A, and gene Ewithin gene D. Sequence analysis of G4DNArevealed an eleventh gene, K, thatspans both genes Aand C; it is also presentin fX174 Shaw et al., 1978). The mRNAsfor overlapping genes are translated fromribosome binding sites within the precedinggenes and read in different frames. Indeed,for five nucleotides that overlap genes A, C,and K, all three possible reading frames areused! Apriori one might expect that thesimultaneous use of a coding region in tworeading frames would severely restrict thesequence differences between the genes infX and G4. In that sense the 22±23% nucleotidechanges observed for overlappinggenes in the two phages are extraordinary;necessarily fewer nucleotide differences arethird position or other conservative changes,leading to higher than average amino aciddifferences Godson et al., 1978).Another translational control mechanismserves to expand the use of the fX genome ingene A. The 37 kDa gene A* protein isformed by reinitiation of translation at anAUG codon within gene A mRNA, whichencodes the entire 56 kDa protein Linney

152 LECLERCand Hayashi, 1974). Proteins of differentfunctions are specified. The same translationalphase is used, so the amino acid sequenceof the A* protein is identical to thecarboxy-terminal half of Aprotein. Accordingly,nonsense mutations that block synthesisof gene A* protein also terminate gene Aprotein, but not the overlapping gene B proteinread in a different frame Weisbeek etal., 1977).Although there can be argument aboutconsidering the Aand A* sequences as separategenes, it is clear that five proteins A,A*, B, C, and K) share a DNA sequencethrough the use of transcriptional and translationalcontrol mechanisms. These extensionsof the ``one gene±one protein''hypothesis display a variety of means forthe frugal usage of nucleotide sequence inthe small phage genomes for ``deriving themost protein from the least DNA'' Pollocket al., 1978). The extra coding capacity isequivalent to nearly 1500 bp, 27% morethan if the DNAwere used only once Weisbeekand van Arkel, 1978). Maybe constraintson DNAcontent imposed by thesize of the icosahedral phage particles, orthe morphogenesis process, led to theexpanded use of available sequence. Mutantsof fX174 have been constructed to test themaximum genome size that can be packaged;inserts greater than 3±4% of the genomewere highly unstable Russell and Miller,1984). It has been noted that in the filamentousphages, with little constraint on genomesize, fX-like overlapping genes do not occurand use is made of a prominent intergenicregion Kornberg and Baker, 1992).Multiple uses for shared DNAsequencealso present interesting puzzles for describingthe evolution of new protein function.For instance, gene E protein, required forhost cell lysis, may have evolved its membrane-relatedfunctions by mutagenesis of apreexisting gene D. The third position of ahigh proportion of gene D codons containsU; the overlapping gene E, its reading frameone nucleotide downstream, then containsthe U residues in the second codon position.Such codons specify amino acids, such asleucine, that are hydrophobic, of the kindfound in polypeptides that interact with thecell membrane Barrell et al., 1976).B. Filamentous PhagesThe ten gene products of the Ff genome canbe assigned to three functional groups, theirgenes arranged counterclockwise on the mapof Figure 4: DNAreplication proteins, theproducts of genes II, X, and V; the coatproteins from genes VII, IX, VIII, III, andVI; and morphogenesis proteins from genes Iand IV. Promoters have been identified invitro in front of all genes except VII andIX, although not all of these promoters areused during infection; apparently the in vivorequirements for transcription are morestringent than those in vitro Smits et al.,1984). Two intergenic regions exist. Asmallone of 59 nucleotides between genes VIII andIII contains the main transcription terminator,so transcription from strong upstreampromoters leads to much more expression ofgene VIII major coat protein) than gene IIIminor coat protein) from its promoter. Thelarge intergenic region between genes IV andII, referred to as IG, contains multiple regulatoryelements for replication, transcription,and phage morphogenesis. Few nucleotidesotherwise separate the genes of Ff phages, oroverlaps of a few nucleotides occur, so thatthe 5 0 regulatory elements promoters andribosome binding sites) for several geneslie within the coding sequences of precedinggenes. In phage Ike 55% identity to theFf genome), the genetic organization ishighly similar, yet the controlling elementsare quite different Stump et al., 1997). InFf phage, for instance, translation of proteinVII is coupled to that of protein V, whilein Ike it is not coupled. Yet, the same levelsof protein are produced Madison-Antenucciand Steege, 1998). It is suggested that thereis a biological or evolutionary significanceto maintaining the same basic genetic arrangementin the filamentous bacteriophages,while control mechanisms haveevolved.

SINGLE-STRANDED DNA PHAGES 153The sequence of the 508-nucleotide intergenicregion, IG, of the Ff phage genomes isgiven in Figure 5, drawn to depict the extensivesecondary structure of the multiregulatoryunit. At its ends are the stop site andstart site for gene IV and gene II proteins,respectively, and within the region are theorigins for both complementary and viralstrand replication, a rho-dependent transcriptionterminator, a sequence requiredfor morphogenesis of the phage, and thegene II promoter. Transcription terminationin the presence of the host rho factor occursin the Ahairpin map position 5565 in Fig.5) on the 5 0 side of IG. Hence no mRNAissynthesized in the rest of the intergenicregion. That IG is the business end of thegenome of phage DNAreplication and packagingis evidenced by the propagation forminiphage, containing IG but no intact cistrons;in the presence of wild-type helperphage to provide phage proteins, these subparticlesoutgrow the wild-type phage Griffithand Kornberg, 1974; Enea and Zinder,1975; Hewitt, 1975). Furthermore insertionof IG into the pBR322 cloning vector yields,again in the presence of helper phage, transducingphage particles that contain singlestranded,chimeric plasmid DNAClearyand Ray, 1980; Dotto et al., 1981). Thereforethe noncoding region contains the cis-actingelements sufficient for controlling the initiationand termination reactions of DNAreplicationand for initiating DNApackagingreviewed by Zinder and Horiuchi, 1985).Although fX-like, out-of-frame overlappinggenes are not found in the Ff phages,reinitiation of translation at an internalAUG codon in gene II mRNA does serveto expand the use of the genome in a manneranalogous to gene Aof fX, and curiously, ina gene encoding a protein of analogous functionin DNAreplication. The carboxy-terminal27% of gene II protein is identical toFig. 5. DNA sequence of the intergenic region IG) of Ff phage DNA, drawn to indicate potentialsecondary structures. The fd DNA sequence is given, with numbering from a unique Hind II cleavage site6408/1), and base exchanges in other Ff phages are indicated as follows: parentheses, exchange in fl only;brackets, exchange in M13 only; no parentheses or brackets, exchanges in both fl and M13. See text fordiscussions of functions encoded in hairpins A±E and initiation sites for complementary strand c-strand) andviral strand v-strand) DNA synthesis. Reproduced from Beck and Zink, 1981, with permission of thepublisher.)

154 LECLERCgene X protein, so that gene products of42 kDa and 13 kDa are encoded in the samesequence and reading frame Yen and Webster,1981). Studying the in vivo functions ofthe products from in-frame overlappinggenes presents an interesting challengeÐinthe overlapping region, a mutation in onegene necessarily affects the other gene, unlikeout-of-frame overlaps. Fulford and Model1984) did a clever analysis by site-specificallymutagenizing the gene X initiationcodon, AUG, to an amber termination)UAG codon so that no gene X protein ismade. They then propagated the phage inamber-suppressing cells carrying gene X ona plasmid. In the absence of the complementinggene X on the plasmid, the amber mutantcould not grow; it produced no singlestrandedDNAfor progeny virions, indicatinga specific requirement for gene X proteinfunction in phage DNAsynthesis.V. PHAGE LIFE CYCLESThe consequences for the host bacteria aredrastically different in the cases of infectionby isometric phages or filamentous phages.Isometric phages are virulent and lyse thehost cell after about 30 minutes of infection,which yields about 200 progeny phage percell. Filamentous phages grow in their hostas parasites, more akin to plasmids, and thehost cells continue to grow, divide, and extrudeseveral hundred phage per cell generation.Avariety of control mechanisms areused throughout the phage life cycles, eitherto ``manage their own course of developmentand streamline the exploitation of the haplesshost'' in the case of isometric phages, orin contrast for the Ff phages, to ``stage arapid coup and then install a stable newregime'' Fulford et al., 1986). Although thephage groups will often be treated togetherin the following description of life cycles, itwill be useful to keep in mind the fundamentallydifferent phage and host relationshipsthat ensure efficient viral reproduction. Forpurposes of the description, stages of the lifecycle will be designated adsorption and penetration,DNAreplication, and phage assemblyand release, but in reality these stagesoverlap and contain several steps.A. Adsorption and PenetrationThe extraordinary accomplishment of phagesis that one infecting particle so quickly andefficiently gains control of the host syntheticapparatus, in competition with the hostgenome a thousand times its size. The adsorptionand penetration stages at the onset ofsuccessful infection unfortunately remainthe least well-understood aspects of the infectionprocess for the small phages. Ageneralmodel was proposed for a ``pilot'' proteinthat guides the phage and its DNAthroughmany stages of development, from cell surfaceinteractions through DNAreplicationJazwinski et al., 1975a). Although propertiesof the proteins encoded by gene H of theisometric phages and gene III of the Ff phagesfulfill many of the criteria for a pilot function,evidence does not strongly support suchmultifunctional roles for one phage proteinsee Tessman and Tessman, 1978; Raschedand Oberer, 1986). But it is still useful toconsider how the phage genome is deliveredto a site available for immediate replication,as a consequence of the phage adhering to thecell surface and penetrating the cell membrane.1. Isometric phagesFor the isometric phages, the 12 spikes of thecapsid are the adsorption organelles, eachcomposed of one molecule of H protein surroundedby 5 molecules of G protein. Asmight be expected, phage mutations thataffect the host range and adsorption ratesmap not only in genes G and H, but also ingene F for the coat protein see Tessman andTessman, 1978). The early work that pointedto adsorption occurring at a spike involvedsome good inference. Hutchison et al. 1967)infected E. coli with both wild-type fXphage and host range mutants in gene H),yielding progeny phage that containedcapsids with varying proportions of wildtypeand mutant subunits. Homogeneousfractions of phage from the mixed progeny

SINGLE-STRANDED DNA PHAGES 155could be prepared by electrophoresis, sothese workers compared the host rangephenotype of hybrid capsids with their subunitcomposition, estimated from electrophoreticmobility. With the assumption thatall subunits must be mutant to adsorb to theextended host, the results nicely conformed,on a statistical basis, with 5 subunits at theadsorption site. Since a fivefold axis of symmetryexists at each of the 12 vertices of thefX icosahedral particle, Hutchison et al.1967) suggested that a spike at the vertex isresponsible for adsorption. Their inferencewas supported by the results of experimentson the adsorption of phages to cell wall fragmentsand their attachment, viewed by electronmicroscopy, at the tip of a vertexBrown et al., 1971). The cellular componentfor phage interaction is lipopolysaccharideLPS) in the outer cell membrane Incardonaand Selvidge, 1973; Bruse et al., 1991), and itis the H protein that interacts specificallywith LPS Suzuki et al., 1999). Differentbacteriophages have adopted a wide varietyof specific cell surface molecules as receptors;even the closely related phages fX174 andS13 differ, in that N-acetylglucosamine onthe lipopolysaccharide chain is required forbinding fX, but not S13 Jazwinski et al.,1975b). S13 adsorbs to specific polysaccharides,requiring either hexose with substitutedalpha-d-glycopyranosyl groups or heptosewith the same groups in a different positionBruse et al., 1989).The adsorption stage of infection is reversible;infectious phage particles can be detachedfrom cells, for instance, by EDTAtreatment. The next complex steps of``injecting'' phage DNAinto the cell arepoorly understood. For most bacteriophagesan eclipse stage has been defined duringwhich infectivity of the phage particles forfresh cells is irreversibly lost, presumablybecause of conformational changes in thecapsid proteins and/or their removal fromviral DNA. The removal of coat proteinsduring eclipse of fX and S13 is coupled toreplication of the viral DNA. Curiously, atleast one of the gene H proteins remainsattached to the DNAenters the cell and isinvolved, as assessed by mutant studies, inthe replication of phage DNAJazwinskiet al., 1975a). The possible multifunctionalroles of H protein led to the aforementionedproposal for its pilot function, but the functions)of the wild-type H protein at eachstep of infection need to be determined inorder to adequately evaluate the model. Anotherintriguing result is that both adsorbedphage and its replicated DNAoccur at zonesof adhesion between the outer and innermembranes of E. coli Bayer and Starkey,1972). Does the cell surface receptor directinfecting DNAto membrane sites withimmediate access to the DNAreplicationmachinery? These observations need clarification,not only for understanding phagebiology but also because the phage systemsprovide exceptional models for exploringhow viruses so effectively gain control ofhost metabolism.2. Filamentous phagesThe specificity of filamentous phages formale host bacteria provides the clue thatadsorption involves the conjugative pili oftransmissible plasmid-bearing strains. Electronmicrographs indeed show adherence atthe tips of a phage filament and a pilus, whileF-specific RNAphages are shown attachedto the sides of F pili. It is considered unlikely,however, that the pilus is used to conductphage DNAin a manner analogous toconjugation see Marvin and Hohn, 1969).Rather, the whole pilus with attached phagemay be retracted into the cell surface, or thephage may be guided down the pilus to areceptor on the cell surface. The ability ofcells lacking pili to propagate phage, for instanceafter transfection using viral DNA,indicates that pili are used solely at the adsorptionand/or penetration stages of infection.Acell envelope protein, encoded by theE. coli fii locus, is required for penetration,but not adsorption, by phages f1 and IKeSun and Webster, 1986). Since fii mutantsare proficient for infection by the isometricDNAphages and RNAphages, as well as for

156 LECLERCconjugation, analysis of Fii function shouldaid in understanding the penetration step forthe filamentous phages.Much information is available on the rolesof Ff phage proteins during the adsorptionand penetration stages; an extensive literatureon the subject is thoroughly discussedby Rasched and Oberer 1986). Asummarypicture that we have, partly based on modelstudies Griffith et al., 1981), is as follows.Gene III protein is located at one end of thephage filament, probably in an ``adsorptioncomplex'' with gene VI protein. During adsorptionto the cell, the complex appearswith a ``knob-on-stem'' structure; the knobis the amino terminus of gene III proteinattached to the pilus tip and the stemanchoring the phage is the carboxyl terminusattached to the phage coat of gene VIII proteinGray et al., 1981). During eclipse, atransition occurs in the gene VIII protein itloses a-helical structure), contracting thephage filament to a spheroid as DNAisejected, its gene III end first, through a rupturedgene III protein end of the particle.The opposite end of the DNAloop, containingIG, remains associated with the phagecoat, possibly attached to the gene VII and/or gene IX proteins. How DNArelease fromthe phage proteins) is accomplished is notknown. The uncoating of the viral DNAiscoupled with its replication to a doublestrandedform on the inner cell membrane.Monomers of gene VIII protein are stored inthe membrane, so essentially the wholephage is resorbed during infection. Sincegene III protein from the phage coat is alsofound in the inner membrane, associatedwith the replicative form RF) of phageDNA, a pilot function for gene III proteinhas been suggested, analogous to that for fXgene F protein Jazwinski et al., 1973,1975a). Some specific objections to the proposedpilot function are that the gene IIIprotein is on the opposite end of the phagefrom IG, which controls the initiation ofDNAreplication, and that gene III proteinis not needed for phage DNAreplication,either in vitro or during transfection withpurified DNAsee Rasched and Oberer,1986). As with fX, the coordination ofevents during eclipse and the involvementof the cell membrane need further investigation.B. DNA ReplicationThe two special features of the genomes ofthe single-stranded DNAphages made themfascinating for studying their modes of reproduction.First, that the genomes are smallsuggests that even when frugally used, theyare mostly devoted to genes specifying phagemorphogenesis and structure, leaving littleroom for DNAreplication enzymes. Second,their single-stranded nature requires them tobe converted to a double-stranded form as aprerequisite for most DNAtransactions, inparticular, transcription for the expression ofany phage-encoded proteins that function inreplication. Necessarily the phages then relyon host enzymes for DNAreplication. Kornbergand Baker 1992) considered the mechanismsof viral replication as ``windows oncellular replication''; indeed, the workcarried out using the fX, G4, and M13 orfd) systems led to the discoveries of most ofthe proteins used for the replication of E. coliDNAsee Firshein, this volume).The usefulness of the small, singlestranded,homogeneous, and biologicallyactive phage DNAs as templates for DNAsynthesis was demonstrated in a dramaticexperiment reported in 1967. The so-calledGoulian, Kornberg, and Sinsheimer experimentshowed that DNApolymerase I fromE. coli, provided primer fragments and deoxyribonucleotideprecursors, could totallycopy pure fX174 circles. Synthesis wasaccurate since the products of the in vitroreaction, sealed by polynucleotide ligase,produced progeny phage in transfectionassays Goulian et al., 1967). The importantseries of enzymological studies that followedwere carried out using reconstituted in vitrosystems, starting with purified phage DNAsas templates for DNAreplication in extractsof uninfected cells R. B. Wickner et al.,1972; W. Wickner et al., 1973). Cell extracts

SINGLE-STRANDED DNA PHAGES 157were then fractionated and the purified componentsadded back together in reconstitutionassays to achieve replication activity.Alternatively, specific components from extractsof wild-type cells were added tomutant cell extracts, devoid of replicationactivity, in order to recover activity in complementationassays, much like genetic complementation.It was in using these assays todetect activity during the purification andanalysis of the DNAreplication proteins,and the template DNAs from fX, G4, orM13 phages that have different requirementsfor the initiation of DNAreplication, thatthe biochemical pathways for the conversionof viral single strands to double-strandedDNAwere reconstructed in vitro. For workingout the enzymology of later steps in replication,which utilize phage-encoded proteins,similar methods were employed starting withextracts of phage-infected cells.Preceding the enzymological work, pioneeringstudies by Sinsheimer and colleaguesdefined the intermediate structures of replicatedfX DNAobserved during the courseof infection see review by Sinsheimer, 1968).The studies relied on marking the infectingDNAmolecules with density isotopes andradioactivity, and then analyzing the in vivoreplication products in cell extracts by centrifugationto equilibrium in CsCl2 gradients,in order to follow the conversion ofsingle-stranded DNAto forms of newly synthesizedDNA. Phage DNA was distinguishedfrom the cellular DNAin extractsby transfection of spheroplasts, a methodthat had recently been worked out for fXDNAGuthrie and Sinsheimer, 1960). Thesewere the studies that defined the circular,double-stranded ``replicative form'' RF) ofDNAas the product of replication on aninfecting viral single strand SS). ParentalRF, consisting of the viral SS DNAtheplus strand) and a complementary nascentstrand the minus strand), is then replicatedin a semiconservative process to produce apool of progeny RF. No evidence for freeprogeny SS DNAin the cell was found; itsappearance only in phage particles impliedthat SS DNAsynthesis is coordinated withDNApackaging in the fX system. Ray andcoworkers, whose studies defined the replicationrequirements and intermediates duringM13 infection, found a pool of intracellularsingle strands late in infection, suggesting adifferent mode of packaging DNAin the Ffphages Ray and Sheckman, 1969; also seereview by Ray, 1977).From a combination of biophysical, genetic,and enzymological studies, we nowrecognize three steps in the replication ofphage DNA, for both the isometric andfilamentous phages: 1) SS!RF, the synthesisof a complementary strand to produceparental RF DNA, is carried out by hostenzymes that exist prior to infection; 2)RF!RF, the replication of parental RF toyield progeny RF DNA, requires phageencodedenzymes in addition to those ofthe host and produces RF DNAadequatefor phage-specific transcription; and 3)RF!SS, the replication of RF DNAto yieldthe progeny viral strands that are packagedinto phage particles. Acomplicating featureof RF replication is that the process is asymmetric,unlike the concerted leading and laggingstrand synthesis of most doublestrandedDNAs. That is, replication on anRF molecule uses the complementary strandas template to produce one progeny viralstrand, thereby forming another RF molecule;the ``old'' viral strand is peeled away,to re-enter the RF pool via the SS!RFmode) or to be sequestered for packaging inphage coats. Which course is followeddepends on the stage of infection and is subjectto elaborate controls. The following onlysummarizes the highlights of each step inphage DNAreplication, with emphasis onthose control mechanisms. All aspects aredescribed beautifully in DNA ReplicationKornberg and Baker, 1992) and detailedreviews of the research can be found in Marians1984) and Baas 1985).1. SS!RFThe reactions used in common by the singlestrandedDNAphages for complementary

158 LECLERCstrand synthesis are chain elongation byDNApolymerase III holoenzyme, gap fillingby DNApolymerase I, ligation of the duplexcircles by polynucleotide ligase, and formationof the superhelical RFI molecules byDNAgyrase DNAtopoisomerase II). Inaddition all of the phage DNAs are coatedwith single-stranded DNA-binding proteinSSB) upon infection; the protein is requiredfor replication, and it may protect the exposedsingle strands from endonucleolyticdegradation. The differences among thephages that made them favorite models forhost DNAreplication are in the initiationreactions for de novo synthesis on infectingsingle strands. As depicted in Figure 6, RNAprimer formation on the three representativephage DNAs is accomplished by three differentenzyme systems. In the case of the filamentousphages, the host RNApolymeraserecognizes the duplex regions of hairpins Band C Fig. 5) and synthesizes approximately30 nucleotides of RNAprimer on hairpin C.The duplex region is the only phage DNAnot coated with SSB, but its binding to theDNAreleased from the hairpin upon primerformation may terminate RNAsynthesis.This simple, efficient system makes use of aplentiful enzyme, RNApolymerase holoenzyme,for priming DNAreplication, and isFig. 6. Pathways for the initiation of complementarystrand synthesis by RNA priming on singlestrandedphage DNAs. SSB, single-stranded DNAbinding protein from E. coli. Reproduced fromBaas, 1985, with permission of the publisher.)well suited for the parasitic lifestyle of thefilamentous phages. In contrast, the virulentisometric phages use enzymes that are inshorter supply. G4 phage as well as St-1,a3, and fK) use E. coli primase, the DnaGprotein, for RNAprimer synthesis in asimple reaction like that of RNApolymerase.The hairpin recognized by DnaG primaseresides in the IR between genes F andG in G4 DNA, while complementary strandorigins of similar sequence are located betweengenes G and H in the DNAs of St-1,a3, and fK. The most complex reaction iscarried out on fX and S13) DNAsee Marians,1984). Aprepriming complex of hostproteins is constructed at the n 0 recognitionsite now called the primosome assemblysite or PAS) in the IR between genes F andG. The prepriming event is necessary forassociation of the DnaG primase, forming aprimosome that functions as a ``mobile promoter''for synthesizing RNAchains onfX174 DNAArai and Kornberg, 1981).Although the synthesis of one primer onfX DNAprobably suffices for the DNApolymerase III-mediated extension to thewhole complementary strand, the primingsystem can generate primers repeatedlyduring processive movement around theDNAcircle. The primosome moves in thetemplate 5 0 ! 3 0 direction, opposite the directionof primer formation and DNAchaingrowth. Since lagging strand synthesis ofchromosomal DNAoccurs in discontinuoussteps away from the replication fork, thesystem is an attractive model for the repeatedpriming needed for initiation of Okazakifragments see Firshein, this volume).The results of in vivo analyses and hostmutant studies are in total agreement withthe well-defined biochemical pathways forSS!RF conversion. What has not beendemonstrated in an in vitro system are functionsfor the products of Ff gene III andisometric phage gene H, implicated by phagemutant studies to have a role in complementarystrand synthesis. Their putative pilotfunctions could well be dispensable in solubleenzyme systems, or the pilot hypothesis

SINGLE-STRANDED DNA PHAGES 159may simply be wrong. It is clear, however,that synthesis of the parental RF occurs atthe membrane, coupled to the uncoating ofphage particles during penetration. Whetherthe discovery of membrane-associated functionschanges the current picture of the SS-RF replication pathway remains to be seen.2 . RF!RFKey participants in RF!RF replication arethe phage initiator proteins, the productsencoded by gene Aof the isometric phagesand gene II of the filamentous phages. Eachof these proteins is a site-specific endonuclease-ligasethat acts on the viral strand ofRF molecules, first to introduce a nick at theorigin of viral strand replication and, afterreplication of the strand, to religate the``old'' viral strand circle. The origin sequenceis located in IG of the Ff phage genome and,strikingly, in gene Aof the isometric phagescf. Figs. 3 and 5). The purpose of the nick isto provide a 3 0 -OH terminus for extension byDNApolymerase III holoenzyme. Otherproteins required for RF replication are thesingle-stranded DNA-binding protein SSB)and another host protein, the product ofthe E. coli rep gene. As a component of thereplication complex, the Rep protein acts asa helicase to unwind the duplex DNAofRF ahead of the replication fork, usingits ATPase activity to provide energy forstrand separation. SSB then coats the separatedstrands. Upon completion of one roundof replication, the viral strand is acted uponagain by the initiator protein, cleavingthe regenerated origin sequence to free theviral strand and ligating it to form a viralstrand circle. Progeny RF continues asymmetricsynthesis in this rolling circle modeGilbert and Dressler, 1968), while the displacedsingle-stranded circle is again incorporatedinto double-stranded DNAby theSS!RF mode. Hence two mechanisms forinitiation are used during RF!RF replication:covalent starts at the cleavage siteson RF viral strands for continuous viralstrand synthesis, and one of the processesof RNApriming for complementary strandsynthesis, which is required anew on eachviral strand.Although the fX gene Aprotein and Ffgene II protein have analogous functions inRF replication, they display several differencesin reaction mechanisms. Gene Aproteinbecomes covalently linked to the 5 0 -phosphate at the origin cleavage site, theenergy of the phosphodiester bond beingconserved throughout the replication cyclefor recircularization of the viral strand. Thereplicative intermediates are described aslooped rolling circles, consistent with the Aprotein-5 0 terminus remaining associatedwith the replication complex at the 3 0 growingend of the chain. This mechanism allowsgene Aprotein to act processively: the Aprotein is transferred to progeny RF duringthe concerted cleavage and ligation reaction,and the replication cycle continues. In contrast,gene II protein action is distributive;the protein is either released after viralstrand cleavage or it forms a weak complexwith the complementary strand. Hence thesubsequent replication step may utilize anothergene II protein, and the energy forcircularization comes from cleavage of theviral strand after replication. How the 5 0end of the linear viral strand becomes associatedwith its 3 0 end for closure is not clear.An event special to the isometric phages islikely a control step late in RF!RF synthesis.About halfway through infection, hostDNAsynthesis ceases, either because oneor more proteins used for replication ofboth phage and host DNAare in limitedsupply or because of the specific action of aphage-encoded protein. The best candidatefor the latter mechanism would be the geneA* protein, which results from the translationalrestart within gene A, because somemutants in the distal part of gene Afail toshut off host synthesis Martin and Godson,1975). As such, gene A* function may bepart of a control system for the switch toRF!SS synthesis; the production of progenyRF from viral strand circles also ceasesafter a burst of RF synthesis, and then DNAreplication is committed to the RF!SS stage

160 LECLERCfor the production of phage. No abruptcessation of RF!RF synthesis occurs duringfilamentous phage infection, and RF!RF synthesis occurs simultaneously withRF!SS synthesis, albeit at a reduced rate.Of course, host macromolecular synthesis iscontinuous for the steady-state productionof filamentous phage.3. RF!SSAlthough RF!SS replication uses theenzyme systems for viral strand synthesis alreadydescribed, phage-encoded proteinshave the predominant roles at this stage. Forthe isometric phages, SS DNAsynthesis iscoupled to phage maturation, wherein viralstrands are immediately encapsidated in aviral protein complex, the prohead, ratherthan being coated with SSB for subsequentdoubling up by complementary strand synthesis.The prohead is comprised of the morphogenesisproteins from genes B, D, F, G,and H, and import of DNAto the proheadutilizes the products of the maturation genesC and J. Since mutant forms of the morphogenesisproteins block SS DNAsynthesis,protein-DNAor protein-protein interactionsthat differ from those in RF!RF synthesismust be operating at the RF!SS stage. In thecase of the filamentous phages, SS DNAsynthesisand phage morphogenesis are notcoupled. Rather than being packed immediately,the nascent SS DNAis coated with aphage-specfic single-stranded DNA-bindingprotein, the product of gene V. Progeny SSDNAaccumulates in the cell in these nucleoproteincomplexes, to be transferred to themembrane for DNApackaging.Since the rolling-circle mode accountsfor viral strand synthesis during both theRF!RF and RF!SS stages of DNAreplication,what controls the switch that is essentialfor efficient viral reproduction? Theanswer probably lies in the absolute andrelative amounts of phage-induced proteinsin the host cell, which vary during the timecourse of infection; note, for instance, thatfor both the isometric and filamentousphages, E. coli SSB protein that coats thenascent SS strands in RF!RF synthesis isreplaced with phage-specific proteins duringRF!SS synthesis. The most developedmodel to explain the fate of SS DNAcomesfrom the work of Fulford and Model 1988a,b) on the control of f1 DNAreplication alsosee review by Fulford et al., 1986). By theirmodel, three multifunctional phage proteinsmodulate the rates of DNAreplication,favoring either RF multiplication or SSDNAdestined for phage particles. The geneII protein that initiates all synthesis on RFmolecules also enhances the doubling of viralstrands, possibly by competing with gene Vprotein for binding the hairpin origin sitesused for SS!RF conversion. Conversely,the cooperative binding of gene V proteinto SS DNAinhibits the conversion to RF,perhaps by melting out the hairpin regions,and sequesters viral strands for packaging.In addition unbound gene V proteinÐathigh concentration late during infectionÐacts as a translational repressor for gene IIprotein synthesis, further thwarting RF productionModel et al., 1982; Yen and Webster,1982). Gene V protein therefore appearsto gauge the ratio of SS to RF moleculesduring the course of infection: if low, unboundprotein made from excess RF) inhibitsgene II protein synthesis and checksthe accumulation of RF; if high, the proteinis bound up in nucleoprotein complexes andgene II protein advances SS!RF andRF!RF conversions. Subsequent experiments,however, have shown that protein Vcontrol of protein II synthesis is dispensible;it may be ancestral Zaman et al., 1992).Finally, the gene X protein, its synthesis tiedto that of gene II protein and also repressedby gene V protein, appears to antagonize geneII protein actions in RF production. Identicalto the carboxyl third of gene II protein,maybe it binds to the same sequence, butwithout catalytic activity or the ability toovercome gene V protein.) The interplay ofthe three phage-encoded proteins involved inDNAreplication, from genes II, X, and V,may thus describe the regulatory circuit thatkeeps phage productionin a long-term steady

SINGLE-STRANDED DNA PHAGES 161the process was reconstituted using purifiedcomponents Aoyama et al., 1983). The reactionsare concerted because the prohead andgene C protein are required for DNAsynthesis;therefore the third stage of phage DNAreplication is more complex than viral strandsynthesis in the second stage. The steps in theprocess are summarized as follows Hayashi,1978; Burch and Fane, 2000):Fig. 7. Model for RF!SS DNA replication and productionof fX174 virions. The nicked replicativeform of DNA, in a complex with gene A proteinRFII-A), is replicated asymmetrically in the presenceof the prohead, DNA polymerase III holoenzyme,and the E. coli rep protein. In a concerted reaction,the ``old'' viral ‡) strand is packaged into the phagehead to produce the mature fX phage particle.Reproduced from Aoyama et al., 1983, with permissionof the publisher.)state, implying that Ff phages have evolvedmeans to limit growth appropriate to the capabilityof their host bacteria. In this regard itmay be worthy to recall the surmised functionof the fX gene A* protein in shutting off hostDNAsynthesis, and possibly in the abruptswitch to phage SS DNAsynthesis. In geneticorganization, gene A* is analogous to the Ffgene X. Since gene X protein functionappears to discourage RF production, theremay be functional similarities too, except thatgene A* protein evolved to fit the virulent lifecycle of the isometric phages.C. Phage Assembly and Release1. Isometric phagesFigure 7 shows a model for RF!SS replicationand packaging reactions on fX174DNA, based on in vitro studies in which1. Five gene B internal scaffolding) proteinsbind to the underside of five moleculesof gene F capsid protein 9Svedberg) andlikely trigger conformational changes in theupper surface of the 9S particle Dokland etal., 1999). This then favors further interactionwith five molecules of gene G spikeprotein 6Svedberg) into a 12S particle, preventingself-aggregation of the 9S particles.2. Twelve of the 12S particles, 240 copiesof the gene D ``external scaffolding'' protein,and twelve copies of the gene H protein formthe 108S prohead.3. The displaced viral strand from RF!SSsynthesis is associated with 60 copies of thegene J protein and introduced into the proheadto form a compact 50S complex. Gene Aand C proteins are required at this step.4. Gene Aprotein catalyzes the DNAcleavage and recircularization steps, freeingthe viral strand circle from RF DNA.5. Maturation of the phage particle requireselimination of the internal scaffoldinggene B) protein to form a 132 S complexand subsequent removal of the external scaffoldinggene D) protein, to form the maturephage as a 114S particle.The mature bacteriophage contains 60copies each of the F protein 426 aminoacids), G protein 175 amino acids), J protein37 amino acids), and twelve copies of the Hprotein 328 amino acids). Most of the protein-proteininteractions arise from the interfacesof the F and G proteins. Pentameric Fand G proteins form the assembly intermediates9Svedberg and 6Svedberg) and ultimatelythe pentameric G protein spikes arestabilized, centered at the fivefold vertices

162 LECLERCof the F protein phage coat McKenna et al.,1994).The COOH-terminus of the gene B productinternal scaffolding protein) is necessaryfor specific coat protein interactions, thoughthe internal scaffolding proteins from thevarious phages in this group can crossfunctionBurch and Fane, 2000).However, eventhough the external scaffolding proteins proteinD) have significant homology among thedifferent phages of the group, the stages ofmorphogenesis outlined above are highly specificfor each of the phages and foreign Dproteins inhibit morphogenesis Burch andFane, 2000). Of two domains in the D protein,one probably blocks formation of theprocapsid while the other inhibits DNApackaging. Therefore the individual phagescan control self-morphogenesis at this level,as has been noted in other bacteriophages andanimal viruses as well Marvik et al., 1995;Spencer et al., 1998).The lytic cycle of isometric phage infectionends with the release of about 200 phageparticles per cell reviewed in Young, 1992).Cell lysis specifically requires the gene Eproduct; when cells were infected with agene E am mutant of fX and lysed artificially,the burst size averaged 2000 particlesper cell Hutchison and Sinsheimer, 1966)!The E protein does not function as a lysozyme.The N-terminal region of the proteinis very hydrophobic, similar to signal peptidesinvolved in membrane transfer, so Eprotein interaction with the membrane islikely Barrell et al., 1976). It is also the N-terminal 35 amino acids that are required forthe lysis activity. The lysis reaction has adependence on the host protein SlyD, amember of the rotamase cis-trans-isomerase)protein family, which may help proteinE accumulate in the membrane Roof et al.,1994). Isolation of phage E mutants thatplated on slyD hosts demonstrated thatSlyD activity is dispensable and led to theidentification of the mraY gene, encodingtranslocase I, as the true target of protein Eaction. Since the translocase forms the firstlipid-linked intermediate in cell wall biosynthesis,cell lysis then appears to result frominhibition of cell wall synthesis Bernhardt etal., 2000). Incidently, it had been noted thatthe mechanism of protein E lysis is quitesimilar to that for penicillin Roof andYoung, 1993).2. Filamentous phagesAn overview of the morphogenesis of filamentousphages is that gene V protein, whichcoats nascent viral DNA, is exchanged forgene VIII protein as the nucleoprotein filamentis extruded through the cell membranereviewed by Russel, 1991; 1995). Twoaspects of this directional membrane processmake it a fascinating model for nucleoproteinmembrane interactions. The first is thenature of the ``morphogenetic signal'' encodedin IG of the phage genome: Howdoes a regulatory sequence direct the initiationof the packaging process at the membrane?Second, the major coat protein fromgene VIII is found associated only with themembrane of infected bacteria. How the proteinis deposited there and then utilized forphage architecture without disrupting thecell are fundamental and yet unansweredquestions in membrane biology.The replacement of gene V protein bycapsid proteins during morphogenesis requiresa specific nucleotide sequence encompassingmuch of hairpin AFig. 5) Dottoand Zinder, 1983). The delineation of themorphogenetic signal came from experimentswith cells that contain the plasmidvector pBR322 carrying segments of IGCleary and Ray, 1980; Dotto et al., 1981).Upon infection with helper phage, progenyparticles that transduce ampicillin resistancewere assessed as a measure of packaged chimericDNA. The efficient production oftransducing particles required both hairpinAand the complete origins for DNAreplication,interpreted to mean that, in the presenceof phage proteins, the chimeric plasmidDNAcan be replicated in the phage mode assingle-stranded DNAand then encapsulated,as is phage DNA. Hairpin A is required forpackaging transducing particles were re-

SINGLE-STRANDED DNA PHAGES 163duced 100-fold in its absence) and the viralstrand sequence plus strand) must be locatedin cis arrangement to the DNAthat ispackaged, although it can be separated fromthe rest of IG by thousands of nucleotides.Finally, the morphogenetic signal is notneeded for either viral or complementarystrand DNAsynthesis Dotto and Zinder,1983). How the morphogenetic signal inhairpin Aworks is not understood. It maybe recognized by membrane-bound gene VIIand/or gene IX proteins during the initiationof morphogenesis, because 1) the Ahairpinis located on the gene VII/IX protein end ofthe phage filament Shen et al,, 1979; Websteret al., 1981) and 2) that is the end of thephage extruded first from the host cell Lopezand Webster, 1983). Three to five copiesof the proteins encoded by gene VII and geneIX are present at the leading end of the virusparticle Russel, 1995). There is also evidencethat protein VII and protein IX may interactwith the packaging signal for bacteriophageDNAand that protein VII interacts withprotein VIII, the major coat protein Endemannand Model, 1995). Indeed, the minorcoat proteins seem to define the ends of thephage particle; the end bearing the gene IIIand VI proteins is these first to penetrate thecell and the last to leave. The functions of theminor coat proteins must be known in orderto describe phage assembly and extrusion.The gene IV protein product is requiredfor phage assembly. It is an integral membraneprotein, usually located in the outermembrane of infected cells. It is rich incharged amino acids and has extensive b-sheet structure, similar to many outer membraneproteins. It is produced as a precursorprotein, and then cleaved as it is transportedto the outer membrane. While membraneintegration is independent of phage assembly,this protein may be vital for extrudingmature phage Brisette and Russel, 1990).Not much is known about another phageprotein essential for morphogenesis, theproduct of gene I.The gene VIII protein has been studied inmuch more detail than the other coat proteinsÐitis abundant in the infected cell andit has been a popular model for intrinsicmembrane proteins see Wickner, 1983).The protein is synthesized in precursorform with an N-terminal signal peptide of23 amino acids, which is cleaved after associationwith the inner membrane by a hostpeptidase. The mature protein of 50 aminoacids is stored as an integral membrane proteinwhen it is not being used as a phage coat.In the membrane the orientation of the geneVIII protein is the same whether it is insertedfrom the inside during synthesis and processing)or from the outside during thepenetration step of infection): the basic C-terminus exposed to cytoplasm, the centralhydrophobic interior anchoring the proteinin the lipid bilayer, and the acidic N-terminusin the periplasm. The N-termini arealso exposed on the phage coat.) The hydrophilicterminus of the protein about 50% oftotal protein) is in a-helical conformationwhen present in the membrane, while almost100% a-helices exist in the phage coat Nozakiet al., 1978). The transition may takeplace during the exchange of gene V proteinfor gene VIII protein ca. 1500 molecules for2800 molecules, respectively). Webster andCashman 1978) picture the gene VIII moleculesstacked in the membrane; a group ofthe C-terminal chains protruding into thecytoplasm interact with the viral DNAduring or after release from gene V protein,taking on total a-helical conformation asthey envelope the DNA; another group ofC-terminal ends is then available to the nextportion of DNAuntil the particle is extrudedfrom the cell. Such a mechanism would allowthe particle size to adjust to different genomelengths, observed in natural infection as thesmall proportion 1%) of miniphages andpolyphages. In this respect it is extraordinarythat the normal phages have maintainedgenome lengths within one nucleotide ofeach other in M13, fd, and fl.Another aspect of the assembly of filamentousphages that deserves comment isthe curious role of a host factor in the process.An E.coli gene called fipA, required

164 LECLERCspecifically for the assembly step as assessedby mutant studies, turned out to encodethioredoxin, the cofactor for reduction ofribonucleoside diphosphates to deoxyribonucleotidesby ribonucleotide reductaseRussel and Model, 1985; Lim et al., 1985;also see review by Fulford et al., 1986). Thehost function of thioredoxin in phage assemblyapparently does not involve redox reactions,however; it is the reduced form of theprotein, or the conformation of the reducedstate, that has been adapted for use in thepackaging pathway. Some mutants of thephage gene I can grow on fipA mutant hosts,taken as evidence for an interaction betweenthioredoxin and the gene I product, but thelatter protein's function is not known either.Since E. coli itself grows with mutant thioredoxins,normally an abundant protein, investigationsinto the role of thioredoxin inphage assembly may reveal an unimaginedproperty of the protein used in host metabolism.VI. RNA BACTERIOPHAGESAlthough this chapter and others in thisvolume have concentrated on bacteriophagescontaining DNAgenomes, the readershould be aware that other classes of bacteriophages,like mammalian viruses, utilizeRNAfor their genetic information. In thissection we discuss bacteriophages that useRNAgenomes in either double-stranded orsingle-stranded form. The discussion will bebrief, but will provide the reader withenough information and references to initiatefurther study. Comprehensive informationcan be found in The Bacteriophages,Vol. 2, edited by Calendar 1992).A. Double-Stranded RNABacteriophagesBacteriophage f6 is a well-characterizedRNAbacteriophage virulent for Pseudomonasspecies of bacteria see de Haas et al.,1999; Mindich, 1999). The phage has agenome consisting of three segments ofdouble-stranded RNAs, small; m, medium;and l, large) and a lipid envelope surroundingthe virus Li et al., 1993). The virions arespherical, 86 nm in diameter, and have surfaceprojections from the membrane as 8 nmlong spikes of protein P3 anchored by proteinP6. The nucleocapsids are isometric and56 nm in diameter. The virus utilizes thespike to attach to male pili of its principalhost, Pseudomonas syringae. As mentionedabove, the strategy for pilus-specific bacteriophagesis to depend on the retractionof the pilus to reach the bacterial surface.There the bacterial outer membrane and bacteriophagemembranes fuse, releasing the nucleocapsidand protein P5, a lysozymelikeprotein, into the periplasm of the bacterialcell. Protein P5 then digests peptidoglycanand permits the nucleocapsid to reach theinner membrane Mindich and Lehman,1979). The nuclecapsid breaches the innermembrane through a step that requires theenergized membrane and protein P8 Romantschuket al, 1988; Ojala et al., 1990).P8 releases from the complex and activatesa four-protein polymerase complex in thenucleocapsid, which replicates the viralgenome. Recently several other dsRNA-containingbacteriophages, differing from f6,have been isolated Mindich et al., 1999).Some of these fuse directly with the outermembrane of the susceptible bacteria, bypassingthe pilus requirement.The assembly of the virus is well documentedand has been replicated in vitroGottlieb et al., 1990). The nucleocapsidenzyme complex shields the replication ofthe virus from the cytoplasm and utilizesphage products. The framework for the complexis made up of protein P1 85 kDa) and isthe site where RNAbinds. The P2 protein75 kDa) is the RNA-dependent RNA polymeraseused for genome duplication. The P4and P7 proteins are involved in ‡) strandpackaging and the fidelity of synthesis. Thethree RNAsegments are packaged successivelyand processively. One model suggeststhat empty particles can bind only the smallsegment of RNA; once s has been packaged,the sites for m and l segments appear insuccession until all three segments have

SINGLE-STRANDED DNA PHAGES 165been packaged. Onodera et al., 1998). Thefilled procapsids then are covered by proteinP8 to become nucleocapsids and then areenveloped by the lipid membrane Gottliebet al., 1990). The morphogenesis of thesebacteriophage resembles in many ways thatof the mammalian reovirus.B. Single-Stranded RNA BacteriophagesThis brief review will concentrate on bacteriophageMS2 see Stockley et al., 1994).MS2 is 275 mm in diameter with no tail andinfects F pilus-displaying bacteria. Thegenome is a ‡) sense single strand of RNAcontaining 3569 nucleotides. This genomeencodes four genes coding for the phagecapsid protein, the maturation protein proteinA), the replicating RNA-dependentRNApolymerase, and a lysis protein. Thereare overlapping genes in this phage with thegene for the lysis protein spanning both thecapsid protein and replicase-specifying genesFiers et al., 1976). Following adsorption tothe F pilus, the single Aprotein in the phageis split proteolytically into fragments of 15and 24 kDa. The fragments remain attachedto the RNAas the pilus is retracted; thecomplex reaches the bacterial surface andpenetrates the membrane. The protein Afragments have no further documented rolein phage maturation Van Duin, 1988). Thephage RNArecruits host ribosomes and, byregulating the level of expression by bothribosome binding and translation controls,produces the required level of proteins forphage assembly and the lysis of the infectedcell.The bacteriophage uses interesting geneticmeans to control the level of expression fromthe RNAgenome during phage maturation.For instance, at the 5 0 end of the RNAmoleculea stem-loop structure forms, which caninteract with a single dimer of the coat proteinLago et al., 2001). As the level of coatprotein rises, this binding is complete andinhibits ribosome binding to sequences inthe 3 0 end that are needed for the expressionof the replicase. Consequently the achievementof the proper level of coat proteinssignals that the replication of RNAmoleculescan cease and further maturationmay ensue. The RNA-coat protein complexthen attracts further capsid dimers and acopy of the maturation protein A. This ultimatelyresults in a self-assembled capsid. Itis suggested that the free energy change thatresults from capsid formation may providethe momentum for this assembly.The replicase for the RNAbacteriophageis inaccurate. This has been studied moreextensively in Qb where the misincorporationrate is between 10 3 and 10 4 per nucleotideper replication cycle Van Duin, 1988).This is due to the lack of a 3 0 to 5 0 proofreadingnuclease in the replicase. Frequentdeletions are also noted, presumably due toregions of homology or perhaps to hairpinloops in the RNA. At any rate, though theremust be selection for a wild-type RNA, thepopulation of Qb phage contains many variants.The genomes of these RNAs phagesexhibit a significant degree of secondarystructure, which may prevent annealing betweenthe ‡) and ) strands of RNA.The genome of MS2 bacteriophage ischaracterized by an overlapping lysis gene.The protein from this gene is responsible forlysing the host bacterium at the end of theinfection. Control of lysis protein expressionis coordinated with that of the coat protein.When coat gene translation terminates,translation of the overlapping lysis gene istriggered Benhardt et al., 2001). Asecondarystructure in the RNAprevents lysis genetranslation in the absence of coat gene translation.It is assumed that the same ribosomeswitches from coat gene to lysis gene translation.With the lysis of the cell, the cycle forthe single-stranded RNAbacteriophage isconcluded.VII. USES INBIOTECHNOLOGYA. Cloning and Sequencing VectorsAs far back as the 1978 Cold Spring Harbormonograph The Single-Stranded DNAPhages Denhardt et al., 1978), six reports

166 LECLERCrevealed the development of the M13, f1, andfd genomes as carriers for foreign DNA, thatis, cloning vectors. Several features of the lifecycle of the filamentous phages commendtheir use for the cloning and subsequent manipulationsof DNA:1. There are few constraints on the size ofinsertions, since the Ff phages allow thepackaging of larger than unit-length DNAalthough inserts < 5 kb are the most stable).2. The phages do not lyse host bacteria,allowing the expression of cloned genes orselective markers on vector DNA. Hencerecombinants can be propagated andhandled either as phage or as phage-infectedcells.3. The replication cycle yields RF DNAthat is like plasmid DNAfor all of its technicaluses in genetic engineering, such as inrestriction enzyme digestion and ligation.4. The abundance of viral DNAÐup to200 RF molecules per cell and phage titers> 10 12 per milliliterÐmeans that adequateDNAfor analysis can be obtained from aslittle as a milliliter of cells.5. The bonus from the phage systems isthat the DNAis packaged as single strands;double-stranded DNAmay be cloned intoRF in either orientation, yielding the separatedstrands in unique viral strand progeny.As such, the DNA from phage particles isideally suited for the chain termination ``dideoxy'')method of DNAsequencing Sangeret al., 1977), so the greatest efforts have goneinto constructing phage vectors for rapidand efficient sequencing technology.Reviews that describe a variety of the earlyphage vectors and methodologies have beenwritten by Zinder and Boeke 1982), Messing1983), and Gelder 1986), and usefulmanuals on cloning and sequence analysisusing the phage systems are available fromthe major biotechnology vendors.The M13mp for Max Planck) phages areprobably the most popular vectors for sequencingcloned DNA. These are the ``blueplaque formers'' developed by Messing andcoworkers, in which a color system is used todetect clones, rather than the more usualcase of inactivating antibiotic resistancegenes Messing et al., 1977; Messing, 1983).An amusing account of the events leading tothe M13mp vectors emphasizes the rich historyof phage biology and lac genetics thatwere brought together in their developmentMessing, 1988; also see Messing, 1996). TheM13mp system is diagrammed in Figure 8.The vector contains an insert of 789 nucleotidesfrom the E. coli lac operon: the regulatoryregion and a lacZ 0 sequence encodingonly the amino-terminal end a-peptide) ofb-galactosidase. The lacZ gene of the hostcell is also defective, containing a deletion of93 nucleotides lacZDM15), which eliminatesamino acids 11 to 41 of b-galactosidase.Functional enzyme is produced upon phageinfection or transfection using phage DNA)by a-complementation, wherein the aminoportion of the enzyme, encoded in hybridphage DNA, complements the lacZM15 proteinof the host cell. When the hybrid phageare plated in a lawn of lacZM15 cells in thepresence of the histochemical dye 5-bromo-4-chloro-3-indoyl-b-D-galactoside ``Xgal''),a-complemented enzyme hydrolyzes the galactoside,the indoyl moiety causing theplaques of infected cells to turn a deep blueFig. 8. The M13mp2vector system for cloning andsequencing DNA. An EcoRl restriction enclonucleasesite in the lacZa) gene of hybrid phage DNA providesthe cloning site for foreign DNA. The openbox represents the universal primer for sequencingcloned DNA; the filled box represents the probeprimer for labeling DNA 5 0 to the cloning site. Thedirection of DNA synthesis for primer elongation isindicated by the arrow on phage DNA.

SINGLE-STRANDED DNA PHAGES 167color. The blue phenotype of M13lacinfectedcells is ideal for cloning experiments:inactivation of the a-peptide by inserting foreignDNAinto the lacZ 0 gene of RF DNAyields colorless plaques of infected cellsÐavivid indicator for screening recombinants.The first usable vector, M13mp2, was developedby mutagenizing the fifth codon ofthe lacZ 0 gene, creating a unique EcoRl sitefor cloning into hybrid phage DNA. Sincethe early codons can be interrupted withsmall in-frame insertions and still encode -complementing activity albeit forminglighter blue plaques), multiple restriction endonucleasesites or polylinkers have beenengineered at the cloning site in successivegenerations of M13mp vectors, making themadaptable to virtually any cloning situation.Their use is so routine that authors of publishedreports have referred to the parentvector as wild-type M13!The principal use of phage vectors is sequencingrecombinant DNA; they have littleadvantage for the expression of clonedgenes. The feature that makes the phagesystems advantageous is the provision ofsingle-stranded template DNAfor the chaintermination method of sequencing. Doublestrandedinserts cloned into phage RF DNAare naturally subjected to strand separationby the asymmetric mode of Ff DNAreplication;the single-stranded viral DNAcarrieseither strand of insert free of the otherstrand, so template DNAcan be obtainedsimply by preparing virions and extractingtheir DNA. To aid in cloning the desiredstrand, pairs of M13mp vectors have beendesigned to contain the same polylinker sequencein opposite orientations shown inFig. 9). The vectors can be cut with twodifferent restriction enzymes to producedifferent ends and then the appropriatelycleaved fragments ``force-cloned'' in eitherorientation. When two clones contain insertsin opposite orientations, the single-strandedDNAs from lysed virions hybridize in thecomplementary region; the structure migratesslower than unannealed moleculesduring gel electrophoresis, providing asimple assay for the orientation of recombinantDNA. Another feature introduced withthe use of phage vectors for sequencing is theuniversal primer, an oligonucleotide thatanneals to the vector DNAat a site flanking,and 3 0 to, the cloning site. Since the cloningsite is at the same map position in all M13mpvectors, one primer sequence is used for allrecombinant clones rather than using manydifferent primers on one recombinant DNA.In order to accommodate the limit of sequencing500 to 1000 nucleotides from asingle primer, numerous methods have beendevised for preparing nested deletions incloned DNA, systematically drawing the distalsequences of long inserts closer to theprimer site. Alternatively, small fragmentsFig. 9. Multiple cloning sites of the Ff phage cloning vectors M13mp18 and M13mp19. The amino acidsequence, given in uppercase letters, corresponds to the sequence of the amino portion of b-galactosidaseencoded in the phage vector. In-frame insertions of restriction endonuclease cleavage sites, and correspondingamino acids given in lowercase letters, are in opposite orientations in the pair of vectors. Courtesy ofResearch Products Division, Life Technologies, Inc., Gaithersburg, MD.)

168 LECLERCare randomly cloned and sequenced ``shotgunsequencing''), and the results are sortedout by computer analysis Messing et al.,1981). Many of these novel methods developedfor sequencing with the Ff vectorsare now amenable to the plasmid vectors,with the introduction of efficient protocolsfor dideoxy sequencing of double-strandedDNAWallace et al., 1981; Chen and Seeburg,1985).The cloning sites for most of the filamentousphage vectors reside in the large intergenicspace, IG, one of the few regions of thephage genome that can be interrupted withoutaffecting viral genes. But IG is hardlydevoid of function, and an insertion such asthe lac insert in M13mp vectors has hadinteresting consequences for phage viability.The lac DNAin the progenitor of the serieswas cloned into domain B of the viral strandreplication origin at position 5867 on thesequence of Fig. 5). Disruption of this replicationenhancer, which lies downstreamfrom the recognition sequence for gene IIprotein, normally drops the activity of theviral strand origin to 1% or less. It turnsout that the viability of the M13mp hybridphages depends on a compensatory mutationwithin gene II; other mutations thatovercome domain B defects lie within geneV or in the regulatory region for gene IImRNA, all causing the overexpression ofthe gene II protein by removing gene V protein±mediatedrepression of its translationDotto and Zinder, 1984a, b). Thus qualitativeor quantitative changes in gene II proteinmake the replication enhancer largelydispensable. Nevertheless, viability is stillaffected in the case of M13mp vectors, sincethey produce titers 5- to 10-fold lower thanwild-type phages or the f1 cloning vectorsthat have insertions upstream from theDNAreplication origins Boeke et al., 1979).B. Hybridization ProbesNucleic acid hybridization is a sensitive andpowerful means to study the structure andfunction of genesÐtheir numbers and sizes,their transcription, and their relatedness.The preparation of single-stranded DNAprobes for hybridization is another usefulapplication of the natural strand separationafforded by filamentous phage vectors. Therecombinant DNAis produced in singlestrandedform, available for annealing withoutdenaturation, and more important, it isstrand specific. The latter feature is mostsignificant for the analysis of gene transcription,since only the coding strand insertedinto the viral DNAhybridizes to mRNA.Labeled probes may be prepared by usingDNApolymerase and radioactive deoxyribonucleotidesto elongate a probe primerthat anneals to vector DNAbeyond the 5 0side of the vector cloning site; complementarysequences in phage DNAare synthesizedand labeled, leaving the recombinantDNAin single-stranded form. For otherapplications, strand-specific recombinantDNAcomplementary to the insert) can belabeled by extending the universal primerand separating it before use. Specific applicationsof these methods are given in thereview by Gelder 1986).C. Site-Directed MutagenesisChanging the regulatory or coding sequencesin DNAat will, once the geneticists' andbiochemists' dream, is now made routine bymethods of site-directed or site-specific)mutagenesis. This technology relies on thesolid phase-supported synthesis of oligodeoxyribonucleotides,so that nucleotide sequencescan be designed to contain thedesired changes. Smith and colleagues pioneeredthe methods for incorporating thechanges into a genome; they used fX174viral DNAas a model system Hutchison etal., 1978). The oligonucleotide, made complementaryto its target sequence except forthe mutation site, was used as a primer for invitro synthesis on the single-stranded DNAso that the mutant sequence could be incorporatedinto the complementary strand ofduplex phage DNA. Upon transfection ofE. coli the two strands of the heteroduplexsegregate by replication; the two types ofprogeny phage produced were the wild-type

SINGLE-STRANDED DNA PHAGES 169and the desired mutant clones. The technologywas quickly adopted for mutating recombinantDNAin the Ff cloning vectorswhen their use became widespread. In additionmethods were developed to improvethe efficiency of the process and to detectmutant clones by probing with labeledoligonucleotides see Zoller and Smith,1983).Transfection studies with heteroduplex f1molecules have shown that the genotypesamong the progeny phage strongly favorthe complementary minus) strand of theoriginal RF molecule; the complementarystrand acts as an ``master template'' so thatits genetic information dominates in the productionof progeny molecules Enea et al.,1975). Although the reasons for this phenomenonare not entirely clear, it impliesthat the progeny RF molecules from transfectingDNAare not used equally duringRF!RF synthesisÐreplication must involvethe asymmetric rolling-circle replication ofFf DNA. In practice, the heteroduplexes derivedfrom site-directed mutagenesis protocolsusually yield only 10±20% progeny fromthe mutant complementary) strand, owingto the inefficiency of the in vitro reactions,repair or mismatch correction during transfection,or unknown causes. Therefore bothbiochemical and genetic approaches havebeen developed to trick host cells intofavoring the mutant strand. Kunkel 1985)devised one of the most effective protocols,which yields 60±80% mutant progeny. Theselection relies on the host excision repairsystem for uracil residues in DNA. Uracilin DNAe.g., from deamination of cytosine)is normally removed by uracil DNA-glycosylaseencoded by the ung ‡ gene), whichexcises the base and leaves an abasic site ondeoxyribose as the first step in base excisionrepair of double-stranded DNAsee Yasbin,this volume). In single-stranded DNA, subsequentphosphodiester bond cleavage at theabasic site destroys the single strand. Uracilcontainingphage DNAcan be obtained bygrowing recombinant phage in the presenceof uridine in an E. coli ung dut mutant strain;the uracil residues are not excised because ofthe ung mutation and the coding propertiesof the DNAare unaffected, since uracil substitutesfor thymine. The additional dut mutationinactivates dUTPase, which cleavesdUTP, also keeping uracil out of DNA.)Used as template DNAin the site-directedmutagenesis procedure, the uracil-substitutedviral strand in the resulting heteroduplexresides opposite a nonsubstitutedcomplementary strand synthesized in vitro.By transfecting an E. coli ung ‡ strain, thehost uracil glycosylase provides a strong selectionfor the desired mutant strand, bothby inactivating the viral strand of the heteroduplexand by destroying unextended templateDNA. Such a high proportion oftransfectants are site-specifically modifiedthat no screening step is requiredÐa fewclones are sequenced to verify the mutantchange.D. Plasmids with Phage Origins ofReplicationAlthough the availability of single-strandedDNAfrom the phage vectors is advantageousfor specific applications, the smallerplasmid vectors are favored for severalreasons, including the stability of largeinserts, means for amplification to obtainlarge amounts of recombinant DNA, RNA,or protein from cloned genes, and simplybecause of their familiarity see Geoghegan,this volume). The advantages of bothsystems have been achieved by the constructionof plasmid vectors that contain theorigins of DNAreplication and morphogeneticsignal ca. 500 bp) from IG of the Ffphage genome Dente et al., 1983; Levinsonet al., 1984; Zagursky and Berman, 1984;review by Cesareni and Murray, 1987). TheIG segment carried on the plasmid is normallysilent in the cell; it is double-stranded,so the complementary strand origin is notfunctional, and the absence of phage-specificproteins precludes RF replication and morphogenesis.Hence the plasmid-IG chimera istreated in all ways like plasmid DNA. Whenan application benefits from the provision of

170 LECLERCsingle-stranded DNA, the cells are infectedwith helper phage, resulting in the extrusionof particles that contain the single-strandedplasmid DNA. Although helper phage arealso produced exceeding the yield of packagedplasmid when the plasmid insert islarge), the phage DNAmay not interferewith particular applications or the plasmidsingle strands may be purified. In additionan M13 derivative has been specially constructedfor use as a helper Vieira and Messing,1987). Owing to inserts in domain B ofthe viral strand origin and the altered gene IIprotein from the M13mp phages, replicationfrom the helper phage origin is less efficientthan from the wild-type phage origin on chimericplasmid DNA, resulting in helperphage titers 10- to 100-fold lower than plasmidparticles.The orientation of the IG segment in plasmidDNAdetermines the plasmid strandthat is packaged, since it is the viral plus)strand that is replicated during RF!SS synthesisand encodes the packaging signal.Therefore either strand of recombinantDNAcan be packaged by cloning inserts inopposite orientations. Use of pairs of plasmidvectors constructed with IG inserted inopposite orientations achieves the same end.Aclever biotechnological use has been madeof the finding that the gene II proteins fromthe Ff and IKe phages are origin specific;that is, the Ff gene II protein does not acton the viral strand origin of IKe, and viceversa Peeters et al., 1986). The plasmidvector pKUN9 contains the M13 replicationorigins and packaging signal on one strandand those from IKe on the other i.e., theywere cloned in opposite orientations). Incells that bear both F-and N-specific episomes,the pKUN9 strand and the recombinantstrand linked to it are packagedaccording to which helper phage is used forinfectionÐeither M13 or IKe.E. Phage Display Technology UsingFilamentous BacteriophageM13 bacteriophage have been modified foruse in displaying foreign proteins on the bacteriophagesurface Smith, 1985; reviewed byWilson and Finlay, 1998; and see Geoghegan,this volume). Both protein VIII majorcoat protein) and protein III at tip of particle)have been used for this purpose seeDavies et al., 2000). As an example, investigatorswere able to display antibodies linkedto protein VIII and then resultant bacteriophagecould be ``biopanned'' with antigensof interest. Kang and coworkers created acombinatorial library of functional antibodiesby combining random kappa) lightchain DNAand Fd heavy chain DNAfusedto gene VIII Kang et al., 1991). They usedprotein VIII since that would allow displayall along the phage particle, rather than justat the tip. The light and heavy chains assemblein the periplasm with the Fd chainanchored in the membrane together withprotein VIII. Phage morphogenesis then assemblesa coat with the antibody on theouter surface. The phage had 1 to 24 copiesof the antibody per bacteriophage. The presenceof functional Fab on the surface of thephage could be confirmed by ELISAassay,and the isolated phage could interact withantigens. Waterhouse and coworkers utilizeda site-specific recombination system frombacteriophage P1 to fuse heavy and lightchain genes on separate replicons prior topackaging Waterhouse et al., 1993), andthen used fd filamentous bacteriophage forthe display.VIII. CONCLUDING REMARKSThe current and novel uses of the filamentousphage genomes for recombinant DNAtechnology are striking examples of the applicationof basic research. Athoroughknowledge of the life cycles of the phagesled to most of the developments; they weredevised and constructed with predictable outcomes.In other cases a chance occurrenceled to unexpected results, and we learnedsomething new about phage biology. It isthe natural progression of science that thebasic research on the processes of infectionby the single-stranded DNAphages has declined;the research groups actively pursuing

SINGLE-STRANDED DNA PHAGES 171fundamental problems have dwindled to perhapsa dozen. The irony is that the numberof laboratories using the phagesÐas tools ofresearchÐis many times that in the heydayof phage research. Several questions remainunanswered. The molecular events at the beginningand at the end of the phage life cyclesare not understood, particularly where thehost cell membranes are involved. The functionsof a few phage gene products are amatter of conjecture. Much can be learnedabout the properties of host proteins by discoveringthe functions that the phagesystems have adopted for their reproduction.The reasons for which Sinsheimer studiedthe small phages, encompassed in his descriptionof fX174 as multum in parvo Sinsheimer,1966), are still applicable for thefundamental problem of how genes encodeinformation for morphogenesis and architecture.These problems will be solved experimentally;others, like describing theevolution of the phage groups convergentor divergent?), will continue to be debated.Even if the current great era of moleculargenetics, that of the human genome, seemsfar removed from the small phages, they areinvolved there too, in new biotechnologicaluses devised for analysis of human DNA.REFERENCESAoyama A, Hamatake RK, Hayashi M 1983): In vitrosynthesis of bacteriophage fX174 by purified components.Proc Natl Acad Sci USA 80:4195±4199.Arai K, Kornberg A 1981): Unique primed start ofphage fX174 DNAreplication and mobility of theprimosome in a direction opposite chain synthesis.Proc Natl Acad Sci USA 78:69±73.Baas PD 1985): DNAreplication of single-strandedEscherichia coli DNAphages. Biochim Biophys Acta825:111±139.Baker R, Tessman 1 1967): The circular genetic map ofphage S13. Proc Natl Acad Sci USA 58:1438±1445.Barrell GB, Air GM, Hutchison CA III 1976): Overlappinggenes in bacteriophage fX174. Nature264:34±41.Bayer ME, Starkey TIN 1972): The adsorption of bacteriophagefX174 and its interaction with Escherichiacoli, a kinetic and morphological study. Virol49:236±256.Beck E, Sommer R, Auerswald EA, Kurz C, Zink B,Osterburg G, Schaller H, Sugimoto K, Sugisaki H,Okamoto T 1978): Nucleotide sequence of bacteriophagefd DNA. Nucleic Acids Res 5:4495±4503.Beck E, Zink B 1981): Nucleotide sequence and genomeorganization of filamentous bacteriophages f1 and fd.Gene 16:35±58.Bernhardt TG, Roof WD, Young R 2000): Geneticevidence that the bacteriophage fX174 lysis proteininhibits cell wall synthesis. Proc Natl Acad Sci USA97:4297±4302.Bernhardt TG, Wang I-N, Stuck DK, Young R 2001):Aprotein antibiotic in the phage Qb Virion: Diversityin lysin targets. Science, 292:2326±2329.Boeke JD, Vovis GF, Zinder ND 1979): Insertionmutant of bacteriophage f1 sensitive to EcoRl. ProcNatl Acad Sci USA 76:2699±2701.Brissette JL, Russel M 1990): Secretion and membraneintegration of a filamentous phage-encoded morphogeneticprotein. J Mol Biol 211:565±580.Brown DT, MacKenzie JM, Bayer ME 1971): Mode ofhost cell penetration by bacteriophage fX174. Virol7:836±846.Bruse GW, Wollin R, Oscarson S, Jansson PE, LindbergAA 1991): Studies of the binding activity of phageS13 to synthetic trisaccharides analagous to bindingstructures in Salmonella typhimurium and Escherichiacoli C core saccharide. Correlation between conformationand binding activity. J Mol Recognit 4:121±128.Burch AD, Fane BA 2000): Foreign and chimeric externalscaffolding proteins as inhibitors of Microviridaemorphogenesis. J Virol 74:9347±9352.Burch AD, Fane BA 2000): Efficient complementationby chimeric Microviridae internal scaffolding proteinsis a function of the COOH-terminus of the encodedprotein.Virol 270:286±290.Calendar R 1992): ``The Bacteriophages''. Vol 1. NewYork: Plenum Press.Cesareni G, Murray JAH 1987): Plasmid vectors carryingthe replication origin of filamentous singlestrandedphages. In Setlow JK ed): ``Genetic Engineering''.New York: Plenum Press, pp 135±153.Chen EY, Seeburg PH 1985): Supercoil sequencing: Afast and simple method for sequencing plasmid DNA.DNA4:165±170.Cleary JM, Ray DS 1980): Replication of the plasmidpBR322 under the control of a cloned replicationorigin from the single-stranded DNAphage M13.Proc Natl Acad Sci USA 77:4638±4642.Davies JM, O'Hehir RE, Suphioglu C 2000): Useof phage display technology to investigate allergenantibodyinteractions. J Allergy Clin Immunol105:1085±1092.deHaas F, Paatero AO, Mindich L, Bamford DH,Fuller SD 1999): Asymmetry at the site of RNApackaging in the polymerase complex of dsRNAbacteriophagef6. J Mol Biol 294:357±372.Denhardt DT 1975): The single-stranded DNAphages,CRC Crit Rev Microbiol 4:161±223.

172 LECLERCDenhardt DT 1977): The isometric single-strandedDNAphages. Compr Virol 7:1±104.Denhardt DT, Dressler D, Ray DS 1978): ``The Single-Stranded DNAPhages''. Cold Spring Harbor, NY:Cold Spring Harbor Laboratory.Dente L, Cesareni G, Cortese R 1983): pEMBL: Anewfamily of single-stranded plasmids. Nucleic Acids Res11:1645±1655.Dokland T, Bernal RA, Burch A, Pletnev S, Fane BA,Rossman MG 1999): The role of scaffolding proteinsin the assembly of the small, single stranded DNAvirus fX174. J Mol Biol 288:595±608.Dotto GF, Enea V, Zinder ND 1981): Functional analysisof bacteriophage f1 intergenic region. Virol114:463±473.Dotto GP, Zinder ND 1983): The morphogenetic signalof bacteriophage fl. Virol 130:252±256.Dotto GF, Zinder ND 1984a): Increased intracellularconcentration of an initiator protein markedly reducesthe minimal sequence required for initiationof DNAsynthesis. Proc Natl Acad Sci USA81:1336±1340.Dotto GP, Zinder ND 1984b): The minimal sequencefor initiation of DNAsynthesis can be reduced byqualitative or quantitative changes of an initiator protein.Nature 311:279±280.Endemann H, Model P 1995): Location of the filamentousphage minor coat proteins in phage and ininfected cells. J Mol Biol 250:496±506.Enea V, Vovis GF, Zinder ND 1975): Genetic studieswith heteroduplex DNAof bacteriophage fl. Asymmetricsegregation, base correction and implicationsfor the mechanism of genetic recombination. J MolBiol 96:495±509.Enea V, Zinder N 1975): Adeletion mutant of bacteriophagef1 containing no intact cistrons. Virol68:105±114.Fiers W 1979): Structure and function of RNAbacteriophages.Compr Virol 13:69±179.Fiers W, Contreras R, Duerink F, Haegaman G, IserentantD, Merregaert J, Min Jou W, Molemans F,Raemaekers A, Van den Berghe A, Volckaert G, YsebaertM 1976): Complete nucleotide sequence of bacteriophageMS2 RNA: Primary and secondarystructure of the replicase gene. Nature 260:500±507Fiers W, Sinsheimer RL 1962): The structure of theDNAof bacteriophage fX174. III. Ultracentrifugalevidence for a ring structure. J Mol Biol 5:424±434.Freifelder D, Kleinschmidt AK, Sinsheimer RL 1964):Electron microscopy of single-stranded DNA: Circularityof DNAof bacteriophage fX174. Science146:254±255.Fujimura FK, Hayashi M 1978): Transcription of isometricsingle-stranded DNAphage. In Denhardt DT,Dressler D, Ray DS eds): ``The Single-StrandedDNAPhages''. Cold Spring Harbor, NY: ColdSpring Harbor Laboratory, pp 485±505.Fulford W, Model P 1984): Gene X of bacterlophage f1is required for phage DNAsynthesis. Mutagenesis ofin-frame overlapping genes. J Mol Biol 178:137±153.Fulford W, Model P 1988a): Regulation of bacteriophagef1 DNAreplication. 1. New functions for genesII and X. J Mol Biol 203:49±62.Fulford W, Model P 1988b): Bacteriophage f1 DNAreplication genes. II. The roles of gene V protein andgene II protein in complementary strand synthesis.J Mol Biol 203:39±48.Fulford W, Russel M, Model P 1986): Aspects of thegrowth and regulation of the filamentous phages.Prog Nucleic Acid Res Mol Biol 33:141±168.Geider K 1986): DNAcloning vectors utilizing replicationfunctions of the filamentous phages of Escherichiacoli. J Gen Virol 67:2287±2303.Gilbert W, Dressler D 1968): DNAreplication: Therolling circle model. Cold Spring Harbor Symp QuantBiol 33:473±484,Godson GN, Barrell BG, Staden R, Fiddes JC 1978):Nucleotide sequence of bacteriophage G4 DNA.Nature 276:236±247.Gottlieb P, Strassman J, Quao X, Frucht A, Mindich L1990): In vitro replication, packaging, and transcriptionof the segmented, double-stranded RNAgenomeof bacteriophage f6: Studies with procapsids assembledfrom plasmid-encoded proteins. J Bacteriol172:5774±5782.Goulian M, Kornberg A, Sinsheimer RL 1967): Enzymaticsynthesis of DNA. XXIV. Synthesis of infectiousphage fX174 DNA. Proc Natl Acad Sci USA58:2321±2325.Gray C, Brown R, Marvin D 1981): Adsorption complexof filamentous fd virus. J Mol Biol 146:621±627.Griffith J, Kornberg A1974): Mini M13 bacteriophage:Circular fragments of M13 DNAare replicated andpackaged during normal infections. Virol 59:139±152.Griffith J, Manning M, Dunn K 1981): Filamentousbacteriophage contract into hollow spherical particlesupon exposure to a chloroform-water interface. Cell23:747±753.Guthrie GD, Sinsheimer RL 1960): Infection of protoplastsof Escherichia coli by subviral particles of bacteriophagefX174. J Mol Biol 2:297±305.Hall CE, Maclean EC, Tessman I 1959): Structure anddimensions of bacteriophage fX174 from electronmicroscopy. J Mol Biol 1:192±194.Hayashi M 1978): Morphogenesis of the isometricphages. In Denhardt DT, Dressler D, Ray DS eds):The Single-Stranded DNA Phages. Cold SpringHarbor, NY: Cold Spring Harbor Laboratory, pp531±547.Hewitt JA1975): MiniphageÐAclass of satellite phageto M13. J Gen Virol 26:87±94.Hill D, Peterson G 1982): Nucleotide sequence of bacteriophagef1 DNA. J Virol 44:32±46.

SINGLE-STRANDED DNA PHAGES 173Hoffman-Berling H, Kaerner HC, Knippers R 1966):Small bacteriophages. Adv Virus Res 12:329±370.Hofschneider PH 1963): Untersuchungen uber ``kleine''E. coli K12 bacteriophagen. 1 und 2 mitteilung. ZNaturforsch 18b:203±210.Hutchison CAIII, Marshall EH, Sinsheimer RL 1967):The process of infection with bacteriophage fX174.XII. Phenotypic mixing between electrophoreticmutants of fX174. J Mol Biol 23:553±575.Hutchison CAIII, Phillips S, Edgell MH, Gillam S,Jahnke P, Smith M 1978): Mutagenesis at a specificposition in a DNAsequence. J Biol Chem253:6551±6560.Hutchison CAIII, Sinsheimer RL 1966): The process ofinfection with bacteriophage fX174. X. Mutations ina fX lysis gene. J Mol Biol 18:429±447.Incardona NL, Selvidge L 1973): Mechanism of adsorptionand eclipse of bacteriophage fX174. II Attachmentand eclipse with isolated Escherichia coli cellwall lipopolysaccharide. J Virol 11:775±782.Jazwinski M, Marco R, Komberg A1973): Acoatprotein of the bacteriophage M13 virion participatesin membrane-oriented synthesis of DNA. Proc NatlAcad Sci USA 70:205±209.Jazwinski MS, Marco R, Kornberg A1975a): The geneH spike protein of bacteriophage fX174 and S13. 11.Relation to synthesis of parental replicative form.Virol 66:294±305.Jazwinski SM, Lindberg AA, Kornberg A 1975b): Thelipopolysaccharide receptor for bacteriophagesfX174 and S13. Virol 66:268±282.Kang, AS, Barbas, CF, Janda KD, Benkovic SJ, LernerRA1991): Linkage of recognition and replicationfunctions by assembling combinatorial antibody Fablibraries along phage surfaces. Proc Natl Acad SciUSA88:4363±4366.Kornberg A, Baker TA 1992): DNA Replication. NewYork: Freeman.Kunkel TA1985): Rapid and efficient site-specificmutagenesis without phenotypic selection. Proc NatlAcad Sci USA 82:488±492.Lago H, Parrott, AM, Moss T, Stonehouse NJ, StockleyPG 2001): Probing the kinetics of formation of thebacteriophage MS2 translational operator complex:identification of a protein conformer unable to bindRNA. J Mol Biol 305:1131±144.Levinson A, Silver D, Seed B 1984): Minimal size plasmidscontaining an M13 origin for production ofsingle strand transducing particles. J Mol Appl Genet2:507±517.Li T, Bamford DH, Bamford JKH, Thomas GJJ 1993):Structural studies of the enveloped dsRNAbacteriophagef6 ofPseudomonas syringae by Raman spectroscopy.I. The virion and its membrane envelope.J Mol Biol. 230:461±472.Lim C, Haller B, Fuchs J 1985): Thioredoxin is thebacteria protein encoded by fip that is required forfilamentous bacteriophage f1 assembly. J Bacteriol161:799±802.Linney E, Hayashi M 1974): Intragenic regulation ofthe synthesis of fX174 gene Aproteins. Nature249:345±348.Loeb T 1960): Isolation of a bacteriophage specific forthe F ‡ and Hfr mating types of Escherichia coli K12.Science 131:932±933.Loeb T, Zinder ND 1961): Abacteriophage containingRNA. Proc Natl Acad Sci USA 47:282±289.Lopez J, Webster R 1983): Morphogenesis of filamentousbacteriophage fl: orientation of extrusion andproduction of polyphage. Virol 127:177±193.Madison-Antenucci S, Steege DA 1998): Translationlimits synthesis of an assembly-initiating coat proteinof filamentous phage IKe. J Bacteriol 180:464±472.Marians KJ 1984): Enzymology of DNAin replicationin prokaryotes. CRC Crit Rev Biochem 17:153±215.Martin DO, Godson GN 1975): Identification of afX174 coded protein involved in the shut off of hostDNAreplication. Biochem Biophys Res Commun65:323±330.Marvik OJ, Dokland T, Nokling RH, Jacobsen E, LarsenT, Lindqvist BJ 1995): The capsid size-determiningprotein sid forms an external scaffold on phage P4procapsids. J Mol Biol 251:59±75.Marvin D 1998): Filamentous phage structure, infectionand assembly. Curr Opin Struct Biol 8:150±158.Marvin D, Pigram W, Wiseman R, Wachtel E, MarvinF 1974): Filamentous bacterial viruses. XII. Moleculararchitecture of the class I fd, f1, IKe) virion. J MolBiol 88:581±598.Marvin DA, Hoffmann-Berling H 1963): A fibrousDNAphage fd) and a spherical RNAphage specificfor male strains of E coli. Z Naturforsch 18b:884±893.Marvin DA, Hohn B 1969): Filamentous bacterial viruses.Bacteriol Rev 33:172±209.McKenna R, Ilag LL, Rossman MG 1994): Analysis ofthe single stranded DNAbacteriophage fX174 at aresolution of 3.0A. J Mol Biol 237:517±543.Messing J 1983): New M13 vectors for cloning.Methods Enzymol 101:20±79.Messing J 1988): M13, the universal primer and thepolylinker. Focus 10:21±26.Messing J 1996): Cloning single-stranded DNA. MolecularBiotechnology 5:39±47.Messing J, Crea R, Seeburg PH 1981): Asystem forshotgun DNAsequencing. Nucleic Acids Res9:309±321.Messing J, Gronenborn B, MuÈller-Hill B, HofschneiderPH 1977): Filamentous coliphage M13 as a cloningvehicle: Insertion of a HindII fragment of the lacregulatory region in M13 replicative form in vitro.Proc Natl Acad Sci USA 74:3642±3646.

174 LECLERCMindich L, Lehma JF 1979): Cell wall lysin as a componentof the bacteriophage f6 virion. J Virol30:489±496.Mindich L 1999): Precise packagaing of the three genomicsegments of the double-stranded-RNAbacteriophagef6. Microbiol Mol Rev 63:149±160.Mindich L, Qiao X, Qiao J, Onodera S, RomantschukM, Hoogstraten, D 1999): Isolation of additionalbacteriophages with genomes of segmented doublestrandedRNA. J Bacteriol 181:4505±4508.Model P, McGill C, Mazur B, Fulford W 1982): Thereplication of bacteriophage fl: Gene 5 protein regulatesthe synthesis of gene 2 protein. Cell 29:329±335.Nozaki Y, Reynolds J, Tanford C 1978): Conformationalstates of a hydrophobic protein: the coat proteinof fd bacteriophage. Biochem 17:1239±1246.Ojala PM, Romantschuk M, Bamford DH 1990): Purifiedf6 nucleocapsids are capable of productive infectionof host cells with partially disrupted outermembranes. Virol 178:364±372.Onodera S, Qiao X, Qiao J, Mindich L 1998): Directedchanges in the number of double-stranded RNAsegmentsin bacteriophage f6. Proc Natl Acad Sci USA95:3920±3924.Peeters BPH, Peters RM, Schoenmakers JGG, KoningsRNH 1985): Nucleotide sequence and genetic organizationof the genome of the N-specific filamentousbacteriophage 1Ke. Comparison with the genome ofthe Ff-specific filamentous phages M13, fd and fl. JMol Biol 181:27±39.Peeters BPH, Schoenmakers JGG, Konings RNH1986): Plasmid pKUN9, a versatile vector for theselective packaging of both DNAstrands into singlestranded DNA-containing phage-like particles. Gene41:39±46.Pollock TJ, Tessman 1, Tessman ES 1978): Potentialfor variability through multiple gene products of bacteriophagefX174. Nature 274:34±37.Pratt D 1969): Genetics of single-stranded DNAbacteriophages.Ann Rev Genet 3:343±361.Rasched L, Oberer E 1986): Ff coliphages: Structuraland functional relationships. Microbiol Rev50:401±427.Ray DS 1968): The small DNA-containing bacteriophages.In Fraenkel-Conrat H ed): ``Molecular Basisof Virology''. New York: Reinhold, pp 222±254.Ray DS 1977): Replication of filamentous bacteriophages.Compr Virol 7:105±178.Ray DS, Sheckman RW 1969): Replication of bacteriophageM13 1. Sedimentation analysis of crude lysatesof M13±infected bacteria. Biochim Biophys Acta179:398±407.Romantschuk M, Olkkonen VM, Bamford DH 1988):The nucleocapsid of bacteriophage f6 penetrates thehost cytoplasmic membrane. Embo J 7:1821±1829.Roof WD, Horne SM, Young KD, Young R 1994):slyD, a host gene required for fX174 lysis, is relatedto the FK506-binding protein family of peptidyl-prolylcis-trans-isomerases. J Biol Chem 269:2902±2910.Roof WD, Young R 1993): fX174 E complementslambda S and R dysfunction for host cell lysis. JBacteriol 175:3909±3912.Russel M 1991): Filamentous phage assembly. MolMicrobiol 5:1607±1613.Russel M 1995): Moving through the membrane withfilamentous phages. Trends Microbiol 3:223±228.Russel M, Model F 1985): Thioredoxin is required forfilamentous phage assembly. Proc Natl Acad Sci USA82:29±33.Russell PW, Muller UR 1984): Construction of bacteriophagefX174 mutants with maximum genomesizes. Virol 52:822±827.Salivar WO, Henry T, Pratt D 1967): Purification andproperties of diploid particles of coliphage M13. Virol32:41±51.Salivar WO, Tzagoloff H, Pratt D 1964): Some physical-chemicaland biological properties of the rodshapedcoliophage M13. Virol 24:359±371.Sanger F, Air G, Barrell BG, Brown NL, Coulson AR,Fiddes JC, Hutchison CA, Slocombe PM, Smith M1977): Nucleotide sequence of bacteriophage fX174DNA. Nature 265:687±695.Sanger F, Coulson AR, Friedmann T, Air GM, BarrellBG, Brown NL, Fiddes JC, Hutchison CAIII, SlocombePM, Smith M 1978): The nucleotide sequenceof bacteriophage fX174. J Mol Biol 125:225±246.Sanger F, Nicklen S, Coulson AR 1977): DNA sequencingwith chain terminating inhibitors. Proc NatlAcad Sci USA 74:5463±5468.Scott JR, Zinder ND 1967); Heterozygotes of phage fl.In Colter JS, Faranchych W eds): The MolecularBiology of Viruses. New York: Academic Press, pp211±218.Shaw DC, Walker JE, Northrop FD, Barrell BG,Godson GN, Fiddes JC 1978): Gene K, a new overlappinggene in bacteriophage G4. Nature272:510±515.Shen CK, Ikoku A, Hearst JE 1979): A specific DNAorientation in the filamentous bacteriophage fd asprobed by psoralen crosslinking and electron microscopy.J Mol Biol 127:163±175.Sinsheimer RL 1959a): Purification and properties ofbacteriophage fX174. J Mol Biol 1:37±42.Sinsheimer RL 1959b): Asingle-stranded deoxyribonucleicacid from bacteriophage fX174. J Mol Biol1:43±53.Sinsheimer R 1966): fX: Multum in parvo. In Cairns J,Stent GS, Watson JD eds): Phage and the Origins ofMolecular Biology. Cold Spring Harbor, NY: ColdSpring Harbor Laboratory, pp 258±264.Sinsheimer RL 1968): Bacteriophage fX174 and relatedviruses. Prog Nucleic Acid Res Mol Biol8:115±169.

SINGLE-STRANDED DNA PHAGES 175Sinsheimer RL 1991): The discovery of a singlestranded,circular DNAgenome. BioEssays 13:89±91.Smith M, Brown NL, Air GM, Barrel BG, Coulson AR,Hutchison CA, Sanger F 1977): DNA sequence atthe C termini of the overlapping genes Aand B inbacteriophage fX174. Nature 265:702±705.Smits M, Jansen J, Konings R, Schoenmakers J1984): Initiation and termination signals for transcriptionin bacteriophage M13. Nucleic Acids Res12:4071±4081.Spencer JV, Newcomb WW, Thomsen DR, Homa FL,Brown JC 1998): Assembly of herpes simplex viruscapsid: preformed triplexes bind to nascent capsid. JVirol 72:3944±3951.Stockley PG, Stonehouse NJ, Valegard K 1994): Themolecular mechanism of RNAphage morphogenesis.Int J Biochem 26:1249±1260.Stump MS, Madison-Antenucci S, Kokoska RJ, SteegeDA1997): Filamentous phage IKe mRNAs conserveform and function despite divergence in regulatoryelements. J Mol Biol 266:51±65.Suzuki R, Inagaki M, Karita S, Kawaura T, Kato M,Nishikawa S, Kashim N, Morita J 1999): Specificinteraction of fused H protein of bacteriophagefX174 with receptor lipopolysaccharides. Virus Res60:95±99.Sun T, Webster R 1986): fii, a bacterial locus requiredfor filamentous phage infection and its relation tocolicin-tolerant tolA and tolB. J Bacteriol 165:107±115.Tessman ES, Tessman 1 1978): The genes of the isometricphages and their functions. In Denhardt DT,Dressler D, Ray DS eds): ``The Single-StrandedDNAPhages''. Cold Spring Harbor, NY: ColdSpring Harbor Laboratory, pp 9±29.Tessman 1 1959): Some unusual properties of the nucleicacid in bacteriophage S13 and fX174. Virol7:263±275.Tessman 1, Tessman ES, Stent GS 1957): The relativeradiosensitivity of bacteriophages S13 and T2. Virol4:209±215.Van Duin J 1988): Single-stranded RNAbacteriophages.In Calendar R ed): The Bacteriophages.New York: Plenum Press, Vol. 1, pp 117±167.van Wezenbeek PMGF, Hulsebos TJM, SchoenmakersJGG 1980): Nucleotide sequence of the filamentousbacteriophage M13 DNAgenome: Comparison withphage fd. Gene 11:129±148.Vieira J, Messing J 1987): Production of single-strandedplasmid DNA. Methods Enzymol 153:3±11.Wallace RB, Johnson MJ, Suggs SV, Miyoshi K, BhattR, Itakura K 1981): Aset of synthetic oligodeoxyribonucleotideprimers for DNAsequencing in theplasmid vector pBR322. Gene 16:21±26.Waterhouse P, Griffiths AD, Johnson KS, Winter G1993): Combinatorial infection and in vivo recombination:Astrategy for making large phage antibodyrepertoires Nucleic Acids Res 21:2265±2266.Webster RE, Cashman JS 1978): Morphogenesis ofthe filamentous single-stranded DNAphages. InDenhardt DT, Dressler D, Ray DS eds):The Single-Stranded DNA Phages. Cold SpringHarbor, NY: Cold Spring Harbor Laboratory, pp557±569.Webster BE, Grant RA, Hamilton LW 1981): Orientationof the DNAin the filamentous bacteriophage fl. JMol Biol 152:357±374.Weisbeek PJ, Borrais WE, Langeveld SA, Baas FD, vanArkel GA 1977): Bacteriophage fX174: Gene Aoverlaps gene B. Proc Natl Acad Sci USA74:2504±2508.Weisbeek PJ, van Arkel GA 1978): The isometric phagegenome: Physical structure and correlation with thegenetic map. In Denhardt DT, Dressler D, Ray DSeds): ``The Single-Stranded DNAPhages''. ColdSpring Harbor, NY: Cold Spring Harbor Laboratory,pp 31±49.Wickner RB, Wright M, Wickner S, Hurwitz J 1972):Conversion of fX174 and fd single-stranded DNAtoreplicative forms in extracts of Escherichia coli. ProcNatl Acad Sci USA 69:3233±3237.Wickner W 1983): M13 coat protein as a model ofmembrane assembly. Trends Biochem Sci 8:90±94.Wickner W, Brutlag D, Schekman R, Kornberg A1973): RNAsynthesis initiates in vitro conversionof M13 DNAto its replicative form. Proc Natl AcadSci USA69:965±969.Wilson DR, Finlay BB 1998): Phage display: applications,innovations, and issues in phage and host biology.Can J Microbiol 44:313±329.Wollin R, Bruse GW, Jansson PE, Lindberg AA 1989):Definition of the phage G13 receptor as structuraldomains of trisaccharides in Salmonella and Escherichiacoli core oligosaccharide. J Mol Recognit2:37±43.Yen TSB, Webster RE 1981): Bacteriophage f1 gene IIand X proteins. Isolation and characterization of theproducts of two overlapping genes. J Biol Chem256:11259±11265.Yen TSB, Webster RE 1982): Translational control ofbacteriophage f1 gene II and gene X proteins by geneV protein. Cell 29:337±345.Zagursky RJ, Berman ML 1984): Cloning vectors thatyield high levels of single-stranded DNAfor rapidDNAsequencing. Gene 27:183±191.Zaman GJR, Kaan AM, Schoenmakers JGG, KoningsRNH 1992): Gene V protein-mediated translationalregulation of the synthesis of gene II protein of thefilamentous bacteriophage M13: Adispensable functionof the filamentous-phage genome. J Bacteriol174:595±600.Zinder ND 1975): ``RNAPhages''. Cold SpringHarbor, NY: Cold Spring Harbor Laboratory.

176 LECLERCZinder ND, Boeke J 1982): The filamentous phageFf) as vectors for recombinant DNA. Gene19:1±10.Zinder ND, Horiuchi K 1985): Multiregulatory elementof filamentous bacteriophages. Microbiol Rev49:101±106.Zinder ND, Valentine RC, Roger M, Stoeckenius W1963): fl, a rod shaped male-specific bacteriophagethat contains DNA. Virol 20:638±640.Zoller MJ, Smith M 1983): Oligonucleotide-directedmutagenesis of DNAfragments cloned into M13vectors. Methods Enzymol 100:468±500.