Initial sequencing and analysis of the human genome - Vitagenes

articlesinclude the DiGeorge/velocardiofacial syndrome region on chromosome22 (ref. 238) and the Williams±Beuren syndrome recurrentdeletion on chromosome 7 (ref. 239).The availability of the genome sequence also allows rapid identi-®cation of paralogues of disease genes, which is valuable for tworeasons. First, mutations in a paralogous gene may give rise to arelated genetic disease. A good example, discovered through use ofthe genome sequence, is achromatopsia (complete colour blindness).The CNGA3 gene, encoding the a-subunit of the conephotoreceptor cyclic GMP-gated channel, had been shown toharbour mutations in some families with achromatopsia. Computationalsearching of the genome sequences revealed the paralogousgene encoding the corresponding b-subunit, CNGB3 (which hadnot been apparent from EST databases). The CNGB3 gene wasrapidly shown to be the cause of achromatopsia in otherfamilies 407,408 . Another example is provided by the presenilin-1and presenilin-2 genes, in which mutations can cause early-onsetAlzheimer's disease 423,424 . Second, the paralogue may provide anopportunity for therapeutic intervention, as exempli®ed byattempts to reactivate the fetally expressed haemoglobin genes inindividuals with sickle cell disease or b-thalassaemia, caused bymutations in the b-globin gene 425 .We undertook a systematic search for paralogues of 971 knownhuman disease genes with entries in both the Online MendelianInheritance in Man (OMIM) database (http://www.ncbi.nlm.nih.gov/Omim/) and either the SwissProt or TrEMBL protein databases.We identi®ed 286 potential paralogues (with the requirement of amatch of at least 50 amino acids with identity greater than 70% butless than 90% if on the same chromosome, and less than 95% if on adifferent chromosome). Although this analysis may have identi®edsome pseudogenes, 89% of the matches showed homology overmore than one exon in the new target sequence, suggesting thatmany are functional. This analysis shows the potential for rapididenti®cation of disease gene paralogues in silico.Drug targetsOver the past century, the pharmaceutical industry has largelydepended upon a limited set of drug targets to develop newTable 26 Disease genes positionally cloned using the draft genomesequenceLocus Disorder Reference(s)BRCA2 Breast cancer susceptibility 55AIREAutoimmune polyglandular syndrome type 1 (APS1 389or APECED)PEX1 Peroxisome biogenesis disorder 390, 391PDS Pendred syndrome 392XLP X-linked lymphoproliferative disease 393DFNA5 Nonsyndromic deafness 394ATP2A2 Darier's disease 395SEDL X-linked spondyloepiphyseal dysplasia tarda 396WISP3 Progressive pseudorheumatoid dysplasia 397CCM1 Cerebral cavernous malformations 398, 399COL11A2/DFNA13 Nonsyndromic deafness 400LGMD 2G Limb-girdle muscular dystrophy 401EVCEllis-Van Creveld syndrome, Weyer's acrodental 402dysostosisACTN4 Familial focal segmental glomerulosclerosis 403SCN1A Generalized epilepsy with febrile seizures plus type 2 404AASS Familial hyperlysinaemia 405NDRG1 Hereditary motor and sensory neuropathy-Lom 406CNGB3 Total colour-blindness 407, 408MUL Mulibrey nanism 409USH1C Usher type 1C 410, 411MYH9 May-Hegglin anomaly 412, 413PRKAR1A Carney's complex 414MYH9 Nonsyndromic hereditary deafness DFNA17 415SCA10 Spinocerebellar ataxia type 10 416OPA1 Optic atrophy 417XLCSNB X-linked congenital stationary night blindness 418FGF23 Hypophosphataemic rickets 419GAN Giant axonal neuropathy 420AAAS Triple-A syndrome 421HSPG2 Schwartz-Jampel syndrome 422.............................................................................................................................................................................therapies. A recent compendium 426,427 lists 483 drug targets asaccounting for virtually all drugs on the market. Knowing thecomplete set of human genes and proteins will greatly expand thesearch for suitable drug targets. Although only a minority of humangenes may be drug targets, it has been predicted that the number willexceed several thousand, and this prospect has led to a massiveexpansion of genomic research in pharmaceutical research anddevelopment. A few examples will illustrate the point.(1) The neurotransmitter serotonin (5-HT) mediates rapid excitatoryresponses through ligand-gated channels. The previouslyidenti®ed 5-HT 3A receptor gene produces functional receptors,but with a much smaller conductance than observed in vivo.Cross-hybridization experiments and analysis of ESTs failed toreveal any other homologues of the known receptor. Recently,however, by searching the human draft genome sequence at lowstringency, a putative homologue was identi®ed within a PAC clonefrom the long arm of chromosome 11 (ref. 428). The homologuewas shown to be expressed in the amygdala, caudate and hippocampus,and a full-length cDNA was subsequently obtained. Thegene, which codes for a serotonin receptor, was named 5-HT 3B .When assembled in a heterodimer with 5-HT 3A , it was shown toaccount for the large-conductance neuronal serotonin channel.Given the central role of the serotonin pathway in mood disordersand schizophrenia, the discovery of a major new therapeutic targetis of considerable interest.(2) The contractile and in¯ammatory actions of the cysteinylleukotrienes, formerly known as the slow reacting substance ofanaphylaxis (SRS-A), are mediated through speci®c receptors. Thesecond such receptor, CysLT 2 , was identi®ed using the combinationof a rat EST and the human genome sequence. This led to thecloning of a gene with 38% amino-acid identity to the only otherreceptor that had previously been identi®ed 429 . This new receptor,which shows high-af®nity binding to several leukotrienes, maps to aregion of chromosome 13 that is linked to atopic asthma. The geneis expressed in airway smooth muscles and in the heart. As theleukotriene pathway has been a signi®cant target for the developmentof drugs against asthma, the discovery of a new receptor hasobvious and important consequences.(3) Abundant deposition of b-amyloid in senile plaques is thehallmark of Alzheimer's disease. b-Amyloid is generated by proteolyticprocessing of the amyloid precursor protein (APP). One ofthe enzymes involved is the b-site APP-cleaving enzyme (BACE),which is a transmembrane aspartyl protease. Computationalsearching of the public human draft genome sequence recentlyidenti®ed a new sequence homologous to BACE, encoding a proteinnow named BACE2 430,431 . BACE2, which has 52% amino-acidsequence identity to BACE, contains two active protease sites andmaps to the obligatory Down's syndrome region of chromosome 21,as does APP. This raises the question of whether the extra copies ofboth BACE2 and APP may contribute to accelerated deposition ofb-amyloid in the brains of Down's syndrome patients. The developmentof antagonists to BACE and BACE2 represents a promisingapproach to preventing Alzheimer's disease.Given these examples, we undertook a systematic effort toidentify paralogues of the classic drug target proteins in the draftgenome sequence. The target list 427 was used to identify 603 entriesin the SwissProt database with unique accession numbers. Thesewere then searched against the current genome sequence database,using the requirement that a match should have 70±100% identityto at least 50 amino acids. Matches to named proteins were ignored,as we assumed that these represented known homologues.We found 18 putative novel paralogues (Table 27), includingapparent dopamine receptors, purinergic receptors and insulin-likegrowth factor receptors. In six cases, the novel paralogue matches atleast one EST, adding con®dence that this search process canidentify novel functional genes. For the remaining 12 putativeparalogues without an EST match, all have long ORFs and all but912 © 2001 Macmillan Magazines Ltd NATURE | VOL 409 | 15 FEBRUARY 2001 | www.nature.com