PLANTBIOTECHNOLOGY:FOODANDFEED

127,000expressedsequencetag(EST)se-quencesfrom19plantspeciesinpublicda-tabasesandthenumberisexpectedtogrowrapidlyduringthenextseveralyears.Thesesequenceswillprovideisomorphismsbe-tweenthemodelgenomesandotherspecies,formingakindoftransectthroughgenomediversityinhigherplantsthatisanchoredincomprehensiveknowledgeofthetworepre-sentativespecies.Thus,asgenesassociatedwithfunctionsortraitsinoneplantarecloned,itusuallywillbepossibletoidentifytheorthologsresponsibleforthetraitinotherplantspeciesbyadatabasesearchorbyusingthesequenceinformationtoclonethecorre-spondinggenefromthespeciesofinterest.

ofprobablegenefunction,ifitisknown,inanyhigherplant.Thisisillustratedbyacomparisonofsequenceidentityofaran-domsampleofputativelyorthologousriceandArabidopsisproteins(Fig.2).Becauseoftherelativelyrecentradiationoftheangiosperms,weconsideritlikelythattherewillbeveryfewprotein-encodingangiospermgenesthatdonothaveor-thologsorparalogsinArabidopsisorrice.Therefore,understandingthegeneticbasisfordiversitymaydevolvetoidentifyingtherelevantdifferencesinthecontrolofex-pressionorthefunctionofessentiallythesamesetofgenes.Indeed,ithasbeenhypothesizedthatthedevelopmentaldiver-sityofhigherplantsmaybelargelyduetochangesinthecis-regulatorysequencesoftranscriptionalregulators(12).

Amajorchallengetounderstandingthegeneticbasisofinterspeciesdiversityisthat,inatleastsomecases,minorchangesinthestructureorexpressionofagenemayleadtomajorchangesinphenotype.Thiswasrecentlyillustratedforthegenesthatcontrolmodificationsoffattyacidsinplants(13).Higherplantscollectivelypro-ducemorethan200fattyacids,whichac-cumulateasstorageoilsinseeds.Thesefattyacidsdifferprimarilybecauseofthepresenceofdoublebonds,hydroxyls,ep-oxygroups,triplebonds,orsecondarymodificationsofthesefunctionalgroupsatvariouscarbonsalongthefattyacylchains.Ithasrecentlybecomeapparentthatthesefunctionalgroupsareproducedbyafamilyofcloselyrelatedfattyacyldesaturase-likeenzymes(14).Theobservationthatasfewasfouraminoacidsubstitutionscancon-vertadesaturasetoahydroxylaseillus-trateshownewchemicalconstituentscanaccumulatewithouttheevolutionofanob-viouslydistinctenzyme(13).Largegenefamilieshavealsobeenobservedforcyto-chromeP450s,enzymesinvolvedinpoly-saccharidebiosynthesis,disease-relatedgenes,transcriptionfactors,proteinkinases,andphos-phatasestonameafew.Thus,amajorchal-lengeassociatedwithexploitingtheexplosionofgenomeinformationwillbeindeducingrulesthatcanpredicttheprecisefunctionofmembersofgenefamilies.

Onepromisingavenue,termedphylo-genomics,exploitstheuseofevolutionaryinformationtofacilitateassignmentofgenefunction(15).Theapproachisbasedontheideathatfunctionalpredictionscanbegreatlyimprovedbyfocusingonhowgenesbecamesimilarinsequenceduringevolutioninsteadoffocusingonthesequencesimilarityitself.Becausethepoweroftheanalysisincreasesinproportiontothenumberofsequencesthatareavailable,thismethodshouldbecomemoreusefulasthedatabaseofplantsequenc-esexpands.

FloweringPlantsContaintheSameGenes

Althoughfloweringplantshaveevolveddur-ingthepast150millionyearsorsoandthereforemightbeexpectedtobeverysimi-laratthegeneticlevel,substantialdevelop-mentalandmetabolicdiversityexists.Under-standingthebasisforthisdiversityisakeytounderstandinghowtoeffectrationalim-provementsintheproductivityandutilityofcropspecies.Knowledgeofthegeneticbasisforintraspeciesvariationinspecifictraitsshouldbeusefulforselectingorcreatingusefulvariationwithinaspecies.

TheavailabilityofextensiveESTinfor-mationformanyspecies,inconjunctionwiththecompletesequencesofriceandArabidopsis,willallowunambiguousin-sightintothequestionofhowsimilarthegenomesofhigherplantsare.WhentheArabidopsisandricesequencesarecom-plete,itwillbepossibletodirectlycomparealltheESTandotheravailablesequencesfromvariousplantswiththegenomicse-quencesfromthemodelgenomes.Ourpre-liminaryanalysisofavailablesequencessuggeststhatmostgeneproductsfromhigherplantsexhibitadequatesequencesimilaritytodeducedaminoacidsequencesofotherplantgenestopermitassignment

AssigningFunctiontoGenes

Oneofthemajorefficienciesthathasemergedfromplantgenomeresearchtodateisthatabout54%ofhigherplantgenescanbeassignedsomedegreeoffunctionbycomparingthemwiththesequencesofgenesofknownfunction(16)(Fig.3).Ineffect,auniversalbiologyhascoalescedfromthecommonlanguageofgeneandproteinsequences.Unfortunately,knowingthegeneralfunctionfrequentlydoesnotprovideaninsightintothespecificroleintheorganism.Forinstance,onthebasisofsequenceanalysis,about13%ofArabidop-sisgenesareinferredtobeinvolvedintranscriptionorsignaltransduction(16).However,knowingthatageneencodesakinaseortranscriptionfactordoesnotpro-videanyusefulinformationaboutwhatpro-cessesarecontrolledbythesegenes.Thus,completionofthegenomesequencesofAra-bidopsisandricewillbefollowedbyasec-ondphaseoflarge-scalefunctionalgenomicsinwhichallofthe20,000to25,000genesthatmakeupthebasicangiospermgenomewillbeassignedfunctiononthebasisofexperimentalevidence.Consideringthatthecombinedeffortsoftheplantbiologycom-

Fig.2.SequenceidentityofArabi-dopsisandriceproteins.Percentsequenceidentityoverthefulllengthoftheproteinswascalcu-latedfor64randomlyselectedproteinsforwhichtheprobablefunctionwasknownandforwhichfull-lengthornear-full-lengthse-quenceswereavailable.Toavoidcomparingmembersoflargemul-tigenefamilies,wedidnotincludeaproteininthecomparisonifse-quenceswereavailableformorethantwoapparentlyrelatedpro-teinsfromeitherofthespecies.Becauseitisuncertainwhether

other,morecloselyrelatedproteinsareencodedintheunsequencedregionsoftheArabidopsisorricegenome,thisanalysisunderestimatesthedegreeofidentitybetweenthesequencesofArabidopsisandriceproteins.

Fig.3.Functionalclassi cationofpredictedgenesina1.9-MbregionoftheArabidopsisgenome.Proteinrelatedreferstogeneproductsinvolvedinsynthesis,degradation,modi ca-tion,storage,andtargetingofproteinsandinintracellulartraf cking.Analysiswasbasedon389predictedorknowngenes.From(16).

SCIENCEVOL28516JULY1999381