微卫星基因组分布假定功能和突变机制.docx

微卫星基因组分布假定功能和突变机制.docx

《微卫星基因组分布假定功能和突变机制.docx》由会员分享，可在线阅读，更多相关《微卫星基因组分布假定功能和突变机制.docx（18页珍藏版）》请在冰豆网上搜索。

微卫星基因组分布假定功能和突变机制

微卫星：

基因组分布，假定功能和突变机制

You-ChunLi*,AbrahamB.Korol,TzionFahima,AvigdorBeilesandEviatarNevo

Microsatellites,ortandemsimplesequencerepeats（SSR）,areabundantacrossgenomesandshowhighlevelsofpolymorphism.SSRgeneticandevolutionarymechanismsremaincontroversial.HereweattempttosummarizetheavailabledatarelatedtoSSRdistributionincodingandnoncodingregionsofgenomesandSSRfunctionalimportance.NumerouslinesofevidencedemonstratethatSSRgenomicdistributionisnonrandom.RandomexpansionsorcontractionsappeartobeselectedagainstforatleastpartofSSRloci,presumablybecauseoftheireffectonchromatinorganization,regulationofgeneactivity,recombination,DNAreplication,cellcycle,mismatchrepairsystem,etc.Thisreviewalsodiscussestheroleoftwoputativemutationalmechanisms,replicationslippageandrecombination,andtheirinteractioninSSRvariation.

Genomicmicrosatellites（simplesequencerepeats;SSRs）,iterationsof1-6bpnucleotidemotifs,havebeendetectedinthegenomesofeveryorganismanalysedsofar,andareoftenfoundatfrequenciesmuchhigherthanwouldbepredictedpurelyonthegroundsofbasecomposition（Tautz&Renz1984;Epplenetal.1993）.Bell（1996）suggestedthattheabundanceandlengthdistributionofSSRsacrossthegenomecouldresultfromunbiasedsingle-steprandom-walkprocesses.SomeauthorsconsideredSSRstobeselectivelyneutralsequencesrandomlyoralmostrandomlydistributedovertheeuchromaticgenome（Schlötterer&Wiehe1999;Schlötterer2000）.Bachtrogetal.（1999）detectedasignificantpositivecorrelationbetweenATcontentand（AT/TA）nSSRdensity,suggestingthatSSRgenesismaybearandomprocess.However,theyalsofoundthat39%ofthecontiguoussequencesanalyseddeviatefromrandomdistributionofSSRsinDrosophilamelanogaster.

ThecontroversialinterpretationofSSRevolutionisreflectedinrecentliterature.NumerousstudieshavedocumentedSSRstructuralpatternswithallelesizeconstraints（Garzaetal.1995;Dermitzakisetal.1998;Samadietal.1998;Lietal.2000c;2002a）,andfunctionalsignificance（reviewedin:

Kashietal.1997;Kingetal.1997;Kashi&Soller1999;King&Soller1999;Gur-Arieetal.2000）.Nevertheless,SSRsareusuallyjustconsideredasevolutionarilyneutralDNAmarkers（e.g.Tachida&Iizuka1992;Awadalla&Ritland1997;Schlötterer&Wiehe1999）.SuchcontroversycallsforfurtherevidenceofSSRfunctionalimportanceandjustifiesacomprehensivediscussionconcerningtheevolutionarysignificanceofgenomicSSRs.AnattempttoanalysethephenomenonofSSRvariationfromtheviewpointofthe'qualitative'alternative,i.e.functionalvs.neutral,doesnotseemtosuittheproblem.Infact,theremaybenofundamentalcontradictionintheoppositeinterpretationsofSSRvariation:

functionalvs.neutral,ifthequestionisformulatedinquantitativeratherthanqualitativeterms.TherichevidenceaccumulatedonSSRsandtheirmanifoldeffectsjustifiesthisapproach.

Thepresentreviewfocusesonthefollowingaspects:

SSRdistributionacrosscodingandnoncodingregionsofthegenome;（ii）evolutionarysignificanceanddynamicsofSSRgenomicdistribution;（iii）SSReffects/functionsingeneexpressionandgeneticdisorder,chromatinorganization,cellcycle,andDNAmetabolicprocesses;and（iv）therelativecontributionofreplicationslippageandassociatedDNArepairmechanismsandrecombinationtoSSRmutation.

SSRsconstitutearatherlargefractionofnoncodingDNAandarerelativelyrareinprotein-codingregions.Forinstance,alloftheobserved101mono-,di-andtetranucleotideSSRswerelocatedinnoncodingregionsacross54plantspecies（Wangetal.1994）.AlltypesofSSRs（frommono-tohexanucleotiderepeats）werefoundinexcess（comparedtorandomappearance）innoncodinggenomicregionsacrossseveneukaryoticclades:

Saccharomycescerevisiae,Caenorhabditiselegans,Schizosaccharomycespombe,Mus,Drosophila,plants,andprimates（Metzgaretal.2000）.Morganteetal.（2002）reportedthatallSSRtypesexcepttripletsandhexanucleotidesaresignificantlylessfrequentinthe25762predictedprotein-codingsequencescomparedwiththenoncodingfractioninsixplantspeciesincludingArabidopsis,rice,soybean,maize,andwheat（Triticumaestivum）.InthegenomeofJapanesepufferfish,Fugurubripes,only11.6%of6042SSRsweredetectedinprotein-codingregions（Edwardsetal.1998）.Thisisattributabletonegativeselectionagainstframeshiftmutationsincodingregions（Metzgaretal.2000）.Previously,asimilardistributionpatternwasfoundfortripletSSRsincodingandnoncodingregionsofgenomesoffungi,protists,prokaryotes,viruses,organelles,plasmidsandhumans（Field&Wills1996;Wrenetal.2000）.However,thedisease-associatedtripletrepeatsaremostlyfoundincodingregionsofthehumangenome（Nadiretal.1996）.Likewise,Morganteetal.（2002）recentlyfoundthattripletSSRsdoubledinfrequencyinthecodingregionoftheabove-mentionedsixplantspecies,asaresultofmutationpressureandpossiblypositiveselectionforspecificsingleaminoacidstretches.Sometri-arraysarenotextensivelyconservedforlongperiodsoftime,evenwhentheyformpartsofprotein-codingsequences,sincelongtri-repeats（e.g.CAGarrays）canbedestabilizedduringmeiosisorgametogenesis（Jankowskietal.2000）.

ThemajorityofSSRs（48-67%）foundinmanyspeciesaredinucleotides（Wangetal.1994;Schugetal.1998）,butinprimatesmononucleotides[mainly,poly（A/T）tracts]arethemostcopiousclassesofSSRs（Tóthetal.2000;Wrenetal.2000）.IncontrasttothetripletSSRs,di-andtetranucleotideSSRsaremuchlessfrequentincodingregionsthaninnoncodingregions.Forexample,dinucleotiderepeatsareabout20timeslessfrequentintheexpressedsequencesthaninrandomgenomicclonesofNorwayspruce,Piceaabies（Scottietal.2000）.Ineightprokaryotesandyeast,longmono-anddi-tractsarealmostexclusivelydistributedinnontranslatedregions（Field&Wills1998）.ForperfectdimericSSRs,Bell&Jurka（1997）foundthatshorterrepeats（3units）incodingsequencesandotherfunctionallyimportantDNAregionscanbepredictedbyaBernoullimodel;and（ii）thelengthdistributionoflonger（5）perfectdimericSSRDNAinanoncodingregionfitstheunbiasedsingle-stepmutationmodel.Inthismodel,repeatsarepermittedtochangelengthbyplusorminusoneunit,withequalprobabilities,andbasesubstitutionsareallowedtodestroylongperfectrepeats,producingtwoshortperfectrepeats.AnalysisofavailableDNAsequencesofhuman,mouse,worm（Caenorhabditiselegans）,andyeastgenomesshowsthatthedistributionfunctionsofallpossibledimericSSRsareexponentialincodingDNA,whereasinnoncodingDNAmostofthedimericSSRshavesurprisinglylongtailsthatbetterfitapower-lawfunction（Dokholyanetal.2000）.Theselong,nonexponentialtailsarehypothesizedtoresultfromahighertoleranceofnoncodingDNAtomutations（Dokholyanetal.2000）.AnumberofgeneswasfoundwithdinucleotideSSRsintheuntranslated5'and/or3'ends,suchasinfivegenesofchannelcatfishIctaluruspunctatus（Liuetal.1999）andinthemammalianheatshockprotein70（hsp70）genes[（GA）6CAG（TC）24tract:

Lisowskaetal.（1997）].DinucleotideSSRsarealsofoundinintrons.Forinstance,inMusmusculustheintronAofAdh-1genecontains（TA）14（TG）8,and（TA）19,andtheintronofIL-5genecontains（AT）17;inthetreeBetulapendutheintronofBVGC34geneincludes（CA）17（TA）14,and（TGTA）3.Thepotentialsizeexpansionofdi-ortetranucleotideSSRsatthe3'and5'regionsandintronscouldleadtodisruptionoftheoriginalproteinand/orformationofnewgenesbyframeshift（Bachtrogetal.1999;Liuetal.1999）.Thesepatternssuggestthatrandomdistributionsofsuchdi-andtetranucleotideSSRswerestronglyselectedagainst（Bachtrogetal.1999;Liuetal.1999）.Foragivennumberofrepeats,thetetranucleotidelocusislongerthanadinucleotide.Thismayaffecttheselectivepressure,ifthestabilityofmeioticprocessesdependsontheabsolutesize（inbasepairs）ofthetargetregion.Lociwithlongerrepeatunitsseemtoexperiencestrongerselectionagainstthedifferenceinsize,especiallyingenomeregionswithhighrecombinationrates（Samadietal.1998）.

ThesefindingsalsosuggestthatthedifferencesbetweencodingandnoncodingSSRfrequenciesarisefromspecificselectionagainstframe-shiftmutationsincodingregionsresultingfromlengthchangesinnontripletrepeats（Liuetal.1999;Dokholyanetal.2000）.Nevertheless,14%ofallproteinscontainrepeatedsequences,withathreetimeshigherabundanceofrepeatsineukaryotescomparedtoprokaryotes（Marcotteetal.1999）.Prokaryoticandeukaryoticrepeatfamiliesareclusteredtononhomologousproteins.Thismayindicatethatrepeatedsequencesemergedafterthesetwokingdomshadsplit.Theeukaryotesincorporatingmorerepeatsmayhaveanevolutionaryadvantageoffasteradaptationtonewenvironments（Marcotteetal.1999;seealsobelowsection5andKashietal.1997;King&Soller1999;Wrenetal.2000）.

Tóthetal.（2000）conductedadetailedanalysisofSSRsinseveraleukaryotictaxa,fromfungitohumans,andrevealedhighlytaxon-specificpatternsinthedistributionofdifferentrepeattypes（frommono-uptohexanucleotides）fordifferentmotifsincodingandnoncodingsequences,inintronsandintergenicregions.Thisspecificitycanpartlybeexplainedbyinteractionofmutationmechanismsanddifferentialselection.TheaccumulatedempiricalevidenceseemstoindicatethatSSRsequencesaremoreabundantandlongerinvertebratesthan