Human evolution: Gradual or Punctuated? Accelerated integration of mtDNA pseudogenes

coincides with emergence of the human genus. A Hypothesis.

Konstantin Gunbin1, Konstantin Popadin2, Leon Peshkin3, Sofia Annis4, Rebecca Ackermann5, and Konstantin Khrapko4*

1 University of Lausanne, Switzerland, 2 Novosibirsk State University, Russia3 Harvard Medical School, Boston USA, 4 Northeastern University, Boston USA, 5 University of Cape Town, South Africa

Fig.2.TheSpecificQues1on:Ishumanevolutionagradualprocessorasuccessionofrapiddiscontinuous changes?Thishas been a focus of intense debatefor decades. Of particular interest has been the mid-Plioceneclimate change event ~2.9-2.5 Ma, which is thought to haveprecipitated the separationof the genusHomoas an independentlineage,i.e.separationofHumansfromnon-Humans(~2.8Ma).Thedebate mostly concerned continuity/punctuality of the fossilrecord,butofcoursetherateoftheunderlyinggeneticchangeisofultimate interest/importance. Did hominid lineage experience anincreasedmutationratewhenalargenumberofdistincthomininsemerged and eventually gave rise to the split betweenAustralopitecus/ParanthropusandHomo?Ref:(deMenocal2004).

Fig.4.Solu1on: NUMTs – mutations with an internal timer. NUMTs areinsertionsofmtDNAsequencesintothenucleargenome.Unlikepointmutation,eachNUMTsactuallyrepresentsabranchonthemtDNA phylogenic tree and thus its time of insertion can bedetermined as precise as their branching point can bepositionedonthetree.Inasense,NUMTsare“mutationswithaninternalclock”,whichissynchronizedwiththewell-establishedmtDNA mutation evolution clock. By determining the NUMTs’insertion time points, one can ask whether NUMTs wereinserted uniformly over time or preferentially during certainperiods of evolution, as implied by the “punctuated evolution”model. AtechnicalpeculiarityisthataNUMTdoesnotnecessarilyinsertintothenucleargenomeatthetimeNUMTbranching.ThemtDNA ancestor of a NUMT might have been evolving as anindependentmtDNAlineagepriortoinsertion,thereforeNUMTbranchmay consist of themitochondrial and the pseudogenicsegments(blackandredintheYigureabove).Theproportionofthemitochondrialsegmentandthusthetimeofintegrationmaybe estimated based of the synonymity of themutations of thebranch, which ought to be a linear combination of highsynonymitymitochondrialmutations,whichareunderselectivepressure and low synonymity of the pseudogene mutations(whicharenot).

Abstract:Fragments of mitochondrial DNA are known to get inserted intonuclear DNA to form NUMTs, i.e. nuclear pseudogenes of themtDNA. The insertion of a NUMT is a rare event. Hundreds ofpseudogenes have been cataloged in the human genome. NUMTsare,inessence,aspecialtypeofmutation,but,unlikeconventionalpoint mutations, NUMTs are mutations with their own internaltimer,which is synchronizedwith an establishedmolecular clock,themtDNA.ThusinsertionofNUMTscanbetimedwithrespecttoevolution milestones such as the emergence of new species. Weasked whether NUMTs were inserted uniformly over time orpreferentiallyduringcertainperiodsofevolution,asimpliedbythe“punctuated evolution” model. Interestingly, the NUMT insertiontimesdoappearnonrandomwithatleastoneclusterpositionedataround 2.8 million years ago (Ma). Interestingly, 2.8 Ma closelycorrespondstothetimeofemergenceofthegenusHomo,andtoawell-documentedperiodofmajorclimatechangeca.2.9–2.5Ma.ItistemptingtohypothesizethattheinsertionofNUMTsisrelatedtothe speciation process. NUMTs could be either “riders”, i.e., theirinsertion could be facilitated by the overall higher genomeYlexibilityduringspeciation,or“drivers”,i.e.theymaymorereadilygetfixedinthepopulationduetopositiveselectionassociatedwithspeciation. If correct, the hypothesis would support the idea thatevolutionofourgenusmayhavehappened ina rapid,punctuatedmanner.Ref:(Gunbinetal.2017).

Fig.3.TechnicalProblem:Imprecisetimingofconventionalmutations.ThedifYicultyinansweringtheabovequestionsliesinthewayancientmutationshavetobetimed.ConventionalpointmutationsareassignedtospeciYicbranchesoftheDNAphylogenetictrees.Theessenceofthe problem is that mutations can be located within branchsegmentsfrombranchingpointtobranchingpoint,buttheexactposition within the segment is principally unknown. Becausethe hominid DNA-derived phylogenetic tree has very lowbranch density, the precision of mutation timing is low, e.g.,human-speciYicmutationscanbepositionedwithin~6Myfromseparation from chimpanzee till the branching of theDenisovans.

Fig.1.TheGeneralQues1on:Is evolutiona processof gradual adaptationand continuouschange of form/function OR a succession of rapiddiscontinuousshifts?Andcanthiscontinuity/discontinuitybeobserved at the level of the rate of genetic change, i.e. is therateof“critical”mutationsconstantorpunctuated?

Fig.5.Results:Hundreds of pseudogenes have been cataloged in the humangenomethathavebeeninsertedoverthelast~60Myofwhichweconsidered the last 6 My. Various quality Yilters resulted in theselection of 18 NUMTs most suitable for phylogenetic analysis.Insertion times of these 18 NUMTs appear non-randomlydistributed with one cluster positioned around 2.8Ma. Whiletiming of insertion of individualNUMTs is imprecise, the overallprobability of forming such a cluster by chance is low, whichmakesthisobservationhighlystatisticallysigniYicant.

Fig.6.Sta1s1calsignificance:SigniYicancewasestimatedbysimulation. We generated10618–point random data sets drawn from the uniform distributionbetween 0and6Ma.Thenwecalculatedthefractionof18-pointsets that were “as highly clustered or more extremely clusteredthan the distribution of NUMT insertion times” (Fig. 5), i.e., thenumberoftimepointsthatlandedwithin2.5-2.9Maintervalwas>= 6. The corresponding probability (p-value proxy) is 0.0006.This impressively low probability, however, depends on theintervalwidthand theassumedHuman/Chimpanzeedivergencetime (because we use it to calibrate the mtDNA clock). Wetherefore repeated these estimates with different widths ofintervals centered at 2.8Ma andwith different divergence times(5,5.5,…,9Ma).TheresultsofthisanalysisareshowninFigure6:p-valuestayedwellbelowthecritical0.05valuewithinawiderange of intervalwidths and human/chimp divergence times. Atthe same time, these results put limitations on our conclusions.For example, if divergence time between humans andchimpanzees were as ancient as 8Ma, then our data are notsufYicient to statistically support temporal association ofacceleratedNUMTinsertionintonuclearDNAwiththeemergenceofHomoand/orclimaticchange2.8Ma.

Discussion:ItistemptingtohypothesizethatacceleratedinsertionofNUMTsissomehowlinkedtotheHomospeciationprocess.NUMTscouldbe either "riders", i.e., their insertion could be facilitated by theoverallhighergenomeYlexibilityduringthe speciationperiod(ifsuch extra Ylexibility indeed exists), or "drivers", i.e. they maymore readily get Yixed in the population during the period ofspeciationduetoincreasedselectivepressures.If correct, the hypothesis of acceleratedpseudogenizationwouldsupport the idea that evolution of our genusmight have been apunctuatedprocess.

References:Gunbin, K.V., Popadin, K., Peshkin, L., Ackermann, R.R., andKhrapko,K.(2017).IntegrationofmtDNApseudogenesintothenuclear genome coincides with critical periods in humanevolution. A Hypothesis. Mitochondrion 34, pp. 20-23, doi:10.1016/j.mito.2016.12.001(2017).deMenocal, P. B. (2004). African climate change and faunalevolutionduring thePliocene-Pleistocene.Earth andPlanetaryScienceLetters,220(1-2),3–24.Elsevier.