LETTER Evidence for Archaic Asian Ancestry on the Human X Chromosome Daniel Garrigan,* Zahra Mobasher,* Tesa Severson,* Jason A. Wilder,* and Michael F. Hammer*(cid:1) *Division of Biotechnology, and(cid:1)Department ofEEB, University of Arizona, Tucson ThehumanRRM2P4pseudogenehasapatternofnucleotidepolymorphismthatisunlikeanylocuspublishedtodate.A gene tree constructed from a 2.4-kb fragment of the RRM2P4 locus sequenced in a sample of 41 worldwide humans clearlyrootsinEastAsiaandhasamost-recentcommonancestorapproximately2Myrbeforepresent.Thepresenceof thisbasallineageexclusivelyinAsiaresultsinhighernucleotidediversityamongnon-AfricansthanamongAfricans.A global survey of a single-nucleotide polymorphism that is diagnostic for the basal, Asian lineage in 570 individuals shows that itoccurs at frequencies upto 53% in south China, whereas only one of 177surveyed Africans carries this archaic lineage. We suggest that this ancient lineage is a remnant of introgressive hybridization between expanding anatomically modern humans emerging fromAfrica andarchaic populationsin Eurasia. Introduction Recently,Hammeretal.(2004)analyzedglobalhuman size,weestimatedtheparameter,h (¼3N l,whereN is ml e e nucleotidevariationat15X-linkedloci,oneofwhichstands theeffectivepopulationsizeandlisthemutationrate),by out as unique in its pattern of polymorphism. In a global generatingasingle-likelihoodcurvecoveringawiderange sample of 41 individuals, sequence variation at the ribo- of possible values. Based on this value of h , we then ml nucleotidereductaseM2subunitpseudogene4(RRM2P4)is estimatedtheN andTMRCA.TheG/Asingle-nucleotide e partitionedintotwodivergent,basallineages.Bothofthese polymorphismatposition2020wasgenotypedinasample lineagesarefoundinAsia,whereasonlyoneisfoundinsub- of570Africans andnon-Africans with theTaqMangeno- Saharan Africa (fig. 1B). The two lineages differ by five typing assay (Applied Biosystems, Foster City, Calif.), fixedmutations,leadingtoanestimateforthetimetoamost- followingthemanufacturer’sprotocol.Thesampleincluded recentcommonancestor(TMRCA)thatisapproximately2 22KhoisanfromNamibia,39DogonfromMali,47Bantu Myrbeforepresent(B.P.).Suchlongbasalbranchesofthe from South Africa, 25 Bakola from Cameroon, 44 Dinka globalRRM2P4genealogyarenotexpectedunderaneutral fromSudan,46Mongolians,40SriLankans,11Tibetans, panmicticmodel(Wall1999,2000).Moreover,agenetree 48 Baining from New Britain, 26 Japanese, 28 Han from that roots in Asia has not been previously observed in northernChina,30YaofromsouthernChina,28Altaians, publishedhumannucleotidepolymorphismdatasets(e.g., 24PapuaNewGuineans,45ItaliansfromcentralItaly,39 Takahata,Lee,andSatta2001)andraisesintriguingques- Dutch,and27Iranians.Allsamplingprotocolswerecarried tionsconcerningtheorigin andexpansion ofanatomically outwiththeapprovaloftheUniversityofArizonaHuman modern humans. In this paper, we explore the possibility SubjectsCommittee. thatpolymorphismatRRM2P4recoversahistoryofadmix- ture between Homo sapiens and an archaic Asian form of Results and Discussion Homo before the latter went extinct in the late Middle Pleistocene(Swisheretal.1996). Figure1Adepictsthe13polymorphicsitesweobserve inatotalof2,385bpofRRM2P4sequence,andtable1lists summarystatisticsforthelocus.Althoughlevelsofnucleo- Methods tide diversity are only slightly higher than the genomic The methodology and samples used to assay nucleo- average (Yu et al. 2002), RRM2P4 is one of the few sur- tide polymorphism at the RRM2P4 locus has been pre- veyed loci that exhibits more variation in non-Africans viously reported by Hammer et al. (2004). Summary thaninAfricans.ItisworthnotingthatthisRRM2pseudo- statistics were calculated with the computer application geneislocatedinaregionwithahighrateofcrossing-over DnaSPversion3.99(Rozasetal.2003).TheTMRCAwas (;3.6cM/Mb)andalowgenedensityonXq27.3(Hammer estimatedbydividingtheobservednetpairwisenucleotide et al. 2004). Despite its genomic context, our 2.4-kb frag- differencesbetweenlineages(D [Nei1987])bytwicethe ment shows no direct evidence of historical recombina- A mutation rate. We also used a maximum-likelihood ap- tionbythefour-gametetestofHudsonandKaplan(1985). proachtoestimatetheTMRCAwiththecomputerprogram This result is not unexpected, because complete linkage GENETREE version 9.0 (Griffiths 2002). Using the null disequilibrium is known to extend beyond 2.4 kb in both demographic model of panmixia and constant population African and non-African populations (Reich et al. 2001). However, this lack of recombination does permit recon- structionofasingle,nonreticulatinggenetree(fig.1B).The Key words: hybridization, human origins, polymorphism, popula- genomic context of RRM2P4 makes it less likely that the tionsubdivision,pseudogene. elongatedbasalbranchesaretheresultoflinkagetoafunc- E-mail:[email protected]. tional site(s) subject to some form of balancing selection. Mol.Biol.Evol.22(2):189–192.2005 Additionally, there are more restricted conditions for doi:10.1093/molbev/msi013 AdvanceAccesspublicationOctober13,2004 a X-linked balanced polymorphism compared with the Molecular Biology and Evolution vol. 22 no. 2 (cid:1) Society for Molecular Biology and Evolution 2005; all rights reserved. Downloaded from https://academic.oup.com/mbe/article-abstract/22/2/189/963863 by guest on 31 January 2018 190 Garriganetal. FIG.1.—(A)Sequencesand(B)genetreefortheeightRRM2P4haplotypes.Scalebardenotestimeinunitsofmillionsofyearsbeforepresent. Mutationsarelabeledwithadotalongtheedgesofthegenealogyandareplacedaccordingtothemaximum-likelihoodestimateoftheirrespectiveages. BelowthetipsofthegenetreearethemultiplicitiesofeachhaplotypeinboththeAfricanandthenon-Africansamples. autosomes (Hedrick and Parker 1997). Indeed, the fre- givenabovewithanaverageof8.25nucleotidedifferences quency spectrum of mutations at RRM2P4 does not reject observed between the two human RRM2P4 lineages, then the null hypothesis of neutral mutation-drift equilibrium we can estimate that the two lineages diverged approx- (Tajima’sDstatistic¼20.206,P¼0.489)(table1). imately 1.96 MYA. This estimate is in reasonable Outgroup sequencing reveals that there are 18 nucle- agreement with a TMRCA of 2.44 Myr B.P. obtained otide substitutions between human RRM2P4 sequences fromthecoalescent-basedmaximum-likelihoodmethodof and one common chimpanzee (Pan troglodytes) sequence Griffiths(2002),whichassumesapanmicticpopulationof and 62 substitutions between human sequences and one constantsize(fig.1B).Thisplacesthemost-recentcommon orangutan (Pongo pygmaeus) sequence. The human- ancestor ofRRM2P4 in the late Pliocene,a time when the chimpanzee comparison yields a neutral mutation rate of genus Homo first appears in the fossil record (Wood and 7.4310210 substitutions per site per year (assuming a 6 Collard 1999). The maximum-likelihood method also Myr human-chimpanzee divergence time). This rate is estimates that the effective population size of RRM2P4 is onlyslightlyslowerthanthemean15-locusratecalculated approximately 25,000. Two other surveyed X-linked loci with the data of Hammer et al. (2004) of 8.4310210 per (DMD44 and PDHA1) yield comparably large TMRCAs siteperyear.Utilizingsequencefromallthreespecies,the andeffectivepopulationsizes(Hammeretal.2004;Harris relativeratetestofTajima(1993)doesnotrejectaconstant and Hey 1999) and may also provide evidence of ancient rate of nucleotide substitution at the RRM2P4 locus (v2¼ population structure (Harding 1999). 0.39, P¼0.532). If we combine the neutral mutation rate The Asian RRM2P4 lineage is found exclusively in three of 31 sampled non-Africans (a Japanese, a Chinese, Table 1 and a Melanesian individual), whereas the other, more Estimates ofPopulation Genetic Parameters forthe diverse, lineage constitutes the entirety of African and the RRM2P4 Locus majority of non-African variation. To obtain a better estimate thefrequency ofthe Asian lineage andto test for Population n S h (%) p(%) D W itspresenceinAfrica,wegenotypedtheSNPoccurringat Global 41 13 0.127 0.119 20.206 position2020(fig.1A;alsoseeMethods)inatotalsample Africans 10 6 0.089 0.090 0.068 Non-Africans 31 11 0.115 0.124 0.224 of 570 African and non-African individuals representing 17populations.Figure2showstheobservedgeographical NOTE.—n is the number of individuals sampled; S is the number of distribution of the Asian lineage. There appears to be a polymorphicsites;hWisWatterson’s(1975)estimateof3Nmpernucleotidesite decreasingfrequencygradientcenteredonsouthernChina, (N is the effective population size and m is the neutral mutation rate); p is the averagenucleotideheterozygosity;D–Tajima’s(1989)neutralitystatistic. wheretheAsianlineagereachesamaximumof53%inthe Downloaded from https://academic.oup.com/mbe/article-abstract/22/2/189/963863 by guest on 31 January 2018 AncientHumanSubdivision 191 FIG. 2.—Geographical distribution of the archaic RRM2P4lineage in 17 populationsamples.Frequencydata canbe foundin Supplementary Materialonline. Yao of southern China and 29% in the Han Chinese. The lineageisfoundatlessthan1%inAfricanpopulations.The Asian lineage is present in only one of 177 sampled distributionoftheAsianlineagestronglysuggestsanAsian Africans (0.6%). This geographical pattern is consistent originbutshouldnotbetakenasdefinitiveproofthatitdid with a Southeast Asian origin of this 2-Myr-old lineage. notoriginate inAfrica. The Asian lineageappears tohave If we accept that the mutations occurring at the diverged from the globally distributed portion of the RRM2P4 locus are selectively neutral (and, therefore, genealogy approximately 2MYA. It is interesting to note influenced only by demographic history), then there are that this estimated divergence time is concordant with the two alternative explanations for its unusual pattern of age of the oldest Homo erectus fossils found outside of polymorphism. The first alternative is that the long basal Africa (Gabunia et al. 2000). Following further lines of branches are simply the chance result of genetic drift in a investigation, if panmixia at the RRM2P4 locus can be large,panmicticancestralpopulation.Itispossiblethatthe rejected,itwouldhaveimportantimplicationsforourview divergent, basal lineage was carried out of Africa and ofHomosapiensasaspecies.Anydegreeofdualancestryin subsequently lost in Africa and/or increased in frequency the modern human genome would either demonstrate that in Asia by genetic drift. Second, it may be the result of thetransitiontoananatomicallymodernformdidnotoccur recent admixture between two divergent populations; that inanisolated,panmicticpopulation(Cann,Stoneking,and is, the expanding anatomically modern human population Wilson 1987) or that replacement of preexisting hominid and Homo erectus. Distinguishing between these two populations was incomplete (e.g., Brauer 1989; Smith, alternatives provesdifficult, giventhelowpowerafforded Falsetti,andDonnelly1989). by any single locus (Nordborg 2000; Wall 2000). To achieve greater single-locus power, it is preferable to examinepatternsofpolymorphismoverphysicaldistances Acknowledgments approximately 20 to 30 kb in length (Wall 2000). WethankJ.WallforhisfruitfulcollaborationandR. Sprissler for assistance with SNP genotyping. We also thank Tanya Karafet, Dashtseveg Tumen, Giovanni Conclusion Destro-Bisol, Himla Soodyall, Andrea Novelleto, Peter Polymorphism occurring at the RRM2P4 locus is de Knijff, Beverly Strassmann, William Wang, Jonathan unique in that it clearly roots in East Asia, has an ancient Friedlaender, and Du Ruofu for providing DNA samples. TMRCA,and also yields higher non-African than African Publicationofthisworkwasmade possible bygrant GM- nucleotidediversity.OurSNPassayestimatesthattheAsian 53566 from the National Institute of General Medical Downloaded from https://academic.oup.com/mbe/article-abstract/22/2/189/963863 by guest on 31 January 2018 192 Garriganetal. Sciences (to M.F.H.). Its contents are solely the respon- Nordborg, M. 2000. On detecting ancient admixture. Genes, sibility of the authors and do not necessarily reflect the fossils, and behaviour: an integrated approach to human official views of the National Institutes of Health. evolution. IOSPress, Amsterdam. Reich,D.E.,M.Cargill,S.Bolketal.(11co-authors).2001.Linkage disequilibriuminthehumangenome.Nature411:199–204. Literature Cited Rozas,J.,J.C.Sanchez-DelBarrio,X.Messeguer,andR.Rozas. Brauer, G. 1989. The evolution of modern humans: a com- 2003.DnaSP,DNApolymorphismanalysesbythecoalescent parison of the African and non-African evidence. Pp. 123– andother methods. Bioinformatics 19:2496–2597. 154 in P. Mellars, and C. Stringer, eds. The human Smith,F.H.,A.B.Falsetti,andS.M.Donnelly.1989.Modern revolution: behavioural and biological perspectives on the humanorigins. Yrbk. Phys.Anthropol. 32:35–68. origins of modern humans. Edinburgh University Press, Swisher,C.C.,W.J.Rink,S.C.Anton,H.P.Schwarcz,G.H. Edinburgh. Curtis, A. Suprijo, and Widiasmoro. 1996. Latest Homo Cann, R. L., M. Stoneking, and A. C. Wilson. 1987. Mito- erectusofJava:potentialcontemporaneitywithHomosapiens chondrial DNA andhumanevolution.Nature 325:31–36. in southeast Asia. Science274:1870–1874. Gabunia,L.,A.Vekua,D.Lordkipanidzeet al.(14co-authors). Tajima, F. 1989. Statistical method for testing the neutral 2000. Earliest Pleistocene hominid cranial remains from mutation hypothesis by DNA polymorphism. Genetics 123: Dmanisi, Republic of Georgia: taxonomy, geological setting, 585–595. andage. Science 288:1019–1025. ———. 1993. Simple methods for testing the molecular evolu- Griffiths, R. C. 2002. Ancestral inference from gene trees. Pp. tionary clockhypothesis.Genetics 135:599–607. 94–117inM.Slatkin,ed.Moderndevelopmentsintheoretical Takahata, N., S. H. Lee, and Y. Satta. 2001. Testing multi- population genetics: the legacy of Gustave Malecot. Oxford regionality of modern human origins. Mol. Biol. Evol. 18: University Press,NewYork. 172–183. Hammer, M. F., D. Garrigan, E. Wood, J. A. Wilder, Z. Wall,J.D.1999.Recombinationandthepowerofstatisticaltests Mobasher, A. Bigham, J. G. Krenz, and M. W. Nachman. of neutrality. Genet.Res. 74:65–79. 2004. Heterogeneous patterns of variation among multiple ———. 2000. Detecting ancient admixture in humans using human X-linked loci: the possible role of diversity-reducing sequence polymorphismdata. Genetics 154:1271–1299. selection in non-Africans. Genetics 167:1841–1853. Watterson, G. A. 1975. On the number of segregating sites in Harding,R.M.1999.MoreontheXfiles.Proc.Natl.Acad.Sci. geneticalmodelswithoutrecombination.Theor.Pop.Biol.7: USA 96:2582–2584. 256–276. Harris,E.E.,andJ.Hey.1999.Xchromosomeevidenceforancient Wood, B., and M. Collard. 1999. The human genus. Science humanhistories.Proc.Natl.Acad.Sci.USA96:3320–3324. 284:65–71. Hedrick,P.W.,andJ.D.Parker.1997.Evolutionarygeneticsand Yu,N.,F.C.Chen,S.Ota,L.B.Jorde,P.Pamilo,L.Patthy,M. genetic variation of haplodiploids and X-linked genes. Ann. Ramsay,T.Jenkins,S.K.Shyue,andW.H.Li.2002.Larger Rev. Ecol. Syst. 28:55–83. geneticdifferenceswithinAfricansthanbetweenAfricansand Hudson,R.R.,andN.L. Kaplan.1985.Statistical propertiesof Eurasians. Genetics 161:269–274. thenumberofrecombinationeventsinthehistoryofasample of DNA sequences. Genetics 111:147–164. Scott Edwards, Associate Editor Nei, M. 1987. Molecular Evolutionary Genetics. Columbia University Press,NewYork. Accepted October 4, 2004 Downloaded from https://academic.oup.com/mbe/article-abstract/22/2/189/963863 by guest on 31 January 2018
Description: