- Email: [email protected]
Integration of mtDNA pseudogenes into the nuclear genome coincides with speciation of the human genus. A hypothesis
MITOCH-01150; No of Pages 4 Mitochondrion xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Mitochondrion journal homepage: www.elsevier.com/locate/mito
Profiles and perspectives
Integration of mtDNA pseudogenes into the nuclear genome coincides with speciation of the human genus. A hypothesis Konstantin Gunbin a, Leonid Peshkin c, Konstantin Popadin b, Sofia Annis e, Rebecca R. Ackermann d, Konstantin Khrapko e,⁎ a
Institute of Cytology and Genetics SB RAS and Novosibirsk State University, Russia University of Lausanne, Center for Integrative Genomics, Swirzerland Harvard Medical School, USA d Department of Archaeology & Human Evolution Research Institute, University of Cape Town, South Africa e Northeastern University, USA b c
a r t i c l e
i n f o
Article history: Received 23 May 2016 Received in revised form 14 October 2016 Accepted 9 December 2016 Available online xxxx Keywords: NUMT Human evolution Mitochondrial DNA Pseudogene Speciation Punctuated evolution
a b s t r a c t Fragments of mitochondrial DNA are known to get inserted into nuclear DNA to form NUMTs, i.e. nuclear pseudogenes of the mtDNA. The insertion of a NUMT is a rare event. Hundreds of pseudogenes have been cataloged in the human genome. NUMTs are, in essence, a special type of mutation with their own internal timer, which is synchronized with an established molecular clock, the mtDNA. Thus insertion of NUMTs can be timed with respect to evolution milestones such as the emergence of new species. We asked whether NUMTs were inserted uniformly over time or preferentially during certain periods of evolution, as implied by the “punctuated evolution” model. To our surprise, the NUMT insertion times do appear nonrandom with at least one cluster positioned at around 2.8 million years ago (Ma). Interestingly, 2.8 Ma closely corresponds to the time of emergence of the genus Homo, and to a well-documented period of major climate change ca. 2.9–2.5 Ma. It is tempting to hypothesize that the insertion of NUMTs is related to the speciation process. NUMTs could be either “riders”, i.e., their insertion could be facilitated by the overall higher genome rearrangement activity during speciation, or “drivers”, i.e. they may more readily get fixed in the population due to positive selection associated with speciation. If correct, the hypothesis would support the idea that evolution of our genus may have happened in a rapid, punctuated manner. © 2016 Elsevier B.V. and Mitochondria Research Society. All rights reserved.
1. Introduction Nuclear pseudogenes of mitochondrial DNA, now called “NUMTs”, were discovered over 30 years ago, first in Neurospora crassa (van den Boogaart et al., 1982) and later in various organisms including humans (Tsuzuki et al., 1983). Currently, over seven hundred NUMTs have been identified in the human genome (Lascaro et al., 2008), (Ramos et al., 2011), (Tsuji et al., 2012). A few different mechanisms of mtDNA insertion into the nuclear genome were proposed, including nonhomologous end joining repair (Hazkani-Covo and Covo, 2008) and retrotransposition (Tsuji et al., 2012). The mechanisms whereby NUMTs become fixed at the population level are not known. It might be expected that a large insertion of highly unusual DNA (mtDNA has a unique strand asymmetric base ⁎ Corresponding author. E-mail address: [email protected] (K. Khrapko).
composition) should disrupt protein binding, tertiary structure, and chromatin configuration of the locus and thus may affect expression patterns even if it does not directly affect coding sequence. Thus, NUMT insertions are expected to be highly non-neutral mutations subject to selective pressure. Indeed, in several instances, insertion of NUMTs has been associated with severe disease (Leister, 2005). Whether NUMTs can be subject to positive selection is not known. It has been recognized for decades that mtDNA pseudogenes represent chronological milestones of genome evolution. By comparing sequences of the pseudogenes to contemporary mitochondrial genomes of various primates (Hu and Thilly, 1994) or the diverse human mtDNA sequences (Zischler et al., 1995), one can estimate the age of divergence of the pseudogene sequence from the mitochondrial genome. In this study, we used a newly developed approach to determine, with considerable precision, the time of insertion of a set of human NUMTs that have formed after the separation of the chimpanzee and human lineages. Surprisingly, insertion times appear to cluster around
http://dx.doi.org/10.1016/j.mito.2016.12.001 1567-7249/© 2016 Elsevier B.V. and Mitochondria Research Society. All rights reserved.
Please cite this article as: Gunbin, K., et al., Integration of mtDNA pseudogenes into the nuclear genome coincides with speciation of the human genus. A hypothesis, Mitochondrion (2016), http://dx.doi.org/10.1016/j.mito.2016.12.001
2
K. Gunbin et al. / Mitochondrion xxx (2016) xxx–xxx
2.8 Ma, i.e. the point of emergence of the human genus. We discuss potential implications of this finding for the role NUMTs might have played in human evolution and for the concept of punctuated evolution. 2. Methods 2.1. Joint phylogenetic analysis of NUMTs and mtDNA 2.1.1. Data selection There are hundreds of established NUMTs in the human genome. Some of the NUMTs were inserted into the nuclear genome tens of millions years ago, before diversification of the new world monkeys, while others are quite recent (Zischler et al., 1995), (Tsuji et al., 2012), (Dayama et al., 2014). Because this study is focused on the period of evolution of the human lineage after its separation from the chimpanzee, we first coarsely selected NUMTs inserted within this time frame among currently cataloged NUMTs ((Gunbin et al., 2016), M&M S.1 and S.2a). Most NUMT sequences, however, are quite short and/or carry a low number of nucleotide changes and, as a result, it is difficult to determine the time of their insertion with sufficient precision. We therefore selected NUMTs that are expected to produce reliable estimates by requiring that mtDNA sequences homologous to each of the NUMTs were capable of supporting the correct topology of the human/chimpanzee/gorilla phylogenic tree ((Gunbin et al., 2016) M&M S.2b). This selection process resulted in 18 NUMTs. Finally, we confirmed that each of the selected NUMTs represented a unique nuclear DNA insertion event ((Gunbin et al., 2016), M&M S.2c). 2.1.2. Calculation of the time of insertion of a NUMT into the nuclear genome To determine the time of insertion of the NUMTs into the nuclear genome we first constructed joint phylogenetic trees of the 18 selected NUMTs with the great ape mtDNA and calculated the branch lengths ((Gunbin et al., 2016), M&M. 3). We assumed that branch lengths are proportional to the evolutionary time elapsed: mtDNA is a good molecular clock. The calibration of the clock was done by assuming that human and chimpanzee diverged 6 Mya. A sketch of a typical joint phylogeny together with the time scale is shown in Fig. 1. Note that the branching point of the NUMT branch in Fig. 1 does not represent the time of insertion of the NUMT into the nuclear genome as one might have assumed. Instead, the NUMT-to-be mtDNA sequence had been evolving and accumulating mutations as a separate mtDNA lineage for some time after branching off the main lineage (off the Homo mtDNA lineage in Fig. 1 example), prior to its insertion into the nuclear genome as has been previously discussed by others (Bensasson et al., 2001), (Baldo et al., 2011). In fact, as we show
below, this time from divergence to insertion may be very substantial and definitely needs to be taken into consideration. Thus, the NUMT branch consists of two segments: the mitochondrial (green) and the nuclear/pseudogenic (red). The time of NUMT insertion is then represented by the endpoint of the green mitochondrial segment of the NUMT branch (Fig. 1). With this in mind, we conclude that to determine the time of insertion we need to find specifically the length of the mitochondrial sub-segment of the NUMT branch. Fortunately, this can be done because mitochondrial and pseudogenic mutations can be easily distinguished. Indeed, the mutations of the nuclear segment are expected to include many more non-conservative, essentially random, nucleotide changes because pseudogenes are not known to be subject to selective pressure (Bensasson et al., 2001), (Baldo et al., 2011). Unlike the previous excellent study (Baldo et al., 2011), we were not able to use the protein coding-based synonymity of mtDNA mutations as a measure of conservatism, because a majority of NUMTs in our selected group originated from the rRNA coding part of mtDNA. We therefore used the GERP score (Davydov et al., 2010) to characterize how conservative a mutation is (see (Gunbin et al., 2016), M&M.S.4 for details of the approach). The GERP score reflects the variability of each genomic position between related species. Less variable positions are considered more conservative. To determine the position of the endpoint of the mitochondrial subsegment on a NUMT branch (e.g., the “NUMT” branch in Fig. 1) we consider that mutations belonging to the NUMT branch are a mix of mutations acquired by the sequence after the lineage has diverged from the contemporary human mtDNA lineage but still resided in the mitochondrion (green), and those acquired after the transfer to the nucleus (red). Therefore, average GERP score of the NUMT branch mutations is a linear combination of the GERP scores of its “mitochondrial” and “nuclear/ pseudogenic“ mutations. Thus to estimate the proportion of each type of mutation on a NUMT branch one needs to determine the proportion of the “mitochondrial” and the “pseudogenic” GERP scores which, when combined, will yield the observed average score of the NUMT branch. To do so, one needs to compare the average GERP score of mutations in the NUMT branch to the GERP scores of two types of mutations, the “mitochondrial” and the “pseudogenic”. As a proxy of “mitochondrial mutations” we have chosen mutations of the Homo branch (Fig. 1). As a proxy of the “pseudogenic” mutations we used a random set of mutations generated on the same DNA fragment. We reasoned that a random set of mutations will best approximate mtDNA pseudogenic mutations not subject to selection. Then a simple linear equation can be used to estimate the proportion of the mitochondrial and the pseudogenic segment within the NUMT branch and thus to position the endpoint of the mitochondrial sub-fragment of the NUMT branch. In our actual calculations this linear approach was modified by using categorical ratios instead of average scores because the real distributions of GERP scores deviates significantly from normal distribution ((Gunbin et al., 2016), M&M S.4). The time of NUMT insertion was then determined by counting time along the corresponding branches from the human/ chimpanzee lineages split 6 million years ago (Ma) to the endpoint of the mitochondrial sub-segment of the NUMT branch (Fig. 1 and (Gunbin et al., 2016), M&M S.5). 3. Results 3.1. Accelerated integration of NUMTs into the nuclear genome correlates with speciation of the human genus
Fig. 1. The time of insertion of a NUMT in to the nuclear genome. An example of a simplified typical joint phylogeny of a NUMT with the great ape mtDNAs (Gorilla, Pan, Homo), with an approximate time scale based on paleontological evidence. Green – mitochondrial portions of the branches, including the mitochondrial portion of the NUMT branch. Red – the pseudogene portion of the NUMT branch (see text for details). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
The estimated times of incorporation of the selected 18 NUMTs into the nuclear genome are shown in Fig. 2. Remarkably, there is a cluster of data points around 2.8 Ma, as if the rate of insertion of NUMTs was higher during this period. Interestingly, the 2.8 Ma timepoint closely corresponds to the time of emergence of the genus Homo and falls within a well documented period of major climate change 2.5–2.9 Ma
Please cite this article as: Gunbin, K., et al., Integration of mtDNA pseudogenes into the nuclear genome coincides with speciation of the human genus. A hypothesis, Mitochondrion (2016), http://dx.doi.org/10.1016/j.mito.2016.12.001
K. Gunbin et al. / Mitochondrion xxx (2016) xxx–xxx
3
factored into our analysis; the cluster of data points at 2.8 Ma can be detected with high confidence despite this variance. 4. Discussion: 4.1. NUMTS as mutations, “riders” or “drivers”?
Fig. 2. The rate of integration of NUMTs into the human nuclear genome appears increased during the climatic event 2.9–2.5 Ma and the time of origin of the human genus (Homo) 2.8 Ma. The 18 selected pseudogenes were ranged by their estimated time of insertion into the nuclear genome to better visualize the insertion rate variability. Diamonds represent the means and the error bars represent 25th and 75th percentiles of the distribution generated using the stringent 37% jackknife randomization procedure. For details, see text and (Gunbin et al., 2016), M&M S.3.
(DeMenocal, 2004). Although the hominin lineage first evolved ~6 Ma upon separating from the chimpanzee lineage, the major branching event that resulted in the emergence of our genus Homo did not occur until circa 2.8 Ma, as recorded by the earliest known Homo fossil specimen, a hemi-mandible (Villmoare et al., 2015). The low number of data points calls for a test of whether the distribution in Fig. 2 could have arisen by pure chance. To address this issue, we used numerical simulation where 106 sets of 18 time points, randomly distributed in the interval 0–6 Ma were generated. Then we calculated the fraction (i.e., probability of occurrence) of a 18-point set that is as clustered or more extreme than the one observed in reality (i.e., one in Fig. 2). The as clustered or more extreme set was defined as a set where the number of points that landed within the interval of the major climate change 2.5–2.9 Ma was equal or higher than 6 (the number of points in this interval in the observed distribution (Gunbin et al., 2016), M&M S.6). The corresponding probability is 0.0006, indicating that it is highly unlikely that the distribution in Fig. 2 has arisen by chance. The above impressively low probability is expected to strongly depend on the size of the interval used in the calculations. We therefore confirmed that this result was not a case of “p-value picking”. Indeed this probability stayed well below the critical 0.05 value within a wide range of interval widths ((Gunbin et al., 2016), M&M S.6, Fig. S3). As a note of caution, the individual estimates of the times of NUMT integration have relatively high “simulated variance” (note error bars in Fig. 2). This variance has been estimated using the 37% jackknife procedure ((Gunbin et al., 2016), M&M S5) and it stems primarily from the small size of some of the selected NUMT sequences, which makes the estimates of the branching points statistically challenging. We believe, however, that despite this high variance it is worth putting forward the hypothesis of a discontinuous rate of NUMT integration and of accelerated integration at c.a. 2.8 Ma. In a sense, this high variance has been
What could have caused an increase of the rate of NUMT insertion at this critical period of human evolution? NUMTs can be considered merely a particular type of nuclear genome mutation. Generally, the rate of mutation accumulation in a species depends on: 1) on the rate of primary mutagenesis and 2) on the rate of fixation of nascent mutations in the population. Accordingly, we propose two possible explanations for the increase of insertion rate. First, it is possible that during certain periods of evolution the genome becomes more susceptible to rearrangements of various types, including insertion of NUMTs. This is the “riders” hypothesis, which postulates that the higher rate of incorporation of mtDNA pseudogenes into the nuclear genome is responsible for the higher rate of NUMT accumulation. In principle, periods of active speciation may indeed result in higher propensity for rearrangements, especially if speciation is associated with crosses between genetically distant individuals. (Mallet, 2007), (Zinner et al., 2011), (Arnold and Meyer, 2006). The prevalence of distant intercrossing at the times of accelerated NUMT insertion is independently supported by the surprisingly long mitochondrial portions of many of the NUMT branches. Long segments imply that the carriers of these (now extinct) mitochondrial lineages have diverged quite far from the carriers of the main mitochondrial lineage before crossing over again (which was necessary to transfer the pseudogene to the carriers of the main mitochondrial lineage), as has been proposed previously for bristletails (Baldo et al., 2011). Second, alternatively (or in addition), it may be that NUMTs that we observe as being clustered in time are in fact important mutations with positive phenotypic consequences that were fixed in the population for that very reason (this is the “drivers” hypothesis). Indeed, there is vast body of evidence that NUMTs are frequently functionally important (Leister, 2005). This is not surprising given that NUMTs are large insertions of highly unusual (both in structure and nucleotide composition) DNA, which is likely to alter local mechanical and biochemical properties of the locus, such as causing changes in the nucleosome positioning pattern, protein binding properties, etc. For example, NUMTs are frequently associated with DNAseI hypersensitivity. The genomic positions of 8 NUMTs considered in this work (NUMTs numbers: 3–7, 9, 10, 11) were tightly linked or overlapped with highly significant self-organizing map clusters deposited in Regulatory Elements Database (Sheffield et al., 2013) that clustered DHSs by their activity across various cell types. 4.2. NUMTs: unique “mutations with internal clock”. Implications for evolutionary theory If NUMTs are phenotypically important mutations, then the increased rate of NUMT insertion during diversification of lineages implies that phenotypically advantageous mutations are in higher demand at critical points in evolution (Flegr, 2010), (Lee et al., 2013). Whether a larger proportion of critical mutations is indeed acquired during relatively short periods of speciation is a matter of debate. On the one hand, there is evidence that the rate of cladogenesis is decoupled from the rate of molecular evolution (Goldie et al., 2011) (Lanfear et al., 2010). On the other hand, there is evidence that critical mutations are acquired during short speciation periods (Pagel et al., 2006), (Flegr, 2010). Moreover, a positive correlation was found between mitochondrial genome evolutionary rates and speciation rates in birds and reptiles (Eo and DeWoody, 2010). More generally, this is related to the concept of punctuated equilibria (Eldredge and Gould, 1973) and/or the “turnover pulse” (Vrba, 1993) (Vrba, 1993) model of how evolution proceeds. The latter concept has been particularly intensively
Please cite this article as: Gunbin, K., et al., Integration of mtDNA pseudogenes into the nuclear genome coincides with speciation of the human genus. A hypothesis, Mitochondrion (2016), http://dx.doi.org/10.1016/j.mito.2016.12.001
4
K. Gunbin et al. / Mitochondrion xxx (2016) xxx–xxx
debated with respect to the emergence of Homo and the possible role major climate change ca. 2.5–2.9 Ma (DeMenocal, 2004) played in driving this speciation event. If our hypothesis of discontinuous rate of NUMT insertion in the human genome around 2.8 Ma is correct, it lends support to the idea that the evolution of our genus may have happened in a rapid, punctuated manner, possibly in response to climatic events altering selective pressures and/or forcing migration and gene exchange. NUMTs offer a unique opportunity to explore this idea, as the time of NUMT creation can be determined with a precision that is unusual for conventional mutations, e.g. point mutations. Normally, mutations can be placed (based on an alignment) on specific phylogenic tree branches between successive bifurcations but cannot be precisely positioned within such a branch. Thus, the exact time when the mutation was created cannot be determined. NUMTs are not subject to this limitation as they possess an internal “timer”, which is synchronized with the mtDNA molecular clock. It would be interesting to further explore whether the rate of NUMT insertion is also coincident with critical periods of evolution in other species. Authors' contributions K.G. and K.K. conceived the project and all algorithms for data analysis. K.G. performed data collection, extraction, checking, most of the data analysis (initial and final), implemented the software intended to pseudogenization dating and randomization statistical tests. L.P. did critical tests allowing to identify clusters of pseudogenization on initial analysis steps. K.P., L.P., and S.A. contributed to the statistical data analysis on final stage. R.R.A. introduced critical paleontological checkpoints for NUMT pseudogenization times clustering. K.K. coordinated the project and wrote most of main manuscript text. All authors contributed to the final manuscript preparation, discussed the results and their implications, and have read and approved the final manuscript. Acknowledgements This research was supported in part by The Ellison Medical Foundation Senior Scholar Award to KK; 13-06-12063-OFI-m grant from the RFBR and grant 14.B25.31.0033 (Resolution No. 220) from the Russian Federation Government to KG. The Siberian Branch of the Russian Academy of Sciences (SB RAS) Siberian Supercomputing Center (CCU “Bioinformatics”) and the Novosibirsk State University High-Performance Computing Center are gratefully acknowledged for providing computer facilities. References Arnold, M.L., Meyer, A., 2006. Natural hybridization in primates: one evolutionary mechanism. Zoology (Jena) 109, 261–276.
Baldo, L., de Queiroz, A., Hedin, M., Hayashi, C.Y., Gatesy, J., 2011. Nuclear-mitochondrial sequences as witnesses of past interbreeding and population diversity in the jumping bristletail Mesomachilis. Mol. Biol. Evol. 28, 195–210. Bensasson, D., Zhang, D., Hartl, D.L., Hewitt, G.M., 2001. Mitochondrial pseudogenes: evolution's misplaced witnesses. Trends Ecol. Evol. 16, 314–321. Davydov, E.V., Goode, D.L., Sirota, M., Cooper, G.M., Sidow, A., Batzoglou, S., 2010. Identifying a high fraction of the human genome to be under selective constraint using GERP++. PLoS Comput. Biol. 6, e1001025. Dayama, G., Emery, S.B., Kidd, J.M., Mills, R.E., 2014. The genomic landscape of polymorphic human nuclear mitochondrial insertions. Nucleic Acids Res. 42, 12640–12649. Demenocal, P., 2004. African climate change and faunal evolution during the PliocenePleistocene. Earth Planet. Sci. Lett. 220, 3–24. Eldredge, N., Gould, S.J., 1973. Models in Paleobiology. Freeman Cooper, San Francisco. Eo, S.H., DeWoody, J.A., 2010. Evolutionary rates of mitochondrial genomes correspond to diversification rates and to contemporary species richness in birds and reptiles. Proc. Biol. Sci. 277, 3587–3592. Flegr, J., 2010. Elastic, not plastic species: frozen plasticity theory and the origin of adaptive evolution in sexually reproducing organisms. Biol. Direct 5, 2. Goldie, X., Lanfear, R., Bromham, L., 2011. Diversification and the rate of molecular evolution: no evidence of a link in mammals. BMC Evol. Biol. 11, 286. Gunbin, K.V., Peshkin, L., Popadin, K., Annis, S., Ackermann, R.R., Khrapko, K., 2016. The time of integration of the human mitochondrial pseudogenes (NUMTs) into the nuclear genome. Data and Methods (Submitted). Data in Brief. Hazkani-Covo, E., Covo, S., 2008. Numt-mediated double-strand break repair mitigates deletions during primate genome evolution. PLoS Genet. 4, e1000237. Hu, G., Thilly, W.G., 1994. Evolutionary trail of the mitochondrial genome as based on human 16S rDNA pseudogenes. Gene 147, 197–204. Lanfear, R., Ho, S.Y., Love, D., Bromham, L., 2010. Mutation rate is linked to diversification in birds. Proc. Natl. Acad. Sci. U. S. A. 107, 20423–20428. Lascaro, D., Castellana, S., Gasparre, G., Romeo, G., Saccone, C., Attimonelli, M., 2008. The RHNumtS compilation: features and bioinformatics approaches to locate and quantify Human NumtS. BMC Genomics 9, 267. Lee, M.S.Y., Soubrier, J., Edgecombe, G.D., 2013. Rates of phenotypic and genomic evolution during the Cambrian explosion. Curr. Biol. 23, 1889–1895. Leister, D., 2005. Origin, evolution and genetic effects of nuclear insertions of organelle DNA. Trends Genet. 21, 655–663. Mallet, J., 2007. Hybrid speciation. Nature 446, 279–283. Pagel, M., Venditti, C., Meade, A., 2006. Large punctuational contribution of speciation to evolutionary divergence at the molecular level. Science 314, 119–121. Ramos, A., Barbena, E., Mateiu, L., del Mar González, M., Mairal, Q., Lima, M., Montiel, R., Aluja, M.P., Santos, C., 2011. Nuclear insertions of mitochondrial origin: database updating and usefulness in cancer studies. Mitochondrion 11, 946–953. Sheffield, N.C., Thurman, R.E., Song, L., Safi, A., Stamatoyannopoulos, J.A., Lenhard, B., Crawford, G.E., Furey, T.S., 2013. Patterns of regulatory activity across diverse human cell types predict tissue identity, transcription factor binding, and longrange interactions. Genome Res. 23, 777–788. Tsuji, J., Frith, M.C., Tomii, K., Horton, P., 2012. Mammalian NUMT insertion is non-random. Nucleic Acids Res. 40, 9073–9088. Tsuzuki, T., Nomiyama, H., Setoyama, C., Maeda, S., Shimada, K., 1983. Presence of mitochondrial-DNA-like sequences in the human nuclear DNA. Gene 25, 223–229. van den Boogaart, P., Samallo, J., Agsteribbe, E., 1982. Similar genes for a mitochondrial ATPase subunit in the nuclear and mitochondrial genomes of Neurospora crassa. Nature 298, 187–189. Villmoare, B., Kimbel, W.H., Seyoum, C., Campisano, C.J., DiMaggio, E.N., Rowan, J., Braun, D.R., Arrowsmith, J.R., Reed, K.E., 2015. Paleoanthropology. Early homo at 2.8 Ma from Ledi-Geraru, Afar, Ethiopia. Science 347, 1352–1355. Vrba, E.S., 1993. Turnover-pulses, the Red Queen, and related topics. Am. J. Sci. 293, 418–452. Zinner, D., Arnold, M.L., Roos, C., 2011. The strange blood: natural hybridization in primates. Evol. Anthropol. 20, 96–103. Zischler, H., Geisert, H., Haeseler von, A., Paabo, S., 1995. A nuclear “fossil” of the mitochondrial D-loop and the origin of modern humans. Nature 378, 489–492.
Please cite this article as: Gunbin, K., et al., Integration of mtDNA pseudogenes into the nuclear genome coincides with speciation of the human genus. A hypothesis, Mitochondrion (2016), http://dx.doi.org/10.1016/j.mito.2016.12.001
Contents lists available at ScienceDirect
Mitochondrion journal homepage: www.elsevier.com/locate/mito
Profiles and perspectives
Integration of mtDNA pseudogenes into the nuclear genome coincides with speciation of the human genus. A hypothesis Konstantin Gunbin a, Leonid Peshkin c, Konstantin Popadin b, Sofia Annis e, Rebecca R. Ackermann d, Konstantin Khrapko e,⁎ a
Institute of Cytology and Genetics SB RAS and Novosibirsk State University, Russia University of Lausanne, Center for Integrative Genomics, Swirzerland Harvard Medical School, USA d Department of Archaeology & Human Evolution Research Institute, University of Cape Town, South Africa e Northeastern University, USA b c
a r t i c l e
i n f o
Article history: Received 23 May 2016 Received in revised form 14 October 2016 Accepted 9 December 2016 Available online xxxx Keywords: NUMT Human evolution Mitochondrial DNA Pseudogene Speciation Punctuated evolution
a b s t r a c t Fragments of mitochondrial DNA are known to get inserted into nuclear DNA to form NUMTs, i.e. nuclear pseudogenes of the mtDNA. The insertion of a NUMT is a rare event. Hundreds of pseudogenes have been cataloged in the human genome. NUMTs are, in essence, a special type of mutation with their own internal timer, which is synchronized with an established molecular clock, the mtDNA. Thus insertion of NUMTs can be timed with respect to evolution milestones such as the emergence of new species. We asked whether NUMTs were inserted uniformly over time or preferentially during certain periods of evolution, as implied by the “punctuated evolution” model. To our surprise, the NUMT insertion times do appear nonrandom with at least one cluster positioned at around 2.8 million years ago (Ma). Interestingly, 2.8 Ma closely corresponds to the time of emergence of the genus Homo, and to a well-documented period of major climate change ca. 2.9–2.5 Ma. It is tempting to hypothesize that the insertion of NUMTs is related to the speciation process. NUMTs could be either “riders”, i.e., their insertion could be facilitated by the overall higher genome rearrangement activity during speciation, or “drivers”, i.e. they may more readily get fixed in the population due to positive selection associated with speciation. If correct, the hypothesis would support the idea that evolution of our genus may have happened in a rapid, punctuated manner. © 2016 Elsevier B.V. and Mitochondria Research Society. All rights reserved.
1. Introduction Nuclear pseudogenes of mitochondrial DNA, now called “NUMTs”, were discovered over 30 years ago, first in Neurospora crassa (van den Boogaart et al., 1982) and later in various organisms including humans (Tsuzuki et al., 1983). Currently, over seven hundred NUMTs have been identified in the human genome (Lascaro et al., 2008), (Ramos et al., 2011), (Tsuji et al., 2012). A few different mechanisms of mtDNA insertion into the nuclear genome were proposed, including nonhomologous end joining repair (Hazkani-Covo and Covo, 2008) and retrotransposition (Tsuji et al., 2012). The mechanisms whereby NUMTs become fixed at the population level are not known. It might be expected that a large insertion of highly unusual DNA (mtDNA has a unique strand asymmetric base ⁎ Corresponding author. E-mail address: [email protected] (K. Khrapko).
composition) should disrupt protein binding, tertiary structure, and chromatin configuration of the locus and thus may affect expression patterns even if it does not directly affect coding sequence. Thus, NUMT insertions are expected to be highly non-neutral mutations subject to selective pressure. Indeed, in several instances, insertion of NUMTs has been associated with severe disease (Leister, 2005). Whether NUMTs can be subject to positive selection is not known. It has been recognized for decades that mtDNA pseudogenes represent chronological milestones of genome evolution. By comparing sequences of the pseudogenes to contemporary mitochondrial genomes of various primates (Hu and Thilly, 1994) or the diverse human mtDNA sequences (Zischler et al., 1995), one can estimate the age of divergence of the pseudogene sequence from the mitochondrial genome. In this study, we used a newly developed approach to determine, with considerable precision, the time of insertion of a set of human NUMTs that have formed after the separation of the chimpanzee and human lineages. Surprisingly, insertion times appear to cluster around
http://dx.doi.org/10.1016/j.mito.2016.12.001 1567-7249/© 2016 Elsevier B.V. and Mitochondria Research Society. All rights reserved.
Please cite this article as: Gunbin, K., et al., Integration of mtDNA pseudogenes into the nuclear genome coincides with speciation of the human genus. A hypothesis, Mitochondrion (2016), http://dx.doi.org/10.1016/j.mito.2016.12.001
2
K. Gunbin et al. / Mitochondrion xxx (2016) xxx–xxx
2.8 Ma, i.e. the point of emergence of the human genus. We discuss potential implications of this finding for the role NUMTs might have played in human evolution and for the concept of punctuated evolution. 2. Methods 2.1. Joint phylogenetic analysis of NUMTs and mtDNA 2.1.1. Data selection There are hundreds of established NUMTs in the human genome. Some of the NUMTs were inserted into the nuclear genome tens of millions years ago, before diversification of the new world monkeys, while others are quite recent (Zischler et al., 1995), (Tsuji et al., 2012), (Dayama et al., 2014). Because this study is focused on the period of evolution of the human lineage after its separation from the chimpanzee, we first coarsely selected NUMTs inserted within this time frame among currently cataloged NUMTs ((Gunbin et al., 2016), M&M S.1 and S.2a). Most NUMT sequences, however, are quite short and/or carry a low number of nucleotide changes and, as a result, it is difficult to determine the time of their insertion with sufficient precision. We therefore selected NUMTs that are expected to produce reliable estimates by requiring that mtDNA sequences homologous to each of the NUMTs were capable of supporting the correct topology of the human/chimpanzee/gorilla phylogenic tree ((Gunbin et al., 2016) M&M S.2b). This selection process resulted in 18 NUMTs. Finally, we confirmed that each of the selected NUMTs represented a unique nuclear DNA insertion event ((Gunbin et al., 2016), M&M S.2c). 2.1.2. Calculation of the time of insertion of a NUMT into the nuclear genome To determine the time of insertion of the NUMTs into the nuclear genome we first constructed joint phylogenetic trees of the 18 selected NUMTs with the great ape mtDNA and calculated the branch lengths ((Gunbin et al., 2016), M&M. 3). We assumed that branch lengths are proportional to the evolutionary time elapsed: mtDNA is a good molecular clock. The calibration of the clock was done by assuming that human and chimpanzee diverged 6 Mya. A sketch of a typical joint phylogeny together with the time scale is shown in Fig. 1. Note that the branching point of the NUMT branch in Fig. 1 does not represent the time of insertion of the NUMT into the nuclear genome as one might have assumed. Instead, the NUMT-to-be mtDNA sequence had been evolving and accumulating mutations as a separate mtDNA lineage for some time after branching off the main lineage (off the Homo mtDNA lineage in Fig. 1 example), prior to its insertion into the nuclear genome as has been previously discussed by others (Bensasson et al., 2001), (Baldo et al., 2011). In fact, as we show
below, this time from divergence to insertion may be very substantial and definitely needs to be taken into consideration. Thus, the NUMT branch consists of two segments: the mitochondrial (green) and the nuclear/pseudogenic (red). The time of NUMT insertion is then represented by the endpoint of the green mitochondrial segment of the NUMT branch (Fig. 1). With this in mind, we conclude that to determine the time of insertion we need to find specifically the length of the mitochondrial sub-segment of the NUMT branch. Fortunately, this can be done because mitochondrial and pseudogenic mutations can be easily distinguished. Indeed, the mutations of the nuclear segment are expected to include many more non-conservative, essentially random, nucleotide changes because pseudogenes are not known to be subject to selective pressure (Bensasson et al., 2001), (Baldo et al., 2011). Unlike the previous excellent study (Baldo et al., 2011), we were not able to use the protein coding-based synonymity of mtDNA mutations as a measure of conservatism, because a majority of NUMTs in our selected group originated from the rRNA coding part of mtDNA. We therefore used the GERP score (Davydov et al., 2010) to characterize how conservative a mutation is (see (Gunbin et al., 2016), M&M.S.4 for details of the approach). The GERP score reflects the variability of each genomic position between related species. Less variable positions are considered more conservative. To determine the position of the endpoint of the mitochondrial subsegment on a NUMT branch (e.g., the “NUMT” branch in Fig. 1) we consider that mutations belonging to the NUMT branch are a mix of mutations acquired by the sequence after the lineage has diverged from the contemporary human mtDNA lineage but still resided in the mitochondrion (green), and those acquired after the transfer to the nucleus (red). Therefore, average GERP score of the NUMT branch mutations is a linear combination of the GERP scores of its “mitochondrial” and “nuclear/ pseudogenic“ mutations. Thus to estimate the proportion of each type of mutation on a NUMT branch one needs to determine the proportion of the “mitochondrial” and the “pseudogenic” GERP scores which, when combined, will yield the observed average score of the NUMT branch. To do so, one needs to compare the average GERP score of mutations in the NUMT branch to the GERP scores of two types of mutations, the “mitochondrial” and the “pseudogenic”. As a proxy of “mitochondrial mutations” we have chosen mutations of the Homo branch (Fig. 1). As a proxy of the “pseudogenic” mutations we used a random set of mutations generated on the same DNA fragment. We reasoned that a random set of mutations will best approximate mtDNA pseudogenic mutations not subject to selection. Then a simple linear equation can be used to estimate the proportion of the mitochondrial and the pseudogenic segment within the NUMT branch and thus to position the endpoint of the mitochondrial sub-fragment of the NUMT branch. In our actual calculations this linear approach was modified by using categorical ratios instead of average scores because the real distributions of GERP scores deviates significantly from normal distribution ((Gunbin et al., 2016), M&M S.4). The time of NUMT insertion was then determined by counting time along the corresponding branches from the human/ chimpanzee lineages split 6 million years ago (Ma) to the endpoint of the mitochondrial sub-segment of the NUMT branch (Fig. 1 and (Gunbin et al., 2016), M&M S.5). 3. Results 3.1. Accelerated integration of NUMTs into the nuclear genome correlates with speciation of the human genus
Fig. 1. The time of insertion of a NUMT in to the nuclear genome. An example of a simplified typical joint phylogeny of a NUMT with the great ape mtDNAs (Gorilla, Pan, Homo), with an approximate time scale based on paleontological evidence. Green – mitochondrial portions of the branches, including the mitochondrial portion of the NUMT branch. Red – the pseudogene portion of the NUMT branch (see text for details). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
The estimated times of incorporation of the selected 18 NUMTs into the nuclear genome are shown in Fig. 2. Remarkably, there is a cluster of data points around 2.8 Ma, as if the rate of insertion of NUMTs was higher during this period. Interestingly, the 2.8 Ma timepoint closely corresponds to the time of emergence of the genus Homo and falls within a well documented period of major climate change 2.5–2.9 Ma
Please cite this article as: Gunbin, K., et al., Integration of mtDNA pseudogenes into the nuclear genome coincides with speciation of the human genus. A hypothesis, Mitochondrion (2016), http://dx.doi.org/10.1016/j.mito.2016.12.001
K. Gunbin et al. / Mitochondrion xxx (2016) xxx–xxx
3
factored into our analysis; the cluster of data points at 2.8 Ma can be detected with high confidence despite this variance. 4. Discussion: 4.1. NUMTS as mutations, “riders” or “drivers”?
Fig. 2. The rate of integration of NUMTs into the human nuclear genome appears increased during the climatic event 2.9–2.5 Ma and the time of origin of the human genus (Homo) 2.8 Ma. The 18 selected pseudogenes were ranged by their estimated time of insertion into the nuclear genome to better visualize the insertion rate variability. Diamonds represent the means and the error bars represent 25th and 75th percentiles of the distribution generated using the stringent 37% jackknife randomization procedure. For details, see text and (Gunbin et al., 2016), M&M S.3.
(DeMenocal, 2004). Although the hominin lineage first evolved ~6 Ma upon separating from the chimpanzee lineage, the major branching event that resulted in the emergence of our genus Homo did not occur until circa 2.8 Ma, as recorded by the earliest known Homo fossil specimen, a hemi-mandible (Villmoare et al., 2015). The low number of data points calls for a test of whether the distribution in Fig. 2 could have arisen by pure chance. To address this issue, we used numerical simulation where 106 sets of 18 time points, randomly distributed in the interval 0–6 Ma were generated. Then we calculated the fraction (i.e., probability of occurrence) of a 18-point set that is as clustered or more extreme than the one observed in reality (i.e., one in Fig. 2). The as clustered or more extreme set was defined as a set where the number of points that landed within the interval of the major climate change 2.5–2.9 Ma was equal or higher than 6 (the number of points in this interval in the observed distribution (Gunbin et al., 2016), M&M S.6). The corresponding probability is 0.0006, indicating that it is highly unlikely that the distribution in Fig. 2 has arisen by chance. The above impressively low probability is expected to strongly depend on the size of the interval used in the calculations. We therefore confirmed that this result was not a case of “p-value picking”. Indeed this probability stayed well below the critical 0.05 value within a wide range of interval widths ((Gunbin et al., 2016), M&M S.6, Fig. S3). As a note of caution, the individual estimates of the times of NUMT integration have relatively high “simulated variance” (note error bars in Fig. 2). This variance has been estimated using the 37% jackknife procedure ((Gunbin et al., 2016), M&M S5) and it stems primarily from the small size of some of the selected NUMT sequences, which makes the estimates of the branching points statistically challenging. We believe, however, that despite this high variance it is worth putting forward the hypothesis of a discontinuous rate of NUMT integration and of accelerated integration at c.a. 2.8 Ma. In a sense, this high variance has been
What could have caused an increase of the rate of NUMT insertion at this critical period of human evolution? NUMTs can be considered merely a particular type of nuclear genome mutation. Generally, the rate of mutation accumulation in a species depends on: 1) on the rate of primary mutagenesis and 2) on the rate of fixation of nascent mutations in the population. Accordingly, we propose two possible explanations for the increase of insertion rate. First, it is possible that during certain periods of evolution the genome becomes more susceptible to rearrangements of various types, including insertion of NUMTs. This is the “riders” hypothesis, which postulates that the higher rate of incorporation of mtDNA pseudogenes into the nuclear genome is responsible for the higher rate of NUMT accumulation. In principle, periods of active speciation may indeed result in higher propensity for rearrangements, especially if speciation is associated with crosses between genetically distant individuals. (Mallet, 2007), (Zinner et al., 2011), (Arnold and Meyer, 2006). The prevalence of distant intercrossing at the times of accelerated NUMT insertion is independently supported by the surprisingly long mitochondrial portions of many of the NUMT branches. Long segments imply that the carriers of these (now extinct) mitochondrial lineages have diverged quite far from the carriers of the main mitochondrial lineage before crossing over again (which was necessary to transfer the pseudogene to the carriers of the main mitochondrial lineage), as has been proposed previously for bristletails (Baldo et al., 2011). Second, alternatively (or in addition), it may be that NUMTs that we observe as being clustered in time are in fact important mutations with positive phenotypic consequences that were fixed in the population for that very reason (this is the “drivers” hypothesis). Indeed, there is vast body of evidence that NUMTs are frequently functionally important (Leister, 2005). This is not surprising given that NUMTs are large insertions of highly unusual (both in structure and nucleotide composition) DNA, which is likely to alter local mechanical and biochemical properties of the locus, such as causing changes in the nucleosome positioning pattern, protein binding properties, etc. For example, NUMTs are frequently associated with DNAseI hypersensitivity. The genomic positions of 8 NUMTs considered in this work (NUMTs numbers: 3–7, 9, 10, 11) were tightly linked or overlapped with highly significant self-organizing map clusters deposited in Regulatory Elements Database (Sheffield et al., 2013) that clustered DHSs by their activity across various cell types. 4.2. NUMTs: unique “mutations with internal clock”. Implications for evolutionary theory If NUMTs are phenotypically important mutations, then the increased rate of NUMT insertion during diversification of lineages implies that phenotypically advantageous mutations are in higher demand at critical points in evolution (Flegr, 2010), (Lee et al., 2013). Whether a larger proportion of critical mutations is indeed acquired during relatively short periods of speciation is a matter of debate. On the one hand, there is evidence that the rate of cladogenesis is decoupled from the rate of molecular evolution (Goldie et al., 2011) (Lanfear et al., 2010). On the other hand, there is evidence that critical mutations are acquired during short speciation periods (Pagel et al., 2006), (Flegr, 2010). Moreover, a positive correlation was found between mitochondrial genome evolutionary rates and speciation rates in birds and reptiles (Eo and DeWoody, 2010). More generally, this is related to the concept of punctuated equilibria (Eldredge and Gould, 1973) and/or the “turnover pulse” (Vrba, 1993) (Vrba, 1993) model of how evolution proceeds. The latter concept has been particularly intensively
Please cite this article as: Gunbin, K., et al., Integration of mtDNA pseudogenes into the nuclear genome coincides with speciation of the human genus. A hypothesis, Mitochondrion (2016), http://dx.doi.org/10.1016/j.mito.2016.12.001
4
K. Gunbin et al. / Mitochondrion xxx (2016) xxx–xxx
debated with respect to the emergence of Homo and the possible role major climate change ca. 2.5–2.9 Ma (DeMenocal, 2004) played in driving this speciation event. If our hypothesis of discontinuous rate of NUMT insertion in the human genome around 2.8 Ma is correct, it lends support to the idea that the evolution of our genus may have happened in a rapid, punctuated manner, possibly in response to climatic events altering selective pressures and/or forcing migration and gene exchange. NUMTs offer a unique opportunity to explore this idea, as the time of NUMT creation can be determined with a precision that is unusual for conventional mutations, e.g. point mutations. Normally, mutations can be placed (based on an alignment) on specific phylogenic tree branches between successive bifurcations but cannot be precisely positioned within such a branch. Thus, the exact time when the mutation was created cannot be determined. NUMTs are not subject to this limitation as they possess an internal “timer”, which is synchronized with the mtDNA molecular clock. It would be interesting to further explore whether the rate of NUMT insertion is also coincident with critical periods of evolution in other species. Authors' contributions K.G. and K.K. conceived the project and all algorithms for data analysis. K.G. performed data collection, extraction, checking, most of the data analysis (initial and final), implemented the software intended to pseudogenization dating and randomization statistical tests. L.P. did critical tests allowing to identify clusters of pseudogenization on initial analysis steps. K.P., L.P., and S.A. contributed to the statistical data analysis on final stage. R.R.A. introduced critical paleontological checkpoints for NUMT pseudogenization times clustering. K.K. coordinated the project and wrote most of main manuscript text. All authors contributed to the final manuscript preparation, discussed the results and their implications, and have read and approved the final manuscript. Acknowledgements This research was supported in part by The Ellison Medical Foundation Senior Scholar Award to KK; 13-06-12063-OFI-m grant from the RFBR and grant 14.B25.31.0033 (Resolution No. 220) from the Russian Federation Government to KG. The Siberian Branch of the Russian Academy of Sciences (SB RAS) Siberian Supercomputing Center (CCU “Bioinformatics”) and the Novosibirsk State University High-Performance Computing Center are gratefully acknowledged for providing computer facilities. References Arnold, M.L., Meyer, A., 2006. Natural hybridization in primates: one evolutionary mechanism. Zoology (Jena) 109, 261–276.
Baldo, L., de Queiroz, A., Hedin, M., Hayashi, C.Y., Gatesy, J., 2011. Nuclear-mitochondrial sequences as witnesses of past interbreeding and population diversity in the jumping bristletail Mesomachilis. Mol. Biol. Evol. 28, 195–210. Bensasson, D., Zhang, D., Hartl, D.L., Hewitt, G.M., 2001. Mitochondrial pseudogenes: evolution's misplaced witnesses. Trends Ecol. Evol. 16, 314–321. Davydov, E.V., Goode, D.L., Sirota, M., Cooper, G.M., Sidow, A., Batzoglou, S., 2010. Identifying a high fraction of the human genome to be under selective constraint using GERP++. PLoS Comput. Biol. 6, e1001025. Dayama, G., Emery, S.B., Kidd, J.M., Mills, R.E., 2014. The genomic landscape of polymorphic human nuclear mitochondrial insertions. Nucleic Acids Res. 42, 12640–12649. Demenocal, P., 2004. African climate change and faunal evolution during the PliocenePleistocene. Earth Planet. Sci. Lett. 220, 3–24. Eldredge, N., Gould, S.J., 1973. Models in Paleobiology. Freeman Cooper, San Francisco. Eo, S.H., DeWoody, J.A., 2010. Evolutionary rates of mitochondrial genomes correspond to diversification rates and to contemporary species richness in birds and reptiles. Proc. Biol. Sci. 277, 3587–3592. Flegr, J., 2010. Elastic, not plastic species: frozen plasticity theory and the origin of adaptive evolution in sexually reproducing organisms. Biol. Direct 5, 2. Goldie, X., Lanfear, R., Bromham, L., 2011. Diversification and the rate of molecular evolution: no evidence of a link in mammals. BMC Evol. Biol. 11, 286. Gunbin, K.V., Peshkin, L., Popadin, K., Annis, S., Ackermann, R.R., Khrapko, K., 2016. The time of integration of the human mitochondrial pseudogenes (NUMTs) into the nuclear genome. Data and Methods (Submitted). Data in Brief. Hazkani-Covo, E., Covo, S., 2008. Numt-mediated double-strand break repair mitigates deletions during primate genome evolution. PLoS Genet. 4, e1000237. Hu, G., Thilly, W.G., 1994. Evolutionary trail of the mitochondrial genome as based on human 16S rDNA pseudogenes. Gene 147, 197–204. Lanfear, R., Ho, S.Y., Love, D., Bromham, L., 2010. Mutation rate is linked to diversification in birds. Proc. Natl. Acad. Sci. U. S. A. 107, 20423–20428. Lascaro, D., Castellana, S., Gasparre, G., Romeo, G., Saccone, C., Attimonelli, M., 2008. The RHNumtS compilation: features and bioinformatics approaches to locate and quantify Human NumtS. BMC Genomics 9, 267. Lee, M.S.Y., Soubrier, J., Edgecombe, G.D., 2013. Rates of phenotypic and genomic evolution during the Cambrian explosion. Curr. Biol. 23, 1889–1895. Leister, D., 2005. Origin, evolution and genetic effects of nuclear insertions of organelle DNA. Trends Genet. 21, 655–663. Mallet, J., 2007. Hybrid speciation. Nature 446, 279–283. Pagel, M., Venditti, C., Meade, A., 2006. Large punctuational contribution of speciation to evolutionary divergence at the molecular level. Science 314, 119–121. Ramos, A., Barbena, E., Mateiu, L., del Mar González, M., Mairal, Q., Lima, M., Montiel, R., Aluja, M.P., Santos, C., 2011. Nuclear insertions of mitochondrial origin: database updating and usefulness in cancer studies. Mitochondrion 11, 946–953. Sheffield, N.C., Thurman, R.E., Song, L., Safi, A., Stamatoyannopoulos, J.A., Lenhard, B., Crawford, G.E., Furey, T.S., 2013. Patterns of regulatory activity across diverse human cell types predict tissue identity, transcription factor binding, and longrange interactions. Genome Res. 23, 777–788. Tsuji, J., Frith, M.C., Tomii, K., Horton, P., 2012. Mammalian NUMT insertion is non-random. Nucleic Acids Res. 40, 9073–9088. Tsuzuki, T., Nomiyama, H., Setoyama, C., Maeda, S., Shimada, K., 1983. Presence of mitochondrial-DNA-like sequences in the human nuclear DNA. Gene 25, 223–229. van den Boogaart, P., Samallo, J., Agsteribbe, E., 1982. Similar genes for a mitochondrial ATPase subunit in the nuclear and mitochondrial genomes of Neurospora crassa. Nature 298, 187–189. Villmoare, B., Kimbel, W.H., Seyoum, C., Campisano, C.J., DiMaggio, E.N., Rowan, J., Braun, D.R., Arrowsmith, J.R., Reed, K.E., 2015. Paleoanthropology. Early homo at 2.8 Ma from Ledi-Geraru, Afar, Ethiopia. Science 347, 1352–1355. Vrba, E.S., 1993. Turnover-pulses, the Red Queen, and related topics. Am. J. Sci. 293, 418–452. Zinner, D., Arnold, M.L., Roos, C., 2011. The strange blood: natural hybridization in primates. Evol. Anthropol. 20, 96–103. Zischler, H., Geisert, H., Haeseler von, A., Paabo, S., 1995. A nuclear “fossil” of the mitochondrial D-loop and the origin of modern humans. Nature 378, 489–492.
Please cite this article as: Gunbin, K., et al., Integration of mtDNA pseudogenes into the nuclear genome coincides with speciation of the human genus. A hypothesis, Mitochondrion (2016), http://dx.doi.org/10.1016/j.mito.2016.12.001