Is the New Primate Genus Rungwecebus a Baboon?

Background In 2005, a new primate species from Tanzania, the kipunji, was described and recognized as a member of the mangabey genus Lophocebus. However, molecular investigations based upon a number of papionins, including a limited sample of baboons of mainly unknown geographic origin, identified the kipunji as a sister taxon to Papio and not as a member of Lophocebus. Accordingly, the kipunji was separated into its own monotypic genus, Rungwecebus. Methodology/Principal Findings We compare available mitochondrial and nuclear sequence data from the voucher specimen of Rungwecebus to other papionin lineages, including a set of geographically proximal (parapatric) baboon samples. Based on mitochondrial sequence data the kipunji clusters with baboon lineages that lie nearest to it geographically, i.e. populations of yellow and chacma baboons from south-eastern Africa, and thus does not represent a sister taxon to Papio. Nuclear data support a Papio+Rungwecebus clade, but it remains questionable whether Rungwecebus represents a sister taxon to Papio, or whether it is nested within the genus as depicted by the mitochondrial phylogeny. Conclusions/Significance Our study clearly supports a close relationship between Rungwecebus and Papio and might indicate that the kipunji is congeneric with baboon species. However, due to its morphological and ecological uniqueness Rungwecebus more likely represents a sister lineage to Papio and experienced later introgressive hybridization. Presumably, male (proto-)kipunjis reproduced with sympatric female baboons. Subsequent backcrossing of the hybrids with kipunjis would have resulted in a population with a nuclear kipunji genome, but which retained the yellow/chacma baboon mitochondrial genome. Since only one kipunji specimen was studied, it remains unclear whether all members of the new genus have been impacted by intergeneric introgression or rather only some populations. Further studies with additional Rungwecebus samples are necessary to elucidate the complete evolutionary history of this newly-described primate genus.


Introduction
In 2005, a new primate species from Tanzania, the kipunji, was described and originally recognized as a member of the mangabey genus Lophocebus, mainly based on its arboreality and noncontrasting black eyelids [1]. However, subsequent molecular studies suggested that the kipunji was more closely related to Papio, rather than to one of the two mangabey genera or any other member of the Papionini tribe [2]. Based on these findings, the kipunji was recently placed into its own genus, Rungwecebus [2]. Besides genetic evidence, the uniqueness of Rungwecebus was also supported by morphological, acoustic, behavioral and ecological characteristics [2][3][4]. Nevertheless, the classification of Rungwecebus as a new genus has been questioned [5].
To place Rungwecebus kipunji phylogenetically, Davenport et al. [2] generated sequences of three mitochondrial (cytochrome oxidase subunit I, COI; cytochrome oxidase subunit II, COII; 12SrRNA) and two nuclear (a 1,3 galactosyltransferase, a 1,3-GT; Y chromosomal testis-specific protein, TSPY) loci of a single voucher specimen. In 2008, another three nuclear loci (autosomal lipoprotein, LPA; autosomal gene encoding CD4; X chromosomal region, Xq13.3) of Rungwecebus became available and the TSPY sequence data were expanded [6]. In both studies, the Rungwecebus data were compared with orthologous sequences of other Old World monkeys, deposited in GenBank. However, the GenBank data did not cover the full taxonomic and geographic range of Papio ( Figure 1). In particular, sequences from Papio of southern Tanzania and neighboring regions, those geographically nearest to the range of Rungwecebus, were not available for both analyses [2,6]. Furthermore, many of the Papio sequences in GenBank are of unknown geographic origin because they derived from captive animals. In the course of a study on the evolution and phylogeography of Papio, we compared the available nuclear and mitochondrial sequence data of Rungwecebus with orthologous data of baboons from known geographic origin and representatives of all other papionin genera.

Results
The main result of our study is that, based on mitochondrial sequence variation, Rungwecebus does not constitute a sister taxon to Papio but clusters within Papio. This result contrasts with those of both Davenport et al. [2] and Olson et al. [6]. However, this relationship becomes obvious only when baboons from parapatric populations to Rungwecebus from south-east Africa are included in the study. The analysis of nuclear data provides limited power due to the extremely low levels of variation and hence, the relationships among baboons and the kipunji cannot be resolved unambiguously using these data.
Our mitochondrial data sets included all seven papionin genera. Furthermore, the 25 baboon individuals represented all five Papio species (P. papio, P. hamadryas, P. anubis, P. cynocephalus, P. ursinus) and covered most of the genus' geographic distribution. Phylogenetic tree reconstructions for the three individual loci (Figures S1, S2, S3) and the concatenated data set with 1486 bp in length ( Figure 2) revealed mainly identical and highly supported relationships, with only a few remaining unresolved or with low support. We found strong support for the division of African papionins into two major clades, one with Mandrillus and Cercocebus, and the other with Lophocebus, Theropithecus, Papio and Rungwecebus. Among the latter, a common origin of Papio and Rungwecebus was highly supported, but the relationship between this clade and either Lophocebus or Theropithecus was not well resolved. Within the Papio+Rungwecebus clade, we found several strongly supported haplogroups. However, these did not correspond to the traditionally recognized baboon species, and with the exception of P. papio, all other baboon taxa were para-or polyphyletic. In contrast, we found a strong geographical signal with local populations forming monophyletic haplogroups irrespective of their species affiliations. This reflects clearly the discordance between mitochondrial phylogeny and baboon morphology. The same is also true for Rungwecebus, which did not represent a sister lineage to the Papio genus [2,6], but instead clustered with yellow baboons (P. cynocephalus) from southern Tanzania, Malawi, Zambia and with chacma baboons (P. ursinus) from Zimbabwe, northern Namibia and northern South Africa. These yellow and chacma baboons from south-east Africa represent local populations that are geographically closest to Rungwecebus (Figure 1). To test for the reliability of the depicted relationships, we evaluated alternative phylogenetic positions of Rungwecebus among papionins with the Kishino-Hasegawa (KH) [7] and Shimodaira-Hasegawa (SH) [8] tests. Accordingly, a sister grouping of Rungwecebus to Papio, Theropithecus, Lophocebus or a Cercocebus+Mandrillus clade was significantly rejected (P,0.001, Table 1).
Each of our nuclear data sets comprised 11 sequences, of which five were derived from the five baboon species and the remaining six from representatives of the other papionin genera. Phylogenetic tree reconstructions for individual loci ( Figures S4, S5, S6, S7, S8) and the concatenated data set of 4486 bp in length ( Figure 3) provided similar tree topologies, although the resolution was relatively low, especially for single loci. As in the mitochondrial phylogeny, the nuclear data strongly supported a major division of  Table S1. doi:10.1371/journal.pone.0004859.g001 African papionins into a Cercocebus+Mandrillus and a Lophoce-bus+Theropithecus+Papio+Rungwecebus clade. Among the latter, the clustering of Papio and Rungwecebus was significant. However, due to the low number of polymorphic sites (Table S4) we were unable to resolve the relationships among the five baboon species and whether Rungwecebus is nested within the genus Papio or represents its sister lineage. Only one transversion (guanine present in all Papio species, thymine in all other papionin genera) at position 534 in the TSPY alignment provided some indications that Rungwecebus might be basal. Alternative tree topology tests significantly rejected a sister grouping of Rungwecebus to Theropithecus, Lophocebus or a Cercocebus+Mandrillus clade (P,0.05, P,0.001, Table 1). Although a tree with Rungwecebus as a sister lineage to Papio represents the most likely relationship, an unresolved polytomy including Rungwecebus and the five Papio species was not rejected (P = 0.578, P = 0.604, Table 1). Therefore, a phylogeny with Rungwecebus nested within the Papio clade cannot be excluded.
Due to the low number of variable sites in the nuclear data set, divergence ages were estimated only for mitochondrial data. Based on our estimates, the initial split separating southern baboon

Discussion
Based on its morphology [1,5], the kipunji was first regarded as a mangabey of the genus Lophocebus. However, subsequent genetic, cranial morphometric and ecological analyses provided evidence for the uniqueness of the kipunji and led to its being placed into a new genus [2,4,6]. Moreover, the then available genetic data suggested a sister taxon relationship of Rungwecebus with Papio and not with Lophocebus or any other papionin genus [2,6].
In contrast to Davenport et al. [2] and Olson et al. [6] we found no clear sister taxon relationship between the kipunji and the baboon clade. Instead, our analyses of mitochondrial sequences of the kipunji demonstrated that this lineage clusters closely with baboon lineages that lie nearest to it geographically (i.e. yellow baboons from southern Tanzania, Malawi and eastern Zambia and chacma baboons from Zambia, Zimbabwe, northern Namibia and northern South Africa). In contrast, nuclear sequence data give some indication that Rungwecebus is a sister taxon to Papio and not nested within the baboon clade. Although, this relationship is only weakly supported by the available nuclear data, morphological and ecological data clearly provide evidence for the distinctiveness of Rungwecebus and its separation from Papio [3,4]. Likewise for Papio, discordance between the mitochondrial phylogeny and the traditional classification into five clearly differentiated species [12,13] was found. Incomplete lineage sorting of mitochondrial DNA could be a possible explanation of such discrepancies, but then Rungwecebus would be expected to cluster with more ancient baboon lineages and not specifically with geographically-adjacent baboon haplotypes. The same is also true for baboon lineages, where local populations form monophyletic groups irrespective of their species affiliations. Hence, the observed discrepancy between mitochondrial data on the one hand and morphological, ecological (and nuclear) data on the other hand raises questions about the evolutionary history of Rungwecebus and Papio.
A possible explanation could be that Rungwecebus is actually a member of the genus Papio as suggested by the mitochondrial data and thus does not warrant placement into its own genus. In support of such a hypothesis, the mitochondrial sequence data yield a relatively recent divergence time (i.e. 0.35 [0.09-0.67] mya) between Rungwecebus and geographically-associated baboon lineages. If Rungwecebus belongs within Papio, its unique morphological and ecological characteristics, which are clearly different from that of baboons [2,5], would reflect recently acquired autapomorphies. Kipunjis are more arboreal than baboons [2,5] and hence, depend on woodland habitats. During the Pleistocene, climatic changes led to shifts in the distribution of forest habitats and savannah biotopes [14][15][16][17][18]. During such events baboon populations might have become isolated and adapted to forest or woodland habitats, resulting in a new and distinct morphotype, the kipunji. However,   Figure 1 and baboons living in forests in other parts of Africa (e.g. olive baboons in eastern and north-eastern Democratic Republic of Congo) did not adapt specifically to a more arboreal life [19]. A more likely explanation is that Rungwecebus is the result of introgressive hybridization. Hybridization and introgression events are well known for African papionins including baboons (within Papio [20]; between the genera Theropithecus and Papio, [21,22]; reviewed by [23]). In fact, Figure 2 shows multiple para-and polyphylies of baboon species, which is consistent with previouslydescribed introgressive hybridization and nuclear swamping events among various baboon lineages [e.g. [9][10][11][24][25][26][27][28]. For a broader view of Papio phylogeny see [9][10][11].
Hybridization and introgression have been considered important in the generation of plant diversity, and an appreciation of their role in the evolutionary diversification of animals has been growing over the past decade [29][30][31][32]. Besides papionins, additional examples of hybridization between other primate species, and even genera, are well documented for a number of clades [reviewed in 23,33]. Introgressive hybridization appears to have likely played a role in the evolution of hominoids, including Homo sapiens [23,30,[33][34][35][36]. There is also evidence that Trachypithecus pileatus, Macaca arctoides and Macaca munzala are the products of hybridization or introgression [37][38][39].
In the case of the kipunji, introgression of the maternallyinherited mitochondrial DNA from baboons into Rungwecebus seems to be likely. A possible scenario that would explain the introgression detected by our study is that a small population of baboons became isolated within the range of kipunjis resulting in female baboons reproducing with male kipunjis until 0.35 (0.09-0.67) mya. As mentioned above, repeated shifts in the extent of forest and savannah habitats during the Pleistocene may have promoted the isolation of local baboon populations. Given backcrossing of the hybrid offspring with kipunji or proto-kipunji males, over several generations, the frequency of baboon nuclear genes within the hybrid population would have decreased sharply. The result of such a process would have been a population with an almost complete kipunji nuclear genome, but with a baboon mitochondrial genome. This is what we have detected in our analyses. Since morphological characteristics would most likely be determined by nuclear genes, members of such a population would be expected to resemble kipunjis rather than baboons.
Based on our findings, the newly described genus Rungwecebus might be congeneric with baboon species and thus, Rungwecebus would be synonymous with Papio. Alternatively, and we would argue more likely, Rungwecebus represents a sister lineage to Papio and experienced later introgressive hybridization. However, since only a single Rungwecebus individual from the population in the southern highlands was studied, it is not clear whether the entire kipunji population possesses 1) a Papio-like mitochondrial DNA in general or 2) an admixture of Papio-like and undefined Rungwecebus haplotypes. Regardless of whether or not the detected mitochondrial haplotype of Rungwecebus turns out to be characteristic for only the examined specimen (or for only a subsample of the individuals belonging to this species) the evolution of Rungwecebus has been reticulate. Defining the extent of the reticulate (i.e. introgressive hybridization) events will require further molecular studies that incorporate more individuals and additional mitochondrial and nuclear loci.

Ethics statement
Our work was conducted according to relevant German and international guidelines, including countries where we collected fecal samples. Fecal samples were collected in a non-invasive way without disturbing, threatening or harming the animals. Blood samples were taken from zoo animals by zoo veterinarians for diagnostic reasons to check the health status of the respective individuals. Blood samples were explicitly not taken for our study.

Sample collection and preservation
In the course of a study on baboon phylogeography [10,11,28] fecal samples from baboons were collected during field surveys from several sites throughout Africa (Figure 1, Tables S1, S2). To ensure that the fecal samples were fresh, of baboon origin, and individually assignable, they were collected directly after baboon individuals were observed to defecate. Samples were preserved following the two-step storage method [40]. Accordingly, fecal

Laboratory methods
DNA from fecal, tissue and blood samples was extracted with the DNeasy and Stool Mini Kit from Qiagen following the supplier's recommendations. For Rungwecebus, sequence data from eight loci, including three mitochondrial (12SrRNA, COI, COII), one X chromosomal (Xq13.3), one Y chromosomal (TSPY) and three autosomal loci (a 1,3-GT, LPA, CD4) were available. To amplify respective fragments in baboons and other papionin species, PCR conditions and primers were identical to those used to amplify the loci from Rungwecebus [2,6] and other primates [41][42][43][44] (Table S3). The Xq13.3 and TSPY sequences were each generated via three overlapping PCR products [2,6]. The results of the PCR amplifications were checked on agarose gels. PCR products were cleaned with the Qiagen PCR Purification Kit and subsequently sequenced on an ABI 3730xl sequencer using the BigDye Terminator Cycle Sequencing Kit (Applied Biosystems). Sequences were deposited in GenBank (Tables S1, S2). To prevent contamination, laboratory procedures followed recommended, standard protocols [37,[45][46][47][48]. Specifically, DNA extraction, PCR, PCR purification and sequencing were performed in separate laboratories and repeated after several months, while always only one individual per species was tested. Finally, all PCR reactions were performed with negative (HPLC-purified water) controls.

Statistical methods
To obtain a comprehensive overview of the phylogenetic position of Rungwecebus among papionins, further orthologous sequences from related taxa deposited at GenBank were included in our study (Tables S1, S2). Each of the five nuclear data sets comprised 11 sequences, representing all seven papionin genera and the five recognized baboon species. Each of the three mitochondrial data sets contained 33 sequences, which represent the seven papionin genera including baboons from most of their geographic range. The macaque sequences (Macaca mulatta or M. sylvanus) were used as an outgroup. Sequences were easily aligned by hand, because few or no indels were present. Due to the low number of polymorphic sites in the nuclear data sets, point mutations, deletions and/or insertions were individually inspected. For phylogenetic tree reconstructions, gaps and poorly aligned positions were manually removed. The sizes of the different alignments and the number of excluded indels are presented in Table S3. Calculations were performed for each locus separately as well as for concatenated nuclear and mitochondrial data sets.
Trees were constructed with maximum-parsimony (MP) and neighbor-joining (NJ) algorithms as implemented in PAUP 4.0b10 [49] as well as with maximum-likelihood (ML) and Bayesian algorithms, using the programs GARLI 0.951 [50] and MrBayes 3.1.2 [51,52], respectively. For MP analyses, all characters were treated as unordered and equally weighted throughout. A heuristic search was performed with the tree-bisection-reconnection (TBR) algorithm with random addition of sequences. The maximum number of trees was set to 100. NJ, ML and Bayesian trees were constructed with the respective best-fitting models as selected under the Akaike information criterion with MODELTEST 3.7 [53] (Table S3). NJ and ML trees from the concatenated nuclear and mitochondrial data sets were analyzed with the TrN+I and TrN+I+G models, respectively. Bayesian analyses for the combined data set were performed in a partitioned framework, allowing locusspecific parameter estimation. Relative support of internal nodes was performed by bootstrap analyses with 1,000 (MP, NJ) or 500 replications (ML). In GARLI, only the model specifications settings were adjusted according to the respective data set, while all other settings were left at their default value. ML majority-rule consensus trees were calculated in PAUP. For Bayesian analyses, four Monte Carlo Markov Chains with the default temperature of 0.1 were used. Four repetitions were run for 10,000,000 generations with tree and parameter sampling occurring every 100 generations. The first 25% of samples were discarded as burnin, leaving 75,001 trees per run. Posterior probabilities for each split were calculated from the posterior density of trees.
To evaluate the reliability of the depicted phylogenetic position of Rungwecebus, alternative tree topologies were evaluated with the KH [7] and SH [8] tests with full optimization and 1,000 bootstrap replications in PAUP. Hypothetical sister group relationships of 1) Rungwecebus to Papio, Lophocebus, Theropithecus and Cercocebus+Mandrillus or 2) that Rungwecebus is nested within Papio were tested.
Due to the low number of variable sites in the nuclear data set, divergence ages were estimated only for mitochondrial data. Therefore, further outgroup taxa were included in the data set (Table S1). After removal of additional indel positions, the final alignment for divergence age estimations comprised 1,479 bp. A Bayesian MCMC method, which employs a relaxed molecular clock approach [54], as implemented in BEAST v1.4.6 [55], was used. We assumed a relaxed lognormal model of lineage variation and a Yule prior for branching rates. Data were partitioned by manually editing the XML file and by applying the respective bestfitting models and parameters [53]. As calibrations we used the divergence between human and chimpanzee, which has been dated at 6-7 mya [56,57], the divergence between Papio and Theropithecus, which is estimated at 3.5-4.0 mya [58,59 and references therein], and the spilt between hominoids and cercopithecoids, which is estimated at 23 mya [for discussion of fossil data see [60]. Instead of hardbounded calibration points, we used the published dates as a normal distribution prior for the respective node. For C1 (Pan/Homo) this translates into a normal distribution with a mean of 6.5 mya and a standard deviation of 0.3 mya, for C2 (Papio/Theropithecus) into a mean of 3.75 mya and a standard deviation of 0.3 ma (95% credibility interval: 3.0-4.5 mya) and for C3 into a mean of 23.0 mya and a standard deviation of 2.5 ma. Four replicates were run for 10 million generations with tree and parameter sampling occurring every 100 generations. The adequacy of a 10% burnin and convergence of all parameters was assessed by visual inspection of the trace of the parameters across generations using the software TRACER v1.3 [61]. Subsequently, the sampling distributions of four independent replicates were combined and downsampled using the software LogCombiner v1.4.6 and the resulting 10,000 samples summarized and visualized using the software TreeAnnotator v1.4.6 and FigTree v1.1.1 [62]. The first two programs are part of the BEAST package [55]. Figure S1 50% majority rule consensus tree (cladogram) based on COI sequences. Numbers on nodes represent bootstrap or posterior probability values (first: MP, second: NJ, third: ML, fourth: Bayesian). Dashes indicate values #50%. Red = P. papio, green = P. anubis, grey = P. hamadryas, yellow = P. cynocephalus, brown = P. ursinus. For abbreviations see Fig. 1 and Table S1. Found at: doi:10.1371/journal.pone.0004859.s001 (0.13 MB TIF) Figure S2 50% majority rule consensus tree (cladogram) based on COII sequences. Numbers on nodes represent bootstrap or posterior probability values (first: MP, second: NJ, third: ML, fourth: Bayesian). Dashes indicate values #50%. Red = P. papio, green = P. anubis, grey = P. hamadryas, yellow = P. cynocephalus, brown = P. ursinus. For abbreviations see Fig. 1 and Table S1. Dashes indicate values #50%. Red = P. papio, green = P. anubis, grey = P. hamadryas, yellow = P. cynocephalus, brown = P. ursinus. For abbreviations see Fig. 1 and Table S1.