Background

The Old World monkeys are traditionally divided into the two subfamilies Cercopithecinae and Colobinae, which differ from each other by numerous morphological, behavioural and ecological characteristics [1–4]. While detailed information on the evolutionary history of cercopithecines (baboons, mangabeys, macaques and guenons) is at hand, knowledge on colobines is still scarce. Although some molecular studies [5–13] exist, they mainly focus on relationships within genera or species groups and not on the general phylogeny of the subfamily. Recently, a first mitochondrial phylogeny of colobine genera was established [14], which confirmed some previous assumptions, but also led to confusions, calling for further research to definitively elucidate their evolutionary history.

Based on distribution and morphology, the colobines are traditionally divided into an African and an Asian clade [3, 4], while Asian colobines are more diverse than their African relatives. Hence, the Asian forms are further split into the odd-nosed monkey (Pygathrix, Rhinopithecus, Nasalis, Simias) and langur (Presbytis, Trachypithecus, Semnopithecus) groups, which are both believed to be monophyletic. Accordingly, langurs were originally combined in the single genus Presbytis [15–17] or Semnopithecus [18], but based on neonatal colouration and cranial morphology, they were split into the three genera Semnopithecus, Trachypithecus and Presbytis [19], and a fourth genus (Kasi) was added [20]. Alternatively, Semnopithecus was separated from Presbytis, with Trachypithecus forming a subgenus of the former [21–23], but recent classifications use a subdivision of langurs into the three genera Presbytis, Trachypithecus and Semnopithecus [3, 4, 24–26].

Within the different langur genera, several species are recognized, which are lumped into species groups due to similar fur colouration, behaviour, ecology or distribution. With five species groups (T. [obscurus], T. [francoisi], T. [cristatus], T. [pileatus] and T. [vetulus]) [4], the genus Trachypithecus is the most diverse of all langurs and posses also the widest distribution, ranging from South India and Sri Lanka through mainland Southeast Asia to the Sundaland (Fig. 1, Fig. 2). Although all of them are morphologically similar, T. [vetulus] was sometimes separated in its own genus Kasi [20], and recent mitochondrial sequence data indicate a closer affiliation of T. [vetulus] and T. [pileatus] to Semnopithecus than to Trachypithecus [7, 9]. Accordingly, the two T. [vetulus] members were recognized as species of Semnopithecus [26]. In contrast to Trachypithecus, the genus Semnopithecus is restricted to the Indian subcontinent (Fig. 1) and traditionally regarded as monotypic with the only species S. entellus [3, 27], although recently several subspecies were elevated to species status [4]. The third langur genus, Presbytis, includes several species, which occur solely in the Sundaland, but are not lumped into distinct species groups [4].

Figure 1
figure 1

Distribution of the genus Semnopithecus and Trachypithecus [ obscurus ]. Genus affiliations and species groups after Groves [4].

Full size image
Figure 2
figure 2

Distribution of the Trachypithecus species groups T . [ vetulus ], T . [ pileatus ], T . [ francoisi ] and T . [ cristatus ]. Genus affiliations and species groups after Groves [4].

Full size image

The phylogenetic relationships among the different Asian colobine genera are disputed. Although a common origin of the odd-nosed monkeys was recently confirmed [14], evidence for monophyly of its putative sister clade, the langur group, is still lacking. Moreover, available data depict Trachypithecus and Presbytis as sister taxa to the exclusion of Semnopithecus [14], which contradicts with traditional classifications, in which Trachypithecus and Semnopithecus are believed to form a clade to the exclusion of Presbytis [4, 21–23]. These findings raise the question of what positions the langur genera occupy among Asian colobines and whether the langurs form a monophyletic clade in general. Moreover, the affiliations of different Trachypithecus species groups, especially T. [vetulus] and T. [pileatus], are disputed, and hence, led to different classifications. Currently, only few genetic data are available [7, 9, 28], so that further information from other markers is required to definitively establish their relationships.

To address all these issues, mitochondrial and Y chromosomal sequence data were phylogenetically related and combined with presence/absence analysis of retroposon integrations. This approach was used to simultaneously analyse paternal-, maternal- and biparental-inherited markers, which allow the detection of incongruence between different gene trees indicating possible hybridization or introgression events between different lineages [29–31]. To determine the phylogenetic position of the langur genera among Asian colobines, a 5 kb fragment of the mitochondrial genome was sequenced from eight colobine genera, and combined with presence/absence analysis of retroposon integrations. Retroposon insertion events are nearly homoplasy-free and precise excision of elements is highly unlikely [32, 33]. Accordingly, retroposon insertions are powerful informative markers, which were already successfully applied to elucidate phylogenetic relationships in various primate lineages [34–39], and comparisons of the data with sequences already deposited at GenBank revealed only intra-species or -generic variation, indicating that no nuclear pseudogenes were amplified.

To determine the phylogenetic relationships among analysed genera, tree reconstructions were conducted with different algorithms, which all led to the same tree topology (Fig. 3). Moreover, most relationships are significantly supported and congruent with previous classifications [1–4], indicating the reliability of the data set. In detail, the reconstructions confirm relationships among cercopithecine genera (Macaca, Papio, Chlorocebus), and reciprocal monophyly of cercopithecines and colobines, as well as of African and Asian colobines. Within the Asian clade, relationships are not resolved, although at least a common origin of the odd-nosed monkeys is depicted. Neither the assumed monophyly of the langur genera, nor the expected close affinity of Trachypithecus and Semnopithecus or the recently indicated sister grou** of Trachypithecus and Presbytis [14] can be verified with significance. Moreover, all alternative tree topologies, in which the three langur genera were regarded as monophyletic, or variously recognized as sister clades to each other, to the odd-nosed monkey clade, or even as basal among Asian colobines, were not rejected (P = 0.097 – 0.776). Accordingly, an unresolved polytomy among the three langur genera and the odd-nosed monkey clade may best reflect the relationships among Asian colobines based on the mitochondrial data.

Figure 3
figure 3

Phylogenetic relationships among colobines and related genera based on mitochondrial data. Numbers on nodes indicate support values >80% (first: ML, second: NJ, third: MP).

Full size image

Several retroposon insertions were detected, which provide insights into the branching pattern of colobines (Fig. 4). Together with seven already published integrations [

Figure 4
figure 4

Phylogenetic relationships among Asian colobine genera based on retroposon integrations. Dark dots represent new generated data, whereas light dots refer to already published data [37]. Numbers in dots indicate single integration events.

Full size image

Although in general the mitochondrial data are suitable to elucidate relationships among the different genera, as indicated by the correct and significantly supported branching patterns among all other studied genera, the relationships among the langurs are unresolved, which is concordant with previous results [14]. In contrast, the presence/absence analysis of retroposon integrations provides evidence for a monophyletic odd-nosed monkey clade and a common origin of Trachypithecus and Semnopithecus, which is in agreement with morphological hypotheses [1, 3, 4, 21–23]. Regardless which markers were used, the phylogenetic position of Presbytis among Asian colobines and accordingly the unity of the langurs remains unclear and needs further investigations.

Species group phylogeny

In order to settle affiliations among the different Trachypithecus species groups and their members, mitochondrial and Y chromosomal sequence data were combined with information on retroposon integrations.

The mitochondrial phylogeny was established on the basis of 573 bp long cytochrome b gene sequences, generated from most species recognized in the genus and its sister genus Semnopithecus. In all tree reconstructions, identical relationships were obtained, with most branches being significantly supported (Fig. 5a). Accordingly, the different species groups are divided into two major groups, with one including solely groups of the genus Trachypithecus, whereas the second one includes representatives of Trachypithecus and Semnopithecus. In the mixed clade, T. [vetulus] and T. [pileatus] members are lumped together with S. entellus. Whereby T. [pileatus] is monophyletic, the members of T. [vetulus] are paraphyletic, with T. vetulus forming a sister clade to S. entellus from Sri Lanka, and T. johnii with S. entellus from South India. Furthermore, a fourth lineage in the mixed clade was detected, which is represented by S. entellus from North India. In contrast, the three species groups (T. [obscurus], T. [francoisi], T. [cristatus]) in the clade comprising solely Trachypithecus groups are all monophyletic. Alternative relationships, in which T. [vetulus] is recognized as monophyletic, either T. [vetulus] or T. [pileatus] belongs to Trachypithecus or even both are members of Trachypithecus, were tested, but all of them were rejected (P < 0.05).

Figure 5
figure 5

Phylogenetic relationships among Semnopithecus and Trachypithecus species groups based on a) mitochondrial data, b) Y chromosomal data, and c) retroposon integrations. Numbers on nodes indicate support values (first: ML, second: NJ, third: MP), and boxed species belong to a species group, with species in bold giving the name of the group.

Full size image

With some exceptions, the Y chromosomal data provide a similar picture (Fig. 5b), but due to the low number of polymorphic sites, support values are in general not as high as in the mitochondrial tree. According to the reconstructions, the species groups are divided into two major clades, with one comprising T. [obscurus], T. [cristatus], T. [francoisi] and T. [pileatus], and the other, T. [vetulus] and Semnopithecus. Relationships among the latter are not resolved. All alternative tree topologies, in which either T. [vetulus] belongs to Trachypithecus or T. pileatus groups with Semnopithecus, were rejected (P < 0.05).

Retroposon insertions further deepened our knowledge on the species group relationships. Altogether, three informative integrations were analysed (Fig. 5c), with one occurring in T. [obscurus], T. [cristatus] and T. [francoisi], and the other two in T. [vetulus] and Semnopithecus. Interestingly, all three integrations are absent in T. pileatus.

With the exception of the varying position of T. [pileatus], the affiliations of the remaining species groups are congruent among different gene trees. Accordingly, all analysed markers relate T. [vetulus] with Semnopithecus, indicating that this species group is a real member of the genus Semnopithecus and not of Trachypithecus as assumed by morphological similarities [4]. These similarities may be the results of adaptations to similar ecological conditions (Semnopithecus is semi-terrestrial and lives in deciduous forest, whereas Trachypithecus including T. [vetulus] is arboreal and occurs in wet evergreen forests). Although the Y chromosomal data allow no resolution within the Semnopithecus – T. [vetulus] clade, the mitochondrial data indicate paraphyly of the two T. [vetulus] species, with T. vetulus clustering with S. entellus from Sri Lanka and T. johnii with S. entellus from South India, which is concordant with their geographical distribution. These findings indicate paraphyly of S. entellus, whereby North Indian representatives form a further distinct lineage. Accordingly, the langurs of the Indian subcontinent should be split into three species groups, with one occurring solely on Sri Lanka, one in Southern India and a third one in Northern India, whereas the Gondavari river seems to be barrier between the latter two.

Monophyly of each of T. [obscurus], T. [cristatus] and T. [francoisi] and their close affiliation is depicted in all gene trees, so that all of them can be regarded as true members of Trachypithecus. These findings confirm previous molecular studies [8–11, 13] and are in general agreement with recent classifications [4, 26].

The only discrepancies between different gene trees were obtained for T. [pileatus]. Whereas the mitochondrial data link the species group with Semnopithecus and T. [vetulus], the Y chromosomal data affiliates it with T. [obscurus], T. [cristatus] and T. [francoisi]. These findings might be explained by incomplete lineage sorting of ancestral mitochondrial or Y chromosomal haplotypes. Accordingly, the ancestor of Trachypithecus, Semnopithecus and T. [pileatus] carried multiple DNA lineages with one lineage being randomly fixed in two taxa, but not in the third. Alternatively, the varying position of T. [pileatus] in different gene trees might be explained by past hybridization between Semnopithecus and Trachypithecus. As depicted by the three retroposon insertions, this putative hybridization event would have occurred between ancestral forms of Semnopithecus and Trachypithecus, before both genera diverged into distinct species groups. The hybridization hypothesis is also supported by some intermediate morphological characteristics [4] and the distribution of T. [pileatus], which is sandwiched between those of Semnopithecus and other Trachypithecus species groups (Fig. 1, Fig. 2).