To elucidate the phylogenetic relationships among the different langur genera and their position among Asian colobines, mitochondrial sequence studies were combined with presence/absence analysis of retroposon insertions.

The herein analysed 5 kb fragment of the mitochondrial genome was assembled from sequences derived from 1–2 kb long and partly overlapping PCR products, whereby no inconsistencies in overlapping sequence fragments were detected. As template material, mainly DNA extracted from feces was used, which minimizes the amplification of nuclear pseudogenes 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 1234, 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 grouping 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.

Several retroposon insertions were detected, which provide insights into the branching pattern of colobines (Fig. 4). Together with seven already published integrations 38, now 14 loci confirm monophyly of the Asian colobines to the exclusion of the African genus Colobus. Among Asian colobines, one insertion indicates a common origin of the odd-nosed monkeys, which is verified by three further loci 38. Most prominent are the five retroposon integrations, which confirm a sister grouping of Trachypithecus and Semnopithecus. No loci were detected, which link Presbytis either with the other two langur genera or the odd-nosed monkeys.

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 134212223. 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.

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).

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 89101113 and are in general agreement with recent classifications 426.

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).

The species analysed in this study are presented in Table 2. All study specimens were identified by fur colouration and other external characteristics. Hanuman langur (S. entellus) samples were collected only from founder animals, of which the area of capture was at least roughly known. Total genomic DNA was extracted from blood, tissue or feces using the DNeasy or Stool Mini Kits from Qiagen. When hair follicle cells were used, 1–3 hairs were directly implemented into the PCR reaction after they were washed with sterile water and 95% ethanol. To prevent contaminations, sample collection and laboratory procedures followed described standard protocols 40414243. In detail, all fecal and hair samples were collected with gloves and stored in sterile tubes or plastic bags before further processing. DNA extraction, PCR, gel extraction and sequencing was performed in separate laboratories and repeated after several months, while always only one individual per species was tested. Moreover, from most specimens two different sample types were available, which both were used as template material. Sequences from independent analyses were identical. Finally, all PCR reactions were performed with negative (distilled water) controls.

To determine the phylogenetic position of langur genera among Asian colobines, a ~5 kb fragment of the mitochondrial genome was sequenced from all colobine genera with the exception of Simias and Procolobus. This region spans the cytochrome b gene, the control region, the 12S rDNA and 16S rDNA, and the intermediate tRNAs. To exclude contaminations of the dataset with nuclear pseudogenes, mainly DNA extracted from feces was used as template material, and ~1–2 kb long and ~200–400 bp overlapping fragments were amplified. PCR products were generated via hot-start technique using a set of 24 primers (Table 3) and PCR conditions comprising a pre-denaturation step at 94°C for 2 min, followed by 40 cycles at 94°C for 1 min, 60°C for 1 min and 72°C for 1–2 min, and a final extension step at 72°C for 5 min. 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 3100-Avant sequencer using the BigDye Terminator Cycle Sequencing Kit (Applied Biosystems). To obtain a comprehensive overview on the phylogeny of colobines, the dataset was expanded with further sequences from related taxa deposited at GenBank (Table 2). Accordingly, the final dataset comprised 15 taxa, including eight colobines, four cercopithecines, two hominoides and a New World monkey, which was used as outgroup taxon. Sequences were aligned with ClustalW 44 and subsequently checked by eye. Gaps and poorly aligned positions were removed with the G-blocks software 45, which reduced the final dataset to 4336 bp. Based on this alignment, phylogenetic trees were constructed with the maximum-parsimony (MP), neighbor-joining (NJ) and maximum-likelihood (ML) algorithms as implemented in PAUP 4.0b10 46 and TREEPUZZLE 5.0 47. 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 and ML trees were constructed with the GTR + I (= 0.2526) + Γ (= 0.4816) model of sequence evolution as it was selected as best-fitting model under the Akaike information criterion with MODELTEST 3.06 48. Relative support of internal nodes was performed by bootstrap analyses with 1,000 replications (MP, NJ), or by the quartet puzzling support values on the basis of 1,000 puzzling steps (ML). Finally, to evaluate the reliability of the depicted phylogenetic position of the langur genera among Asian colobines, alternative tree topologies were evaluated with the Kishino-Hasegawa 49 and Shimodaira-Hasegawa 50 tests in PAUP, both performed with full optimization and 1,000 bootstrap replications. Therefore, all three langur genera were regarded as monophyletic or variously recognized as sister clade to each other, to the odd-nosed monkey clade or even as sister clade to all other Asian colobines.

To determine phylogenetic affiliations of species groups, a 573 bp long fragment of the mitochondrial cytochrome b gene was analysed from all species of the different groups (Table 2). The generation of sequences followed laboratory methods as described 91113. The final alignment, which was easily generated by eye due to the absence of insertions or deletions, comprised 19 individuals including the outgroup taxon Presbytis melalophos. Phylogenetic trees were constructed as described above. As best-fitting model, MODELTEST selected the TIM + I (= 0.5977) + Γ (= 2.3137) model, which was applied for NJ and ML reconstructions. As for the 5 kb fragment, several alternative tree topologies, 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.

The SRY gene was selected as it represents a single-copy gene and as it is proved to be informative in reconstructing the Y chromosomal evolutionary history of macaques 51. PCR conditions and primers were applied as described 51. To amplify the SRY gene from fecal material, two overlapping fragments were generated with published primers 51 and the newly generated internal primers 5'-TGGGCGGAGTTGAGAGGGGT-3' and 5'-TAGCGGTCCCGTTGCTGCGG-3'. The final alignment of 777 bp comprised 15 taxa representing all species groups of the genera Semnopithecus and Trachypithecus as well as Presbytis melalophos, which was used as outgroup. To reconstruct NJ and ML trees, the K80 model of sequence evolution, determined with MODELTEST, was used. MP trees were generated as described above. The reliability of the depicted position of T. [vetulus] and T. [pileatus] was tested in PAUP by using alternative tree topologies, in which either T. [vetulus] belongs to Trachypithecus or T. pileatus groups with Semnopithecus.

Due to their high copy number and relative small size (~300 bp), the primate specific Alu elements were selected as cladistic markers. The presence or absence of Alu elements in different colobines at specific loci was tested via PCR using primers occupying the flanking region of the insertion site. Details on analysed loci, primers and studied species are listed in Table 4.

To detect new loci, a subtractive hybridization approach 52 was performed with some modifications. As tracer and driver, different colobine genera were selected. Genomic DNA of tracer and driver was digested with RSAI (Fermentas), and subsequently, the adapters AdapA1/AdapAA1 (5'-TGTAGCGTGAAGACGACAGAAAGGGCGTGGTGCGGAGGGCGGT-3'/5'-ACCGCCCTCCG-3') and AdapA2/AdapAA2 (5'-TGTAGCGTGAAGACCTGTCTTAGGGCGTGGTGGCCAGGGCCGT-3'/5'-ACGGCCCTGGC-3') were ligated to the tracer fragments. Each of ~15 ng tracerA1 and tracerA2 were hybridized with ~1,500 ng driver for 20 h at 60°C. 2 μl of the hybridization result was used as template to amplify solely tracer fragments using the adapter primers A1 (5'-TGTAGCGTGAAGACGACAGAA-3') and A2 (5'-TGTAGCGTGAAGACCTGTCTT-3'). The PCR program consisted of a pre-extension step at 72°C for 6 min to fill in adaptor ends, followed by 25 cycles, each with a denaturation step at 95°C for 1 min, annealing at 60°C for 1 min and extension at 72°C for 2 min. To enrich fragments with Alu insertions, a semi-nested PCR was added using either primer A1 or A2 and the Alu-specific AluY (5'-GGAGAATGGCGTGAACCCGGGA-3') oligonueclotide. The PCR products were separated on agarose gels and fragments over 500 bp were excised from the gel. After purification, the fragments were cloned into the pGEMTeasy vector (Promega) and transformed into electro-competent TOP10 cells (Invitrogen). Bacterial clones were collected in 96-well microtiter plates and re-screened via PCR with the primers A1 or A2 and AluY. Positive clones were sequenced and analysed with REPEATMASKER and BLAST as implemented in NCBI and EMBL. Based on the generated alignments, locus-specific primers were constructed, with the forward primer occupying a region 5'-end upstream of the insertion site, which is conserved among the colobine and human or chimp sequences. Due to the absence of the 3'-end downstream sequence of the tested colobine species, reverse primers were constructed solely on the basis of human or chimp sequences. Subsequently, the presence or absence of respective Alu insertions in different colobine species was tested via PCR. The orthology of insertions was confirmed by sequencing of at least one species per genus or species group. Sequences were deposited at GenBank and are available under the accession numbers EU004484–EU004537 and EU006662–EU006692.