Chromosome preparation: Peripheral blood samples were collected from all animals with previously heparinized disposable syringes. Lymphocytes were cultured for 72 h at 37 °C following Mudry (1990). At least 50 metaphases were analyzed to determine the diploid number (2N) at 1000×. Metaphase spreads were treated with G-Wright banding (Steinberg et al. 2007). At least 10 G-banded metaphases with the species diploid number (2N) were photographed with a Leica DFC 340 FX camera. Chromosomes were arranged according to previously described karyotypes using Photoshop CS (Adobe) and the species assignation of each specimen was corroborated.

Analysis of homeologies: For Alouatta caraya and Alouatta guariba clamitans, the homeologies with human chromosomes and the homeologies with the other South American howlers are well known (Consigliere et al. 1998, Mudry et al. 2001, de Oliveira et al. 2002, Stanyon et al. 2011). The G-banded chromosomes of Alouatta pigra and Alouatta palliata were first compared with those of Alouatta caraya and Alouatta guariba clamitans. We took Alouatta caraya’s karyotype as the reference for the comparisons with Mesoamerican howlers (Mudry et al. 2001, Szapkievich and Mudry 2003). To compare homeologies, the G-banded metaphases obtained for Alouatta caraya, Alouatta guariba clamitans, Alouatta pigra and Alouatta palliata were also compared with those published for Alouatta guariba guariba (de Oliveira et al. 2002, Stanyon et al. 2011), Alouatta macconnelli (de Oliveira et al. 2002), Alouatta seniculus arctoidea (Consigliere et al. 1996), Alouatta belzebul (Armada et al. 1987, Consigliere et al. 1998) and Alouatta sara (Consigliere et al. 1996).

FISH analysis with human chromosome painting probes #3 and #15 was used as a tool to confirm the identity of the sex chromosome systems in howlers. Whole chromosome painting probes for human chromosomes #3 (red), #15 (green), #21 (green), X (green) and Y (red) (PCT3 Cy3, PCT15 FITC, PCT21 FITC, PCTX FITC, PCTY Cy3, LEXEL S.R.L., Buenos Aires, Argentina) were used for FISH analysis on the metaphases of Alouatta pigra, Alouatta caraya, Alouatta guariba clamitans and Alouatta palliata. Homo sapiens (HSA) metaphases were used as a positive control of hybridization. The HSA3/21 syntenic association, considered ancestral in mammals and conserved in most primate species (Müller et al. 2000), was analyzed simultaneously as a control of synteny conservation. Human X and Y chromosomes were also tested.

FISH was performed according to the supplier’s instructions (LEXEL S.R.L., Buenos Aires, Argentina). Slides were counterstained with DAPI (Sigma) and analyzed with a Leica DMLB fluorescence microscope. Chromosome images were obtained with a Leica DFC 340 FX camera. Images were processed with Image Pro-Plus 4.5 (Media Cybernetics Inc.).

Our results were compared with those previously described (Consigliere et al. 1996, 1998, Mudry et al. 2001, de Oliveira et al. 2002, Stanyon et al. 2011).

Chromosomal dataset: We used data obtained from the comparisons of G-banding patterns and the analysis of chromosomal syntenic associations, both from the present study and from previous reports (Consigliere et al. 1996, 1998, García et al. 2001, 2002, Mudry et al. 2001, de Oliveira et al. 2002, Amaral et al. 2008, Stanyon et al. 2001, 2011). We considered the structural changes as characters. The pattern observed before and after their occurrence, i.e. their presence or absence, was considered as the character states. The matrix (see Appendix 1) was produced taking into consideration the characters proposed by Neusser et al. (2001) and modified for howlers by de Oliveira et al. (2002). These authors used an abbreviated nomenclature for ancestral Platyrrhini chromosome forms with a correspondence in the human karyotype. In the present contribution, for the character nomenclature, we refer directly to the human G band ideogram (Table 2). New characters were obtained from our karyological comparisons and introduced in the chromosomal dataset.

Molecular dataset: The sequences available in GenBank for the same species used in the G-banding pattern and FISH comparisons were taken into to choose the molecular marker. The only molecular marker that fullfiled all the requirements was cyt b. The sequences used were (Genbank Accession Numbers): Alouatta belzebul (AY374348.2), Alouatta caraya (AY374359.2), Alouatta seniculus arctoidea (AY065886.1), Alouatta sara (AY065887.1), Alouatta macconnelli (AY065888.1), Alouatta guariba guariba (AY065899.1), Alouatta guariba clamitans (DQ679782.1), Alouatta pigra (AY065884.1), Alouatta palliata (AY065879.1) (Cortés-Ortiz et al. 2003, Harris et al. 2005, Lorenz et al. 2005, Nascimento et al. 2005, Casado et al. 2010). Cebus apella Linnaeus, 1758 (FJ529102.1) and Lagothrix lagotricha Humboldt, 1812 (AY671799.1) were used as outgroups. Cebus apella, from the Cebidae family, was taken as an outgroup species, since it is accepted that this species presents the most ancestral karyotype within Platyrrhini (Clemente et al. 1990, García et al. 2000). Lagothrix lagotricha, also a member of the Atelidae family, was chosen as the second outgroup to test the monophyly of the group. All sequences were aligned using CLUSTALW (Thompson et al. 1994).

Phylogeny: A Maximum Parsimony phylogeny using the exhaustive search option was obtained with PAUP 4.0 software (Phylogenetic Analysis Using Maximum Parsimony, (Swofford 2002)), for each separate partition and the combination of both the chromosomal and molecular datasets. All characters had the same weight, based on the premise that chromosome rearrangements occur by equal chance (de Oliveira et al. 2002, Dobigny et al. 2004). The relative stability of nodes was assessed by bootstrap estimates (Felsenstein 1985) based on 200 iterations. Each bootstrap replicate involved a heuristic parsimony search with 10 random taxon additions and tree-bisection reconnection (TBR) branch swapping.

Karyological analysis: The cytogenetic characterization of the Alouatta specimens showed diploid numbers, sex chromosome systems and G-bandings patterns in agreement with the ones previously described for each species. Figures 1a and 1b show all the comparisons performed.

Chromosomal homeologies between Alouatta caraya and Alouatta palliata: The chromosomal rearrangements that could explain the homeologies were grouped in two categories: 1) Alouatta palliata chromosomes with no rearrangements with respect to Alouatta caraya chromosomes: 3, 5, 6, 7, 8, 9, 10, 13, 14, 16, 17, 19, 20, 21, 22, 24, 25 and X₁; 2) Alouatta palliata chromosomes with more than one rearrangement with respect to Alouatta caraya chromosomes: 1, 2, 4, 11, 12, 15, 18 and 23. No homeologies were allocated for Alouatta palliata chromosome 26 and chromosome arms 4q and 2p using the level of resolution of the classical cytogenetic techniques applied. The rearrangements detected between the Alouatta caraya and Alouatta palliata karyotypes included at least seven fissions/fusions, two paracentric inversions and one deletion. Alouatta caraya chromosome 7 (X₂ in males) shares homeology with two Alouatta palliata chromosome pairs, 23 and 18, which are not the ones involved in the sex chromosome system in Alouatta palliata. The Alouatta palliata chromosomal pair 19 (X₂ in males) shares homeology with chromosome 14 of Alouatta caraya.

Chromosomal homeologies between Alouatta caraya and Alouatta pigra: The chromosomal rearrangements that could explain the homeologies were grouped in two categories: 1) Alouatta pigra chromosomes with no rearrangements with respect to Alouatta caraya chromosomes: 2, 5, 6, 7, 8, 10, 15, 16, 17, 20, 22, 23, 25, 28 and X₁; 2) Alouatta pigra chromosomes with more than one rearrangement with respect to Alouatta caraya chromosomes: 1, 3, 4, 9, 11, 13, 18, 19, 24, 26 and 27. No homeologies were allocated for Alouatta pigra chromosomes 4p_prox, 12, 14 and 21 using the level of resolution of the classical cytogenetic techniques applied. The rearrangements detected between the Alouatta caraya and Alouatta pigra karyotypes included at least 12 fissions/fusions, two paracentric inversions, two translocations and one deletion. Alouatta caraya chromosome 7 (X₂ in males) shares homeology with two Alouatta pigra chromosome pairs, 26 and 19, which are not the ones involved in the sex chromosome system in Alouatta pigra. Alouatta pigra chromosome 17 (X₂ in males) shares homeology with chromosome 14 of Alouatta caraya (which in turn has homeology with HSA7).

Chromosomal homeologies among all howlers: The chromosomal homeologies found among all howlers are shown in Table 3 and Figures 1a and 1b. Results show that Mesoamerican howlers share several human chromosomal syntenic associations with South American ones: HSA15q13-q21.2/4q23-q31.2 and HSA16p/10q, shared with all howlers; HSA15q21.3-q24/14, shared with all howlers except Alouatta seniculus arctoidea and Alouatta macconnelli, and HSA8p/18, shared with all howlers except Alouatta seniculus arctoidea. Two new chromosomal syntenic associations, HSA4p_ter-q22/9/11 and HSA15q21.3-q24/14/21q, were found for Alouatta pigra in chromosomes 1 and 4q, respectively.

In the Homo sapiens metaphases, the hybridization signals on chromosomes HSA3, HSA21, HSA15, HSAX and HSAY for chromosome painting probes #3 (red), #21 (green), #15 (green) (Figure 2a), X and Y (data not shown) were corroborated.

In Alouatta guariba clamitans, the signal for HSA21 was observed in 18q_ter, the signal for HSA3 was observed in 18q_prox (thus corroborating the HSA3/21 synteny in Alouatta guariba clamitans), 16q, 7q (X₂ in males) and Y₂, and the signal for HSA15 was observed in 1_int, 2_int, 7p (X₂) and Y₁. This corroborates the HSA3/15 syntenic association to the multiple sex chromosome system X₁X₁X₂X₂X₃X₃/X₁X₂X₃Y₁Y₂ of this species (Figures 2b and 3b).

In Alouatta pigra, the signal for HSA3 was observed in 16q and 19q, while that for HSA21 hybridized in 4p_ter, thus indicating that the HSA3/21 synteny is not present in Alouatta pigra (Figures 2c and 3c). The probe for HSA15 hybridized in Alouatta pigra metaphases in 4q_ter, 22q_ter and 26q, showing that the HSA3/15 syntenic association is also absent. None of these Alouatta pigra chromosomes is involved in the sex chromosome system of this species (Figures 2d and 3c).

In Alouatta caraya, the signal for HSA21 was observed in 21q_ter, whereas that for HSA3 was observed in 21q_prox, thus confirming the conservation of the HSA3/21 synteny. HSA15 hybridized in 7p (X₂ in males) and Y_1ter, and HSA3 in 7q and Y_1prox, exhibiting the HSA3/15 syntenic association in the sex chromosome system X₁X₁X₂X₂/X₁X₂Y₁Y₂ (Figure 3a).

Alouatta palliata showed a pattern similar to that of Alouatta pigra (therefore Figure 2 illustrates only the latter). HSA3 hybridized in 16q and 18q, HSA21 hybridized in 1p_ter and HSA15 in 2q_ter, 21q_ter and 23q (Figure 3d). Both the HSA3/21 and HSA3/15 syntenic associations are absent in Alouatta palliata and chromosomes with homeology to HSA3 and HSA15 are also not involved in the sex chromosome system of this species.

The probe for the human X chromosome showed positive hybridization signal in X₁ of all the species analyzed. The probe for the human Y chromosome did not hybridize in any of the howler species (data not shown).

The data obtained from the G-banding pattern and FISH homeologies, together with cyt b sequences obtained from previous reports, were used as the basis to perform a cladistic analysis. The HSAY/7 association, corresponding to the Y-autosome translocation that gave rise to the multivalents observed in Alouatta pigra and Alouatta palliata, was added as an extra character to the original list (de Oliveira et al. 2002). The syntenic associations HSA4p_ter-q22/9/11 observed in chromosome 1 of Alouatta pigra and HSA15q21.3-q24/14/21q observed in chromosome arm 4q were not included in the analysis, because, as autopomorphies for Alouatta pigra, they are considered non-informative.

Three data matrices were obtained: one including only chromosomal data, another including only molecular data and the last one including both types of characters (chromosomal and molecular) in a single matrix (see Appendix 1).

Chromosomal partition: The analysis of chromosomal data resulted in 36 informative characters, 23 constant characters and 40 non-informative characters. After analyzing 704 trees, PAUP retained the two most parsimonious trees (Appendix 2: Figures Sa and Sb), both with a length of 87 (L = 87). The analysis using only the partition of chromosomal data did not resolve the node ((Alouatta palliata, Alouatta pigra), (Alouatta caraya, Alouatta belzebul), ((Alouatta guariba clamitans, Alouatta guariba guariba), (Alouatta macconnelli (Alouatta sara, Alouatta seniculus arctoidea))), since it was established in a polytomy (Appendix 2: Figure Sc).

Molecular partition: Heuristic analysis of cyt b gene sequences, made ​​from a total of 800 characters, produced 109 informative characters, 551 constant characters and 140 non-informative characters. After analyzing 916 trees, PAUP retained a single most parsimonious tree (Appendix 2: Figure Sd), with a length of L = 366. The analysis using only molecular data did not resolve the node (Alouatta sara, Alouatta macconnelli, Alouatta seniculus arctoidea, Alouatta caraya), which was established as a polytomy different from that described from chromosomal data.

Combined analysis: The heuristic analysis of the combined data showed a total of 899 characters, 145 of which were informative, 180 non-informative and 574 constant. After analyzing 684 trees, PAUP retained only one, with a length of L=460 (Figure 4). This type of analysis allowed us to solve all the nodes, resulting in a fully resolved tree.

The karyotypes of Alouatta pigra and Alouatta palliata share more syntenic associations with those of Alouatta caraya and Alouatta belzebul than with those of the “Alouatta seniculus group” (Alouatta seniculus arctoidea, Alouatta sara, Alouatta macconnelli, denominated as such because they were once all subspecies of Alouatta seniculus together with Alouatta seniculus seniculus Linnaeus, 1766, and Alouatta seniculus stramineus Hill, 1962). This supports the basal grouping of the Alouatta pigra-Alouatta palliata Mesoamerican clade and the basal grouping of Alouatta belzebul among South American howlers.

The chromosomal comparisons showed that Alouatta pigra and Alouatta palliata conserved the HSA8/18 and HSA14/15 syntenies, considered ancestral for Platyrrhini (Stanyon et al. 2008), as well as the HSA10/16/10/16 syntenic association, ancestral for Atelidae (de Oliveira et al. 2002), but lost the HSA3/21 synteny, ancestral for mammals (Müller et al. 2000, Müller 2006).

According to our combined phylogeny, the HSA2/20 and HSA5/7/5/7 syntenic associations, previously considered as synapomorphies of the Alouatta caraya-Alouatta belzebul group (de Oliveira et al. 2002), would be homoplasies (parallelism). The HSA16/4 syntenyc association would be ancestral for the genus and might either be absent in Alouatta pigra, Alouatta guariba and the Alouatta seniculus group or might constitute a parallelism among Alouatta palliata, Alouatta caraya and Alouatta belzebul.

Like the HSA3/21 synteny, the HSA3/15 syntenic association, involved in the sex chromosome systems in South American howlers, is not present in Mesoamerican ones. This syntenic association of human 3/15 chromosomal segments has been described in other Atelidae species such as Ateles geoffroyi Kuhl, 1820 and Ateles belzebul hibridus Geoffroy, 1829, although not associated with the sex chromosome system (Morescalchi et al. 1997), but not observed in Lagothrix Geoffroy, 1812, and Brachyteles Spix, 1823 (Stanyon et al. 2001, de Oliveira et al. 2005). This association has not been observed in other genera of Neotropical Primates such as Cebus libidinosus Spix, 1823, or Saimiri boliviensis boliviensis Geoffroy & Blainville, 1834 (Mudry et al. 2001). Therefore, the HSA3/15 syntenic association either could be interpreted as the ancestral condition for the family Atelidae, where the association with multiple sex chromosomes would be an evolutionary novelty (apomorphy) in howlers and the loss of the association a apomorphy for the Lagothrix and Brachyteles group, or could have appeared independently in Alouatta, involved in the Y-autosome translocation, and in Ateles, not involved in the sex chromosome system (de Oliveira et al. 2005). However, our results suggest that the HSA3/15 syntenic association is not an ancestral condition for Alouatta, since the most basal species (see Figure 4) Alouatta pigra andAlouatta palliata (this contribution) and Alouatta belzebul (Consigliere et al. 1996) do not possess this association.

Taking into consideration the data obtained, a hypothesis can be proposed regarding the origin of the sex chromosome systems in the genus. Within the family Atelidae, with the exception of Alouatta, all genera have an XX/XY sex chromosome system. Therefore, it can be considered that the Alouatta ancestor possessed a chromosomal sex determination XX/XY, prior to the biogeographic separation of Mesoamerican and South American groups (see below). After this separation, both groups independently acquired the multiple sex chromosome systems currently observed through independent Y-autosome translocations.

The sex chromosome system X₁X₁X₂X₂/X₁X₂Y would have arisen independently in the lineages of Meso and South American howlers by a Y-autosome translocation (Figure 5a). In males, two fissions, one in Yp_ter and another in q_prox of the autosomal pair involved (Aq_prox), followed by translocation of Aq_prox to Yq-p_prox, formed the new chromosome Y₁. The Yp_ter segment is lost and the proximal region of the fissioned autosome either is lost or, in certain howler species, could have given rise to microchromosomes (e.g.: Alouatta seniculus (Yunis et al. 1976, Lima and Seuánez 1991, Torres and Leibovici 2001), Alouatta sara (Minezawa et al. 1985) and Alouatta macconnelli (Lima et al. 1990)). The homologous autosomal pair involved in the translocation is the one now denominated X₂. In the case of South American howlers, the autosomal pair involved would share homeology with HSA3, whereas in the Mesoamerican species it would share homeology with HSA7.

From this X₁X₁X₂X₂/X₁X₂Y sex chromosome system, an X₁X₁X₂X₂/X₁X₂Y₁Y₂ system could have arisen from a new translocation (Figure 5b). Under this hypothesis, simultaneous breaks in X₂p_prox and q_prox of another autosome (A´q_prox), followed by the translocation of most of the A´q arm to X₂p_prox, gave rise to the new X₂ chromosome. The X₂p_prox acentric fragment is lost and the rest of the autosome (A´) either is lost or could have remained as a microchromosome in some howler species (see above). The chromosome homologous to the autosome in question (A`) became Y₂. In the case of South American howlers, the new autosomal pair involved in the sex chromosome system would share homeology with HSA15. A further translocation between Y₁ and Y₂ (Figure 5c) would explain the hybridization pattern of the segments with homeology to human chromosomes 3 and 15 observed in the sex chromosome systems X₁X₁X₂X₂/X₁X₂Y₁Y₂ in South American howlers (Figure 5d).

On the other hand, in the Mesoamerican species, the X₁X₁X₂X₂/X₁X₂Y₁Y₂ sex chromosome system could have arisen either as described in Figure 5b (with the autosomal pair involved sharing homeology with a human chromosome not yet identified by G-banding pattern) or by a fission in Y₁ that would have given rise to two chromosomes, the new Y₁ (containing the segment corresponding to the ancestral Y chromosome) and Y₂ (containing a portion of the autosomal pair with homeology to HSA7). This last hypothesis would require a centromeric activation in Y₂.

However, considering the observation of the independent origin of the multiple sex chromosome systems in these two groups of howlers, the possibility of an independent origin of the X₁X₁X₂X₂/X₁X₂Y and X₁X₁X₂X₂/X₁X₂Y₁Y₂ sex chromosome systems within the Meso and South American groups cannot be ruled out until further studies.

It can be considered that multiple sex chromosome systems would be an extremely rare phenomenon due to complication in meiosis. Extreme cases are platypus and echidna, with a large number of sex chromosomes (Bick and Jackson 1967, Renz et al. 2007). In primates, multiple sex chromosome systems are even more infrequent. Moreover, Alouatta would be the first case where an independent origin of multiple sex chromosome systems is described. In other taxa, such as Drosophila (Flores et al. 2008), Erythrinidae fishes (Cioffi et al. 2013) and mole-rats (Deuve et al. 2006), a few cases have been observed, but these descriptions are still scarce.

The chromosomal homeologies and FISH analysis were used to construct a data matrix for the phylogenetic analysis. For comparison purposes, independent phylogenetic reconstructions were performed with each type of partition (Appendix 2: Figure Sa, b, c and d), along with the combined analysis of the two datasets (Figure 4). The chromosome partition grouped Alouatta caraya and Alouatta belzebul as sister taxa, in agreement with that reported by de Oliveira et al. (2002), a relationship that was not observed in the other two analyses, which grouped Alouatta caraya with Alouatta sara, Alouatta seniculus arctoidea and Alouatta macconnelli (although in the case of the molecular partition this relationship constituted a polytomy). This last species arrangement was also proposed by Nascimento et al. (2005) and Cortés-Ortiz et al. (2003) using molecular characters. In our molecular data partition (Figure Sd), Alouatta belzebul was grouped with the clade of Alouatta guariba, in agreement with that reported by Bonvicino et al. (2001) and Cortés-Ortiz et al. (2003). The three types of analyses agreed to place Alouatta sara, Alouatta seniculus arctoidea and Alouatta macconnelli into a single group, although the molecular data partition did not resolve the relationships between them, as they formed a polytomy. The grouping of all those species (“Alouatta seniculus group”) was observed in all phylogenetic studies performed so far (see above). Another coincidence was that the Mesoamerican species were placed as a separate clade from other South American species and, as expected, the two subspecies of Alouatta guariba in one group. All sets of taxa analyzed in this new approach were solved without polytomies only with the combined analysis, demonstrating the usefulness of incorporating more than one source of data for a more accurate elucidation of the relationships among current taxa.

The grouping of South American species as a separate group of the Mesoamerican group coincides with previous phylogenetic analyses using only molecular characters (Cortés-Ortiz et al. 2003, Ellsworth and Hoelzer 2006) and with the hypothesis of monophyletic origin of the Mesoamerican howlers previously proposed by Smith (1970). Smith’s hypothesis holds that Mesoamerican howlers originated by an expansion of the geographic distribution of South American howlers after the formation of the Isthmus of Panama, estimated to be completed about 3 million years ago (Coates et al. 2004). However, other studies indicate that the rise of the isthmus was a process rather than an event (Knowlton and Weigt 1998), resulting in intermittent periods with connected and divided lands during the past 18 million years. Given this last fact, another hypothesis was postulated to explain the current geographic distribution of the species of the genus. Instead of a single colonization event, various founder events, either across the Isthmus of Panama during one of the periods in which the two Americas were connected or across islands in the Caribbean archipielago, could have occurred (Cortés-Ortiz et al. 2003, Ellsworth and Hoelzer 2006). Primate fossils have been found in Cuba and Jamaica, but the origin and relationships of these specimens with modern Platyrrhini are still under debate (Fleagle 1999, Gutiérrez Calvache and Jaimez Salgado 2007, Rosenberger et al. 2009, Cooke et al. 2011).

Independently of the biogeographic scenario under consideration, it is clear that the evolutionary history of Mesoamerican howlers is different from that of South American howlers, an assertion that would be supported by the evidence provided by our new data.

This contribution provides new useful information for the systematics of the genus Alouatta, while supporting the hypothesis of chromosomal evolution in primates as a speciogenic strategy. The combined analysis resolved the phylogenetic relationships between howler species of both American origins, as a first approach to the “Total Evidence” concept and towards clarifying the controversies related to the Taxonomy and Evolution of Ceboidea.