Molecular data highlight hybridization in squirrel monkeys (Saimiri, Cebidae)

Abstract Hybridization has been reported increasingly frequently in recent years, fueling the debate on its role in the evolutionary history of species. Some studies have shown that hybridization is very common in captive New World primates, and hybrid offspring have phenotypes and physiological responses distinct from those of the "pure" parents, due to gene introgression. Here we used the TA15 Alu insertion to investigate hybridization in the genus Saimiri. Our results indicate the hybridization of Saimiri boliviensis peruviensis with S. sciureus macrodon, and S. b. boliviensis with S. ustus. Unexpectedly, some hybrids of both S. boliviensis peruviensis and S. b. boliviensis were homozygous for the absence of the insertion, which indicates that the hybrids were fertile.

Saimiri populations occupy ample geographic areas ( Figure 1), with many potential zones of contact that pro-vide opportunities for hybridization between neighboring taxa (Hershkovitz, 1984). Thorington Jr (1985) reported cases of hybridization between S. ustus and S. sciureus on the east bank of the Tapajós River. Silva et al. (1992) investigated 49 specimens from a region in Peru occupied by both S. b. peruviensis and S. s. macrodon. By analyzing biochemical markers, these authors found clear evidence of admixture in approximately 45% of the individuals. Costello et al. (1993) also reported hybrids between S. ustus and S. sciureus from a region between the Madeira and Tapajós rivers.
Natural hybridization is the subject of a great deal of debate due to its potential importance as an evolutionary mechanism, especially for speciation, in addition to its relevance for taxonomy, conservation and species extinction (Mallet, 2005(Mallet, , 2007Genovart, 2009;). Hybridization is known to have played a role in the evolutionary history of at least one quarter of plants and 10% of animal species (Rieseberg, 1997;Seehausen, 2004). Arnold and Meyer (2006) concluded that reticulate evolution is a common process in the evolutionary history of animals, with numerous examples of the formation of new taxa as a consequence of introgressive hybridization. In primates, this phenomenon has been reported in both captivity and the natural environment (Schreiber et al., 1998;Zinner et al., 2009;Matauschek et al., 2011). However, the exact role of hybridization in the evolutionary history of an organism is usually unclear, and reticulate evolution represents a potential pitfall for phylogenetic reconstructions. Arnold and Meyer (2006) suggested that the accuracy of some phylogenetic constructs of New World monkeys is probably weakened by hybridization events that occurred in the past. While it is difficult to detect hybridization events, Osterholz et al. (2008) described the integration of an Alu element in S. boliviensis, which is absent in S. sciureus.
In the human genome, Alu elements are the most abundant transposable features (Kriegs et al., 2007), and these elements are now known to comprise approximately 10% of the primate genome (Batzer and Deininger, 2002;Zhang et al., 2002). Once inserted into the genome of a species during its evolutionary history, Alu insertions will be present in all the descendants of that species. An Alu insertion is thus a single and irreversible event (Hamdi et al., 1999;Shedlock and Okada, 2000;Salem et al., 2003), and represents a marker free of homoplasies. The present study investigated the potential occurrence of hybridization in free-living populations of S. boliviensis, based on the presence or absence of AluTA15, as described by Osterholz et al. (2008).
We examined 107 samples of Saimiri: two S. sciureus macrodon, 16 S. collinsi, 17 S. ustus, 22 S. boliviensis peruviensis and 50 S. b. boliviensis (Table 1). All the individuals sampled were born in the wild, although in some cases, the blood samples were collected in captivity. The samples of S. collinsi were collected from animals captured during the rescue operation of the UHE Tucurui hydroelectric reservoir in Para, Brazil (La Rovere and Mendes, 2000), and those of S. ustus at UHE Samuel, in Rondonia (Fearnside, 2005). The samples of S. b. boliviensis, S. b. peruviensis and S. s. macrodon were obtained from two captive facilities, the "Centro de Reproducción y Conservación de Primates No Humanos" (CRCP/IVITA) in Iquitos, Peru, and the "Centro Argentinode Primates" (CAPRIM) in Corrientes, Argentina. The species were identified based on the morphological characteristics described by Hershkovitz (1984). S. b. boliviensis has a white zone around the eyes exhibiting sparse white hairs and a flattened arch over the eyes (roman arch) while in S. s macrodon the arch formed above each eye is more evident and has been named as a "gothic arch".
While S. b. boliviensis and S. b. peruviensis have an arch that is less pronounced over the eyes (roman arch), S. b. peruviensis has a crown pattern on the head which is less eumelanized than that of S. b. boliviensis. The specimens held at CRCP/IVITA were classified as S. boliviensis peruviensis (roman arch) and those from the vicinity of Iquitos (Figure 1) as S. sciureus macrodon (gothic arch), 540 Carneiro et al.  . Some of the animals at CAPRIM were born in captivity. Further details on the specimens and the geographical distribution of each population are presented in Table 1 and Figure 1. The material analyzed in the present study was part of the sample bank maintained by the Molecular Phylogenetics Laboratory at the Bragança campus of the Federal University of Para. The total DNA was extracted using the Wizard Genomic kit (Promega, Madison, WI, USA) following the manufacturer's recommendations. The region of interest (AluTA15) was amplified using the primers and the protocol described by Osterholz et al. (2008). The initial denaturation step was 2 min at 94°C, followed by 40 cycles of denaturation (1 min at 94°C), annealing (1 min at 58°C), and extension (1 min at 72°C), with a final extension step of 5 min at 72°C. After amplification, the PCR products were separated electrophoretically in a 2% agarose gel at 60 V, 150 mA for 60 min together with a 1 kb plus DNA ladder (Invitrogen, Carsbad, CA, USA). All the fragments were stained with GelRed, as recommended by the manufacturer (Biotium, Hayward, CA, USA). Sequence reactions were conducted with a Big Dye v.3.1 kit (ABI BigDye® Terminator Mix; Applied Biosystems, Carlsbad, CA, USA), conducted in an ABI 3500xL sequencer (Applied Biosystems), to confirm that the region amplified by PCR was the fragment of interest (AluTA15). The sequences were aligned and edited manually in the BioEdit program (Hall, 1999).
The primers designed by Osterholz et al. (2008) amplify fragments of distinct sizes depending on the presence or absence of the Alu insertion (AluTA15). When the AluTA15 insertion is present, a fragment of approximately 750 base pairs (bps) is generated, but when it is absent, a fragment of only 450 bps is generated. As the insertion is only present in S. boliviensis (Figure 2), in hybrids between this species and other Saimiri species, two fragments will be amplified, one with 750 bps and another with 450 bps.
The AluTA15 insertion was not detected in any of the individuals identified as S. ustus (n=17) from Rondonia, S. collinsi (n=16) from Para or S. sciureus macrodon (n=2) from Peru. All 35 individuals presented only one band of approximately 450 bps (Table 2). By contrast, 50 specimens from Santa Cruz de La Sierra, Bolivia, identified as S. b. boliviensis, presented the insertion, of which 90% were homozygous (+/+) and 10% (five individuals) were heterozygous (+/-) showing both bands (750 bps and 450 bps). This configuration was unexpected because Osterholz et al. (2008) proposed that the AluTA15 element was inserted into the lineage that originated the extant species S. boliviensis, which implies that all S. boliviensis should be homozygous for AluTA15 (+/+). Interestingly, all three possible combinations were found in the population previously identified as S. b. peruviensis from Peru (CRCP), with six individuals (28%) being homozygous for the insertion (+/+), eight (36%) being homozygous for its absence (-/-), and the other eight being heterozygous (+/-), showing both bands (750/450 bps) in the gel (Figure 2). Osterholz et al. (2008) also found three possible patterns of bands (+/+; +/-; -/-) for specimens that were previously identified as S. b. peruviensis. So again, if the AluTA15 was inserted into the ancestral lineage of S. boliviensis, as proposed by Osterholz et al. (2008), it is unclear how specimens of this species could lack the insertion (-/-).
It is well known that Alu elements are replicated in a copy-and-paste way in the primate genome, and once inserted into a genome, they cannot be excised. Given this, individuals phenotypically typical of Saimiri b. peruviensis, but heterozygous for the insertion (+/-), must be the result of natural hybridization, which would presumably have involved the geographically closest taxon, S. sciureus macrodon. Furthermore, the absence of the insertion (-/-) in morphologically typical S. b. peruviensis can only be accounted for by the crossing of hybrid (+/-) Saimiri b. peruviensis or crosses between a hybrid and S. sciureus macrodon (-/-). These crosses would generate 25% or 50% of descendants without the insertion (-/-) and with dubious or intermediate morphological characteristics, which would represent conclusive evidence that hybridization between S. boliviensis and Saimiri sciureus macrodon produces fertile offspring. However, only 10% of the 50 Saimiri b. boliviensis specimens were heterozygous (+/-), and probably originated from crosses with Saimiri ustus, due to the proximity of the geographical distribution of these species (Figure 2). It is interesting to note that S. b. peruviensis and S. s. macrodon occur sympatrically in the region between the Marañón and Tapiche rivers in the Peruvian Amazonia, whereas S. b. boliviensis is parapatric with S. s. macrodon and S. ustus, which are separated by the Juruá and Purus-Guaporé Rivers, respectively (Hershkovitz, 1984). However, these rivers do not constitute an effective geographic barrier to gene flow in lizards (Souza et al., 2013), primates, and other organisms (Gascon et al., 2000), which implies that there may be gene flow between the present-day ranges of the three Saimiri species, resulting in hybridization between Saimiri boliviensis and Saimiri sciureus or S. ustus, as suggested by previous authors (Hershkovitz, 1984;Thorington Jr, 1985;Silva et al., 1992Silva et al., , 1993Osterholz et al., 2008) based on morphological data.
Using chromosomal data, Jones and Ma (1975) were able to distinguish between S. b. peruviensis and S. s. macrodon from the vicinity of Iquitos (Peru) and Leticia (Colombia), respectively. Both species revealed a diploid number of 2n=42, with 10 meta/submetacentric, 22 acrocentric and 10 telocentric chromosomes in S. b. 544 Carneiro et al.