Intraspecific variation in Rhinophylla pumilio
Our G-, C-, and Ag-NOR banding analyses have shown two distinct karyotypes for specimens of Rhinophylla pumilio from localities ranging more than 1000 km. The differences between these karyotypes may be caused by a pericentric inversion in the chromosome pair 16 or, alternatively, an amplification of rDNA cistrons accompanied with a faint block of heterochromatin in Rhinophylla pumilio with FN=64 (Fig. 3a). This segment is coincident with CMA3 positive staining for NOR and DAPI positive to the heterochromatic block (Fig. 3b).
Comparative analysis of karyotypes from different geographic localities (Table 2) allows discussing the morphology and number of chromosomes. Since only data of conventional staining or karyotype formula were described in the literature we had to restrict our comparisons to number and basic morphology of chromosomes. In this way, specimens of Rhinophylla pumilio collected on the Marajó island and northeastern Pará (Fig. 1, triangles 1, 2, 3, 4, 5, and 6) in the left side of the Amazon basin on Pará and Amazonas (triangles 7, 8 and 12) and Bahia (triangle 14) have 2n=34 and FN=62. Meanwhile, the samples from western Pará (triangles 9, 10 and 11) and Mato Grosso (triangle 13) presented the same fundamental number as specimens collected from Suriname, with 2n=34, FN=64 (Honeycutt et al. 1980, Baker et al. 1981, square 1).
Karyotype with 2n=26 and FN=48 described by Toledo (1973) (Fig. 1, Bahia, square 4) was found only in 100 km from the collection site of our sample with 2n=34 and NF=62. Varella-Garcia et al. (1989) suggested that the chromosome differences between populations of Rhinophylla pumilio described by Toledo (1973) and Baker and Bleier (1971) would be enough to reach the reproductive isolation between them. Nevertheless, analysis of mithocondrial DNA did not reveal sufficient genetic distance (0,3%) between two specimens from Northeastern Brazil (Pernambuco and Bahia) (Ditchfield 2000). Such distance is commonly observed within a breeding population. A re-analysis of the chromosome data from Toledo (1973) showed a disagreement with respect to the small size of the X chromosome and discordant number of chromosomes in mitotic and meiotic cells.
Another cytogenetic study on specimens of Rhinophylla pumilio from Colombia described a karyotype with 2n=36 and FN=62, (Baker and Bleier 1971, Fig. 1, square 3), differing from populations with 2n=34 and FN=62 probably by a chromosome fusion/fission event. Bats with karyotypes 2n=34, FN=56 (Baker and Bickham 1980, square 2) and 2n=34, FN=64 (Honeycutt et al. 1980, Baker et al. 1981, square 1) could be probably found in sympatry on the territory of Suriname.
Intergeneric comparative analysis
Comparative analysis of chromosome banding patterns of Rhinophylla pumilio was undertaken with representatives of two other subfamilies of Phyllostomidae bats: Phyllostomus hastatus, Phyllostomus discolor, Mimon crenulatum (Phyllostominae) and Glossophaga soricina (Glossophaginae). Karyotypes of these species supposed to be ancestral for their respective subfamilies (Patton and Baker 1978, Baker and Bass 1979, Baker and Bickham 1980, Haiduk and Baker 1982, Baker et al. 1989) and karyotype of Rhinophylla pumilio with 2n=34 and FN=56 described by Baker and Bickham (1980) revealed several characters shared with the above mentioned species.
Comparative analysis revealed that there are an extensive number of conserved chromosomes shared among these species. However, Rhinophylla pumilio shared more characters with Phyllostominae species than Glossophaga soricina (Fig. 4b). Based on outgroup comparisons, Baker and Bickham (1980) proposed that the most primitive karyotype for the family Phyllostomidae is identical to that of Macrotus waterhousii Gray, 1843. This hypothesis together with the basal position of Mimon waterhousii in recent phylogenies (Baker et al. 2000, 2003b, Datzmann et al. 2010) allows to suppose the most basal nature of chromosome pairs 12 and 8q of Glossophaga soricina because they are homologous to the acrocentric element 22 and to short arm of the biarmed element 1/2 of Mimon waterhousii, respectively (in Baker and Bass 1979). However, we suggest that in the basal branch that led to peculiarity of chromosome pairs 11 and 12 of Phyllostomus hastatus, Phyllostomus discolor, Mimon crenulatum and Rhinophylla pumilio, the same chromosomes (12 and 8q of Glossophaga soricina) could be involved in a simple translocation from a segment on the long arm of pair 8 to short arm of the pair 12 of Glossophaga soricina. Alternatively, the same chromosomes would be synapomorphic in Glossophaga soricina, as well as in some species of the Glossophaginae subfamily, and symplesiomorphic in other species analyzed here.
Furthermore, other differences among karyotypes (Fig. 4b) are a pericentric inversion on pair 7 of Phyllostomus hastatus (Patton and Baker 1978) and a simple translocation involving the pairs 4 and 13 of this species as was observed by Pieczarka et al. (2005). Such events are symplesiomorphic in Glossophaga soricina, synapomorphic in Phyllostominae species and probably autoapomorphic in Rhinophylla pumilio (pair 15). Integration of data derived from multidirectional chromosome painting with chromosome probes of Carollia brevicaudaSchinz, 1821 and Phyllostomus hastatus on metaphase spreads of Glossophaga soricina and chromosome map using probes of human chromosomes in the last species (Volleth et al. 1999) have shown that the basal position of Glossophaga soricina is supported by the fact that the pair 6 of human chromosomes was not disrupted. This chromosome has been assumed to be disrupted and subsequently fused with chromosome 13 of the Phyllostominae group, whereas this small segment forms an independent pair 15 in Rhinophylla pumilio (unpublished data).
Another interesting problem in our comparative analysis is the pair 16 in Rhinophylla pumilio, which has two chromosomal traits similar to those observed within representatives of genus Phyllostomus Lacépède, 1799. The difference between the karyotypes of Phyllostomus hastatus and Phyllostomus discolor consists of a pericentric inversion of the pair 15 (Patton and Baker 1978, Rodrigues et al. 2000). This chromosome is biarmed in Phyllostomus discolor and acrocentric in Phyllostomus hastatus, Phyllostomus elongatus Geoffroy, 1810, Phyllostomus latifoliusThomas, 1901 and Phylloderma stenops Peters, 1865 (Baker 1979, Baker and Bickham 1980, Honeycutt et al. 1980, Santos et al. 2002). Rodrigues et al.(2000) suggested that the biarmed state of pair 15 of Phyllostomus discolor could be most basal, because it has been shared with Mimon crenulatum, considered the most basal for the genus, and because this chromosome seems to be the result of a fusion of two acrocentric chromosomes of Mimon waterhousii (Patton and Baker 1978). The other species of Phyllostomus along with Phyllostomus stenops form a clade supported by the acrocentric form of the pair 15. However, the three species analyzed in this work showed different forms of the biarmed pair 15 (16 in Rhinophylla pumilio). The short arm of Mimon crenulatum represents a block of heterochromatin followed by the NOR, whereas in Rhinophylla pumilio the NOR appears before the heterochromatin. On the other hand, in Glossophaga soricina the NOR is represented at the long arm near the centromeric region accompanied by a heterochromatic block. Figure 3 shows the pattern of G- C and NOR sequential staining of pair 15 (16 in Rhinophylla pumilio) as well as the pattern of A/T-G/C evidenced by double staining with fluorescence DAPI and CMA3. The more plausible explanation is that the biarmness appeared in different branches of Phyllostomidae bats by amplification of rDNA cistrons accompanied or not with addition of heterochromatin, and possibly with other types of rearrangements.
Baker et al. (1972) defined three morphological types (submetacentric, acrocentric and subtelocentric) for the 5th chromosome pair of Mimon crenulatum at localities encompassing a wide geographic distance (Trinidad, Peru and Colombia). In this work, we have collected two specimens geographically apart from sites studied by Baker et al. (1972). We have found similar morphological types but G-banding analysis revealed that the acrocentric chromosome belonged to the 5th pair and the subtelocentric – to the 6th pair. That means that this polymorphism is defined by two pairs of chromosomes instead of one as it was suggested earlier.
Among species of genus Carollia karyotypes are highly rearranged and after the reciprocal chromosome painting Pieczarka et al. (2005) found only two chromosomes conserved in toto between Carollia brevicauda (pairs 7 and 9) and Phyllostomus hastatus (pairs 11 and 14). This finding suggests that they represent probably a part of the ancestral karyotype of Phyllostomidae, since they are preserved in such phylogenetically remote species. In the genus Rhinophylla these shared chromosomes are also presented by pairs 11 and 14 and can be also observed in others species studied herein except for the 8th pair of Glossophaga soricina that is partially homologous to the 11th pair of Rhinophylla pumilio. Therefore an analysis of the chromosomes homology among other species, especially those closely related to the genus Carollia, will be necessary to corroborate the sister group relationships of the genus Carollia and Rhinophylla.
Finally, we believe that variation of karyotypes along the area of Rhinophylla pumilio is correlated with intraspecific variation where the karyomorphs would be derived from ancestral karyotype with 2n=34, FN=62, since this karyotype is similar to other close related species at the chromosome level. However, additional analyses will be necessary to elucidate the biogeographical patterns related to the chromosome variation in Rhinophylla pumilio.