The karyotype of three Brazilian Terrarana frogs (Amphibia, Anura) with evidence of a new Barycholos species

A recent substantial rearrangement of the 882 described eleutherodactyline frog species has considerably improved the understanding of their systematics. Nevertheless, many taxonomic aspects of the South American eleutherodactyline species remain unknown and require further investigation using morphological, cytogenetic and molecular approaches. In this work, the karyotypes of the Brazilian species Ischnocnema juipoca (Atibaia and Campos do Jordão, SP), Barycholos cf. ternetzi (Uberlândia, MG, and Porto Nacional, TO), and Pristimantis crepitans (Chapada dos Guimarães and São Vicente, MT) were analyzed using Giemsa staining, Ag-NOR labeling, and C-banding techniques. All individuals had a diploid number of 22 chromosomes, but the Fundamental Numbers were different among species. The herein described low chromosome number of Pristimantis crepitans is unique within this genus, suggesting that cytogenetically this species is not closely related either to its congeneric species or to Ischnocnema. In addition, karyotype differences, mainly in the NOR position, clearly distinguished the two Barycholos populations, besides indicating the existence of a so far undescribed species in this genus. A taxonomic review could clarify the systematic position of P. crepitans and verify the hypothetic new Barycholos species.

Based on DNA sequences from mitochondrial and nuclear genes of 344 species, Hedges et al. (2008) placed the 882 described species of Brachycephalidae into a new taxon, Terrarana, and classified them into four families, four subfamilies, 24 genera and 11 subgenera. Of those, two families, three subfamilies, six genera, and two subgenera were proposed and named as new taxa. The genera Brachycephalus and Ischnocnema remained in the Brachycephalidae family, a group restricted to the southeastern region of Brazil, as previously suggested by Heinicke et al. (2007). Pristimantis and Barycholos were allocated to the family Strabomantidae, subfamilies Strabomantinae e Holoadeninae, respectively, which are new taxa proposed by Hedges et al. (2008).
In the present work, we analyzed two samples of Ischnocnema juipoca, two of Barycholos cf. ternetzi and two of Pristimantis crepitans. This latter species was not yet studied by molecular techniques, being included in genus Pristimantis, family Strabomantidae, based only on its geographic distribution (Hedges et al., 2008). We aimed to increase the number of karyotyped Brazilian Terrarana species and further understand their taxonomy and evolutionary relatedness.  (Table 1). The sampling locations where the specimens were surveyed are displayed in Figure 1.

Specimens of
Mitotic chromosomes were obtained from suspensions of intestinal epithelium and testicular cells from animals pre-treated with 2% colchicine for at least 4 h, as described by King and Rofe (1976) and Schmid (1978). Conventional chromosome staining was performed with 10% Giemsa solution, Ag-NOR labeling (Howell and Black, 1980), and C-banding (Sumner, 1972), as modified by Siqueira et al. (2008). The slides were examined with a BX60 Olympus microscope and images were captured using the Image Pro-plus 4.5.1 and QCapture 2.81.0 softwares. Chromosomes were measured and classified according to Green and Sessions (1991). Siqueira et al. 471

Results
In all analyzed individuals, the diploid number was 22 chromosomes, but the Fundamental Numbers (FN) discriminated the three species. The FNs were determined as 40 in I. juipoca, 38 in Barycholos cf. ternetzi, and 44 in Pristimantis crepitans (Figures 2-5).

Ischnocnema juipoca
The karyotypes of the two analyzed I. juipoca populations consisted of five pairs of metacentric chromosomes (1, 6, 7, 9 and 10), four pairs of submetacentrics (2, 3, 4 and 5) and two pairs of telocentrics (8 and 11) (Figures 2A-C). Secondary constrictions were present on the short arm of pair 3 and occasionally adjacent to the centromere of pair 11 ( Figure 2A). Blocks of heterochromatin were detected in the centromeric region of all chromosomes, and in several metaphases there was a faint C-band adjacent to the centromere of pair 11, coinciding with the secondary constriction, and on the telomere of the short arm of pair 2 as well (Figure 2B). In the telocentric pair 11, the Ag-NOR sites were adjacent to the centromere and coincided with the secondary constriction and the pericentromeric block of heterochromatin (Figures 2A-C).
Unusual size variation of the telocentric pairs was observed, both among metaphases of the same specimen and among different specimens. This size variation was probably due to differences in chromosome compaction and has hampered the positioning of these chromosomes in the karyogram, since they could be placed in any position among the last four pairs. Size variation of the telocentric chromosomes was also observed in the Barycholos populations.

Barycholos cf. ternetzi (Uberlândia, MG)
The karyotype of Barycholos cf. ternetzi consisted of six metacentric (1, 2, 4, 6, 7 and 9), one submetacentric (pair 3), one subtelocentric (pair 5) and three telocentric (8, 10 and 11) pairs ( Figures 3A-E). In several metaphases, secondary constrictions were found on the telomere of the NOR-carrying telocentric chromosome 8 ( Figures 3A-E). Three distinct NOR patterns were detected: (1) in three specimens, three Ag-NOR sites were observed on the 472 Cytogenetics of Brazilian Terrarana frogs  telomere of one pair 7 homologue, and two on one pair 8 homologue, one on the centromere and the other on the telomere ( Figure 3C); (2) in four specimens, three Ag-NOR sites were found, two on the telomeres of pair 7 and one adjacent to the centromere of one chromosome of pair 8 (Figure 3D); and (3) in two specimens, four Ag-NORs were found, one on the telomere of one pair 6 homologue and one on the centromere of one pair 8 homologue, and two labels on the centromeres of pair 11 ( Figure 3E). Heterochromatic blocks were detected in the centromeric region of all chromosomes, and a faint C-band was observed near the telomere on the long arm of pair 4 ( Figure 3B).

Pristimantis crepitans
The P. crepitans karyotype consisted of eigth pairs of metacentric (1, 2, 5, 6, 8, 9, 10 and 11) and three of submetacentric chromosomes (3, 4 and 7). Interstitial secondary constrictions were observed on the long arms of pair 7, where the NOR sites were detected ( Figures 5A and C). Blocks of strongly stained heterochromatin were located in the centromeric region of all chromosomes ( Figure 5B).

Discussion
The diploid number of 22 chromosomes observed in I. juipoca, Barycholos cf. ternetzi and P. crepitans has also been described for 28 other Terrarana frog species. In this anuran group, P. crepitans is the only species with such a low chromosome number occurring in the midwest of Brazil. The species with 2n = 20 and 22 are typically distributed in southeastern and southern Brazil, while most Brazilian species of the northern and northeastern regions have 2n = 30 and 34 (Bogart, 1973;DeWeese, 1975;Siqueira et al., 2004, Siqueira et al., 2008. The chromosome morphology and C-banding patterns found in the karyotype of the Barycholos cf. ternetzi specimens from Uberlândia were very similar to those previously described in Gurinhatã specimens (Campos et al., 2008). These two sampling sites are located in the State of Minas Gerais, not very distant from each other. Some small karyotype differences may have resulted from the use of different chromosome preparation techniques and from the classification methods used for karyotype description. Moreover, in the Barycholos specimens from Uberlândia, Siqueira et al. 473  the only heterochromatic block detected on the long arm of pair 4 corresponded to a band on pair 4 of the C-banded karyotype of Barycholos ternetzi (Gurinhatã), as shown by Campos et al. (2008), although the authors did not explicitly mention this band. In contrast, the karyotype of Barycholos cf. ternetzi from Uberlândia showed variation in number and position of NORs, as also reported for B. ternetzi from Gurinhatã by Campos et al. (2008). These authors suggested that the fixed NOR must be that on pair 11. In the present work, we found three additional NOR patterns. Of these, only one was observed on pair 11, indicating that the principal NOR-bearing chromosome is still uncertain. The variation in number and position of NORlabeling in the two studied Barycholos populations, one by Campos et al. (2008) and the Barycholos cf. ternetzi presented herein, suggests the occurrence of chromosomal rearrangements involving mainly the telomeric regions. In anuran species, transposition of mobile genetic elements, ribosomal cistron amplification, and rDNA reinsertion errors during extra chromosomal amplification of ribosomal cistrons have been suggested as responsible for such NOR variations (Wiley et al., 1989;King et al., 1990;Foote et al., 1991;Schmid et al., 1995;Kaiser, 1996;Lourenço et al., 1998). The variation in NOR number and location may represent an incipient process of species differentiation, and a taxonomic review, including additional methodologies, could clarify if this is indeed a new taxon.
Regarding the Barycholos cf. ternetzi specimens from Porto Nacional, the Ag-NOR labeling was fixed in two chromosome pairs (10 and 11) and located in an interstitial position, differently from the telomeric position observed in the Barycholos cf. ternetzi specimens from Uberlândia. The four analyzed Barycholos cf. ternetzi specimens from Porto Nacional also differed from the B. ternetzi from Gurinhatã (Campos et al., 2008). Additionally, the specimens from Porto Nacional did not show the heterochromatic block on the long arm of pair 4.
In spite of the similar chromosome morphology presented by the two Barycholos cf. ternetzi populations studied, the karyotypic differences, mainly in NOR position and C-banding, indicate the possible existence of two species and call attention to the need of a taxonomic review of these populations. The NOR locations and heterochromatin pattern have been useful to distinguish among different species such as, for example, Scythrophrys (Lourenço et al., 2003a(Lourenço et al., , 2008, Paratelmatobius (Lourenço et al., 2003b(Lourenço et al., , 2008 and Pristimantis dundeei and Pristimantis. aff. dundeei (Siqueira et al., 2008). Moreover, since vocalization and chromosomal features of the topotypical population are unknown, none of the already studied populations can be assigned with certainty to the nominal Barycholos ternetzi.
The data obtained for I. juipoca from Atibaia and Campos do Jordão are in agreement with those described by Campos et al. (2008) for populations from Itatiba and Santa Branca, also in the State of São Paulo. The other karyotyped species of genera Ischnocnema and Brachycephalus, both within family Brachycephalidae, were substantially diverse from those described herein. Those other species, such as I. guentheri, I. parva and B. ephippium, have the same chromosome number as I. juipoca (2n = 22), but no telocentric chromosomes (Siqueira et al., 2004;Ananias et al., 2006), whereas I. holti and I. lactea present a diploid number of 20 chromosomes . In addition, in B. ephippium the NOR was located interstitially on the metacentric pair 8, while in I. juipoca it was on the telocentric pair 11.
Pristimantis crepitans was removed from the synonymy of Eleutherodactylus fenestratus, (Heyer and Muñoz, 1999), where it was originally placed (Lynch, 1980). Hedges et al. (2008) allocated P. crepitans to the P. peruvianus group, and J. M. Padial (pers. comm. to Hedges et al., 2008) indicated the possibility that both P. crepitans and P. dundeei might belong to the P. conspicillatus group. However, the low chromosome number and the ecological characteristics of P. crepitans and P. dundeei indicate that these species are not closely related to the other Pristimantis. The putative taxonomic position of the former within genus Pristimantis, as proposed by Heinicke et al. (2007) and Hedges et al. (2008), was based solely on its geographical distribution, since it was never included in any molecular analysis. The low chromosome number (2n = 22) of P. crepitans is commonly found in Ischnocnema and is highly divergent from other known Pristimantis karyotypes, which typically have high chromosome numbers (2n = 30 to 34). Unfortunately, the only other Pristimantis species with a low diploid number, P. altae with 2n = 26 (DeWeese, 1975), was not yet submitted to molecular analysis. On the other hand, a preliminary analysis of spermatozoa ultrastructure indicated great differences between P. crepitans and the other Pristimantis species, as well as the studied Ischnocnema species (S. Siqueira S., unpublished data). Pristimantis crepitans is also unique by living in open and xeric habitats in the Cerrado biome, among granitic or arenitic outcrops. These divergences strongly suggest a need of complementary molecular analysis to reassess the recently proposed allocation of P. crepitans to genus Pristimantis (Heinicke et al., 2007;Hedges et al., 2008). Therefore, further studies are necessary to clarify the systematic position of these species.

Concluding Remarks
Major contributions to the understanding of the molecular phylogeny of the South American eleutherodactyline species were recently brought by Frost et al. (2006), Heinicke et al. (2007) and Hedges et al. (2008). However, for Brazilian species there are many unresolved taxonomic aspects which require further investigation us-ing a combination of morphological, cytogenetic and molecular techniques. Based on previously reported molecular data (Frost et al., 2006), behavioral studies (Caramaschi and Pombal, 2001), and karyotypes (Campos et al., 2008) it seems conceivable that Barycholos and the already karyotyped Ischnocnema species are close relatives. However, the additional molecular studies of Hedges et al. (2008) indicate that Barycholos is phylogenetically distant from both Ischnocnema and Haddadus binotatus. Therefore, the observed chromosomal similarities might be symplesiomorphies or the result of convergence generated by chromosomal rearrangements, thus not substantiating a hypothesis of close evolutionary relationships between the species I. juipoca and Barycholos.
In the new systematic arrangement for "eleutherodactyline" frogs proposed by Heinicke et al. (2007) and Hedges et al. (2008), mostly based on molecular data, there are still indications of divergences in relation to the available karyological data. Most likely, the divergences are due to the lack of molecular analysis of many of the Brazilian karyotyped species, such as Pristimantis crepitans and P. altae. Further molecular and chromosome analyses of Terrarana frogs, including these divergent species, should provide a broader understanding of their evolutionary relatedness and systematic status.