Allopatric chromosomal variation in Nematocharax venustus Weitzman , Menezes & Britski , 1986 ( Actinopterygii : Characiformes ) based on mapping of repetitive sequences

Characiformes is the most cytogenetically studied group of freshwater Actinopterygii, but karyotypical data of several taxa remain unknown. This is the case of Nematocharax, regarded as a monotypic genus and characterized by marked sexual dimorphism. Therefore, we provide the first cytogenetic report of allopatric populations of Nematocharax venustus based on distinct methods of chromosomal banding and fluorescence in situ hybridization (FISH) with repetitive DNA probes (18S and 5S rDNA). The karyotype macrostructure was conserved in all specimens and populations, independently on sex, since they shared a diploid number (2n) of 50 chromosomes divided into 8m+26sm+14st+2a. The heterochromatin was mainly distributed at pericentromeric regions and base-specific fluorochrome staining revealed a single pair bearing GC-rich sites, coincident with nucleolar organizer regions (NORs). On the other hand, interpopulation variation in both number and position of repetitive sequences was observed, particularly in relation to 5S rDNA. Apparently, the short life cycles and restricted dispersal of small characins, such as N. venustus, might have favored the divergence of repetitive DNA among populations, indicating that this species might encompass populations with distinct evolutionary histories, which has important implications for conservation measures.


Introduction
The Neotropical freshwater ichthyofauna is remarkable by their richness, complexity and high endemism (Vari & Malabarba, 1998).A great majority of fish species in this region belongs to the family Characidae, which includes at least 1,107 valid species (Eschmeyer & Fong, 2015).However, the phylogenetic relationships and systematics of genera in this family are controversial, being several taxa placed as incertae sedis (Weitzman & Fink, 1983).In addition, the identification of species complexes or cryptic species hinders a precise characterization of the diversity in this group (Kavalco et al., 2009;Castro et al., 2014a).
Accordingly, cytogenetic analyses have been carried out to address this issue, since chromosomal markers have proved to be useful for both systematic and evolutionary inferences in Neotropical fish (Blanco et al., 2010;Almeida et al., 2013;Piscor et al., 2015), including taxonomically problematic groups (Bellafronte et al., 2010;Mendes et al., 2011).Moreover, the availability of refined cytogenetic methods have allowed identifying population polymorphism and unique evolutionary units, even within morphologically similar groups (Bitencourt et al., 2011;Utsunomia et al., 2014).
Nematocharax Weitzman, Menezes & Britski represents one of the several genera in Characidae whose taxonomic status is still under debate.This genus has been considered monotypic since the validation of Nematocharax venustus Weitzman, Menezes & Britski (1986), originally described for the Jequitinhonha River basin, eastern Brazil.Recently, Bragança et al. (2013) reported a second species, named N. costai, for the Contas River basin, northeastern Brazil, which was not widely accepted among ichthyologists.Hence, Menezes et al. (2015) carried out a detailed analysis of meristic, morphometric, osteological and coloration patterns of specimens from several hydrographic basins throughout the range of Nematocharax.According to these authors, N. costai is actually a junior synonym of N. venustus once overlapped morphological features and polymorphism of secondary sex traits in males were observed for both putative species, thereby supporting Nematocharax as a monotypic genus.
Based on the complex systematics and morphological variation of N. venustus, we performed comparative chromosomal analyses in allopatric populations of this species, including those previously recognized as N. costai in order to elucidate the taxonomic status of this fish group and infer biogeographic aspects in ichthyofauna from coastal basins of eastern Brazil.Therefore, distinct cytogenetic methodologies, ranging from basic banding techniques up to fluorescence in situ hybridization (FISH) experiments with repetitive DNA probes, were used to identify the karyotypic structure and possible population differences through the range of this species.

Material and Methods
A total of 71 males and 70 females of Nematocharax venustus were collected in seven localities along the Contas, Almada and Jequitinhonha River basins in the states of Bahia and Minas Gerais (Brazil) (Fig. 1).The data about collection sites and samples are shown in Table 1.Gongogi 1, 2, and 3), and Jequitinhonha River basins (Jequitinhonha 1 and 2) in the states of Bahia and Minas Gerais.
The specimens were transported to aerated tanks, and mitotic chromosomes were obtained from kidney cells according to Netto et al. (2007) and Blanco et al. (2012) after mitotic stimulation for 48 h (Lee & Elder, 1980).All experimental procedures and euthanasia of specimens were previously authorized by the Comitê de Ética em Uso de Animais da Universidade Estadual do Sudoeste da Bahia (CEUA/UESB, number 32/2013).
Two repetitive DNA sequences isolated from Hoplias malabaricus (Bloch, 1794) were used as probes in FISH.The first probe comprised the 5S rDNA, which included 120 base pairs (bp) of coding sequence and 200 bp of non-transcribed spacer (Martins et al., 2006).The second probe corresponded to a segment of 1,400 bp of 18S rDNA obtained by Polymerase Chain Reaction (PCR) (Cioffi et al., 2009a).Both probes were cloned in plasmidial vectors and propagated in competent cells of Escherichia coli DH5α (Invitrogen, San Diego, CA, USA).The 5S and 18S rDNA probes were labeled by nick translation with biotin-16-dUTP and digoxigenin-11-dUTP, respectively, following manufacturer's instructions (Roche, Mannheim, Germany).

Results
All specimens, independently on collection site or sex, shared a diploid number of 2n=50 with karyotypes composed of 8m+26sm+14st+2a (Fig. 2).The heterochromatin was preferentially distributed at pericentromeric regions of submetacentric and subtelocentric chromosomes in most populations (Fig. 3a).However, the samples from the Upper Contas River were characterized by the presence of C-bands at terminal regions of pairs 11 and 18 (Fig. 3b).Invariably, Ag-NORs in N. venustus were detected at terminal regions on short arms of a single pair.Yet, the NORbearing pair differed among populations, since the Ag-NORs in specimens from the Upper Contas River were interspersed with C-bands on pair 18 (st) while the remaining samples revealed active ribosomal regions on pair 8 (sm) (Fig. 4).Likewise, GC-rich regions (CMA 3 + and DAPI -) were observed in a single pair, being equivalent to Ag-NORs (Fig. 4).The FISH with ribosomal probes confirmed the occurrence of a single NOR system in most populations.The only exception refers to one sample in the Jequitinhonha River basin (named Jequitinhonha 2), which presented additional 18S rDNA sites on long arms in one homologous from pair 8 and on short arms of a single chromosomes from pair 10.This procedure was also informative in revealing four distribution patterns of 18S and 5S rDNA in N. venustus (Fig. 4).
The first pattern, shared by specimens from the Almada River and some populations from the Contas (Gongogi 1) and Jequitinhonha River basins (Jequitinhonha 1), includes 18S rRNA genes at terminal region of short arms of a sm pair (equivalent to Ag-NORs) and 5S rRNA genes at interstitial region on short arms of two pairs (16 -sm, and 21 -st).The population from Gongogi 3 differs from this pattern by presenting heteromorphic 5S rDNA cistrons  in pair 16, located at interstitial position on short arms of one chromosome and at pericentromeric region in the homologous (Fig. 4).The second pattern was similar that abovementioned, being observed in the population named as Gongogi 2. The only difference in this case refers to the presence of a single pair (16) bearing 5S rDNA at interstitial position on short arms (Fig. 4).
The population identified as Jequitinhonha 2 was characterized by a unique distribution of ribosomal genes, including bitelomeric 18S rDNA signals in one chromosome from pair 8.Moreover, this sample presented syntenic location of 18S and 5S rDNAs on short arms of one homologous from pair 10, being the latter closer to the centromeric region (Fig. 4).Another differential pattern was verified in the population from the Upper Contas River, since the specimens from this locality showed 18S rDNA sites on short arms of pair 18 while 5S rDNA regions were located interstitially on short arms of a single sm pair (16) (Fig. 4).The double-FISH metaphases for each different pattern observed, followed by the same metaphases stained with DAPI only, are shown in Figure 5.
In the case of taxa with similar chromosomal number and morphology, the analysis of heterochromatin distribution might be helpful to provide cytotaxonomic or population markers, as previously shown in other Characiformes (Galetti Júnior et al., 1991;Mantovani et al., 2000;Jacobina et al., 2009).However, the C-bands were restricted to pericentromeric region and NORs in most populations of N. venustus, following a common trend in fish (Imai, 1991;Aguilar & Galetti Júnior, 2008;Rosa et al., 2009).The only exception was the population from the Upper Contas River, which presented terminal heterochromatic segments in two pairs (Fig. 3b).This unique C-banding pattern indicates that this population is genetically divergent, as further discussed.On the other hand, microstructural interpopulation differences were observed by mapping the ribosomal genes in N. venustus, particularly in relation to the samples from Jequitinhonha 2 and Upper Contas.Even though most populations presented active rDNA sites interspersed with GC-rich sites (CMA 3 + ), a common feature in fish (e.g., Verma et al., 2011;Santos et al., 2012), additional 18S rDNA cistrons were identified by FISH in specimens from Jequitinhonha 2, characterizing a multiple NOR system.In the case of individuals from Upper Contas, the NORs were located on the first st pair while submetacentric NOR-bearing pairs were observed in the other populations of N. venustus (Fig. 4).Thus, the distribution of 18S rDNA cistrons apparently represents cytogenetic markers for the populations from the Contas River basin.
The remarkable variation of both position and number of ribosomal sites in characins and other Characiformes might reflect the polyphyletism of this fish group (Peres et al., 2008) and/or the association of rDNA with transposable elements (TE) that mobilize adjacent ribosomal sequences (Cioffi et al., 2010).Reinforcing the polymorphic nature of rRNA genes in Characidae, the mapping of 18S and 5S rDNA by FISH revealed clear interpopulation cytogenetic divergence in N. venustus.
Usually, the 18S and 5S rDNA sites in Neotropical fish are located on distinct chromosomal pairs (Galetti Júnior & Martins, 2004).This pattern was observed in most populations but Jequitinhonha 2, since double-FISH revealed syntenic 18S and 5S rDNA cistrons in a homologous from pair 10 (Fig. 4).However, this characteristic may be an intrapopulation variation found in Jequitinhonha 2, since the double-FISH was obtained for a single individual from this collection site.Other reports of synteny between both families of ribosomal genes are available in some genera of Characiformes, like Astyanax Baird & Girard (Castro et al., 2014b), Prochilodus Agassiz (Vicari et al., 2006), Triportheus Cope (Diniz et al., 2009), and Hoplias Gill (Cioffi et al., 2009b).It should be pointed out that active Ag-NORs in the population from Jequitinhonha 2 were observed only in pair 8. Nonetheless, syntenic ribosomal genes were detected by FISH in one chromosome from pair 10.The lack of hybridization signal in the other homologous could indicate differences in the number of copies for this gene between chromosomes of pair 10 possibly related to unequal exchanges.
In addition, the location of 18S rDNA sites at terminal chromosomal regions is widespread in teleosteans.Apparently, this behavior favors the rearrangement of NORs without interfering with other genetic linkage groups (Hanson et al., 1996), besides promoting the redistribution of ribosomal genes by the proximity of telomeres in the interphasic nuclei (Gornung, 2013).Such distribution of 18S rRNA genes in N. venustus could account for the putative transposition of some copies of this rDNA in one homologous of pair 8 in the population from Jequitinhonha 2, giving rise to the bitelomeric NORs (Fig. 4).Again, the presence of 18S rDNA sites in both telomeres have already been reported in other Characiformes characterized by accentuated chromosomal variation, such as A. fasciatus (Cuvier, 1819) (Fernandes et al., 2009), A. scabripinnis (Mantovani et al., 2005), A. hastatus Myers, 1928 (Kavalco et al., 2009), and H. malabaricus (Cioffi et al., 2009a).
Differently from 18S rDNA, the 5S rDNA sites in fish are usually located at interstitial and euchromatic chromosomal regions (Martins & Galetti Júnior, 2001), as reported in N. venustus.This condition seems to reduce the genomic dispersal and reorganization of 5S rRNA genes when compared to NORs (Martins & Wasko, 2004).Nonetheless, a possible heterozygous pericentric inversion, encompassing the 5S rDNA cluster, is suggested for the heteromorphic pair 16 in individuals from Gongogi 3 (Fig. 4), thereby explaining the differential position of 5S rDNA signals between homologous.Moreover, numerical variation of 5S rDNA signals was also detected in the present study.While most populations of N. venustus shared two pairs bearing 5S rRNA genes, as commonly reported in characins (Kavalco et al., 2011), the individuals from Gongogi 2 and Upper Contas were characterized by a single 5S rDNA-bearing pair.These data show the importance of refined methodologies of chromosomal mapping in revealing chromosomal rearrangements that otherwise would remain imperceptible, mainly due to the homogeneous macrokaryotypic structure in this species.
Along with the cytogenetic particularities, the specimens from the Upper Contas River, in Diamantina Plateau, are also morphologically distinctive once they present slender bodies and short rays in pelvic fins of males (data not shown), being likely to correspond to a new and undescribed species of Nematocharax.Reproductive isolation and speciation in the Upper Contas River are favored by the geological processes in the eastern region of Brazil, which includes the Diamantina Plateau and the Espinhaço Hills (Derby, 1906).In fact, divergent species associated to the isolation promoted by these biogeographic barriers have been reported even in animals of high dispersal potential, like birds (Vasconcelos et al., 2012).
Furthermore, karyotypic novelties can evolve independently within relatively short periods in populations of small characins, like N. venustus, because their restricted dispersal and short life cycles (Castro, 1999) lead to reduced gene flow and rapid establishment of chromosomal rearrangements (Centofante et al., 2006;Pazza & Kavalco, 2007).Therefore, isolated and small populations in headwaters, like the one from the Upper Contas River, are more susceptible to genetic drift effects, potentially leading to the fixation of unique chromosomal features at random.Similarly, the populations of N. venustus from the Jequitinhonha River basin are endangered due to the increased human activities (Menezes & Lima, 2008), which might have reduced their effective population sizes.Thus, inbreeding and genetic drift effects related to bottleneck events might have increased the probability of fixing distinct chromosomal variants (Hedrick, 1981).
In conclusion, the present results increase the karyotypic reports in characins from coastal basins in Eastern Atlantic, a region recognized by high levels of endemism combined with severe human impacts (Camelier & Zanata, 2014).The comparative cytogenetic analysis provided evidence of allopatric chromosomal variation in N. venustus along its range, suggesting the presence of populations with distinct evolutionary histories that should be further investigated using other genetic markers.Otherwise, species or structured populations can be lost because they lack appropriate conservation management plans.