Cytogenetic characterization of four species of the genus Hypostomus Lacépède, 1803 (Siluriformes, Loricariidae) with comments on its chromosomal diversity

Abstract Cytogenetic analyses were performed on fishes of the genus Hypostomus (Hypostomus ancistroides (Ihering, 1911), Hypostomus strigaticeps (Regan, 1908), Hypostomus regani (Ihering, 1905), and Hypostomus paulinus (Ihering, 1905)) from the seven tributaries of the Paranapanema River Basin (Brazil) by means of different staining techniques (C-, Ag-, CMA3- and DAPI-banding) and fluorescence in situ hybridization (FISH) to detect 18S rDNA sites. All species showed different diploid numbers: 2n=68 (10m+26sm+32st-a) in Hypostomus ancistroides, 2n=72 (10m+16sm+46st-a) in Hypostomus strigaticeps, 2n=72 (10m+18sm+44st-a) in Hypostomus regani and 2n=76 (6m+16sm+54st-a) in Hypostomus paulinus. Ag-staining and FISH revealed various numbers and locations of NORs in the group. NORs were usually located terminally on the subtelocentric/acrocentric chromosomes: on the long arm in Hypostomus strigaticeps (2 to 4) and Hypostomus paulinus (2); and on the short arm in Hypostomus ancistroides (2 to 8) and Hypostomus regani (2 to 4). Conspicuous differences in heterochromatin distribution and composition were found among the species, terminally located in some st-a chromosomes in Hypostomus ancistroides, Hypostomus strigaticeps, and Hypostomus paulinus, and interstitially dispersed in most st-a chromosomes, in Hypostomus regani. The fluorochrome staining indicated that different classes of GC and/or AT-rich repetitive DNA evolved in this group. Our results indicate that chromosomal rearrangements and heterochromatin base-pair composition were significant events during the course of differentiation of this group. These features emerge as an excellent cytotaxonomic marker, providing a better understanding of the evolutionary mechanisms underlying the chromosomal diversity in Hypostomus species.

Most species of this family have a wide distribution in Central and South America. They usually dwell in the rapids, but may be present in different aquatic habitats and in sand banks or rocky rivers. The species of Hypostominae are restricted to freshwater habitats, with the exception of Hypostomus watwata Hancock, 1828, which is a benthic species that lives in estuarine waters. Most of these animals have twilight habits and during daylight hours remain under stones or trunks of dead trees (Weber 2003).

Material and methods
Cytogenetic analysis was performed on a total of 148 specimens of four Hypostomus species collected at different sites of the Paranapanema River Basin (southern Brazil) ( Table 1). The specimens were deposited in the Museu de Zoologia of the Universidade Estadual de Londrina (MZUEL), Londrina, Paraná State, Brazil.
Chromosome banding. C-banding was performed according to Sumner (1972). The silver staining of the nucleolus organizer regions (Ag-NORs) was performed according to Howell and Black (1980). The GC-and AT-rich bands were detected by staining with Chromomycin A 3 (CMA 3 ) and 4'6-diamidin-2-phenylindole (DAPI), respectively, according to Schweizer (1980). The slides were stained with 0.5 mg/mL CMA 3 for 1 h, washed in distilled water and sequentially stained with 2 µg/mL DAPI for 15 min. Slides were mounted with a medium composed of glycerol/McIlvaine buffer (pH 7.0) 1:1 supplemented with 2.5 mM MgCl 2 .
Images were acquired with Leica DM 4500 B microscope equipped with a DFC 300FX camera and Leica IM50 4.0 software.

Specimens of
Hypostomus ancistroides showed a diploid number 2n=68 and a fundamental number (FN) of 104, with a karyotype formula of 10m+26sm+32st-a. One chromosome of pair 26 showed size heteromorphism (Fig. 1a). Silver nitrate staining (Fig. 1a left box) and FISH (Fig. 1a right box) revealed up to four pairs of subtelocentric/acrocentric NOR-bearing chromosomes. CMA 3 marked the terminal region of the long arms of pair 26, the pericentromeric region of the second pair of metacentric chromosomes, and probably the NOR-bearing chromosomes (Fig. 2a). No fluorescent staining was observed after DAPI staining (Fig. 2b). Heterochromatin was distributed in the pericentromeric region of the second pair (m) of the complement and in the terminal region of the long arm (pair 26) (Fig. 3a).
Hypostomus strigaticeps presented a diploid number 2n=72 and a FN of 98, with a karyotype formula of 10m+16sm+46st-a (Fig. 1b). The Ag-NOR site numbers ranged from two to four marked chromosomes (st-a) located in the terminal region of the long arm (pairs 18 and 28) (Fig. 1b left box), similar to the number observed in FISH (Fig. 1b right box). CMA 3 marked four chromosomes, possibly the Ag-NOR sites, and the pericentromeric regions of most subtelocentric/acrocentric chromosomes (Fig. 2c). Staining with DAPI revealed large blocks in the terminal regions of four-eight subtelocentric/acrocentric chromosomes (Fig. 2d). C-banding revealed the occurrence of heterochromatic blocks in the pericentromeric region of the third pair of metacentric chromosomes and of up to eight large blocks in the terminal regions of the long arms of subtelocentric/acrocentric chromosomes. In one of those chromosome pairs, the heterochromatic block was adjacent to the secondary constriction (Fig. 3b).
Hypostomus regani had 2n=72 with a karyotype formula of 10m+18sm+44sta and FN of 100 (Fig. 1c). Ag-NORs were located in the terminal position on the short arms of four subtelocentric/acrocentric chromosomes (pairs 26 and 27) ( Fig. 1c left box). The same number of NOR-bearing chromosomes was observed after FISH ( Fig. 1c right box) and CMA 3 -staining (Fig. 2e). Interstitial CMA 3 -negative blocks were observed in most of the subtelocentric/acrocentric chromosomes, which, in contrast, were positive after DAPI staining (Fig. 2f ). Heterochromatin was distributed in the interstitial region of most st-a chromosomes and in the pericentromeric region of one metacentric pair (Fig. 3c).
Hypostomus paulinus showed 2n=76, FN=98 and a karyotype formula of 6m+16sm+54st-a (Fig. 1d). NORs were located in the terminal position on the long arms of chromosome pair 16 (Fig. 1d left box), similar to the chromosomes observed in FISH (Fig. 1d right box). CMA 3 -banding marked up to eight chromosomes (st-a) with large GC-rich blocks, and one st-a pair, probably corresponding to NOR-bearing chromosomes, and in the pericentromeric region of the first (m) pair (Fig. 2g); after DAPI staining, eight fluorescent bands were observed (Fig. 2h). Heterochromatin was distributed in the pericentromeric region of the first pair of metacentric chromosomes, in the terminal region of the long arms of eight pairs of Figure 1. Karyotypes of a H. ancistroides b H. strigaticeps c H. regani d H. paulinus arranged from Giemsa-stained chromosomes. In the insets, partial karyotypes of the NOR-bearing chromosome pairs after Ag-staining (left) and FISH with 18S rDNA probe (right). Bar = 10 µm.  H. ancistroides a, b H. strigaticeps  c, d H. regani e, f H. paulinus g, h. The arrows indicate the NOR-bearing chromosomes. Bar = 10 µm. subtelocentric/acrocentric chromosomes, one of which was the NOR-bearing pair. In this pair, a heterochromatin block was located at the proximal portion of the secondary constriction, whereas three heterochromatin blocks, which occupied almost the entire long arm, were observed in a pair of subtelocentric/acrocentric chromosomes (pair 12) (Fig. 3d).

Discussion
All species differed with respect to their diploid chromosome number and/or karyotype, as follows: 2n=68 (10m+26sm+32st-a) in H. ancistroides (Fig. 1a), 2n=72 (10m+16sm+46sta) in H. strigaticeps (Fig. 1b), 2n=72 (10m+18sm+44st-a) in H. regani (Fig. 1c), and 2n=76 (6m+16sm+54st-a) in H. paulinus (Fig. 1d). This variability is consistent with the chromosomal data previously reported in the genus Hypostomus, which showed a wide variation in 2n (from 52 to 84) ( Table 1). The available cytogenetic studies showed that the species that possess the same 2n have different karyotypes. In the same way as the features observed in H. ancistroides (2n=68) but with different fundamental numbers (FN) among different populations, i.e. 106, 102 and 96 (Michele et al. 1977, Artoni and Bertollo 1996, Alves et al. 2006) and the characteristics found in H. regani, the cytogenetic analysis showed the same diploid number (2n=72) and a FN of 102 and 116 (Artoni and Bertollo 1996, Alves et al. 2006, Mendes Neto 2008, also differing from those analyzed herein (Table  1). On the other hand, studies conducted by Michele et al. (1977) in H. paulinus and H. strigaticeps showed differences in both 2n and FN. This difference may be ascribed to the existence of different cytotypes in these species, the occurrence of cryptic species, problems with the species identification or with chromosomal classification.
According to Artoni and Bertollo (2001), 2n=54 is considered as a basal condition for the family Loricariidae. In a phylogenetic study of Loricariidae using morphological data, the genus Hypostomus was considered the most derived (Armbruster 2004), representing a group with more derived karyotypic forms, consisting mostly of st-a chromosomes with a high diploid number. It seems that there was a divergent karyotypic evolution among the Hypostomus species; on the other hand, two main chromosome rearrangements appear in the evolution of the genus: i) an increase in the diploid number (2n) in several species, probably due to centric fissions and ii) the same 2n but with a difference in the karyotype formula, probably accounted by pericentric inversions.
The same variability found in 2n and in karyotypes was also detected in NORs. Our data showed different phenotypes among the Hypostomus species, observed after silver staining and FISH. All species showed Ag-NORs and 18S rDNA sites located in the terminal regions of st-a chromosomes, but with a significant variation in number and location among them. H. ancistroides showed up to 8 NOR sites, all located on the short arms (Fig. 1a left and right boxes, respectively). H. strigaticeps showed NORs on the long arms and H. regani, NORs located on the short arms, and both species with up to 4 sites ( Fig. 1b and 1c left and right boxes, respectively), and H. paulinus evidenced only two NOR-bearing chromosomes located on the long arm ( Fig. 1d left and right  boxes), which could be considered as species-specific characteristics.
The presence of one pair of NOR-bearing chromosomes, and also its interstitial location seems to be a widespread condition for Loricariidae fish, since this occurs among the Neoplecostominae and Hypoptopomatinae species (Alves 2000). However, in Hypostomini, the occurrence of multiple NORs and their location in the terminal position is most common, as observed here and recorded by other authors (Artoni and Bertollo 1996, Kavalco et al. 2005, Alves et al. 2006. But the exact location and number of ribosomal sites are confirmed only by the FISH technique. With regard to the genus Hypostomus, the available molecular cytogenetic data on the location of ribosomal genes are few and restricted to 18S rDNA sites of H. affinis (Kavalco et al. 2005). These data are very important to prompt more discussions about the evolution of ribosomal DNA in this group.
In the four species presently studied, the NORs were positive for CMA 3 staining (Fig. 2), a feature that has been conserved among all Neoteleostei (Ráb et al. 1999). In addition, some other chromosomal regions were also considered GC-rich in the four species, mainly in H. ancistroides (Fig. 2a) and H. paulinus (Fig. 2g). H. strigaticeps, H. regani, and H. paulinus (Fig. 2d, f, h respectively) are three species that also showed several positive markers for DAPI staining, indicating AT-rich regions that were not found in H. ancistroides (Fig. 2b).
Some other studies carried out in Hypostomus (Artoni and Bertollo 1999, Kavalco et al. 2004, Cereali et al. 2008) also showed that this fish group may possess different classes of GC and/or AT-rich repetitive DNA families, as observed in the species analyzed in the present report. AT-rich regions are also rare among fishes, and have been reported mainly in some Hypostomini species (Artoni and Bertollo 1999, Kavalco et al. 2004, Rubert et al. 2008, some zebrafish species (Gornung et al. 1997, Phillips andReed 2000), and gobiid fishes (Canapa et al. 2002).
The chromosome banding performed in all species analyzed showed a variation in the heterochromatin distribution pattern. However, the presence of heterochromatin in some chromosomes was constant, as observed in the pericentromeric region of a metacentric pair in H. ancistroides (pair 2), H. strigaticeps (pair 3), and H. paulinus (pair 1) (Fig. 3a, b, d, respectively), also reported in H. nigromaculatus by Rubert et al. (2008). An additional characteristic is the presence of some conspicuous blocks in the terminal regions of some st-a chromosomes of the karyotype. The same banding profile, organized in blocks, was also observed by others researchers: in Hypostomus sp. B from the Mogi Guaçu River (Artoni and Bertollo 1999), H. affinis (Kavalco et al. 2004), H. cochliodon (Cereali et al. 2008), and H. nigromaculatus (Rubert et al. 2008). Interestingly, in H. paulinus, pair 12 proved to be well differentiated, with the long arm almost entirely heterochromatic, a feature observed only in this species. On the other hand, H. regani showed a more distinct heterochromatin distribution in relation to the other species, with a preferential location in the interstitial regions of st-a chromosomes (Fig. 3c).
The presence of a marker chromosome that seems conserved for most Hypostomus species, corresponding to the NOR-bearing chromosome pair, which shows a heterochromatin block adjacent to this site (e.g. Artoni and Bertollo 1999, Kavalco et al.  2004, Rubert et al. 2008), was also observed. It can be inferred from all data on the heterochromatin composition and distribution that each species has its own peculiarities, i.e., each species has a unique banding pattern.
Karyotypes, banding patterns, number and location of ribosomal DNA sites, and repetitive DNA are important tools for the cytotaxonomy of Hypostomus species. Since these characteristics do not vary among the different populations of the same species, they are significant cytogenetic markers at the species level.
Further data on other Hypostomus species from different rivers, as well as detailed studies of satellite DNA sequences may clarify important issues of genome organization, be used as genetic markers, and provide interesting insights for the comprehension of the evolution of this genus.