Introduction

Sex determination mechanisms in vertebrates are varied and may involve environmental and genetics factors. Among the species with genetic sex determination, there are variations that go from those that have the implication of only one locus to those having a morphological sex chromosome system (^{Ezaz et al., 2006}).

Fish is the most interesting group to study sex chromosomes because they have a large diversity of systems, including XY and ZW and their derivatives, with several stages of differentiation (^{Devlin and Nagahama, 2002}). This evidences multiple origins of these systems in the group, even among closely related species (^{Ezaz et al., 2006}). Among the multiple sex chromosome systems registered in fish, the X₁X₁X₂X₂/X₁X₂Y is the most common (^{Devlin and Nagahama, 2002}; ^{Kitano and Peichel, 2012}). This type of system arises as a consequence of rearrangements involving autosomes and sex chromosomes or fission of a sex chromosome (^{White, 1973}; ^{Kitano and Peichel, 2012}).

Gymnotiformes is the order that includes the electric fishes of the Neotropical region. These species have a high karyotypic variability that is reflected in diploid numbers (2n), karyotypic formula (KF), number and distribution of the nucleolar organizer regions, occurrence of B chromosomes and sex chromosome systems (^{Almeida-Toledo et al., 1984}, ¹⁹⁹³, ^2000a,^b, ²⁰⁰¹, ²⁰⁰²; ^{Fernandes-Matioli et al., 1998}; ^{Fernandes et al., 2005}; ^{Silva and Margarido, 2005}; ^{Fonteles et al., 2008}; ^{Milhomem et al., 2007}, ²⁰⁰⁸, ^2012a,^b; ^{Silva et al., 2009}; ^{Cardoso et al., 2011}; ^{Scacchetti et al., 2011}; ^{Mendes et al., 2012}; ^{Silva et al., 2013}). The group presents simple and multiple sex chromosome systems and heterogamety of males and females (Table 1). This diversity evidences different origins and allows the identification of alternative mechanisms involved in the evolution of the sex chromosomes in this group (^{Henning et al., 2008}, 2011; ^{Almeida-Toledo et al., 2002}; ^{Silva et al., 2009}).

Brachyhypopomus Mago-Leccia 1994 is one the most diverse genera of Gymnotiformes (^{Crampton, 2011}), but cytogenetic data are known only for two populations of Brachyhypopomus gauderio from the Paraná river basin (^{Almeida-Toledo et al., 2000b}; ^{Mendes et al., 2012}). In the present work we describe the karyotypes of two species of electric fish of this genus, which have an X₁X₁X₂X₂/X₁X₂Y sex chromosome system. We also discuss the process that may be involved in the evolution of these systems in Gymnotiformes.

Material and Methods

We analyzed 23 specimens of Brachyhypopomus pinnicaudatus Hopkins, 1991 (10 males and 13 females) and 12 specimens of Brachyhypopomus n. sp. FLAV (an undescribed species; Crampton, pers. Comm.; 6 males and 6 females; “FLAV” indicates the first four letters of the specific epithet that the species will receive), both from Reserva de Desenvolvimento Sustentável Mamirauá (Amazonas state, Brazil). All samples were identified by Dr. William G. R. Crampton and their collection was authorized by IBAMA (Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis) permit 020/2005 (IBAMA Registration: 207419). The animals were deposited in the ichthyologyic collection of Museu Paraense Emílio Goeldi (MPEG 22740, MPEG 22743, MPEG 27107, MPEG 27108, MPEG 27109) and in the ichthyologyic collection of the Instituto de Desenvolvimento Sustentável Mamirauá (IDSMIctio000731, IDSMIctio000737, IDSMIctio000738, IDSMIctio000792). Eugenol was used to anesthetize the animals. Metaphase chromosomes were obtained following standard procedures (^{Bertollo et al., 1978}) and were analyzed after sequential C-banding (^{Sumner, 1972}) and classified according to ^{Guerra (1986)}. Meiotic chromosomes were obtained following standard procedures (^{Gross et al., 2009}).

Results

Males of Brachyhypopomus pinnicaudatus presented 2n = 41 (1m-sm/40st-a), and females had 2n = 42 (42st-a) (Figure 1A,C). Males and females Brachyhypopomus n. sp. FLAV showed 2n = 43 (1m-sm/42st-a) and 2n =44 (44st-a), respectively (Figure 2A,C). The reduced diploid number in males and the chromosomes morphologies strongly suggest a fusion between an autosome and the Y chromosome giving origin to an X₁X₁X₂X₂/X₁X₂Y sex chromosome system. In B. pinnicaudatus, the Y chromosome is metacentric and the X₁ and X₂ chromosomes are acrocentric with similar sizes. On the other hand, in B. n. sp. FLAV, the Y chromosome is submetacentric and the X₁ and X₂ chromosomes are acrocentric with different sizes.

Figure 1 Karyotypes of males (A, B) and females (C, D) of Brachyhypopomus pinnicaudatus after conventional staining (A, C) and C-banding (B, D). 

Figure 2 Karyotypes of males (A, B) and females (C, D) of Brachyhypopomus n. sp. FLAV after conventional staining (A, C) and C-banding (B, D). 

Constitutive heterochromatin was detected in the centromeric region of all chromosomes of males and females of the two species and in the terminal region of pair 15 in males and females of B. n. sp. FLAV (Figure 2B,D). Furthermore, small heterochromatic bands can be seen in the interstitial and distal regions of pair 1 in males and females of B. pinnicaudatus (Figure 1B,D).

Analysis of meiosis allowed the identification of 19 bivalents and 1 trivalent in diplotene cells of B. pinnicaudatus (Figure 3A) and of 20 bivalents and 1 trivalent in B. n. sp. FLAV (Figure 3B). The trivalents showed two types of association: chiasmatic between X₂-Yq and end-to-end in the PAR region of X₁-Yp (Figure 3A,B).

Figure 3 Meiotic chromosomes of Brachyhypopomus pinnicaudatus (A) and B. n. sp. FLAV (B) after conventional staining. Black arrows indicate the sexual trivalent. Boxes: red arrows indicate chiasmata; blue arrows indicate end-to-end pairing. 

Discussion

Our data indicate sex-related differences in the karyotypes of Brachyhypopomus pinnicaudatus and B. n. sp. FLAV. These differences result from the presence of a biarmed chromosome exclusively in the males’ karyotypes, which is part of a trivalent during meiosis. The fact that it shows chiasmatic and end-to-end associations suggests that a rearrangement involving a sex chromosome occurred. If the large chromosome had resulted from the rearrangement between two autosomes, only chiasmatic associations would be present (^{Ueno and Takai, 2008}). The association of these variations with the sex and the involvement of a pair of autosomes plus the sex chromosomes indicate that the two species have a X₁X₁X₂X₂/X₁X₂Y sex chromosome system.

An X₁X₁X₂X₂/X₁X₂Y system has also been previously identified in a population of B. pinnicaudatus from the Tietê river (^{Almeida-Toledo et al., 2000b}). This population was later identified as B. gauderio in a different report dealing with populations of B. gauderio from the upper Paraná river (^{Mendes et al., 2012}). However, in a report describing this species, the distribution area of B. gauderio did not include the region where the population of the Tietê river was found (^{Giora and Malabarba, 2009}). According to Campos-da-Paz (pers. comm.), this species was introduced into the upper part of the Paraná river (including the Tietê river) (^{Graça and Pavanelli, 2007}). Furthermore, Crampton (pers. comm.) analyzed samples of several localities and confirmed that the population of the Tietê river (^{Almeida-Toledo et al., 2000B}) is B. gauderio and that B. pinnicaudatus occurs only in the basins of the rivers of the Amazon and Guiana Shield, although B. gauderio is endemic to the Paraná-Paraguay system. Therefore, since the population from the Tietê river is not B. pinnicaudatus but B. gauderio, this is the first record of the X₁X₂Y sex chromosome system in B. pinnicaudatus, and also in B. n. sp. FLAV.

A review on the occurrence of X₁X₁X₂X₂/X₁X₂Y systems in fishes suggested that a fusion between an autosome and the ancestral Y was the major mechanism for its origin (^{Kitano and Peichel, 2012}). The same mechanism may explain the origin of the multiple sex chromosome systems identified in B. pinnicaudatus, B. n. sp. FLAV and B. gauderio (^{Almeida-Toledo et al., 2000b}; ^{Mendes et al., 2012}). In the ancestral karyotype with a simple XX/XY sex chromosome system, the X and Y chromosomes were morphologically undifferentiated (but likely divergent in genetic content) and paired only through the pseudoautosomal region (PAR) during meiosis I. A centric fusion between the Y chromosome and an acrocentric autosome would have originated the X₁X₁X₂X₂/X₁X₂Y system (Figure 4A), resulting in a meiotic trivalent involving the sex chromosomes and with two types of association: chiasmatic between X₂-Yq and end-to-end in the X₁-Yp PAR region (Figure 3 and 4B). The regions involved in the chiasmatic association are undifferentiated or poorly differentiated, similar to those found in autosomal bivalents (Figure 3). This suggests a recent origin for these multiple sex chromosome systems, differently from those that only present the end-to-end association, which would be more differentiated (^{Suzuki et al., 1988}).

Figure 4 Probable mechanism of evolution of the X₁X₂Y sex systems in Brachyhypopomus: an ancestral homomorphic XX/XY sex system (A) underwent a fusion between an autosome and the Y chromosome (B) originating a multiple X₁X₂Y system. 

A similar system was found in the spider M. ferruginea, in which an X₁X₂X₃X₄X₅Y system arose through a rearrangement between the homomorphic sex chromosome pair and an autosome. The constitutive heterochromatin pattern of the original sex chromosomes did not differ from those of autosomes and the multiple X chromosomes were generated by non-disjunctions of the sex pair. The structural differentiation of the newly formed X chromosomes was facilitated by their heterochromatinization, observed in the sex chromosomes bivalents during prophase I of females (^{Král, 2007}).

The occurrence of X₁X₁X₂X₂/X₁X₂Y systems in three species of the same genus suggests a common origin. However, whereas in B. pinnicaudatus and B. gauderio the neo-Y is metacentric and the X₁ and X₂ are acrocentric of similar size, in B. n. sp. FLAV, the neo-Y is submetacentric and the X₁ and X₂ are acrocentric of different sizes (Figure 5). Moreover, meiotic chromosome pairing allowed us to conclude that in B. n. sp. FLAV, the X₁ is smaller than X₂ because the neo-Y has a chiasmatic association with the larger X chromosome and only an end-to-end association with the smaller X. Meiotic chromosome pairing also confirmed that the X₁ and X₂ have similar sizes in B. pinnicaudatus. These findings suggest that the ancestral sex chromosomes (XY) and the autosomes that evolved to form the X₁X₁X₂X₂/X₁X₂Y systems in B. pinnicaudatus and B. gauderio had a similar size. On the other hand, in B. n. sp. FLAV, the sex chromosomes were small acrocentrics and the autosomes were large acrocentrics. These data indicate that B. pinnicaudatus and B. gauderio likely share a common ancestral sex chromosome system. This hypothesis is reinforced by the fact that they are considered sister-species (^{Crampton, 2011}) and also because they likely share the same karyotype (same 2n and KF). On the other hand, the sex chromosome system of B. n. sp. FLAV likely had an independent origin. Other differences in the karyotype of this species (2n and KF) result from an event of fusion/fission involving autosomes. This may suggest that this species is less related with B. pinnicaudatus and B. gauderio than these two are between themselves, which may reinforce the hypothesis of an independent origin for the B. n. sp. FLAV sex chromosome system. Independent origins of sex chromosome systems in related species are well documented in fish (^{Devlin and Nagahama, 2002}). In Gymnotiformes, several kinds of systems have been registered, and their disagreement with the phylogeny is consistent with this pattern (^{Silva and Margarido, 2005}; ^{Henning et al., 2010}). Despite the present evidence for the independent origins of X₁X₁X₂X₂/X₁X₂Y systems in Brachyhypopomus, we cannot discard the possibility of a common origin followed by alternative processes of differentiation.

Figure 5 Morphology of the sex chromosomes in the genus Brachyhypopomus. *Species analyzed by ^{Almeida-Toledo et al. (2000b)}. 

It is important to point out that the emergence of X₁X₁X₂X₂/X₁X₂Y systems in Gymnotiformes is more recurrent than that of other types of systems (Table 1). This suggests an advantage in the fusion between the primitive Y and an autosome. Theoretical models suggest that sex chromosomes evolve by suppression of recombination in regions that harbor sexually antagonistic genes (genes with alleles that confer differential fitness in males and females) tightly linked to the sex determination locus, and that this sexual antagonism may be the driving force behind this process (^{Charlesworth et al., 2005}; ^{van Doorn and Kirkpatrick, 2007}; ^{Kikuchi and Hamaguchi, 2013}). In proto-simple sex systems (XY, ZW), inversions play an important role in the suppression of recombination and chromosome differentiation (^{Charlesworth et al., 2005}). On an other hand, when the sexually antagonistic gene is unlinked to the sex chromosomes, other mechanisms such as translocations or fusions involving an autosome and the sex chromosomes may create a linkage between them (multiple sex chromosome systems) and suppress crossing-over in the regions around the breakpoint in heterozygous (^{Charlesworth et al., 2005}; ^{Kitano and Peichel, 2012}). Moreover, in this last situation, the generation of neo-sex chromosomes as a neo-Y may avoid the degeneration of the ancestral Y caused by the suppression of recombination that promotes the accumulation of deleterious mutations and the expansion of the non-recombining segment, which may cause haploinsufficiency (^{Volff et al., 2007}; ^{Blaser et al., 2012}; ^{Livernois et al., 2012}). The recurrent emergence of the X₁X₁X₂X₂/X₁X₂Y system may represent events of sex chromosomes turnover that are driven by the process explained above. Future analyses aiming to detect possible closely related species with different sex chromosome systems and to understand the relationship between these systems may help to elucidate this question.