Comparative cytogenetics of some marsupial species (Didelphimorphia, Didelphidae) from the Amazon basin

Abstract We investigated the karyotype of 18 didelphid species captured at 13 localities in the Brazilian Amazon, after conventional staining, C-banding, Ag-NOR and fluorescent in situ hybridization (FISH) using the 18S rDNA probe. Variations were found in the X chromosome, heterochromatin distribution and the 18S rDNA sequence. The main variation observed was in the position of the centromere in the X chromosome of Caluromys philander Linnaeus, 1758 and Marmosa murina Linnaeus, 1758. For both species, the X chromosome showed a geographical segregation in the pattern of variation between eastern and western Brazil, with a possible contact area in the central Amazon. C-banding on the X chromosome revealed two patterns for the species of Marmosops Matschie, 1916, apparently without geographic or specific relationships. The nucleolus organizer region (NOR) of all species was confirmed with the 18S rDNA probe, except on the Y chromosome of Monodelphis touan Shaw, 1800. The distribution of this marker varied only in the genus Marmosa Gray, 1821 [M. murina Thomas, 1905 and M. demerarae Thomas, 1905]. Considering that simple NORs are seen as a plesiomorphic character, we conclude that the species Marmosa spp. and Didelphis marsupialis Linnaeus, 1758 evolved independently to the multiple condition. By increasing the sample, using chromosomal banding, and FISH, we verified that marsupials present intra- and interspecific chromosomal variations, which suggests the occurrence of frequent chromosomal rearrangements in the evolution of this group. This observation contrasts with the chromosomal conservatism expected for didelphids.


Introduction
In the Americas, subclass Metatheria Huxley, 1880 is represented by the three marsupial orders: Didelphimorphia Gill, 1872, Paucituberculata Ameghino, 1894and Microbiotheria Ameghino, 1889. The largest of the three American orders is Didelphimorphia, which is represented by the family Didelphidae Gray, 1821, whose species are widely distributed throughout the continent. Didelphidae is the only marsupial group present in Brazil. Together with rodents, they make up an important part of the mammalian fauna of the Amazon region Jansa 2009, Wilson andReeder 2011). Currently, 14 genera and 39 species are recorded in the Amazon basin. Although moderate in terms of species richness, didelphids are abundant in the region (Brandão et al. 2015).
Historically, the first cytogenetic data on American marsupials were recorded by Jordan (1911; cited in Reig et al. 1977), on the spermatogenesis of Didelphis virginiana Kerr, 1792. Since then, our knowledge of cytogenetics of American and Australian marsupials has grown significantly. Hayman (1990) reported the karyotype of 178 species of American and Australian marsupials and Svartman (2008) reported 45 karyotypes for American marsupials.
Unlike other mammal orders, such as Rodentia Bowdich, 1821, marsupials show relatively little chromosomal variation (Nagamachi et al. 2015). Chromosomal stability in marsupials was first verified in conventional staining karyotypes that revealed the existence of three main diploid numbers in species from both continents: 14, 18 and 22 chromosomes.
Among all the metatherian families, Macropodidae Gray, 1821 (order Diprotodontia Owen, 1866) is the most diverse in diploid number, varying from 2n=10 to 32. While the American Didelphidae has only the three main diploid numbers, with the most frequent being 2n=14 (Reig et al. 1977, Hayman 1990, Palma and Yates 1996, Carvalho et al. 2002, which has been suggested as the ancestral diploid number of all marsupials (Reig et al. 1977, Westerman et al. 2010. Further comparisons using chromosome banding in American and Australian marsupial species revealed that chromosomal stability is verified not only on the diploid number but also on longitudinal banding patterns that show intense conservation on chromatids (Yonenaga-Yassuda et al. 1982, Rofe and Hayman 1985, Casartelli et al. 1986, Souza et al. 1990, Svartman and Vianna-Morgante 1999. Limited sampling effort has hampered the estimation of species richness in the Amazon, leaving large gaps in our knowledge of the mammalian fauna Emmons 1996, da Silva et al. 2001). Currently, of the 39 species of Amazonian marsupials (Brandão et al. 2015) only 17 have associated cytogenetic data (Nagamachi et al. 2015). However, considering the taxonomic instability of Amazonian marsupials, this representation might not be accurate, since new phylogenetic studies will probably change the current classification of several taxa. Furthermore, the earlier literature often lacks a connection between the karyotype of putative species and the analyzed specimens, making it difficult to verify a posteriori the taxonomic identification attributed to a given karyotype.
The number of taxa analyzed to date is also limited, and existing cytogenetic analyses have been usually restricted only to the diploid and fundamental numbers (Nagamachi et al. 2015). New advances in the taxonomic classification of Amazonian marsupials, complementary techniques of cytogenetic analysis (banding, in situ hibridization), and added sampling efforts (more specimens, new localities) are necessary to improve current knowledge on the cytogenetics of these animals.
In this study, we analyze the main morphological differences in the sex chromosomes and the C-band pattern of 18 didelphid species from the Brazilian Amazon. In addition, we describe for the first time karyotype for six species (Monodelphis touan, Monodelphis aff. adusta, Monodelphis sp., Marmosops impavidus, Marmosops bishopi and Marmosops pinheiroi) and discuss these patterns in a broader geographical context, including other regions of Brazil and South America.

Material and methods
We cytogenetically analyzed 111 individuals in 18 species and 8 didelphid genera, collected in 13 localities in the Amazon (Table 1 and ( ( The metaphases were obtained from bone marrow by in vivo method according to Ford and Harmerton (1956). Each animal received 1mL/100g weight of a 0,0125% colchicine solution for 30 minutes, the cells were exposed for 20 minutes to a 0,075M KCl solution, fixed 3:1 in methanol and acetic acid and stored at -20 °C. The C-band and Nucleolus Organizing Regions (NORs) patterns were determined according to the techniques described by Sumner (1972), and Howell and Black (1980), respectively. Chromosome pairing considered morphology in decreasing order of size and the chromosomes were classified as metacentric (m), submetacentric (sm), subtelocentric (st) and acrocentric (a) according to the ratio of chromosome arms and the position of the centromere, according to Patton (1967). 18S rDNA sequences were mapped by fluorescence in situ hybridization (FISH) according to Pinkel et al. (1986), whose probe was obtained by polymerase chain reaction (PCR) using the following primers designed by Gross et al. (2010): 18SF (5'-CCG CTT TGG TGA CTC TTG AT-3') e 18SR (5'-CCG AGG ACC TCA CTA AAC CA-3') and labeled with digoxigenin (DIG-Nick translation, ROCHE) or Biotin (Bio-Nick translation, ROCHE), following manufacturer's instructions.
Two C-band patterns were present in the X chromosome for species of Marmosops. In pattern 1, X was entirely heterochromatic except for a proximal band in the long arms ( Fig. 2g -box); in pattern 2, the heterochromatin was in the short arms and the centromere (Fig. 2g -box). Both patterns were present in M. parvidens and M. bishopi, while only pattern 1 was observed in M. cf. pakaraimae, M. impavidus and M. pinheiroi ( Table 2). The Y chromosome was entirely heterochromatic in all species.
In the species with 2n=18 chromosomes, the heterochromatin was centromeric in G. venusta (Fig. 3a-II NORs confirmed by FISH using the 18S rDNA probe were present in the short arms of pair 6 in all 2n=14 species and G. venusta (2n=18). In M. demerarae and M. murina sites were also detected in the terminal position of the long arms of pair 5 (Fig. 2, a-III e d-III). In M. emiliae (2n=18) the NOR was positioned on the short arms of pair 7 (Fig. 3d-III), and in M. touan in the X and Y chromosomes, although no 18S site was detected in Y ( Fig. 3c-III). Only Monodelphis brevicaudata exhibited multiple NORs (Fig. 3e-III), whose sites were in the terminal region of the long arms of pair 7 and the short arms of X and Y.
In D. marsupialis, both the 18S rDNA probe and silver were detected in three chromosome pairs. In two pairs, the sites were located in the terminal region of the long arms, while in one pair they were bitelomeric (Fig. 3g-III). However, in regards to activity, there was a variation of four to eight markings.

Discussion
In the last decade, advances in systematic and taxonomic studies of the family Didelphidae introduced changes in the taxonomy and nomenclature of several of its taxa (Jansa and Voss 2000, Voss and Jansa 2003, Voss and Jansa 2009, Rossi et al. 2010, Gutiérrez et al. 2010. We used the phylogenetic tree of Jansa and Voss (2014) to map the cytogenetic data of the 18 species we have analysed in order to gain an understanding of chromosome evolution in the group. This work represents the most updated phylogeny of the intergeneric relationships of didelphid marsupials, making our interpretation of the cytogenetic data more integrative than a mere consideration of chromosomal data, and more accurate in light of an independently generated phylogenetic hypothesis. The autosome structure observed here corroborates karyotypic conservation in the diploid number and chromosomal formula (NFa) as previously described in the didelphid species Didelphis marsupialis, Marmosa demerarae, Metachirus nudicaudatus, Monodelphis touan (previously named M. brevicaudata), Monodelphis aff. adusta (previously named as M. cf. emiliae) and for species of Marmosops (Reig et al. 1977, Yonenaga-Yassuda et al. 1982, Casartelli et al. 1986, Hayman 1990, Souza et al. 1990, Palma and Yates 1996, Svartman and Vianna-Morgante 1998, 1999, Carvalho et al. 2002. Although didelphids are generally considered to have a conserved karyotype, by comparing the karyotypes among different genera, it was possible to associate them with certain species due to the presence of diagnostic characters. For example, M. demerarae and M. murina differ in their FNa, morphology, and sex chromosome size. In species of the genus Monodelphis, morphological variation in chromosomes was restricted to pair 6, which grants an FNa varying between 30 (as observed in M. aff. adusta, M. touan and, M. brevicaudata) and 32 arms (Monodelphis sp.). However, the same does not occur for the genus Marmosops, in which the five species analysed, present a very similar chromosome macrostructure.
The genus Marmosa has a complex taxonomy and recently underwent great taxonomic changes, with all species of Micoureus, formerly treated as a separate genus, now considered as a subgenus of Marmosa. Considering the taxonomic instability in Didelphidae, with individuals being reclassified, and some complex of species being divided into two or more valid taxa, even purportedly karyotyped species may in fact have their karyotypes still unknown. Thus, our knowledge as to how many and which species among didelphids were karyotyped remains unstable. A revision of the literature for species with reported karyotypes is required. Souza et al. (2013) observed different forms of the X chromosome in Caluromys philander, and our data contribute to show their wide geographic distributions. The acrocentric X are found in northeastern and southeastern Brazil (Fig. 4, localities 16 and 17), as well as in central (Fig. 4, locality 6) and eastern Amazon (Fig. 4, localities 10, 11 and 12). While submetacentric form is located in Venezuela (Fig. 4, locality  14) and areas in the western, central and eastern Amazon (Fig. 4, localities 4, 6, 12 and 15) (Reig et al. 1977, Svartman and Vianna-Morgante 1999, Pereira et al. 2008. Interestingly, both homozygote and heterozygote females were recorded in central Amazonia near Manaus (Fig. 4, locality 6). It is not clear how often this condition is found in natural populations. Indeed, so far, the few heterozygous records for X might be related to the low number of captured and cytogenetically analyzed individuals.

X chromosome variations
Apparently, there is a likely geographical structure in the distribution of the morphological forms of the X chromosome in Marmosa murina, with the metacentric X so far found in the northern and western parts of its distribution, the submetacentric X prevailing in the Amazon basin of Brazil and the acrocentric forms prevailing in the other known localities in central and eastern Brazil (Fig. 5). According to Brito et al (2015), this species is currently under revision and is likely to be split into three species. It remains to be seen if there will be a correspondence between those species and the karyotypic forms depicted here.
Among the Amazonian marsupials analyzed here, the variation in centromere position and heterochromatin patterns of the X chromosome is noteworthy. Souza et al. (2013) suggested that pericentric inversions in the X chromosome of Caluromys  Pereira et al. 2008. m=metacentric;sm=submetacentric;a=acrocentric;d=dot-like. philander altered its morphology, and our results support their findings. In contrast, in M. murina, the different morphologies (m, sm, and a) of chromosome X might be due to centromeric shift without the presence of rearrangements. Such reorganization was already observed in other mammals and might be related to three main regions of the chromosome: subtelomeric, proximal and in the boundary between heterochromatin and euchromatin (Rocchi et al. 2012, Burrack andBerman 2012).

Heterochromatin distribution
We observed chromosomal conservatism in the heterochromatin pattern in eight didelphid species: (C. lanatus, G. venusta, D. marsupialis, M. touan, M. aff. adusta, M. emiliae, G. emiliae and M. nudicaudatus). C. philander presented heterochromatic pattern different from the heterochromatic distribution reported in the literature for this  Pagnozzi et al. 2002;(• 29) Espírito Santo State, Paresque et al. 2004. m=metacentric;sm=submetacentric;a=acrocentric. species (Souza et al. 1990, Souza et al. 2013. In Marmosops spp., the C-band patterns of the X chromosome are widespread throughout the Amazon basin, but are found in sympatry in the area between the confluences of the Negro-Purus and the Trombetas-Tapajós Rivers, forming pattern 1 to the west and pattern 2 to the east (Table 2). It remains to be seen if there is a correspondence between these patterns with possible cryptic species to be uncovered by broader molecular systematics and morphological studies of these taxa.
Thus, heterochromatin distribution patterns can serve as a cytotaxonomic character, as well as shedding light on chromosomal evolution and regulation of gene expression. However, our results demonstrate that, except for Marmosops spp., the other species under study presented little heterochromatin intraspecific variation, including the X chromosome. Thus, this character alone does not allow for distinguishing among  Hayman and Martin 1974, Reig et al. 1977, Pereira et al. 2008, Carvalho et al. 200222;25;24;26;27;28;29 14 22 5q;6p 5q;6p a/ d A Carvalho et al. 2002, Pagnozzi et al. 2002, Lima 2004, Paresque et al. 2004  didelphid populations, although heterochromatin distribution may be an effective character for distinguishing between certain species pairs. This is the case for M. demerarae and M. murina, with the former presenting larger centromeric heterochromatic blocks than the latter, and between C. philander and C. lanatus, both with 2n=14 and NF=24, but with distinct heterochromatic patterns.

Nucleolus organizer regions (NORs) and their evolution
The NOR in Didelphidae can be simple or multiple. According to Hsu et al. (1975), the single NOR would be an ancestral character in mammals, with subsequent rearrangements leading to multiple NORs in derived groups. The presence of NOR in sex chromosomes also could be considered a derived character since originally it was present in autosomes and ended up in the X chromosome due to rearrangements such as translocation or transposition.  (Svartman and Vianna-Morgante 1999, Merry et al. 1983, Carvalho et al. 2002. In M. touan and M. brevicaudata there are simple NORs on the X and Y chromosomes, a condition previously identified in Monodelphis domestica Wagner, 1842 (Merry et al. 1983, Pathak et al. 1993. Hsu et al. (1975) reported ribosomal genes in mammal sex chromosomes of the bat species Carollia castanea. These authors emphasize that NOR in the X chromosome can generate problems with dosage compensation in mammals.
In the Y chromosome of M. touan, FISH did not confirm the marking. This situation was verified in other organisms, where precipitation in the heterochromatic regions took place but could lead to an erroneous interpretation of the distribution of this marker (Schneider et al. 2012). Thus, the marking observed (Fig. 3c III) was not a ribosomal site but a heterochromatic block with silver affinity.
When mapping the NOR character on the phylogenetic tree of Jansa and Voss (2014, fig. 01) (not shown here), we verified that multiple NORs are distributed in two distinct lineages: the first in species of the genus Marmosa and the second in species with 22 chromosomes of the genera Didelphis and Philander Brisson, 1762. The mapping of the simple condition onto the phylogenetic tree depicts a wide distribution for this character, present at the base of the tree (Caluromys philander, C. lanatus, Glironia venusta) and in at least one or more species of the remaining major clades (Gracilinanus emiliae, Marmosops spp., Metachirus nudicaudatus, Monodelphis touan, Monodelphis kunsi, and Monodelphis dimidiata) (Souza et al. 1990, Palma and Yates 1996, Carvalho et al. 2002, Svartman and Vianna-Morgante 2003, Pereira et al. 2008, Souza et al. 2013). This distribution of NOR character on the didelphid phylogeny is thus congruent with the hypothesis advanced by Hsu et al. (1975) that the single NOR is an ancestral state.
When mapping the NOR character on the phylogenetic tree of Pavan et al. (2014) for the genus Monodelphis, we verified that M. emiliae, Monodelphis sp. and Monodelphis aff. adusta seem to have retained the plesiomorphic condition of a simple NOR. Conversely, this condition became variable in M. domestica and in the M. brevicaudata species complex, which in addition to the NOR identified in the autosomal pair 7, also presents NORs in both chromosomes of the sex pair, indicating a duplication of this site.
In M. murina, intraspecific geographic variation in NORs were detected. Specimens from the Purus River have multiple NORs, those collected in the state of Goiás have simple NOR in the short arms of pair 6 (Palma andYates 1996, Carvalho et al. 2002) and those from the state of Pernambuco present additional markings in the long arms of pairs 3 and 5 (Souza et al. 1990). Furthermore, both specimens from the Purus River differed from the others regarding sex chromosomes.
Our results indicate geographic variation in NORs for M. demerarae. Amazonian specimens analysed did not present ribosomal cistrons in the short arms of the fifth pair, as recorded for specimens from the Atlantic forest in the Rio Grande do Sul and São Paulo states of southern Brazil (Carvalho et al 2002, Svartman and Vianna-Morgante 2003, Svartman 2008. Several studies have shown that considerable genetic variation exists among referred populations of this taxon (Voss and Jansa 2003, Dias et al. 2010, Gutiérrez et al. 2010. Therefore, several nominal taxa previously considered synonyms are now treated as valid species. Currently, M. demerarae is considered to occur in south to northern and central Brazil, and to southern Bahia (Gardner 2008, Dias et al. 2010 and Marmosa paraguayana Tate 1931 occurs from northern border of Espírito Santo state, south to Rio Grande do Sul, and east to Misiones (Argentina), and eastern Paraguay (Gardner 2008). However, some authors consider it to go as far north as Pernambuco state in northeastern Brazil (Voss and Jansa 2003). Thus, considering the geographic dis-tribution of this taxon, the 18S rDNA data presented for locations in northern and eastern Brazil possibly belong to specimens of M. paraguayana. As such, this character would have a cytotaxonomic value, and rearrangements involving the ribosomal sites could be related to speciation events related to this sister-species pair (Gutiérrez et al. 2010).
In Didelphis marsupialis from several Amazonian sites, only NOR activity varied, as was already reported in specimens from the Atlantic forest (Yonenaga-Yassuda et al. 1982, Svartman andVianna-Morgante 2003). Dutrillaux (1979) suggested that a small sample size would be inadequate for the knowledge of species karyotypes. Heeding this admonition, we used a relatively large number of individuals for each species analysed to uncover a range of variations that most likely would not have been detected had we used fewer individuals per species. The use of integrative analyses and new methodologies, such as taxonomy, phylogeny, and molecular cytogenetics could improve our understanding of the significance of these chromosomic variations. However, for the Amazon region, a significant limitation for cytogenetic studies is still the restricted collection effort, the vast geographical extent of the region and the difficulty of access to remote areas.

Conclusion
The cytogenetic data presented here shows that didelphid marsupial karyotypes present intraspecific variation in the morphology of sex chromosomes and in chromosomic markers (C-band and NOR) and present some geographic variation in the distribution of these features for several species. Furthermore, there are many areas in the Amazon, including the transition zone between the Amazon and the Cerrado biomes, which do not have cytogenetic records for any didelphid species. This situation seriously undermines our understanding of the significance of the recorded variation, whether it is part of a continuous gradient, or whether it represents intraspecific gradations, or whether it is related to new lineages or cryptic species still uncovered. Thereby, despite the chromosomal stability related to diploid numbers and chromosomal formula in marsupials across continents, didelphids present some intra-and interspecific chromosomal variations, probably related to frequent chromosomal rearrangements. Additional systematic sampling and analyses will be required for a better understanding of the karyotypic evolution of this group.