Advances in cytogenetics of Brazilian rodents: cytotaxonomy, chromosome evolution and new karyotypic data

Abstract Rodents constitute one of the most diversified mammalian orders. Due to the morphological similarity in many of the groups, their taxonomy is controversial. Karyotype information proved to be an important tool for distinguishing some species because some of them are species-specific. Additionally, rodents can be an excellent model for chromosome evolution studies since many rearrangements have been described in this group.This work brings a review of cytogenetic data of Brazilian rodents, with information about diploid and fundamental numbers, polymorphisms, and geographical distribution. We point out that, even with the recent efforts on cytogenetic studies in this group, many species lack karyotypic data. Moreover, we describe for the first time the karyotype of Carterodon sulcidens (Lund, 1838) (Family Echimyidae), a new fundamental number for an undescribed species of Neacomys Thomas, 1900 (Family Cricetidae, Subfamily Sigmodontinae), and illustrate the karyotype of a Brazilian specimen of Mus musculus Linnaeus, 1758 (Family Muridae). This review compiles the cytogenetic data on Brazilian rodents reported in the last three decades, after the last revision published in 1984, including synonyms, chromosomal variations, and geographic distribution. Additionally, it also reinforces that Brazilian biodiversity is still poorly known, considering the new data reported here.


Introduction
More than three decades after the last revision of cytogenetics of Brazilian rodents , in which the karyotypes of approximately 60 species were reported, several new karyotypes and chromosomal rearrangements have been described. In the last 30 years, huge progress has been made, and up to this date, new species have frequently been described. However, as we shall explore herein, there still remain gaps in knowledge about many species.
Cytogenetic information on Brazilian rodents was firstly described by Cestari and Imada (1968) for the species referred to as Akodon arviculoides cursor Thomas, 1913. From then on, cytogenetic data confirmed the great chromosomal variability in rodents, especially after the advent of banding techniques in the beginning of the 1970s.
Throughout the following decades, several Master dissertations and PhD theses have addressed cytogenetic studies on Brazilian rodents. It became evident that karyotypic data could contribute to accurate taxonomic information, since different names were applied to groups that shared the same karyotype, and very distinct karyotypes were attributed to a single species. Additionally, major fieldwork efforts in Brazil (especially in unexplored areas) have led to the discovery of many new species.
The increasing number of cytogenetic studies on rodents resulted in the characterization of banding patterns, recognition of sex chromosomes, identification of supernumerary chromosomes, pericentric inversions and Robertsonian rearrangements, variations in the amount and localization of constitutive heterochromatin, and recognition of species (cytotaxonomy). These discoveries have led researchers to consider that rodents have undergone a "karyotypic explosion" process and that they stand out as an excellent group for chromosomal evolution studies, since they present many examples of chromosome rearrangements. These rearrangements may have played an important role in karyotype diversification and speciation, with the reduction of gene flow due to meiotic problems (King 1993, Rieseberg 2001, Patton 2004, Faria and Navarro 2010. Previously, chromosome evolution studies were essentially based on the comparison of banding patterns (Yonenaga-Yassuda et al. 1975, 1987a, Leal-Mesquita et al. 1992, Silva and Yonenaga-Yassuda 1999. Later, the association of cytogenetics with molecular biology allowed for a new important approach for studying karyotype evolution. Notwithstanding, molecular cytogenetics allows the localization of specific DNA sequences in the chromosomes based on DNA denaturation and its subsequent annealing with complementary sequences. In Brazilian rodents, localization of specific sequences using fluorescence in situ hybridization (FISH) was specifically applied in the Akodontini and Oryzomyini tribes of the Family Cricetidae, Subfamily Sigmodontinae, which is traditionally divided into 10 tribes and one incertae sedis group (Pardiñas et al. 2015a). Nevertheless, this kind of approach is still lacking for the other tribes of Sigmodontinae, and the remaining rodent families, mainly because of the difficulty in obtaining specific probes.
FISH was first performed using telomeric sequence probes, revealing that, besides the telomeric position itself, the sequences could also be detected at telomeric interstitial sites (ITS), such as those present in the Sigmodontinae genus Akodon Meyen, 1833, Thaptomys Thomas, 1916, and Cerradomys Weksler, Percequillo & Voss, 2006(Fagundes et al. 1997a, Silva and Yonenaga-Yassuda 1998a, Andrades-Miranda et al. 2002a, Ventura et al. 2004, 2006. These ITS were correlated with components of constitutive heterochromatin, amplification of TTAGGG n sequences, telomeres remnants after chromosomal rearrangements or reservoirs for future fission rearrangements. On the other hand, the absence of ITS in other Sigmodontinae species with chromosome polymorphisms, such as Oligoryzomys Bangs, 1900, andRhipidomys Tschudi, 1845, was also described Yonenaga-Yassuda 1997, 1999).
More recently, probes from entire chromosomes were obtained by microdissection or flow sorting, representing a breakthrough in evolutionary studies. The first Brazilian study employing this technique was published by Fagundes et al. (1997b), in which the largest pair (pair 1) of the karyotype of the rodent Akodon cursor (Winge, 1887) (Subfamily Sigmodontinae, tribe Akodontini) was obtained in order to investigate regions of homology between chromosomes of this species and Akodon montensis Thomas, 1913. More than one decade later, Hass et al. (2008), using Mus musculus commercial chromosome probes, established chromosomal homology maps between five species of the tribe Akodontini, plus one Oryzomyini species. One year later, Ventura et al. (2009) performed chromosome painting using Akodon species-specific probes.
The role of cytogenetics in species recognition (cytotaxonomy) has been know for a while, considering that many rodents' species are morphologically similar (Bonvicino and Weksler 1998, Percequillo et al. 2008). In addition, molecular phylogenetics improved the possibility of recognizing monophyletic clades. In fact, the proper identification of undescribed species is only possible with the association of morphology, cytogenetics, geographic distribution and molecular phylogeny. Altogether, these different approaches are essential not only for identifying the cryptic Brazilian biodiversity but also for public health programs, since some rodents' species are Hantavirus reservoirs (Souza et al. 2002, Lemos et al. 2004). Therefore, the aim of this review is to compile all the cytogenetic data available  Bonvicino et al. 2008, Gava et al. 2011, Walker et al. 2014 Cavia porcellus  Maia 1984 Hydrochoerus hydrochaeris (Linnaeus, 1766) -66 102 -All Brazilian States, except CE Wurster et al. 1971 Kerodon acrobata Moojen, 1997 -52 92 -Northeastern GO Bonvicino et al. 2008, Zappes et al. 2014 Kerodon rupestris Pericentric inversion From PI and CE to Northern MG Maia 1984, Lessa et al. 2013 Family Variation in the Y morphology; deletion of the X long arm Southern Brazil Kasahara and Yonenaga-Yassuda 1984, Vitullo et al. 1986, Sbalqueiro 1989, Pardiñas et al. 2015b Akodon cursor (Winge, 1887) Akodon arviculoides

Cytogenetic preparation
Chromosome preparations of Carterodon sulcidens, the five samples of Mus musculus, four Neacomys a. amoenus, and a specimen of Neacomys from Vila Rica, Mato Grosso State, were obtained in vivo from bone marrow and spleen, following Ford and Hamerton (1956) or in vitro from fibroblast culture (Freshney 1986). Conventional Giemsa staining was performed to determine the diploid and fundamental numbers, and Cbanding and Ag-NOR were performed according to Sumner (1972) and Howell and Black (1980), respectively.

Molecular phylogeny analyses of Neacomys
DNA was extracted from the liver or muscle with Chelex 5% (Bio-Rad) (Walsh et al. 1991) of five specimens of Neacomys. DNA of the specimen from Vila Rica, Mato Grosso State, was extracted from fibroblast cell culture using DNeasy Blood and Tissue kit (Qiagen, catalog number 69506). PCR was performed in a thermal cycler (Eppendorf Mastercycler ep Gradient, Model 5341) using primers MVZ05 (5-CGA AGC TTG ATA TGA AAA ACC ATC GTT G-3) and MVZ16 (5-AAA TAG GAA RTA TCA YTC TGG TTT RAT-3) (Irwin et al. 1991, Smith andPatton 1993, respectively). PCR mixture contained 30 ng of DNA, 25 pmol of each primer, 0.2 mM of dNTP, 2.52 µL of reaction buffer (50 mM KCl, 2.5 mM MgCl 2 , 10 mM Tris-HCl; pH 8.8) and 0.2 units of Taq DNA polymerase (Invitrogen). Thirty-nine amplification cycles were performed, consisting of denaturation at 94 °C for 30 s, annealing at 48 °C for 45 s, extension at 72 °C for 45 s and the final extension at 72 °C for 5 min. The PCR products were separated using 1% agarose gel in TAE buffer. Sequencing was conducted using BigDye (DNA "Big Dye Terminator Cycle Sequencing Standart," Applied Biosystems) and an ABI PRISM 3100 Genetic Analyzer (Applied Biosystems). All sequences were submitted to a comparative similarity search on BLAST (Basic Local Alignment Search Tool) before the alignment. Alignments were performed by using Muscle (Edgar, 2004) implemented in Geneious 4.8.5 (Biomatters). GenBank access numbers are provided in Suppl. material 1.

Results
The current review encompasses all rodent species which up to the present have been reported in Brazil, comprising 271 species from 10 families (Musser and Carleton 2005, Fabre et al. 2016). Diploid number ranges from 2n = 9, 10 in Akodon sp. n. to 2n = 118 in Dactylomys boliviensis Anthony, 1920 (Table 1). It is noteworthy that 38 species (14%) lack any cytogenetic data. Besides, nine species present only the diploid number with no information about the fundamental number.
Many species show chromosome rearrangements leading to variation in diploid and fundamental numbers. Also, more than one diploid number was associated with one single species, suggesting that they could represent species' complexes. Additionally, new karyotypes were assigned to 22 species highlighting them as candidate species, which have not been formally described yet.
All comments below refer to the data compiled and presented in Table 1.

Tribe Akodontini
This is the second most diverse tribe in the subfamily Sigmodontinae. Only five out of 42 species (D'Elía and Pardiñas 2015) that occur in Brazil lack diploid number information (Table 1). However, for one species, Akodon toba Thomas, 1921, such information is available only for Paraguayan specimens. In addition to the species on which there is no information on the diploid number, four species of the genus Oxymycterus Waterhouse, 1837 have not had their fundamental number established, yet.
Cytogenetic studies have proved to be a useful tool in the recognition of species, mainly in the case of the cryptic and sympatric species as Akodon cursor and A. montensis. On the other hand, karyotype was less variable in some other Akodontini genus (for instance Brucepattersonius and Oxymycterus), and in this case, they could not be distinguished cytogenetically. This reveals the need for gathering cytogenetic, molecular and morphological data in taxonomic studies.

Tribe Oryzomyini
Comprising 73 species up to now, this tribe alone comprises about 47% of the Sigmodontinae diversity. Notwithstanding, it is one of the best cytogenetically studied taxa of Brazilian rodents, and cytogenetic information on fundamental number lacks for only one species: Neacomys guianae Thomas, 1905. In Brazilian representatives the diploid number varied from 2n = 34 in Neacomys musseri Patton, da Silva & Malcolm, 2000 to 2n = 86 in Zygodontomys brevicauda (J. A. Allen & Chapman, 1893).
Pericentric inversion (n = 13) and Robertsonian rearrangements (n = 8) are common rearrangements, as well as sex chromosomes variations, that were described in 12 species and correlated to addition/deletion of constitutive heterochromatin and pericentric inversions.
Besides, Oryzomyini is also the tribe with more species having supernumerary chromosomes (n = 6). Remarkably, B chromosomes in this tribe present different morphology and composition, not only between, but also within the same species. For instance, Nectomys squamipes Brants, 1827 presents from one to three supernumeraries that could be large/medium submetacentric or medium acrocentric, with interstitial or entire long arm C-banded, with late or early replication and with or without interstitial telomeric sites (Silva and Yonenaga-Yassuda 1998b). Differences were also described in Bs of Holochilus brasiliensis, Nectomys rattus Pelzeln, 1883, and Oligoryzomys flavescens (Waterhouse, 1837) . Recently, FISH with Holochilus brasiliensis probes of sex chromosomes (X and Y) and both supernumeraries (B1 and B2) were performed, revealing positive signal on sex chromosome of 12 Oryzomyini species and Bs of Holochilus brasiliensis, Nectomys rattus and N. squamipes . No signal was observed in Bs of Oligoryzomys flavescens and Sooretamys angouya (G. Fischer, 1814), though, corroborating that supernumeraries in this group may have had independent origins .
Just as in all hierarchical levels of rodents' taxonomy, cytogenetic diversity is underestimated in this tribe. For instance, recently, Silva et al. (2015) described two new cytotypes for Neacomys: 2n = 58, FN = 64, from samples collected in Marabá, and 2n = 58, FN = 70, from samples collected in Chaves, Marajó Island, localities from Pará State. According to the authors, both cytotypes differed in the number of biarmed pairs due to amplification/deletion of constitutive heterochromatin in the short arms of pairs 24, 26, and 27 (from Marajó Island) and pericentric inversion involving pairs 28 (metacentric) and 24 (acrocentric) from Marajó Island and Marabá, respectively. These karyotypes could not be assigned to any species described so far, and molecular phylogeny of these samples corroborates the cytogenetic data that it might be a new species .
Herein, we describe the same diploid (2n = 58), but with a different fundamental number (66) to Neacomys collected in Vila Rica, Mato Grosso State (approximately 700 km from those samples described by Silva et al. 2015). The karyotype comprises 23 acrocentric pairs decreasing in size (pair 1 is the largest of the complement), and five small biarmed pairs. The X chromosome is a large submetacentric, and the Y is a small submetacentric (Fig. 1a). The C-banding pattern shows constitutive heterochromatin at the pericentromeric regions of all autosomes, and in the short arm of both X and Y (Fig. 1b).
For phylogenetic analyses, the best model selected for the mitochondrial gene (cytb) was GTR+I+G. Our molecular phylogeny suggests that this specimen with 2n = 58, FN = 66, from Vila Rica may be an undescribed species that belongs to the same one reported by Silva et al. (2015) with 2n = 58, FN = 64, but with a new fundamental number, probably due to pericentric inversions (Fig. 2). Two structured clades of Neacomys with 2n = 58 were recovered: one with samples with FN = 70, and the other with FN = 64 and 66. Additionally, a sample from Igarapé-Açu (MTR12842), Rio Abacaxis (Amazonas, Brazil) was recovered as the sister group of these two clades. Although the phylogenetic reconstruction lacks N. tenuipes Thomas, 1900 (because the unique sequence available in GenBank has only 177pb), it is unlikely that samples with 2n = 58 belong to N. tenuipes once this species is distributed in Colombia and Venezuela and did not nest in the clade of N. tenuipes of the molecular phylogeny presented by Silva et al. (2015). In addition, our phylogenetic reconstruction recovered Neacomys as monophyletic with high support values (1PP/ 99ML). ML and IB analyses recovered the same topology.

Tribe Phyllotini
In Brazil, this tribe was initially composed only of the genus Calomys Waterhouse, 1837. However, due to sampling efforts, a new genus was recently added, Calassomys Pardiñas, Lessa, Salazar-Bravo & Câmara, 2014. The diploid number varied from 2n = 36 in Calomys cerqueirai to 2n = 66 in Calomys tener and Calomys expulsus, although the latter presents two different diploid numbers and karyotypes associated to its name, therefore highlighting the need for further investigation Almeida 2000, Mattevi et al. 2005). Cytogenetic information is available for all the representatives,  and it is an important tool for the recognition of species (cytotaxonomy). One species presents centric fusion (Calomys cerqueirai) (Colombi and Fagundes 2014).

Tribe Reithrodontini
In Brazil, the only representative of this tribe is Reithrodon typicus Waterhouse, 1837. This species possesses a low diploid number (2n = 28) and occurs on the border of Uruguay (Freitas et al. 1983, Pardiñas et al. 2015c (Table 1).

Tribe Sigmodontini
Only one species of this tribe can be found in Brazil, Sigmodon alstoni (Thomas, 1881). Voss (1992) karyotyped 11 specimens from three localities at Venezuela with 2n = 78, 80 and 82, but the picture of the karyotypes and the fundamental numbers were not reported. Also, the author suggested that Robertsonian rearrangement is a plausible explanation for the variation observed. There have been no Brazilian representatives of this species karyotyped so far.

Tribe Thomasomyini
This tribe is represented by only two genera in Brazil: Rhipidomys Tschudi, 1845 and Rhagomys Thomas, 1886. The diploid number varied from 2n = 36 in Rhagomys rufescens (Thomas, 1886) to 2n = 50 in Rhipidomys nitela Thomas, 1901. Apart from R. nitela, which possesses 2n = 48 (samples from Roraima State) or 50 (samples from Manaus, Amazonia State), in general, the karyotype is not informative for Rhipidomys, since nine species present the same diploid number (2n = 44), and two species lack karyotype data Yonenaga-Yassuda 1999, Tribe 2005). In fact, Tribe (2015) provisionally inserted the 2n = 50 samples in R. nitela but reiterated that they need taxonomic revision. Pericentric inversion, found in six species, plays an important role in the genus, and this is reflected in the variation of the fundamental number. Two species lack cytogenetic data: Rhipidomys ipukensis R. G. Rocha, Costa &R. wetzeli A. L. Gardner, 1990.

Tribe Wiedomyini
This tribe is composed of two species: Wiedomys pyrrhorhinos (Wied-Neuwied, 1821) and W. cerradensis P. R. Gonçalves, Almeida & Bonvicino, 2005. Both occur in Brazil with disjunctive distribution (W. pyrrhorhinos at Caatinga, and W. cerradensis at Cerrado) and possess different karyotypes (2n = 62 and 60, respectively) Langguth 1987, Gonçalves et al. 2005). Recent molecular studies indicate that W. pyrrhorhinos, may represent a species complex with Rio São Francisco acting as a barrier to the populations from both river banks (Di-Nizo in prep.). Pericentric inversions have also been described for this species.

Family Ctenomyidae
This family comprises a single genus, Ctenomys, which presents a great variation in diploid numbers, especially C. lami T. R. O. Freitas, 2001, C. minutus Nehring, 1887and C. torquatus Lichtenstein, 1830 for which Robertsonian rearrangements and in tandem fusions were described Lessa 1984, Fernandes et al. 2009). The diploid number varied from 36 in Ctenomys nattereri Wagner, 1848 to 58 in C. lami. Only one species out of eight lacks karyotype information. Cytogenetic data was useful for recognizing Ctenomys bicolor Miranda-Ribeiro, 1914, C. ibicuiensis T. R. O. Freitas, Fernandes, Fornel & Roratto, 2012 and C. nattereri, because it presents exclusive karyotype (Stoulz 2012). Pericentric inversion has been described for C. lami and in tandem fusions for C. minutus.

Family Dasyproctidae
This family comprises two genera: Dasyprocta Illiger, 1811, with nine species, and Myoprocta Thomas, 1903, with two species . There is no cytogenetic data known for three species (Table 1). The diploid number in the Family varied from 62 to 65, and in the genus Dasyprocta, from 64 to 65, due to the presence of B chromosomes in four species (Ramos et al. 2003).

Family Echimyidae
Even being the second largest Brazilian rodent family, a remarkable gap regarding cytogenetic data of this family still remains, with 14 species out of 68 lacking such information. This represents about 37% of all the unknown karyotypic information of all Brazilian rodents.
In this work, the karyotype of Carterodon sulcidens is being described for the first time, showing 2n = 66. Since the animal was a female, it was not possible to recognize the X chromosomes and the exact morphology of the small pair, so we could not establish the fundamental number. Karyotype is composed of 32 acrocentric pairs decreasing in size and presumably one biarmed pair (pair 33). Also, the fourth largest pair possesses a remarkable secondary constriction (Fig. 3a). Constitutive heterochromatin is located in the pericentromeric region of all autosomes (Fig. 3b). Ag-NOR showed signals in the secondary constriction of pair 4 (Fig. 3b inset).
Within the Echimyidae Family, the only other species with 2n = 66 described so far is Makalata didelphoides, but its karyotype presents 20 pairs of metacentric chromosomes, which clearly differs from the karyotype of Carterodon sulcidens.

Family Muridae
This family (represented by the genera Mus and Rattus) was introduced from Europe, and even though it is not a native, it is currently widespread throughout Brazil (Musser and Carleton 2005).
Little is known about the cytogenetics of the Mus musculus Brazilian populations because this species seems to be negglected. The present paper features the first picture of Mus musculus karyotype from Brazil. This species presented 2n = 40, FN = 38, with all chromosomes acrocentrics. C-banding was restricted to the centromeric region of all chromosomes (Fig. 4). Sex chromosomes could only be recognized after G-banding (not showed) because they have similar morphology compared to the autosomes.
For the black rat Rattus rattus Linnaeus, 1758, diploid number of South America population is the same as those from Oceania (2n = 38), and Kasahara and Yonenaga-Yassuda (1981) described pericentric inversion for individuals from São Paulo, Brazil.

Family Sciuridae
Cytogenetic data is unknown for almost the entire family. For the two species to which chromosome information is known, diploid number is 2n = 40, and pericentric inver-sion has been described for one of them, Guerlinguetus brasiliensis (Gmelin, 1788) (Lima andLangguth 2002, Fagundes et al. 2003) (Table 1).

Advances since the last revision
The last cytogenetic revision on Brazilian rodents, published in 1984, described the karyotype of 62 species, mainly from South and Southeast Brazil . This paper compiles the karyotype of 271 species distributed throughout Brazil, representing an increase of more than 300%.
Since 1984, many species' names have been redescribed or validated (e.g. Zygodontomys lasiurus was named as Bolomys lasiurus for a long time, and nowadays is Necromys lasiurus -see synonyms of Table 1). Also, due to the progress of molecular biology during the 1990, associated to morphological information, the number of species described has increased exponentially. It is important to emphasize that molecular phylogeny hitherto has contributed to better understand the cryptic diversity of Brazilian rodents, recognizing monophyletic clades. For instance, new candidate species of Akodon Yonenaga-Yassuda 1998a, Silva et al. 2006), Oecomys (Suárez-Villota et al. under revision), Oligoryzomys (Andrades- Miranda et al. 2001a, Miranda et al. 2008, Neacomys (Silva et al. 2015, present paper), Thaptomys (Ventura et al. 2004(Ventura et al. , 2010, etc. were recognized based on new karyotypes associated to the monophyly of the samples. Even new genera were described based on multidisciplinary approaches: Drymoreomys (Percequillo et al. 2011) and Calassomys (Pardiñas et al. 2014).
Technological advances with fluorescent in situ hybridization (developed at the end of 1980's but more used during 2000's to date), made it possible to characterize chromosome rearrangements more precisely.
In this paper, we provide a new fundamental number for an undescribed species of Neacomys. The karyotype presented here (FN = 66) is similar to the one described by Silva et al. (2015) with FN = 64, except that we found five biarmed pairs and the distribution of constitutive heterochromatin in autosomes was restricted to pericentric regions. We suggest that differences in fundamental numbers are due to pericentric inversions in a small pair, since C-banding evidenced constitutive heterochromatin at the pericentromeric regions, and the morphology of chromosomes was accurately defined. Sex chromosomes presented the same morphology, although the Y was heterochromatic in the short arm (present paper), while it was entirely heterochromatic in the samples described by Silva et al. (2015).
Karyotype information was the first to point out that this specimen may represent a new species, since 2n = 58, FN = 66, has never been described for any Neacomys species. Although we used only one molecular marker (incomplete cyt-b), which was the same used by Silva et al. (2015), the phylogeny corroborates this information, since all samples with 2n = 58 clustered in a monophyletic high supported the clade. This included two well-supported structured clades, one with samples with FN = 70 (Chaves, Marajó Island) and the other with samples with FN = 64 and 66 (Marabá, Pará State and Vila Rica, Mato Grosso State, respectively), both sister clade to the sample from Igarapé-Açu, Amazonas State. Whether these samples belong to the same undescribed entity with strong population structure or whether they represent at least three different species must be clarified with further phylogeographic and morphological studies, including samples from other localities. This shows the importance of integrative approaches.
In fact, Neacomys have a greater diversity than previously known. Recently, based on morphology and molecular phylogeny, Hurtado and Pacheco (2017) demonstrated that Neacomys spinosus is a species complex and considered the subespecies Neacomys spinosus amoenus a valid species. After this revision, Neacomys spinosus is restricted to populations from Peruvian Amazon, and Neacomys amoenus encompasses two subspecies: Neacomys a. amoenus (from Brazilian Cerrado and Bolivia) and Neacomys a. carceleni (from Amazon basin of Ecuador, Brazil and Peru). Thus, sequences related to N. spinosus from central Brazil, and transition areas of Cerrado and Amazonia correspond to N. amoenus. Also, a new species, N. vargasllosai, from southern Peru and Bolivia was described. In this same revision, authors recovered three new species pending formal description (the first from Pará, Brazil, the second from Amazonas, Brazil, and the third from Peru and Ecuador). The one from Pará corresponds to the clade composed of samples with 2n = 58 (Fig. 2), reiteraiting the lack of knowledge in this genus.
The description of the karyotype of Carterodon sulcidens (a rare species) also corroborates the lack of knowledge for some species, and the importance of fieldwork in discovering new data.
We also show the picture of the karyotype of the exotic species Mus musculus for the first time. Despite the noteworthy variation in diploid numbers in Western Europe and Mediterranean populations because of Robertsonian rearrangements (Nachman et al. 1994), in Brazil, the only diploid number described was the standard one (2n = 40).

Progress in cytogenetics: the molecular era
During the beginning of the 1970s (although banding techniques had already been described), karyotypes of Brazilian rodents were studied mainly through conventional staining and the description was limited to diploid and fundamental numbers. Even so, the idea of a wide chromosomal variability already existed. From the 1980s until now, comparative cytogenetics with chromosome banding persists and contributed for elucidating these variations, being that G and C-banding and Ag-NORs are the commonest and cheapest banding techniques.
In fact, the distribution of constitutive heterochromatin and Ag-NORs can be markers in some species. For example, large blocks of constitutive heterochromatin were detected in Clyomys laticeps (family Echimyidae) (Souza andYonenaga-Yassuda 1984, Bezerra et al. 2012) and a huge heterochromatic arm in Pseudoryzomys simplex (family Cricetidae, subfamily Sigmodontinae, tribe Oryzomyini) (Moreira et al. 2013). C-band pattern is also an important technique for recognizing sex chromosomes, especially within the subfamily Sigmodontinae (Silva 1994, Di-Nizo 2013. Regarding the nucleolus organizer region, it seems that secondary constriction is a characteristic feature of the family Echimyidae and, as with other vertebrates, may be an important marker. However, chromosomal comparison is now passing from banding patterns to the use of higher resolution innovation of molecular cytogenetics using FISH. FISH using chromosome painting allows a comparison in a wide genomic scale, revealing a greater number of chromosome changes, unrevealed by the commonest banding techniques, especially in the tribes Akodontini and Oryzomyini of the Subfamily Sigmodontinae. For instance, G-banding pattern showed several rearrangements between Akodon species (Tribe Akodontini) (Geise et al. 1998), but much more complex rearrangements within this genus were observed after crossspecies chromosome painting ).
Extensive chromosomal rearrangements such as Robertsonian, in tandem fusion/ fission and pericentric inversion, were also observed within the genus Oligoryzomys (Tribe Oryzomyini), after chromosome painting. Using a molecular phylogeny as a reference, it was also possible to detect the direction of the rearrangements and to infer that fission events were as common as fusion events (Di-Nizo et al. 2015). Moreover, Robertsonian rearrangement between O. rupestris Weksler & Bonvicino, 2005 (referred as Oligoryzomys sp. 1), 2n = 46, FN = 52, and Oligoryzomys sp. 2, 2n = 46, FN = 52 was firstly detected by using classic cytogenetic and FISH with telomeric probes (Silva and Yonenaga-Yassuda 1997) and later corroborated by chromosome painting (Di-Nizo et al. 2015). However further studies with molecular phylogeny and morphology are necessary to clarify if both entities represent a single species (with a polymorphism spread in the population) or two different species (in the case of this rearrangement resulted in reproductive incompatibilities leading to the speciation of ancestral population).
The advent of chromosome painting made it possible to compare not only related species but also distant ones, something which is difficult to achieve with banding patterns. Hass et al. (2008) compared Mus musculus (family Muridae) to Akodon species (family Cricetidae); Nagamachi et al. (2013) compared two different, unrelated genera of the tribe Oryzomiyni (Cerradomys and Hylaeamys) and Suárez et al. (2015) and Pereira et al. (2016) compared homologies between the tribes Akodontini and Oryzomyini.
Despite the 'modern cytogenetics era', chromosome banding is still an important tool for animal cytogenetic studies, not only because FISH cannot reveal chromosome inversions, but also because it is still a difficult and expensive technique to use.

Chromosome rearrangements and speciation
Rodents proved to be a good model for chromosome evolution studies. Cytogenetics associated with molecular or morphological phylogenetic reconstruction broke cytogeneticist paradigms that fusion rearrangement is more common than fission, and that the reduction in 2n is the expected pattern (e.g. Di-Nizo et al. 2015).
Chromosomal rearrangement could possibly be the cause of reproductive isolation in many Brazilian rodents' species, leading to speciation. The main rearrangements that lead to species formation are Robertsonian, in tandem fusion/fission and pericentric inversion, while the variability in constitutive heterochromatin does not seem to create a reproductive barrier and consequent speciation (King 1993, Romanenko andVoloboeuv 2012).
For a long time, it was thought that chromosomal structural rearrangements promoted speciation by generating gametes with duplications and deficiencies, therefore, causing less adaptability of the heterozygotes, but this model was rejected because it lacked theoretical support (Rieseberg 2001, Patton 2004, Jackson 2011. Recently, a different model of chromosome speciation was proposed in which the gene flow is reduced because of recombination-suppression in rearranged regions (Noor et al. 2001, Rieseberg 2001.
In fact, normal meiotic behavior with suppression of crossing over in inverted segments of heteromorphic chromosomes caused by pericentric inversions of Akodon cursor and Oligoryzomys nigripes was observed, with non-selective disadvantages in het-erozygous carries , Bonvicino et al. 2001a. Some genetic mechanisms seem to be responsible for overcoming meiotic errors in heterozygous individuals, such as the occurrence of heterosynapsis and the low frequency of chiasm between the inverted segments.
A remarkable chromosome variation can be found in the semi-and fossorial Brazilian rodents Blarinomys breviceps (in which molecular phylogeny demonstrated two structured clades -see Ventura et al. 2012), Clyomys laticeps and Ctenomys minutus. Their species status, and whether their chromosome variation is adaptative and correlated with ecological patterns should be evaluated.
For example, a very well-known case of chromosome speciation due to population adaptation to climatic stress and ecological unpredictability was described in the subterranean rodent Spalax ehrenbergi (Family Spalacidae) found in Israel, in which diploid numbers increase coincidently with geographic regions of high aridity (Wahrman et al. 1969). The weak dispersion pattern of this fossorial rodent may have contributed to the fixation of adaptative chromosome change (Árnason 1972).

Cytotaxonomy
Cytotaxonomy is the use of chromosome data as the first clue in the identification of species. Since many Brazilian rodent species present species-specific karyotype and show morphological similarities, chromosome information showed to be useful in the diagnosis of species.
On the other hand, since rates of karyotype evolution differ in distinct branches of the rodents' phylogeny, some species present identical diploid and fundamental numbers, and they cannot be identified solely through chromosome data. This is the case of the following species: (i)  (Table 1).

Interdisciplinarity
Since the beginning of the cytogenetic studies in Brazilian rodents, there have been cases in which different karyotypes were assigned to one species or the same karyotype was referred to in different species. In fact, many of these cases were solved after the integration of different disciplines. For instance, for many years cytogenetic information indicated that the previous "Oryzomys subflavus" could, in fact, be more than one species, since nine different karyotypes were attributed to a single taxonomic entity (Maia and Hulak 1981, Almeida and Yonenaga-Yassuda 1985, Svartman and Almeida 1992, Silva 1994). Nowadays, after interdisciplinary studies with morphology and molecular phylogeny, it is possible to recognize eight species (Weksler et al. 2006, Percequillo et al. 2008, Tavares et al. 2011, Bonvicino et al. 2014. Moreover, for a long time Nectomys was represented by only one species in Brazil, with two diploid numbers (2n = 52 + 1 to 3 Bs and 2n = 56 + 1 to 3 Bs). Nevertheless analyses of the spermatogenesis in hybrids and the sterility of crosses between both cytotypes indicated that Nectomys should be considered two distinct species: Nectomys rattus (2n = 52) and Nectomys squamipes (2n = 56) (Bonvicino et al. 1996).
Some of these cases persist until today, for instance, more than one karyotype was described for Euryoryzomys macconnelli and E. lamia (Table 1). Molecular phylogeny and morphology corroborate the species complex status of both entities (Almeida 2014, Percequillo 2015a. Similarly, Oecomys roberti, O. paricola, and O. catherinae are probably species complexes, not only because of their variability in diploid number, but also because of phylogenetic reconstruction and morphological studies (Suárez-Villota et al. 2017). Ctenomys minutus, C. torquatus, Hylaeamys yunganus, Rhipidomys nitela, Sigmodon alstoni and Zygodontomys brevicauda also deserve taxonomic attention because they may represent cases in which different diploid numbers are attributed to the same names. Similarly, Blarinomys breviceps has a variable diploid number and two geographic structured clades were recovered in the molecular phylogeny (Ventura et al. 2012), indicating that a morphological revision is needed.
Remarkably, such examples can also be found in the family Echimyidae. The need to use different approaches for taxonomic revision is clear in order to investigate whether Phyllomys blainvillii, Phyllomys pattoni, and Proechimys guyannensis represent species complexes, given the fact that they have more than one karyotype associated.
Interdisciplinary approaches, including cytogenetic, molecular phylogeny, morphology and geographic distribution are essential for accessing the limits of Brazilian rodents' species. One of the best-known examples was the old genera Oryzomys, considered the most complex and composing almost half of the species of the tribe Oryzomyini (Musser and Carleton 1993). The current genera Melanomys, Microryzomys, Nesoryzomys, Oecomys, and Oligoryzomys, were first considered a subgenus of Oryzomys and later elevated to the category of genus after morphology, chromosomal and molecular analyses (Myers et al. 1995, Smith and Patton 1999, Bonvicino and Moreira 2001. Another outstanding example of an integrative approach was the study in which ten new genera were described for species that were previously referred to as Oryzomys (Weksler et al. 2006), corroborating the cryptic diversity in Oryzomyini previously indicated by cytogenetic data.
Within the Family Echimyidae, the association of morphology and molecular analysis was essential for elevating Trinomys (considered subgenus of Proechimys) to the genus category (Lara et al. 1996, Leite and.

Perspectives
Despite the new technological approaches, chromosome characterization with conventional staining and banding pattern is still important, mainly because 38 species lack any karyotype information (Table 1). From this amount, 16 are distributed in the Amazonian biome, evidencing the lack of knowledge for this region. The fieldwork is very important and must be encouraged not only because new species and even genera are constantly being described but also because cytogenetic and distribution information of several species are poorly known.
Concerning the family Echimyidae, it is noteworthy that cytogenetic information is lacking for more than 20% of its species. Eleven out of 17 echimyid genera which occur in Brazil are arboreal (Galewski et al. 2005. The issues for sampling small arboreal mammals and the consequent low number of studies with this approach have already been highlighted in the literature (Malcolm 1991, Taylor and Lowman 1996, Graipel et al. 2003. In this sense, it can be inferred that this deficiency in echimyid cytogenetic knowledge may be related to sampling scarcity. The future of molecular biology is promising, with next-generation sequencing (NGS) technology and mitogenomics hopefully providing more robust phylogenetic studies. This new approach was performed with the Family Echymyidae, revealing new supported nodes and clarifying some aspects of the group's taxonomy (Fabre et al. 2016).
However, it is important to reiterate the heterogeneity of characters since DNA, chromosomes, morphology, and behavior are not evolving at the same rate. This particularity may imply in different taxonomic interpretations, with a population being identified as a unique species by one character and two or more species by another, especially in the cases of recent or ongoing speciation. The consequences can be taxonomic inflation or underestimation of the biodiversity, and that is why interdisciplinary approaches are crucial to better understand the biological diversity of rodents.