Cytogenetic analyses in Trinomys (Echimyidae, Rodentia), with description of new karyotypes

Trinomys Thomas (1921) is a terrestrial genus of spiny rats endemic to the Brazilian areas of Atlantic Forest and the transitional areas of Cerrado and Caatinga. Although most species have been already karyotyped, the available cytogenetic information is mostly restricted to diploid and fundamental numbers. We analyzed the chromosomes of two Trinomys species: Trinomys moojeni (2n = 56, FN = 106) and Trinomys setosus setosus (2n = 56, FN = 106 and 2n = 56, FN = 108). Our analyses included GTG- and CBG-banding, silver-staining of the nucleolar organizer regions, and chromosome mapping of telomeres and 45S rDNA by fluorescent in situ hybridization (FISH). Comparative GTG- and CBG-banding suggested that the interspecific variation may be due to rearrangements such as pericentric inversions, centromere repositioning, and heterochromatin variation. We report two new karyotypes for T. s. setosus and describe for the first time the banding patterns of the two Trinomys species.

Within Echimyidae, the Atlantic spiny rats of the genus Trinomys Thomas, 1921, allocated within Euryzygomatomyinae (Lara & Patton, 2000;Fabre et al., 2017), are amongst the most taxonomically complex genera. Trinomys comprises ten extant species endemic to Brazilian areas of Atlantic Forest and transitional areas of Cerrado and Caatinga . Most species have few morphological synapomorphies, with many primitive and few derived features (Dalapicolla & Leite, 2015), which led different authors to consider several of them as subspecies in different taxonomic arrangements (Lara, Patton & Da Silva, 1996;Lara & Patton, 2000;Pessôa et al., 2015). Three species, Trinomys eliasi, Trinomys moojeni, and Trinomys yonenagae, are considered near threatened or endangered due to forest fragmentation and habitat destruction (http://www.iucnredlist.org).
The karyotypes of all recognized species of Trinomys have already been described, with the exception of Trinomys mirapitanga (Table 1). Nevertheless, most reported cytogenetic data are restricted to the description of the 2n and fundamental numbers (FN), without information on banding patterns or FISH. The 2n ranges from 2n = 54 in T. yonenagae to 2 n = 60 in Trinomys albispinus, Trinomys dimidiatus, and Trinomys iheringi (Table 1). Some specimens of T. iheringi presented a higher 2n due to the presence of minute supernumerary chromosomes (Yonenaga-Yassuda et al., 1985;Fagundes, Camacho & Yonenaga-Yassuda, 2004). Comparisons of the GTG-banded chromosomes of T. iheringi, T. albispinus minor (2n = 60, FN = 116), and T. yonenagae (2n = 54, FN = 104), the only Trinomys species analyzed after banding, revealed very conserved karyotypes (Leal-Mesquita et al., 1992). Closely related species of Trinomys seem to share the same karyotype, as is the case of the sister taxa Trinomys paratus and T. eliasi (both with 2n = 58, FN = 112), and of T. dimidiatus and T. iheringi (both with 2n = 60, FN = 116). In fact, it has been suggested that the divergence time among Trinomys species was not sufficient to produce great karyotypic changes (Souza, Corrêa & Pessôa, 2006;Lazar et al., 2017). We comparatively analyzed the karyotypes of T. moojeni and T. s. setosus, including GTG-and CBG-banding, silver staining of the nucleolar organizer regions (Ag-NORs), and FISH with telomeric and 45S rDNA probes. Two new karyotypes are described for T. s. setosus and this is the first description of banding patterns for both species.

MATERIAL AND METHODS
We studied five specimens of Trinomys, collected in the state of Minas Gerais, southeastern Brazil ( on skull, dental and skin characters described by Moojen (1948), Iack-Ximenes (2005), Dalapicolla & Leite (2015) and Pessôa et al. (2015). The morphological diagnosis of each specimen is given in Supplemental Table S1. We plotted the sampling sites of the specimens used in this study against the known range of Trinomys species using QGIS 2.18.16 ( Fig. 1; QGIS Development Team, 2018). Spatial datasets containing the known range of Trinomys species were obtained from IUCN (http://www.iucnredlist.org). Chromosome preparations were obtained directly from bone marrow (Ford & Hamerton, 1956). GTG-and CBG-banding patterns and silver-staining of the nucleolar organizer regions (Ag-NORs) were performed according to Seabright (1971), Sumner (1972), and Howel & Black (1980), respectively. FISH with a biotinylated telomeric sequence (Invitrogen, Carlsabad, CA, USA) and with the R2 45S rDNA probe labeled by nick translation with digoxigenin-11-dUTP (DIG-Nick Translation mix; Roche Applied Science, Penzberg, Germany), followed Araújo et al. (2017); Araújo et al. (2014), respectively. Immunodetection was carried out with neutravidin and antidigoxigenin, both conjugated with rhodamine (Roche Applied Science). The analyses and image acquisition were performed under a Zeiss Axioimager 2 epifluorescence microscope using the AxioVision software (Zeiss, Oberkochen, Germany), Adobe Photoshop CS3 Extended was used for image edition. For each specimen, at least 20 metaphases of each experiment were analyzed.
Ordination and phylogenetic methods were employed in order to check the assignment of MCN-M 2587, MCN-M 3296, and MCN-M 3297 to T. setosus. In order to do this, we sequenced the 401 bp-long segment of the mitochondrial cytochrome b (cytb) of each specimen and included sequences from nine Trinomys species retrieved from GenBank in a phylogenetic analysis. The sequences of the specimens MCN-M2816 (T. moojeni) and UFMG 6024 (T. s. setosus) were previously deposited in GenBank after assembly of their mitochondrial genomes (Araújo et al., 2016). Euryzygomatomys spinosus, Fischer, 1814, was used as outgroup. Total genomic DNA of each Trinomys specimen was extracted from liver and their cytb was amplified by polymerase chain reaction (PCR) with primers MVZ 05 and MVZ 04 (Smith & Patton, 1993). The PCR products were purified using the Wizard SV Gel and PCR Clean-up System kit (Promega, Madison, WI, USA) and sequenced on the ABI3130 platform (Myleus Biotechnology). The GenBank accession numbers of the sequences generated in this study, as well as those included in the analyses are presented in the Supplemental Information 1.
The sequences obtained and those from GenBank were aligned using the Muscle (Edgar, 2004) algorithm. MEGA 7 (Kumar, Stecher & Tamura, 2016) was used to build a Kimura-2-parameter corrected distance matrix in which our ordination analysis was based. Ordination methods were used since they are useful tools to perform dimensionality reduction and to represent the distance between sequences in a coordinate (Cartesian) space where the distances are preserved (Higgins, 1992;Ramette, 2007;Zhang et al., 2011). Principal Coordinate Analysis (PCoA) was used to explore the similarity among our specimen's sequences and other Trinomys species. The analysis was conducted in R (R Core Team, 2017) using the ''pcoa'' function in package APE (Paradis, Claude & Strimmer, 2004) and Lingoes procedure for correcting for negative eigenvalues.
Phylogenetic relatedness was used as a way of determining the most probable identity of the subject sequences. Thus, two methods of phylogenetic reconstruction were employed: maximum likelihood (ML) and bayesian inference (BI), which were carried out in RaxML 8 (Stamatakis, 2014) and MrBayes 3.2 (Ronquist et al., 2012), respectively. ML search comprised optimizations over 100 randomized maximum parsimony starting trees using the rapid hill-climbing algorithm under the GTRGAMMA model. As a measure of branch support, information on frequencies of 1,000 replicates of non-parametric bootstrap were annotated on the best-scoring ML tree. Bayesian inference comprised two independent runs composed of four chains each. A reversible jump MCMC sampling was used in order to explore different substitution schemes. Parameters and trees were sampled every 1,000 generations along a total of 20 million generations. After discarding a quarter of samples as burn-in, parameters and trees were summarized and the following metrics were used to assess MCMC convergence: standard deviation of split frequencies, effective sample size and potential scale reduction factor for each parameter.  Fig. 2; Figs. S1 and S2). This was further supported by the PCoA results, that showed samples of the same species clustering together on the graph (Fig. S3). Morphological characters analyses corroborated the phylogeny and allowed assigning the T. setosus specimens as T. s. setosus (Table S1).

Bayesian inference and ML trees recovered
The female T. moojeni had a complement with 2n = 56 and FN = 106, similar to that described by Corrêa et al. (2005), composed of 26 pairs of biarmed (pairs 1-26) and one pair of acrocentric (pair 27) autosomes, and submetacentric X chromosomes (Fig. 3). The autosomes of the female T. s. setosus (2n = 56, FN = 106) collected in Serro included 26 biarmed pairs decreasing in size from large to small (pairs 1-25 and 27) and a small acrocentric pair (pair 26). The X chromosomes were large acrocentrics (Fig. 4A). The other three specimens of T. s. setosus had karyotypes with 2n = 56 and FN = 108 ( Fig. 4C and Fig. S4), similar to the other cytotype of T. s. setosus, but with pair 26 as a biarmed element. Their X chromosome was a large acrocentric and the Y was a small acrocentric ( Fig. 4 and Fig. S4).
After GTG-banding, it was possible to identify all chromosomes of each species (Figs. 3A and 4) and to verify that the complements of T. s. setosus with FN = 106 and FN = 108 (Fig. 5) differed in relation to pair 26, which was acrocentric or biarmed, in the animals with FN = 106 and FN = 108, respectively. CBG-banding revealed weak centromeric constitutive heterochromatin in pairs 1-5, 9, 10, 15, 17-27, and the X chromosome of T. moojeni (Fig. 3B); pairs 1 , 11, 15, 16, 19-27, and Fig. 4D). Both species had a large interstitial secondary constriction on the long arm of pair 6, which bears the NORs (Fig. 6). Hybridization with the telomeric probe showed signals only at the termini of all chromosomes of the two species studied (Figs. 6C, 6F and 6I).

DISCUSSION
The interspecific grouping of Trinomys, recovered by the phylogenetic analyses, was congruent with previous studies (Lara & Patton, 2000;Tavares, Pessôa & Seuánez, 2015;Lazar et al., 2017). Our phylogenetic analysis is also supported by the specimens' distribution ( Fig. 1), morphology (Supplemental Table S1) and karyotypes. The collecting locality of T. moojeni (MCN-M 2816), Serra do Caraça Private Reserve, is the same of the specimens studied by Cordeiro-Júnior & Talamoni (2006) and the karyotype was similar to that described for this species (Corrêa et al., 2005; Fig. 3). T. s. setosus, in turn, which occurs from the coastal area of the Brazilian states of Sergipe, Bahia, and Espírito Santo to the interior of Minas Gerais , were collected in three municipalities of Minas Gerais.
In order to establish the exact mechanisms involved, further experiments including FISH with specific sequences from the regions of interest are necessary. The karyotypes described herein for T. s. setosus differed in 2n, FN, and/or the sex chromosome morphology from those already published for this genus (Yonenaga-Yassuda et al., 1985;Leal-Mesquita et al., 1992;Corrêa et al., 2005;Pessôa et al., 2005;Souza, Corrêa & Pessôa, 2006;Lazar et al., 2017; Table 1). Trinomys gratiosus bonafidei also has 2n = 56 and FN = 108, but differently from our specimens, has a metacentric Y chromosome (Pessôa et al., 2005). The most recent revision on Trinomys divided T. setosus into the subspecies T. s. setosus and T. s. elegans . The diploid number was reported only for T. s. elegans and without banding patterns (2n = 56, FN = 104;Corrêa et al., 2005).  mentioned that the karyotype of T. s. setosus from Almenara, Minas Gerais state, has 2n = 56 and FN = 108, but no figure was provided. Our T. s. setosus had karyotypes with FN = 106 and 108 and differed from that described by Corrêa et al. (2005) by the presence of additional short arms on pair 27 and pairs 26 and 27 of our specimens, respectively. These differences may be real or may reflect variations in chromosome condensation between both samples, as poorly elongated small chromosomes could prevent the detection of short arms. If real, these differences between T. s. setosus and T. s. elegans may be correlated with their subspecies allocation or may be due to interpopulational variation, as seems to be the case of T. s. setosus. Our phylogenetic analyses did not distinguish between T. s. setosus and T. s. elegans (Fig. 2, Figs. S1 and S2), but the morphological analysis allowed to recognize these subspecific taxa. The karyotype information was also relevant in species identification, revealing karyotypes that differed from those of other species of the genus. The T. s. setosus karyotypes described herein also differed from others previously reported for Trinomys in the morphology of the X chromosome. With the exception of T. setosus and T. yonenagae, which presented acrocentric X chromosomes, all Trinomys species had a submetacentric X (T. albispinus, T. dimidiatus, T. eliasi, T. gratiosus, T. iheringi, T. moojeni, and T. paratus). Based on our cytochrome b phylogeny (Fig.  2), we suggest that a pericentric inversion or a centromere shift on the X chromosome occurred in the lineage that gave rise to T. setosus and T. yonenagae. The change in X chromosome morphology in the common ancestor of both species may be related to karyotype differentiation from other taxa and reproductive isolation. It has been suggested that chromosome rearrangements may affect chromatin structure (Johnson & Lachance, 2012) and, consequently, play a role in hybrid incompatibility. The change in gene expression after chromosome rearrangements was also suggested to contribute to the speciation process (Potter et al., 2017).
T. moojeni and T. s. setosus had a large interstitial secondary constriction on the long arm of pair 6, which bears the NORs (Fig. 6). A chromosome pair with a large secondary constriction bearing the single NOR is a marker of echimyids, as already reported for T. iheringi and other Echimyidae genera (Fagundes, Camacho & Yonenaga-Yassuda, 2004;Silva et al., 2012;Araújo et al., 2014). The comparison of the GTG-banded karyotypes suggests that the NOR-bearing chromosome is the same in our specimens and in T. albispinus, T. iheringi, and T. yonenagae (Leal-Mesquita et al., 1992;Fagundes, Camacho & Yonenaga-Yassuda, 2004) and is probably conserved in the genus.
Hybridization with the telomeric probe showed signals only at the extremities of all chromosomes (Figs. 6C, 6F, and 6I). This pattern of hybridization was similar to that described for T. iheringi (Fagundes, Camacho & Yonenaga-Yassuda, 2004). Bolzán (2017) suggested that the absence of interstitial telomeric sequences indicates the evolutionary status of the chromosomes of a species. Accordingly, species without or with only a few interstitial telomeric sequences would have more conserved chromosomes, as seems to be the case of Trinomys.

CONCLUSIONS
In summary, based on the available data, it is clear that the Trinomys species present conserved karyotypes with small variation in diploid numbers (2n = 54 to 2n = 61-65) and mostly composed of biarmed autosomes. The X chromosomes are usually large submetacentrics and all the species analyzed presented one marker chromosome pair with a secondary constriction corresponding to the NOR, which is also typical for the other echimyid genera. The great conservation extends to the GTG-and CBG-banding patterns in the few species which had these patterns described. As previously proposed by Leal-Mesquita et al. (1992), pericentric inversions, centromere repositioning, and other minor rearrangements seem to be responsible for the chromosome evolution in this genus. Further analyses, including a robust phylogenetic hypothesis, cytogenetic studies with high resolution banding patterns and molecular data of a larger array of Trinomys species, are needed to improve our understanding of the chromosome evolution and genome organization of this genus. It should be stressed that Trinomys species, especially those from Minas Gerais, need more thorough morphological and molecular analyses, as their cytogenetic information alone is insufficient for taxonomic identification. In fact, several different species present very similar karyotypes (Lazar et al., 2017).