Chromosomal Signatures Corroborate the Phylogenetic Relationships within Akodontini (Rodentia, Sigmodontinae).

Comparative chromosome-painting analysis among highly rearranged karyotypes of Sigmodontinae rodents (Rodentia, Cricetidae) detects conserved syntenic blocks, which are proposed as chromosomal signatures and can be used as phylogenetic markers. In the Akodontini tribe, the molecular topology (Cytb and/or IRBP) shows five low-supported clades (divisions: “Akodon”, “Bibimys”, “Blarinomys”, “Oxymycterus”, and “Scapteromys”) within two high-supported major clades (clade A: “Akodon”, “Bibimys”, and “Oxymycterus”; clade B: “Blarinomys” and “Scapteromys”). Here, we examine the chromosomal signatures of the Akodontini tribe by using Hylaeamys megacephalus (HME) probes to study the karyotypes of Oxymycterus amazonicus (2n = 54, FN = 64) and Blarinomys breviceps (2n = 28, FN = 50), and compare these data with those from other taxa investigated using the same set of probes. We strategically employ the chromosomal signatures to elucidate phylogenetic relationships among the Akodontini. When we follow the evolution of chromosomal signature states, we find that the cytogenetic data corroborate the current molecular relationships in clade A nodes. We discuss the distinct events that caused karyotypic variability in the Oxymycterus and Blarinomys genera. In addition, we propose that Blarinomys may constitute a species complex, and that the taxonomy should be revised to better delimit the geographical boundaries and their taxonomic status.


Introduction
Chromosome painting (Fluorescence In Situ Hybridization-FISH-with whole-chromosome probes) can be used to investigate genomic changes among different groups of vertebrates. This approach has improved our understanding of the evolutionary processes ranging from speciation [1] to ancestral karyotype hypotheses [2,3]; it allows the comparison of homeologies between species with evolutionary distances of about 55 Ma (million years ago). Comparative chromosome-painting analysis also Int. J. Mol. Sci. 2020, 21, 2415 2 of 17 identifies chromosome segments that have preserved synteny throughout evolution [4], without the need for complex mapping for each species [5], even allowing the determination of putative ancestral karyotypes [6]. For instance, in Xenarthra (sloths, armadillos, and anteaters), the hybridization of human whole-chromosome probes (HSA) in Bradypus torquatus and B. variegatus revealed that the syntenic association of HSA 17/19 is an exclusive trait of the Bradypus genus, leading to the proposal of a hypothetical Xenarthran karyotype with 48 chromosomes distinct from the ancestral Eutherian karyotype. Moreover, trait HSA 7/10 and the fragmentation of HSA 8 into three blocks were suggested to support the monophyly of Xenarthra [3].
In general, chromosomal rearrangements are rare events with low levels of homoplasy [14]. Despite this, in some vertebrate lineages, chromosomal number and morphology are highly variable [2,12,13,15]. Several studies using chromosome painting were able to detect conserved syntenic associations shared among distinct taxa (chromosomal signatures). These signatures have been used in the construction of chromosomal topologies and as phylogenetic markers, as they are often group-specific, evolutionarily conserved, and phylogenetically informative, and corroborate the phylogenetic relationships obtained from molecular topologies [2,3,9,10,[12][13][14][15][16][17]. As an example, the monophyly of Afrotheria was debated due to the lack of morphological evidence, but the identification of syntenic associations HSA 1/19 and 5/21 provided support for the Afrotheria monophyly [5]. Understanding how these rearrangements occur and what they mean in evolutionary terms can be answered in the near future by a chromosomics approach, in which sequencing and epigenetics can be used to comprehend why some rearrangements are conserved while others are not [18].
Here, we focused on two Akodontini genera that present distinct chromosomal evolutionary patterns: Oxymycterus and Blarinomys. The genus Oxymycterus has 16 valid species [20,35]. Cytogenetic data are available for seven of them, which have a consistent 2n (54) but fundamental autosomal number (FN) ranging from 60 to 64 ( Table 2). The cytogenetic studies in Oxymycterus have been limited to classical cytogenetics (conventional staining and chromosomal banding); in most cases, the 2n and FN were reported without showing the karyotype. The Blarinomys genus is considered monotypic (B. breviceps), with two main well-structured clades identified by phylogenetic reconstructions (Maximum Likelihood, Maximum Parsimony, and Bayesian Inference analysis of Cytb gene); it has eight chromosomal forms with 2n ranging from 28 to 52, a consistent FN of 50, and a varied number of B chromosomes (from 0 to 4). Blarinomys breviceps karyomorphs were analyzed by C-banding, G-banding, R-banding, and FISH with telomeric probes that showed the presence of interstitial telomeric sequences (ITS), but these methods were not employed in all specimens. For example, the karyotype with 2n = 28, FN = 50 was assessed only by G-banding and telomeric FISH techniques [42].
Taking into account that comparative cytogenetic data are often phylogenetically informative [14] and chromosomal signatures are maintained in rodent lineages regardless of the high rate of chromosomal change that may occur within each group [9,25], we set out to detect chromosomal signatures that can be used as phylogenetic markers to elucidate the phylogenetic relationships of some Akodontini members. Towards this, we performed comparative chromosome painting using HME whole-chromosome probes [23] on representatives from three Akodontini divisions [27]: Oxymycterus amazonicus (Oxymycterus division; present work), Blarinomys breviceps (Blarinomys division; present work), Akodon sp., A. montensis, Necromys lasiurus, and Thaptomys nigrita (Akodon division) [24,25]; we also compared these species with data obtained using HME probes in other taxa [8,9,23,26].
Here, we describe new cytogenetic results for Oxymycterus and Blarinomys collected in distinct localities of Brazil (Figure 1), and use the chromosomal signatures detected to elucidate the phylogenetic relationships within Akodontini clade A. Bayesian Inference analysis of Cytb gene); it has eight chromosomal forms with 2n ranging from 28 to 52, a consistent FN of 50, and a varied number of B chromosomes (from 0 to 4). Blarinomys breviceps karyomorphs were analyzed by C-banding, G-banding, R-banding, and FISH with telomeric probes that showed the presence of interstitial telomeric sequences (ITS), but these methods were not employed in all specimens. For example, the karyotype with 2n = 28, FN = 50 was assessed only by G-banding and telomeric FISH techniques [42]. Taking into account that comparative cytogenetic data are often phylogenetically informative [14] and chromosomal signatures are maintained in rodent lineages regardless of the high rate of chromosomal change that may occur within each group [9,25], we set out to detect chromosomal signatures that can be used as phylogenetic markers to elucidate the phylogenetic relationships of some Akodontini members. Towards this, we performed comparative chromosome painting using HME whole-chromosome probes [23] on representatives from three Akodontini divisions [27]: Oxymycterus amazonicus (Oxymycterus division; present work), Blarinomys breviceps (Blarinomys division; present work), Akodon sp., A. montensis, Necromys lasiurus, and Thaptomys nigrita (Akodon division) [24,25]; we also compared these species with data obtained using HME probes in other taxa [8,9,23,26].
Here, we describe new cytogenetic results for Oxymycterus and Blarinomys collected in distinct localities of Brazil ( Figure 1), and use the chromosomal signatures detected to elucidate the phylogenetic relationships within Akodontini clade A.

Results
It should be noted that the heterochromatic regions do not exhibit hybridization signals by FISH analysis with Hylaeamys megacephalus (HME) whole-chromosome probes.

Results
It should be noted that the heterochromatic regions do not exhibit hybridization signals by FISH analysis with Hylaeamys megacephalus (HME) whole-chromosome probes.

Blarinomys Breviceps (BBR; 2n = 28, FN = 50)
The karyotype of Blarinomys breviceps (BBR) has 2n = 28 and FN = 50, with autosomes comprising 12 meta/submetacentric pairs (pairs 1 to 12) and one acrocentric pair (pair 13); the X chromosome is a middle-sized acrocentric, and the Y chromosome is a small acrocentric. The constitutive heterochromatin is distributed at the centromeric region of all autosomes and the X chromosome; the Y chromosome is almost entirely heterochromatic (Figure 2b).

New Cytogenetic Data for Oxymycterus and Blarinomys
In the last two decades, the genus Oxymycterus has undergone several changes in taxonomy, with increases in the number of species and many synonyms for its representatives [20]. Although the 16 valid species of Oxymycterus exhibit variability in diagnostic morphological features [20,35], some studies use obsolete names for some taxa that make it difficult to assign karyotypes to species, given that the group has very low cytogenetic variability (2n = 54, FN = 60-64). Here, we adopt the taxonomic classification proposed by Patton, Pardiñas, and D'Elía [20] (Table 2).
In relation to sex chromosome morphology, we observed three types of X:  [29] show the presence of a large heterochromatic block in the short arm; this suggests that amplification/deletion of constitutive heterochromatin probably accounts for the difference between subtelocentric and submetacentric morphologies, while the medium metacentric is probably due to pericentric inversion or centromeric repositioning [44]. The Y chromosome is described as acrocentric for five species (O. caparaoe, O. dasytrichus, O. paramensis, O. quaestor, and O. rufus), while the Y-chromosomes of Oxymycterus sp. and O. amazonicus have a submetacentric morphology that is probably due to pericentric inversion or centromeric repositioning [44].
Among the eight karyomorphs of Blarinomys breviceps (2n = 28-52, FN = 50), more one-armed chromosomes are seen in those with a higher 2n, whereas more bi-armed chromosomes are seen in karyotypes with a lower 2n [42]. Ventura et al. [42] recovered two major clades in a phylogenetic reconstruction of B. breviceps specimens: the northeastern lineage comprised samples with 2n = 52, while the southeastern lineage exhibited a higher diversity of 2n (ranging from 28 to 45). Although the authors did not speculate on the direction of chromosomal change in the southeastern lineage, we believe that this group experienced several Robertsonian translocations that caused the lower 2n. This is consistent with a previous proposition regarding the evolutionary trend of Akodontini members [34], as seen in Akodon [21].
We also compared the distribution of constitutive heterochromatin and noticed that four previously reported karyomorphs (2n = 52, 43, 37, 31) had heterochromatin signals in the pericentromeric regions of only some autosomes [42], while our karyotype (2n = 28; MN68882) presented heterochromatin in all autosomes. Differences in the amount of constitutive heterochromatin in karyotypes of the same species or genus are frequent in rodents [45], as observed in Neacomys that showed heterochromatin blocks in from three to five bi-armed pairs among samples from distinct localities in eastern Amazon [9,46].
In Blarinomys, the genetic divergence ranges from 4.8% to 8.4% [42], and several morphological traits vary between the northeastern and southeastern clades [20]. This level of genetic divergence is consistent with the range established for potentially undescribed species (over 5%) [47,48]. This suggests that Blarinomys constitutes a species complex, and that the taxonomy should be revised to take into account geographical boundaries.

Speciation Hypothesis in Oxymycterus and Blarinomys
Oxymycterus presents distinct morphological features and molecular variation among its representatives [20], with low chromosomal variability (2n = 54, FN = 60-64; Table 2). It also exhibits great adaptability, being widely distributed in distinct environments of South America (Amazon Forest, Atlantic Forest, Caatinga, Cerrado, Pampas, Chaco, and Andes), and at elevations ranging from sea level to about 3500 m in the montane forests [20]. Some species have isolated distributions, while others may occur in sympatry with two or even six species [20].
The literature does not contain any phylogeographic study that covers all representatives of the Oxymycterus genus. Peçanha et al. [49] investigated the phylogeographic history of Oxymycterus nasutus, which occurs in the Atlantic Forest and Pampas biomes. The authors recovered six major structured populations that exhibited morphological and molecular differences. They proposed that there had been a population expansion with late retraction during the postglacial period, and concluded that the observed structure was a result of vicariance events with refuge isolation.
Similarly, we hypothesized that the speciation process of Oxymycterus was developed by vicariance, occurring through ecological adaptations to the distinct and complex biogeographic dynamics that occurred in South America [50]. Ecological studies in Oxymycterus demonstrated that the species of this genus are not affected by habitat fragmentation [51]; this could explain the chromosomal stability of this genus, since its members tend to be organized in larger populations. This suggests that chromosomal rearrangements have not played a crucial role in the speciation process of Oxymycterus. However, we emphasize that only after detailed phylogeographic studies of the Oxymycterus genus will it be possible to fully comprehend the speciation process of this understudied group [52].
The Blarinomys genus occurs in the Atlantic Forest biome ( Figure 1) and is separated into two lineages: the northeastern and southeastern clades [42]. This separation is in accordance with two phylogeographic regions that are recognized in the Atlantic Forest [53] and are attributed to the presence of the Rio Doce [54]. Similar separations have been described for other groups of terrestrial vertebrates (e.g., amphibians, birds, and small mammals) in this region [55][56][57].
Although the Atlantic Forest is considered a hotspot for biodiversity [58], it is one of the most anthropized biomes in the world [59]. Due to five centuries of human expansion and degradation [60], this biome is now represented by small and isolated fragments that comprise only 11.7 % of its original cover [61].
The literature lacks any population dynamic data for the Blarinomys genus. Haag et al. [62] used microsatellite and mitochondrial markers to investigate the impact of Atlantic Forest fragmentation on Panthera onca samples. Although P. onca has a high dispersal capability, Haag et al. [62] identified decreases in population size and genetic diversity and probable impediments in gene flow. Thus, there appears to be limited dispersion across separate fragments that contain small and isolated populations, which are suffering from the effects of genetic drift. The impact of this fragmentation has been observed in large carnivores and is likely to be much more intense in rodents. As discussed before [11], some biological features of rodents allow the isolation of populations, since many generations can be produced in a short period of time, due to an elevated reproductive rate and a short pregnancy with the birth of large numbers of individuals per gestation. In addition, the low vagility of rodents favors endogamy [63]. This increases the probability of interbreeding between individuals heterozygous for a rearranged chromosomal form, which could lead to the development of homozygous subpopulations within a few generations [64,65].
Blarinomys exhibits high levels of morphological and molecular divergence between, but not within, the northeastern and southeastern clades [42], indicating that the Rio Doce has acted as an allopatric barrier to gene flow between these two lineages [54]. Carnaval et al. [55] demonstrated that the portion of the forest northeastern of the Rio Doce is highly stable, while the portion to the southeast is unstable. The northeastern populations exhibit karyotypic stability, which is compatible with a stable environment. However, the southeastern lineage exhibits a high level of chromosomal variability, with 2n ranging from 28 to 45. This indicates that chromosomal rearrangements played an important role in the separation of populations, and that the fragmentation (instability) of the Atlantic Forest reinforced the isolation of distinct populations with established karyotypes.

Chromosomal Signatures in Akodontini Reinforce Clade A Monophyly
Above, we discussed the chromosomal evolution of Oxymycterus and Blarinomys using only classical cytogenetics data to predict the events that may have occurred within both genera. Here, by analyzing OAM and BBR chromosome-painting data (Table 3) and comparing them with karyotypes available in the literature [19,21,33,[38][39][40]42], we reaffirm that Oxymycterus and Blarinomys had distinct and independent chromosomal evolutionary processes: the former exhibits chromosomal stability with few inversions and/or translocations, while the latter is marked by numerous Robertsonian translocations that decrease the 2n in this group.
The Akodontini topology recovered by D'Elía [27] using a Maximum Parsimony strict consensus tree (Cytb and/or IRBP) exhibited five weakly supported divisions (Akodon, Bibimys, Blarinomys, Oxymycterus, and Scapteromys). The authors noted that further analysis would be needed to clarify the phylogenetic relationships among the genera, but that two major clades were recovered with a jackknife support value of 97%: clade A (Akodon, Bibimys, and Oxymycterus divisions) and clade B (Blarinomys and Scapteromys divisions) (Table 1, Figure 4).
In an attempt to understand the evolutionary process from the karyotypic changes observed in A. montensis (AMO), Thaptomys nigrita (TNI) [24], Akodon sp. (ASP), Necromys lasiurus (NLA) [25], Oxymycterus amazonicus (OAM), and Blarinomys breviceps (BBR) (present study), we detected chromosomal signatures that could clarify relationships within clade A. We focused on this clade because we had data from five species (ASP, AMO, TNI, NLA, OAM) from four (Akodon, Thaptomys, Necromys, and Oxymycterus) of the nine genera of clade A, but only one species (BBR) of five genera from clade B. Thus, any assumptions made regarding clade B would be only weakly supported.
One branch comprises Necromys and Thalpomys, which exhibit loss of HME (13,22)/11, and another branch is composed of Castoria and Thaptomys, wherein multiple events have eliminated HME 18/2 and 4/11/ (16,17). In the most recent divergent branch of Akodon and Deltamys, another ancestral trait (HME 18/25/3) is present as a derived character (HME 18/25) that arose via a fission event (Node D, Figure 4). We also identified a chromosomal signature (HME (9,10)/5) shared only by TNI and NLA, that reinforces the clade ((Castoria + Thaptomys) (Necromys + Thalpomys)) (Node E, Figure 4), and five specific traits for the Akodon genus (HME (16,17) Similar approaches have helped in the understanding about the evolution of other rodent species. As an example, da Silva et al. [46] plotted the chromosomal information (2n and FN) on a Cytb phylogeny of the Neacomys genus; the authors postulated that two increases and one reduction from the hypothetical ancestral 2n (56) occurred, which was in agreement with the molecular relationships retrieved by molecular data.
Samples were collected using Pitfall traps [66] and deposited at the zoological collections of Museu de Zoologia da Universidade Federal do Pará (UFPA), Belém, Pará, and the Museu Nacional da Universidade Federal do Rio de Janeiro (UFRJ), Rio de Janeiro, Rio de Janeiro. All institutions are in Brazil.

Image Capture and Analysis
Digital images of C-and G-banded karyotypes were obtained using an Olympus BX41 microscope with a CCD 1300QDS digital camera and analyzed using GenASIs software version 7.2.7.34276 from ASI (Applied Spectral Imaging). FISH images were obtained using a Nikon H550S microscope, a DS-Qi1Mc digital camera, and the Nis-Elements software. The karyotypes were organized according to established chromosomal morphology [70]. The final images were edited using Adobe Photoshop CS6.

Conclusions
We herein report new cytogenetic information for Oxymycterus and Blarinomys and distinct chromosomal evolutionary patterns for these genera. Our results show that while Oxymycterus exhibits a stable karyotype with few rearrangements (inversions and/or translocations), Blarinomys presents multiple Robertsonian translocations that reduced the 2n of this group. We propose that the speciation process of Oxymycterus was caused by vicariance events, and that chromosomal rearrangements did not play a crucial role in this process. We also suggest that the population structure of Blarinomys breviceps was established by the fragmentation of the Atlantic Forest and reinforced by chromosomal rearrangements. We used chromosomal signatures as an additional tool to elucidate the phylogenetic relationships among Akodontini clade A lineages, which reinforce the current low-supported arrangement of the branches retrieved by molecular data. Our comparative chromosome painting of Akodontini expands the analysis and may help improve our understanding of the evolutionary process and phylogenetic relationships in this extremely diverse group of rodents.