Extensive Chromosomal Reorganization in the Evolution of New World Muroid Rodents (Cricetidae, Sigmodontinae): Searching for Ancestral Phylogenetic Traits

Sigmodontinae rodents show great diversity and complexity in morphology and ecology. This diversity is accompanied by extensive chromosome variation challenging attempts to reconstruct their ancestral genome. The species Hylaeamys megacephalus–HME (Oryzomyini, 2n = 54), Necromys lasiurus—NLA (Akodontini, 2n = 34) and Akodon sp.–ASP (Akodontini, 2n = 10) have extreme diploid numbers that make it difficult to understand the rearrangements that are responsible for such differences. In this study we analyzed these changes using whole chromosome probes of HME in cross-species painting of NLA and ASP to construct chromosome homology maps that reveal the rearrangements between species. We include data from the literature for other Sigmodontinae previously studied with probes from HME and Mus musculus (MMU) probes. We also use the HME probes on MMU chromosomes for the comparative analysis of NLA with other species already mapped by MMU probes. Our results show that NLA and ASP have highly rearranged karyotypes when compared to HME. Eleven HME syntenic blocks are shared among the species studied here. Four syntenies may be ancestral to Akodontini (HME2/18, 3/25, 18/25 and 4/11/16) and eight to Sigmodontinae (HME26, 1/12, 6/21, 7/9, 5/17, 11/16, 20/13 and 19/14/19). Using MMU data we identified six associations shared among rodents from seven subfamilies, where MMU3/18 and MMU8/13 are phylogenetic signatures of Sigmodontinae. We suggest that the associations MMU2entire, MMU6proximal/12entire, MMU3/18, MMU8/13, MMU1/17, MMU10/17, MMU12/17, MMU5/16, MMU5/6 and MMU7/19 are part of the ancestral Sigmodontinae genome.


Introduction
Muroids are the most diverse group of extant rodents, with approximately 1500 species distributed in six families [1]. The families Cricetidae and Muridae are the most species rich [1][2][3].
G-banding is useful for the accurate identification of chromosomal homologies in karyotypes with few rearrangements, but is not useful in highly rearranged karyotypes, which makes it difficult to understand species with extensive chromosomal variation. Sigmodontinae have diploid numbers ranging from only 9-10 in species of genus Akodon to 92 in Neusticomys ferreirai, Anotomys leander and Ichthyomys pittieri [1,[14][15][16]. This large variation is problematic when trying to identify the chromosomal rearrangements between the extreme karyotypes in Sigmodontinae. However, chromosome painting has been very successful in demonstrating such rearrangements. This has been shown in Akodon species with diploid numbers varying from 10 to 44 by Ventura et al. [17], in Akodon and Thaptomys by Suarez et al [18] and by Swier et al [19] in Sigmodon genomes, which are quite stable, with few or no chromosome rearrangements. Nagamachi et al. [20] have used the same strategy to demonstrate that the Oryzomyini Hylaeamys megacephalus (2n = 54) and Cerradomys langguthi (2n = 46) are also highly rearranged. In addition, mouse whole chromosome probes were used to compare the karyotypes of the six Sigmodontinae species (five Akodontini and one Oryzomyini), and this enabled the reconstruction of chromosomal phylogeny and phylogenetic relationships [21][22]. However, not all segments had their homeology detected in some genomes (e.g.: Necromys lasiurus, Thaptomys nigrita, Oligoryzomys flavescens, Akodon cursor, A. montensis, A. paranaensis and A. serrensis; [21][22]). Recently Di-Nizo et al. [23] using whole chromosome probes of the Oligoryzomys moojeni (2n = 70), demonstrated that five species of the genus Oligoryzomys (Oryzomyini) have a high degree of chromosomal reorganization; not all existing homeologous were detected. The use of probes from different species, and the gaps left by these studies, make it difficult to comprehend all the mechanisms involved in the reconstruction of the ancestral Sigmodontinae karyotype (See [24]).
In this study, we constructed chromosomal homology maps between Akodontini Akodon sp. (2n = 10) and Necromys lasiurus (2n = 34) using cross species chromosome painting with Oryzomyini chromosomal probes from Hylaeamys megacephalus (2n = 54) to assess the mechanisms leading to the abrupt evolutionary rearrangements between species. We also compared our findings with those from the literature for species already mapped with H. megacephalus probes. Finally, we were able to compare our results on NLA using HME probes with some published results on NLA that used MMU probes. This allowed the identification of some corresponding regions of chromosome homology in studies made by different investigators using different probes. Our results reveal new findings for this important group of rodents and indicate new paths towards the reconstruction of the putative ancestral Sigmodontinae karyotype.

Ethics Statement
JCP has a permanent field permit, number 13248 from "Instituto Chico Mendes de Conservação da Biodiversidade". The Cytogenetics Laboratory from UFPa has a special permit number 19/2003 from the Ministry of Environment for the transport of samples and permit 52/2003 for using the samples in research. The Ethics Committee (Comitê de Ética Animal da Universidade Federal do Pará) approved this research. The specimens were captured using a live capture method designed for small mammals (traps type Sherman, Tomahawk and pitfalls [25]). Specimens were maintained in the lab with food and water, free from stress, until their euthanasia, made with the IP injection of barbiturates after local anesthetic (Ketamine HCl in combination with Diazepam).

Specimen characteristics and Chromosome preparations
The specimens Necromys lasiurus (NLA, two males and one female) and Akodon sp. (ASP, one female and two males) were collected from the municipality of Parauapebas, Pará State, northern Brazil ( Table 1). The sample was collected between October 2009 and January 2010. The identification of the specimens was made on the characteristics of skull and skin, and the voucher material deposited in the Mastozoology Collection of the Museu de Zoologia da Universidade Federal do Pará (MZUFPA). The chromosomal preparations were obtained from bone marrow after Colchicine treatment following Ford and Hamerton, [26]. We also obtained metaphases from a fibroblast cell culture of Mus musculus (MMU) in order to define some hybridizations not described previously in the literature.

Chromosomal banding
Conventional staining was used for diploid (2n) and fundamental number (FN) determination. G-banding followed the saline solution (2xSSC) incubation method [27]. The metaphases were stained with Wright's solution after treatment with 2xSSC. C-banding was carried out according to Sumner [28].
Chromosome painting was performed following the protocol previously described [20,29], with some adaptations. Briefly, the slides were incubated in pepsin solution, and dehydrated in an ethanol series (70%, 90% and 100%), air-dried and aged in a 65°C incubator for two hours. Chromosomal DNA was denatured in 70% formamide/2xSSC at 70°C for 60 seconds, followed by preannealing the probes for 30 minutes at 37°C. The slides immersed immediately in cold 70% ethanol for 4 minutes followed by the ethanol series described above. After hybridization for 48-72 hours at 37°C (72-96 hours at 37°C for Mus musculus) and washing the slides (2x formamide 50%, 2x (2xSSC), 1x (4xSSC)/Tween at 38-40°C), the metaphases were stained with DAPI. Images were captured using the Axiovision 3.0 software with a CCD camera (Axiocam) coupled on a Zeiss-Axiophot 2 microscope or with a software Nis-Elements on a Nikon H550S microscope. Adobe Photoshop CS4 software was used for image processing.

Analysis of shared syntenic blocks
We compare our results of cross-species painting to results from the species already mapped for H. megacephalus probes, namely Cerradomys langguthi (CLA; [20]), Akodon montensis (AMO) and Tapthomys nigrita (TNI, [18]), in order to demonstrate shared syntenic blocks in Sigmodontinae. The existing regions of homeology between the karyotypes of Necromys lasiurus and Mus musculus were taken from Hass et al. [22] and Guilly et al. [30]. This permitted a more complete comparative analysis between the karyotype of NLA and other species, which had been painted with MMU probes ( [21-22, 24, 30-46]; S1 Table).

Karyotypes and distribution of heterochromatin (HC)
Necromys lasiurus presented 2n = 34 and FN = 34, consisting of fifteen acrocentric and one small metacentric pair. The X and Y are a medium and small acrocentrics, respectively.
Akondon sp. presented 2n = 10 and FN = 14, consisting of two large metacentric pairs, a large acrocentric and a small metacentric pair. The X and Y are a medium and small acrocentrics, respectively.
The Mus musculus karyotype was standard with 2n = 20, in which all pairs are acrocentric. In species NLA, ASP and MMU HC is located in the pericentromeric region of all chromosome pairs, the exception being the Y that is fully heterochromatic (data not shown).

FISH with HME probes on ASP
All probes from HME hybridized to metaphases of ASP (2n = 10 and FN = 14), revealing 45 homologous segments (Fig 2a-2e and S2 Fig). Almost all chromosome pairs of ASP hybridized with more than one HME probe, the exceptions being ASP4 and X, corresponding to HME26 and HMEX, respectively (Fig 2a).  (Fig 2a-2e). HME The FISH of probes HME14, HME (16,17) and HME26 on MMU Since the homologies between NLA2proximal+medial, NLA16 and the genome of MMU were not defined previously [22], we hybridized HME14, HME (16,17) and HME26 to MMU chromosomes to demonstrate that these regions are homologous to NLA2 and NLA16 (Fig 1a).
The probe HME14 showed two signals, in MMU10 and MMU16 (Fig 1g); HME(16,17) also showed two signals, with HME16 homeologous to MMU12 and HME17 to MMU5 by FISH and G-banding (Fig 1f and S3 Fig); HME26 did not show any signal in MMU.

Discussion
The genome of NLA The karyotype of NLA (2n = 34/FN = 34) mapped here with HME probes is the same as previously mapped with MMU probes [22]. In that study not all MMU probes hybridized to NLA, because the long time after the separation of the two species, between 23.3 and 24.7 Mya [3,22], probably resulted in a high divergence of DNA sequences.
In the present study we found all syntenies between the karyotypes of HME and NLA, including rearrangements not found before in mapping between MMU and NLA [22], such as the insertion or inversion that led to NLA15 (HME19/14/19 ; Fig 1a) or the translocation that led to NLA14 (HME5/22; Fig 1a). In the map of Hass et al. [22], chromosome NLA16 did not show any homology with MMU. Here NLA16 was homologous to HME26, showing conserved synteny by both FISH and GB (Fig 1a and S3 Fig). Furthermore, we find that MMU2 is homologous only to NLA2medial+distal, and not to all NLA2 as suggested by Hass et al. [22] (Fig 1e); the MMU2proximal is homologous to MMU5 and MMU16 (Fig 1a and 1e-1g). Therefore, we redefined the existing homology between MMU and NLA2, where NLA2 is homologous with the association MMU5/16/2 (Fig 1a and 1e-1g).
Our comparative analyses for FISH and G-banding among NLA, HME and MMU revealed that NLA1 is the result of an in tandem fusion between MMU5 and MMU9 and a Robertsonian translocation between MMU9 and MMU14 (S3  (Fig 1a and S3 Fig). These in tandem fusions and translocations support the Barros et al. [47] hypothesis that NLA reduced its diploid number during the Akodontini radiation.
The karyotype of Akodon sp. (2n = 10) studied here has the same diploid number, a similar morphology and the same syntenic groups of Akodon sp. (2n = 10) as were reported by Ventura et al. [17]. The extrapolation of our mapping in ASP to the mapping performed by Ventura et al. [17] in ASP (S1 Table) revealed that the two karyotypes have syntenic groups distributed in a different order (Fig 3). We observe that the chromosome pairs are homologous, but the chromosome ASP1 differs in three complex rearrangements, such as inversions and/or insertions involving large syntenic blocks (Fig 3). ASP2 differs by two inversions and/or insertions surrounding the block HME18/16/4/11/16/4/18 (ASP2, this study; Fig 3) or HME4/11/ 16/18 (ASP2, [17] ; Fig 3). ASP3 differs by an inversion involving HME8/22/5/17 (Fig 3) and an inversion and/or insertion occurring in block HME24/22/24/5 (ASP3; this study) or HME5/ 24/22 ([17] ; Fig 3). ASP3 differs also for a possible translocation, since in our sample there is a segment of HME25, not found in the sample of Ventura et al. [17]. The fourth pair and the sex chromosomes do not show any differences. These rearrangements suggest that the animals in the two reports are from different species, despite their similar karyotypes. The samples were collected from localities that are one thousand kilometers apart. It is noteworthy that the diploid number is the same and that most of the rearrangements are intrachromosomal. It may be that there is a mechanism that maintains a stable diploid number. These rearrangements may contribute to reproductive isolation, since they can cause meiotic problems during gametogenesis in eventual hybrids generated from these two cytotypes [48][49][50].
The genome of NLA vs. ASP Our comparison between NLA and ASP revealed that the three major pairs of ASP (ASP1, ASP2 and ASP3) originated from complex rearrangements involving multiple insertions, translocations, fusions in tandem and inversions involving NLA pairs (Fig 4).
Our data suggest that ASP had its diploid number reduced when compared to NLA, in agreement with Ventura et al. [17]. Barros et al. [47] suggested that the ancestral karyotype in Akodontini had a high diploid number (2n = 52) and had a tendency to reduction in many species, as in ASP and NLA. When compared to the Oryzomyini HME (2n = 54) our data are in agreement with this proposition.

The genome of Sigmodontinae
The 11 syntenic blocks in HME that are shared with other Sigmodontinae can be potential markers in phylogenetic analyses ( Table 2). The associations HME2/18, HME3/25 and HME18/  [17]. The existing homology between the karyotype ASP (2n = 10) studied by Ventura et al. [17] and HME was determined and based upon extrapolation of our data to data in the literature, as established in S1  25 seem to be shared only by Akodontini, suggesting that these characters are unique to that tribe (Table 2). HME 26, HME1/12, HME6/21 and HME20/13 are shared for almost all species studied here (TNI does not share HME1/12; Table 2) and possibly are part of the ancestral genome of Sigmodontinae. HME7/9 is found in CLA (Oryzomyini), is shared with NLA (HME7/9 in NLA) and AMO (Akodontini) and probably is part of the ancestral genome of Sigmodontinae too ( Table 2). HME4/11/16 is found in NLA (HME4/11/16 in NLA; Table 2), ASP and AMO, while in CLA and TNI only HME11/16, is found ( Table 2). This difference can be the result of a translocation of HME4 in CLA and TNI to another region of its genome, or the translocation of HME4 to HME11/16 in NLA, ASP and AMO. Probably HME11/16 is the ancestral form, since it is found both on Oryzomyini and Akodontini, while HME4/11/16 can be an Akodontini trait (  (Table 2). This difference may be the result of an inversion in CLA or ASP. In NLA there is HME5/22 probably because of a translocation of HME17 to another region. It is not possible to determine the ancestral position of this block.
The huge variability of the genome structure of Sigmodontinae makes the reconstruction of the ancestral karyotype by classic cytogenetics (and so the understanding of the evolutionary history of its genome) an impossible task. However, the chromosome painting technique allows the precise visualization of the homology of syntenic blocks in many species. At the moment there are only a few studies in Sigmodontinae that use this approach, but these studies already have provided some relevant information. We have used the MMU genome as a reference for rodents and can now suggest that MMU2entire, MMU6proximal/12entire, MMU3/18,  [21] in NLA. Arrow indicates the region, which does not hybridize to any MMU. Invers: inversion. b) Chromosomes of NLA (NLA9, NLA14 and NLA15) which homologous segments were identified by FISH but are not G-banding conserved. Dark lines delimit the conserved regions. Adapted from Nagamachi et al. [19] and Guilly et al. [29]. (JPG) S1 Table. Homeologies among Akodon paranaensis (APA; [17]) and Hylaeamys megacephalus (HME; [19]) whole chromosome probes according to Fig 3A in Suarez et al. [18].