Mitochondrial DNA and karyotypic data confirm the presence of Mus indutus and Mus minutoides (Mammalia, Rodentia, Muridae, Nannomys) in Botswana

Abstract We use a combination of cytochrome b sequence data and karyological evidence to confirm the presence of Mus indutus and Mus minutoides in Botswana. Our data include sampling from five localities from across the country, including one site in northwestern Botswana where both species were captured in syntopy. Additionally, we find evidence for two mitochondrial lineages of M. minutoides in northwestern Botswana that differ by 5% in sequence variation. Also, we report that M. minutoides in Botswana have the 2n=34 karyotype with the presence of a (X.1) sex-autosome translocation.


Introduction
Delineating geographic distributions of African Mus (subgenus Nannomys Peters, 1876) in Sub-Saharan Africa has been especially challenging due to a combination of incomplete taxon sampling throughout the region as well as uncertainties in species of M. minutoides have recently been confirmed for Angola and Namibia (Lamb et al. 2014), providing additional support for the extended range map proposed by Britton-Davidian et al. (2012) Regarding chromosomal rearrangement in southern Africa, M. minutoides from South Africa exhibit Robertsonian fusions with two major monophyletic groups showing either a diploid number of 2n=18 -where all of the acrocentric chromosomes are fused to produce metacentric elements, or a 2n=34 -where sex-chromosome translocations have been reported (Veyrunes et al. 2010). Additionally, WARTs have been documented in several populations exhibiting the 2n=18 karyotype in South Africa, which has contributed significantly to reported chromosomal variation, with at least four different cytotypes within this clade (Veyrunes et al. 2007). Currently, the geographic distributions of the 2n=18 and 2n=34 forms of M. minutoides are not known outside of the country of South Africa (Veyrunes et al. 2010).
Our objective was to utilize material from recent collecting efforts and molecular techniques to accurately delimit which species of Nannomys occur within Botswana. Further, we describe karyotypes for individuals from this region and make comparisons with previously published data from South Africa.

Materials and methods
Our mitochondrial phylogeny was generated from combining previously published sequences deposited on GenBank (Appendix) with those derived from sequencing new specimens collected during field trips to Botswana conducted in 2008, 2009. We collected 16 specimens of Mus from five localities in Botswana including: Gcwihaba Caves (20°00.99'S; 21°15.89'E); Kang (23°32.10'S; 22°32.76'E); Koanaka Hills (20°09.60'S; 21°11.61'E); Lepokole Hills (21°49.59'S; 28°23.94'E); and Tsabong (25°56.57'S; 22°25.44'E) (Fig. 1a-c). Specimens were collected using Sherman live traps, pitfall traps, or Museum Special snap traps. Standard external measurements were recorded in the field (Table 1). Specimens were preserved as skins with complete skeletons (SSPS), skulls only, or as whole bodies in alcohol (alc.) and deposited at the at the Natural Science Research Laboratory (NSRL) at the Museum of Texas Tech University, Lubbock, Texas, USA or the Botswana National Museum, Gaborone, Botswana. Tissue samples were preserved in 95% ethanol, lysis buffer, or flash frozen in liquid nitrogen for future genomic analyses (2011 material) and deposited in the NSRL. Field collecting methods followed taxon specific guidelines for wild mammals (Sikes et al. 2012) as outlined by the Animal Care and Use Committee of the American Society of Mammalogists (Gannon et al. 2007;Sikes et al. 2011).
Genomic DNA was extracted using a DNeasy Blood and Tissue Kit (Qiagen Inc., Chatsworth, California). The complete cytochrome b gene (cytb, 1140 nucleotides) was amplified following methods outlined in Veyrunes et al. (2010). Cycle sequencing reactions were performed with BigDye terminator version 3.1 and were electrophoresed on an ABI 3100-Avant (Applied Biosystems, Foster City, California). Sequences were edited and aligned using SEQUENCHER version 4.9 (Gene Codes Corporation, Ann Arbor, Michigan). Novel sequences (GenBank accession nos. KF184308-KF184323) were aligned with previously published sequences deposited on GenBank using only individuals that exhibited unique haplotypes (Appendix). The final alignment was trimmed to exclude regions with large amounts of missing data due to the large number of GenBank sequences in the alignment that were partial cytb sequences. Therefore, a total of 741 base pairs of the cytb gene (the first 7 codons and last 126 codons were removed from the analysis) were used in the final alignment for the phylogenetic analysis including 125 individuals.
Appropriate models of evolution were examined using MEGA version 5 (Tamura et al. 2011). Phylogenetic relationships were estimated using Bayesian inference with the program MRBAYES version 3.2 (Huelsenbeck and Ronquist 2001). Four independent Markov chains were run for 50 million generations and trees were logged every 1000 th iteration. Log-likelihood values were examined in the program TRACER version 1.5 (Rambaut and Drummond 2007) and the first 5,000 trees were discarded as burn-in. An additional phylogeny was estimated using the Maximum-likelihood method with the program PhyML version 3.0 (Guindon et al. 2010) with a BIONJ starting tree (Gascuel 1997) and 1,000 bootstrap replicates. Kimura 2-parameter genetic distances were calculated using MEGA version 5 (Tamura et al. 2011). were recorded in the field using a handheld Garmin GPS Rino 120 unit using the datum WGS84. Elevations given in meters.

Genbank
No. Specimens were karyotyped in the field using bone marrow after 1 h of in vivo incubation with Velban (Sigma-Aldrich, St. Louis, Missouri), following the methods described in Baker et al. (2003). Mus indutus males were not karyotyped in this study because both males captured died in snap traps. Fluorescent in situ hybridization (FISH) experiments were performed using Star*FISH © biotin-labeled mouse chromosome X paints (Cambio), following the manufacturer's instructions and using Cy3-conjugated streptavidin (Invitrogen) for signal detection.

Species
In order to assess the nature of the X-autosome translocation of the specimens that exhibited the translocation, we compared the X-chromosome of our specimens with those from South Africa using images of inverted DAPI-banding, and G-banding (Seabright 1971). Images were captured using the GENUS SYSTEM version 3.7 (Applied Imaging Systems, San Jose, California) through an Olympus BX51 epi-fluorescence microscope. Cy3 and DAPI (4',6-diamidino-2-phenylindole) signals were pseudocolored yellow and red, respectively.

Results
The model with the lowest AICc (Akaike Information Criterion, corrected) and BIC (Bayesian Information Criterion) scores was the General Time Reversible (GTR) model using a discrete gamma distribution (+G) and a fraction of invariable sites (+I). Overall, the two methods of phylogenetic analysis resulted in similar tree topologies, except that the Maximum-likelihood analysis recovered weak support for the south + east M. minutoides clade (Fig. 2). Additionally, the relationship between M. indutus, M. sp., M. mattheyi, M. haussa, and the portion of the phylogeny that includes M. minutoides and M. musculoides was unresolved in the Maximum-likelihood analysis, though it was well-supported using Bayesian inference.
Sixteen cytb sequences were generated from specimens from Botswana, corresponding to two species. Seven individuals are phylogenetically related to M. minutoides from South Africa and nine individuals cluster with M. indutus. Five individuals, captured from the same locality in the Koanaka Hills region of northwestern Botswana, represent two clades within M. minutoides that are 5% different in cytb sequence variation (Fig. 2). Six of the individuals of M. indutus were collected in the Koanaka Hills alongside both of these lineages of M. minutoides (Fig. 1c).
Karyotypes for individuals in the M. minutoides clade exhibited a diploid number of 34 and fundamental number (as defined by Veyrunes et al. 2004 as the total number of chromosomal arms per diploid genome, instead of number of autosomal arms) of FN=36 ( Fig. 3a-d, Table 2). All autosomes were acrocentric in morphology, including the pair 13, which presented a small short arm in some metaphase spreads. The metacentric X chromosome is the largest element of the chromosome complement, followed by the subtelocentric Y chromosome, which is comparable in size with the first autosomal pair. Individuals in the M. indutus clade exhibited diploid and fundamental numbers of 36 ( Fig. 3e-f, Table 2). All chromosomes had an acrocentric morphology. Due to the lack of male karyotyped specimens, the Y chromosome morphology could not be determined. The FISH with Mus X whole chromosome probe allowed the detection of an X-autosome translocation on the karyotypes of M. minutoides specimens (Fig. 3b, d), but not for individuals of M. indutus (Fig. 3e). Banding results indicate that individuals of M. minutoides from Botswana share the same sex-chromosome translocations (X.1) and (Y.1) as M. minutoides from South Africa, although differential condensation of the South African chromosomes makes direct comparison difficult ( Fig. 3a and b).

Discussion
Efforts to resolve the geographic distributions of African pygmy mice remain in a state of flux. Regional studies involving DNA sequence data and karyotypes, such as presented here, contribute to a broader understanding of this complex genus. Historical (see Schmidt et al. 2008) and recent (Ferguson et al. 2010) bioinventories have resulted in extensive collections of Mus from Botswana, but there has been little consensus as to whether both M. minutoides and M. indutus occur in the country.
Mitochondrial sequence and cytogenetic data confirm the presence of both M. minutoides and M. indutus in Botswana. These specimens represent the first DNA sequences for these two species in Botswana, which we also made available for use in a recent paper by Lamb et al. (2014). Despite previous suggestions that M. minutoides and M. indutus occur in allopatry, our results confirm that these two species occur in sympatry and even syntopy in northwestern Botswana. Interestingly, we also found two lineages of M. minutoides in northwestern Botswana (Koanaka Hills) that were 5% different in cytb sequence variation. We hypothesize that these two mitochondrial lineages were separated in the past and have now come back together in a region of secondary contact in the arid savannah region near the Kalahari Desert, a hypothesis that should be tested with broader sampling and using additional genetic markers.
Also of interest is the fact that no M. setzeri were collected from either the Koanaka Hills or Gcwihaba Caves although their current range -as delimited by Monadjem and Coetzee (2008) and Skinner and Chimimba (2005) -includes this region of Botswana. We compared our specimens with M. setzeri deposited at the National Museum of Natural History, Smithsonian Institution, Washington D.C., USA and found no evidence that any of our individuals correspond to this conspicuous form. Our failure to capture M. setzeri, in spite of concerted trapping efforts in this region (> 2600 Sherman trap nights, > 280 pitfall trap nights during June 2008 and July 2009 seasons), is in agreement with Monadjem (2013b) who pointed to the scarcity of this species in collections as evidence for true ecological rarity. Further sampling is clearly warranted to more accurately delimit the exact geographic boundaries of Nannomys species both within Botswana and throughout the broader Southern African Subregion (Skinner and Chimimba 2005).
Mus minutoides in Botswana exhibit the 2n=34 karyotype with the diagnostic (X.1) and (Y.1) sex-autosome translocations that have also been documented in specimens from South Africa (Veyrunes et al. 2010), Zambia, Kenya (Castiglia et al. 2002(Castiglia et al. , 2006, Central African Republic, and Ivory Coast (Jotterand-Bellomo 1984, 1986. Veyrunes et al. (2004) propose that 2n=34 with the 1 sex chromosome translocation is the ancestral karyotype for M. minutoides and our results provide further support for an early (X.1) translocation before the radiation of M. minutoides over a large geographic area. Furthermore, the 2n=34 cytotype is reported in several locations in northern South Africa, but not in southern South Africa or in other countries to the north, including Botswana. The fact that our sampling localities included individuals from the easternmost and northwestern regions of Botswana might be an indicator that this is the predominant cytotype in the country, likely extending into the bordering countries of Zambia, Zimbabwe, and Namibia.
We found that three of our gender identifications made in the field (Table 2, "Gender Field") did not match the identifications made from karyotype assessments ( Table 2, "Gender Lab") indicating the potential for X*Y females. Therefore, we attempted to Table 2. Individuals of Mus (Nannomys) collected in Botswana including GenBank number, final species identification, gender determined in the field, museum preparation type (Alcoholic=alc; skin, skull, postcranial skeleton=SSPS; or Skull only), collection date, total length (TL), tail length (T), hind foot (HF), ear (E), weight in grams, karyotype, and sex-chromosome.
Genbank No. XY examine these specimens for the possibility of sex reversal in M. minutoides, which has been documented in other countries ). Although we have tried to identify the X* chromosome in our samples through X chromosome morphology assessment as well as DAPI banding patterns, the particular high degree of condensation of the chromosomes in our in vivo bone marrow preparations did not allow us to ascertain the nature of the X chromosomes of two of these three specimens. For one of the individuals, both the morphology and banding patterns of the X chromosome do not seem to correspond to those of the derivative X* chromosome (Fig. 3a), indicating that field misidentification of sex might have been the case for that specimen (the reproductive organs can no longer be clearly seen on the prepared skin of this specimen).

Species
Additionally, there were no evident X chromosome polymorphisms in the XX female specimens, which would be expected in populations where X*Y females were present.
Due to our small sample, and the relative low frequency of the X* found in populations outside South Africa, we were not able to rule out the presence of the X polymorphism in Botswana. Further collecting efforts, together with an in depth sex determination study, including high quality chromosome preparations suitable for G-banding studies, will be needed to shed further light on this issue. Our data presented here agree with previous molecular phylogenies of Nannomys, with well-defined clades representing M. minutoides and M. indutus exhibiting diploid and fundamental numbers consistent with those reported in the literature. Veyrunes et al. (2010) detected a wide range of chromosomal variation for M. minutoides in South Africa, with one particular clade presenting 2n=34, FN=36. Our M. minutoides samples display chromosome conservation as well as sequence similarity to the South African clade bearing karyotypic stasis, indicating that these specimens might be part of a widespread group chromosomally and genetically isolated from the karyotypically diverse 2n=18 M. minutoides clade. Mus indutus on the other hand exhibits a karyotype not very divergent from the proposed ancestral karyotype for Nannomys (2n=36 with all acrocentric chromosomes; Veyrunes et al. 2004), similar to many of the basal lineages included in recent molecular phylogenies (see Britton-Davidian et al. 2012).
of Environment,Wildlife,and Tourism [permit number: EWT 8/36/4 XVI (46)]. Specimens were collected under SHSU IACUC issued to MLT (permit number: 08-04-03-1005-3-01). For assistance with field support and logistics during the 2008 and 2009 trips we thank M. Gabadirwe, J. Marenga, B. Mokotedi and for 2011 we thank the Botswana National Museum, including K. Lemson and Jaime Castro Navarro of the Universidad Nacional Autónoma de México. For assistance with logistics, field support, and field supplies for all years we thank J. Marais. We would especially like to thank Mr. N. E. Mosesane of the Botswana National Museum for his support of this work. A RAPID grant award 0910365 from the United States National Science Foundation as well as Faculty Research Grants and Exploratory Grants for Research from Sam Houston State University to MLT and PJL were used to fund 2008-2009 expeditions. An Explorers Club-Exploration Fund Research Grant and Texas Academy of Science Research Grant to MMM provided partial funding for 2011 fieldwork.