Rhinolophus clivosus

Uvizl, Marek, Varadínová, Zuzana Kotyková & Benda, Petr, 2024, Phylogenetic relationships among horseshoe bats within the Rhinolophus ferrumequinum group (Mammalia, Chiroptera), Zoologica Scripta 53 (3), pp. 249-266 : 259-260

4.2 | Relationships between Rhinolophus clivosus and R. ferrumequinum

Previous studies have indicated a rather complex relationship between R. ferrumequinum and R. clivosus (Benda & Vallo, 2012; Demos et al., 2019; Dool et al., 2016). In our study, we analysed mitochondrial and nuclear sequences across the whole distribution range of R. clivosus and most of the distribution range of R. ferrumequinum (Figure 1). Notably, we extensively sampled the area where both species are in close proximity. Although we found R. ferrumequinum and R. clivosus to be separate clades in the nuclear tree, this was not the case for the mitochondrial dataset. The mitochondrial tree showed four major lineages of R. clivosus (ETH, ESA, ARS, NAL) , with sequences from the Sahara and the Levant (NAL) mixed as a sub-lineage within the R. ferrumequinum clade. The mitochondrial tree agrees with the topology recovered by Benda and Vallo (2012) and Dool et al. (2016), where R. clivosus individuals from Egypt and Jordan, and Algeria, respectively, were mixed with R. ferrumequinum . Additionally, the ARS clivosus lineage was found sister to the mixed lineage of R. ferrumequinum and NAL clivosus in the mitochondrial tree. On the other hand, all Palaearctic R. clivosus (NAL + ARS lineage) sequences were ascertained to be monophyletic in the nuclear tree. We suggest that the mitochondrial genome of R. clivosus from southern Arabia is the original genome set and it was replaced by the mitochondrial genome of R. ferrumequinum in the northern part of the range (Levant), where both species live in parapatry or limited sympatry ( Benda et al., 2010; Mendelssohn & Yom-Tov, 1999; own unpubl. data). It is noteworthy that two groups of clivosus, NAL and ARS , were found partially connected in the phylogenetic network indicating that a phylogenetic signal is still present in mtDNA.

This discordance between mitochondrial (mtDNA) and nuclear (nucDNA) genomes has increasingly been reported as the use of both types of markers has become a standard ( Toews & Brelsford, 2012). This discordance could be caused by several mechanisms. If the differences between mitochondrial and nuclear markers show a geographic pattern, the discordance is mostly caused by an introgression of a genome part of one species into the genome of another species ( Mao & Rossiter, 2020; Toews & Brelsford, 2012). On the other hand, if no evidence for a geographic signal is apparent in the discrepancy between the genomes, it could be due to incomplete lineage sorting (e.g. Funk & Omland, 2003). Moreover, the mtDNA introgressions occur more often than the nucDNA introgressions ( Bachtrog et al., 2006; Klymus et al., 2010), and are typically asymmetric from a donor species to a receiving species ( Mao et al., 2010), and/or from a native species to an invading species ( Currat et al., 2008). In bats, the assumed evidence for mtDNA introgression is increasing and is even suggested from the family Rhinolophidae (e.g. Mao & Rossiter, 2020; Mao et al., 2010, 2013; Sun et al., 2016; Taylor et al., 2018), as well as from other bat families such as Vespertilionidae (e.g. Berthier et al., 2006; Juste et al., 2013; Vallo et al., 2012), Mormoopidae ( Méndez-Rodríguez et al., 2021) , or Pteropodidae ( Nesi et al., 2013) , and other mammals such as hares, deer or bears ( Edwards et al., 2011; Melo-Ferreira et al., 2009; Senn & Pemberton, 2009); for a review see Toews and Brelsford (2012).

Therefore, we suggest that the mtDNA introgression occurred at the present parapatric contact zone of R. ferrumequinum and R. clivosus , as only one clivosus population (NAL clivosus ) showed a closely related mtDNA with R. ferrumequinum . Additionally, according to Currat et al. (2008), the invading species would be R. clivosus which was supposedly moving northward until it encountered the local R. ferrumequinum . The distribution of R. ferrumequinum now follows the Mediterranean and Irano-Turanian bioclimatic zones whereas R. clivosus is more commonly found in the Saharo-Arabian climatic zone ( Asouti et al., 2015; Miebach et al., 2019; Zohary, 1973). The Irano-Turanian biome expanded especially during colder periods (supposedly during the Pleistocene glacials) and the Saharo-Arabian biome expanded during dry and warmer periods ( Miebach et al., 2019). Therefore, it is only possible to hypothesise that R. clivosus expanded northward and encountered the resident R. ferrumequinum with which it hybridised during interglacials. Alternatively, the asymmetric introgression could result from the body and genitalia size differences as was demonstrated with Chinese Rhinolophus ( Mao et al., 2013) . In China, the smaller R. sinicus and R. thomasi introgressed the mtDNA from the larger R. septentrionalis which resembles our case where the mtDNA of the larger R.ferrumequinum introgressed into the smaller R. clivosus . Altogether, in our case, the introgression of mtDNA between R. ferrumequinum and R.clivosus would be deeply historical because (1) their distribution ranges do not overlap currently, and (2) they do not share haplotypes when the smallest genetic distance between ferrumequinum and NAR clivosus is 1.0% of the Cyt -b gene. When the genetic distances were compared with both time-calibrated trees, we could speculate that the introgression might occur around 0.3–0.6 Ma, that is, in the last second to last fourth interglacial period.