Dorcadion axillare

Dascălu, Maria-Magdalena, Caba, Florina-Georgiana & Fusu, Lucian, 2022, DNA barcoding in Dorcadionini (Coleoptera, Cerambycidae) uncovers mitochondrial-morphological discordance and the hybridogenic origin of several subspecies, Organisms Diversity & Evolution (New York, N. Y.) 22 (1), pp. 205-229 : 220-221

Dorcadion axillare , D. murrayi and D. pusillum

Dorcadion axillare , D. murrayi and D. pusillum represent an interesting case of hybridisation followed by mitochondrial capture. The mtDNA introgressed massively across species boundaries even if nuclear gene flow seems to be restricted. In the absence of extensive nuclear marker data, we base this conclusion on the morphological distinctiveness of the three species. This is also confirmed by the discrepancy between the combined 28S and ITS2 trees and the COI tree that show a clear case of cytonuclear discordance. Both the 28S and ITS2 sequences are distinct between D. murrayi and D. axillare axillare , D. axillare moldavicum having the nuclear sequences clustering with those of the nominotypical subspecies, unlike the COI sequences (Figs. 1 and S2). The 28S sequence is identical for nominotypical D. pusillum and D. murrayi likely reflecting their recent split, and thus uninformative for the affinity of D. pusillum ochrolineatum and D. pusillum vasiliscus ( Fig. S2 ). The ITS2 sequences, however, are identical/very similar for all D. pusillum subspecies, and distinct from D. murrayi ( Fig. S2 ). Hence, both the morphology and the analysis of the nuclear genes (Fig. 1) confirm that D. axillare moldavicum , D. pusillum ochrolineatum and D. pusillum vasiliscus are indeed subspecies of D. axillare and D. pusillum as originally described, and not of D. murrayi .

The range of D. axillare extends from Bulgaria to S Romania and continues to NE Romania where a separate subspecies, D. axillare moldavicum , was described ( Dascălu & Fusu, 2012) ( Fig. 5 ). In our COI trees, the populations of the two subspecies did not cluster together: D. axillare axillare forms a distinct barcode cluster while D. axillare moldavicum is nested within both D. murrayi and D. pusillum . Most populations have D. murrayi mitochondria except the population from Chirceşti which has D. pusillum pusillum mitochondria (Fig. 3). This last population was considered transitional by Dascălu and Fusu (2012). In all the localities of D. axillare moldavicum , including the type locality, D. murrayi is missing. The historically heavily forested Bârlad Plateau ( Fig. 5 ) acts as a geographic barrier that impedes the spread of D. murrayi to the north ( D. axillare crossed it somehow). In this context, the probability of D. axillare moldavicum being a first-generation hybrid is zero. Knowledge about the introgression also elucidates the origin of the characters used by Dascălu and Fusu (2012) to differentiate D. axillare moldavicum from the nominotypical subspecies. The bigger and more elongated body and the longer pronotal spines were most likely taken from D. murrayi . Because of its hybridogenic origin we prefer to consider this taxon as a subspecies rather than as a species ( Dorcadion moldavicum ) following Danilevsky (2020).

In most animal species, mitochondria are passed to the next generation by females. A possible explanation of the unidirectional mitochondrial transfer lies in the contrasting mobility of the two sexes in Dorcadionini . Males are more mobile than females, as shown for I. fuliginator by Baur et al. (2005), and hence, the likelihood for them to cross a geographic barrier such as the Danube River is higher. Our hypothesis is that stranded males of D. axillare axillare migrating from what is now Bulgaria encountered and mated with D. murrayi females, a species distributed north of the Danube. Even if some D. axillare axillare females also migrated to the north, the few mitochondria they passed to the next generations were likely lost by genetic drift or are rare and we could not detect them. Most documented hybridisation cases in animals are explained by the so-called Hubbs principle, stating that the acceptance rate of heterospecifics as mating partners is correlated to the rarity of conspecifics (Willis, 2013). Additionally, as shown by Richmond (2014) for Drosophila , it is possible that crosses between D. axillare axillare females and D. murrayi males are not possible due to differences in genitalia size, D. axillare axillare being a much smaller species.

In contrast to D. axillare , the two subspecies of D. pusillum endemic to Romania have their distribution area completely overlapping with D. murrayi ( Fig. 5 ); however, they are frequently separated at the landscape level. Dorcadion pusillum prefers lowland habitats with slightly salty soils along small rivers, while D. murrayi prefers more steppelike habitats on hill slopes ( Dascălu, 2018 and author’s unpubl. data). That is why in the type locality of D. pusillum ochrolineatum , D. murrayi was not found, while in the type locality of D. pusillum vasiliscus , it is present in the area but not in the same place ( Dascălu, 2018 and author’s unpubl. data). However, at Spătaru Forest ( Fig. 5 ), both species share the same habitat and populations here could represent a hybrid swarm with introgression being an active phenomenon. The hybridisation was probably accelerated by the human-mediated encounter of otherwise ecologically distinct species, following regulation of water levels.

Similar to D. axillare moldavicum , in D. pusillum vasiliscus and D. pusillum ochrolineatum , the morphological characters used by Dascălu (2018) to distinguish them from the nominotypical subspecies are most likely derived from D. murrayi : reduction of dorsal and humeral elytral stripes and spots and a darker antenna.

Dorcadion murrayi and D. pusillum pusillum that generated the hybridogenic taxa D. pusillum vasiliscus and D. pusillum ochrolineatum are remarkably close genetically. The pairwise distance between D. murrayi and nominotypical D. pusillum of only 0.77% is the smallest of all species pairs that we examined. However, they are not particularly close based on morphology and we suspect that another hybridisation took place before the ongoing one, leading to the high similarity of their DNA sequences. Our best estimation in BEAST indicates that the COI sequences of the two species diverged only about 0.47 Myr ago. Within the current Quaternary glaciation, this would correspond to five interglacial periods ago, since the Quaternary is characterised by interglacials separated by periods of 0.1 Myr ( Mudelsee & Stattegger, 1997; Jahn et al., 2003). Surprisingly, the end of the Marine Isotope Stage 13 (MIS 13, a warm interglacial period) is also dated at about 0.47 Myr ago ( Railsback et al., 2015). The cooling of the climate after this stage and the onset of a glacial period are the likely driving forces behind the observed genetic divergence. Also, the MIS 14 glacial period that preceded MIS 13 was much warmer than other glacial epochs and this blended MIS 13 with the MIS 15 interglacial, resulting in a long and warm ‘super-interglacial style’ climate 0.621 –0.478 Myr ago ( Hao et al., 2015). This certainly led to considerable range expansion of the species followed by range contraction at its ending. One species captured the mitochondria of the other, and they diverged afterwards in allopatry until they met once again during the present interglacial and hybridised once more. Even if our molecular clock estimates are not accurate, it is still highly likely that the distribution areas of D. murrayi and D. pusillum overlapped at some another time in the past during their population retreat and expansion pursuing climatic oscillations.

Given the apparent ease with which these three species can hybridise, one can ask whether they are truly distinct species. A pro argument is the endophallus, characteristic and distinct for each of them ( Dascălu, 2018; Dascălu & Fusu, 2012); combined with differences in habitus and colour, this clearly indicates that they are distinct species. Based on biogeographic data, D. axillare is a Balcanic species ( Dascălu & Fusu, 2012), and most of the distribution range of D. pusillum coincides with the Pontic-Caspian steppe ( Dascălu, 2018), while D. murrayi is known almost exclusively from Romania ( Fig. 5 ). Therefore, it is likely that the three species evolved in allopatry and their range overlap is recent, caused by colonisation of new areas after the end of the last Ice Age (Taberlet et al., 1998). Since they were initially allopatric, no assortative mating mechanisms to prevent hybridisation evolved, but they may appear by reinforcement ( Mallet, 2005) now that the species are partly sympatric. The continuous introgression between the three species is delayed by two geographic barriers: the Prut and Danube rivers. Even if they were crossed by D. pusillum and D. axillare , the third species, D. murrayi , mostly did not cross them and hence it did not colonise the areas south and east of Romania. Outside Romania, this species is known from one locality in Serbia, close to the Danube ( Ilić & Ćurčić, 2015) ( Fig. 5 ).

Figures

Fig. 2 Maximum clade credibility ultrametric tree generated with Bayesian inference in BEAST from 162 Dorcadionini COI sequences (152 from this study, 10 from GenBank). Clades with numerous sequences and no delimitation issues are collapsed, with numbers

Fig. 5 World distribution of D. murrayi (black circles) and of D. axillare (blue circles; violet shading is subspecies moldavicum) and western limit of the distribution of D. pusillum (red circles; red shading is subspecies ochrolineatum, and yellow shading is subspecies vasilis-