Chromosomal and DNA barcode analysis of the Polyommatus (Agrodiaetus) damone (Eversmann, 1841) species complex (Lepidoptera, Lycaenidae)

Abstract The Polyommatus (Agrodiaetus) damone (Eversmann, 1841) species complex comprises from 5 to 8 species distributed in southeastern Europe and southern Siberia. Here we used chromosomal and DNA-barcode markers in order to test the taxonomic hypotheses previously suggested for this complex. We revealed that all taxa within this group demonstrate chromosomal stasis and share the same or very similar haploid chromosome number (n = 66 or n = 67). This finding is unexpected since the karyotypes are known to be very diverse and species-specific within the other taxa of the subgenus Agrodiaetus Hübner, 1822. Analysis of the mitochondrial gene COI revealed six diverged clusters of individuals within the complex. Each cluster has a specific geographic distribution and is characterized by distinct morphological features in the wing pattern. The clusters mostly (but not always) correlate with traditionally recognized species. As a result of our study, we describe a new subspecies P. (A.) iphigenides zarmitanussubsp. nov. from Uzbekistan and Tajikistan and show that the taxon originally described as Lycaena kindermanni var. melania Staudinger, 1886 represents a subspecies P. (A.) iphigenides melanius (Staudinger, 1886). Polyommatus (A.) samusi Korb, 2017 (syn. nov.) and P. (A.) melanius komarovi Korb, 2017 (syn. nov.) are considered here as junior subjective synonyms of P. (A.) iphigenides iphigenides (Staudinger, 1886).

Here we analyzed karyotypes and mitochondrial DNA-barcodes of all species of the P. (A.) damone complex in order to test the taxonomic hypotheses previously suggested for this group (see the references above).
The taxa P. (A.) damone walteri Dantchenko et Lukhtanov, 1993, P. (A.) damone fabiani Bálint, 1997 and P. (A.) damone bogdoolensis Dantchenko et Lukhtanov, 1997 are not considered in this paper since neither chromosomal nor molecular data are available. This also applies to P. (A.) carmon altaiensis (Forster, 1956), recently treated by Eckweiler and Bozano (2016) as a separate species. All these taxa represent the most eastern populations of the P. (A.) damone complex distributed in Mongolia, Altai and southwestern Siberia. Morphologically they are close to other populations of P. damone or to P. mediator Dantchenko et Churkin, 2003. Their study will become possible in the future as soon as the material suitable for molecular and chromosomal analyses becomes available.

Molecular methods and DNA barcode analysis
Standard COI barcodes (658-bp 5' segment of mitochondrial cytochrome oxidase subunit I) were studied. COI sequences were obtained from 44 specimens representing the P. damone species group and from two samples [P. damon (Denis et Schiffermüller, 1775) and P. icarus (Rottemburg, 1975)] which were selected as outgroup (Table 1). Legs were sampled from these specimens, and sequence data from the DNA barcode region of COI were obtained at the Canadian Centre for DNA Barcoding (CCDB, Biodiversity Institute of Ontario, University of Guelph) using protocols described in Hajibabaei et al. (2005), Ivanova et al. (2006) anddeWaard et al. (2008). Specimens examined are deposited in the Zoological Institute of the Russian Academy of Sciences, St. Petersburg, Russia and in the McGuire Center for Lepidoptera and Biodiversity We also used 31 published COI sequences (Wiemers 2003;Kandul et al. 2004Kandul et al. , 2007Lukhtanov et al. 2005Lukhtanov et al. , 2009Vodolazhsky et al. 2011;Vodolazhsky and Stradomsky 2012) which were downloaded from GenBank (Table 1).
Sequences were aligned using the BioEdit software (Hall 1999) and edited manually. Phylogenetic hypotheses were inferred using Bayesian inference as described previously (Vershinina and Lukhtanov 2010;Przybyłowicz et al. 2014;Lukhtanov et al. 2016). Briefly, the Bayesian analysis was performed using the program MrBayes 3.2 (Ronquist et al. 2012) with default settings as suggested by Mesquite (Maddison and Maddison 2015): burn-in = 0.25, nst = 6 (GTR + I + G). Two runs of 10,000,000 generations with four chains (one cold and three heated) were performed. We checked runs for convergence and proper sampling of parameters [effective sample size (ESS) > 200] using the program tracer v1.7.1 (Rambaut et al. 2018). The first 25% of each run was discarded as burn-in. The consensus of the obtained trees was visualized using FigTree 1.3.1 (http://tree.bio.ed.ac.uk/software/figtree/).

Chromosomal analysis
Karyotypes were studied in 16 adult males representing four species (Table 2) and were processed as previously described Vishnevskaya et al. 2016). Briefly, gonads were removed from the abdomen and placed into freshly prepared fixative (3:1; 96% ethanol and glacial acetic acid) directly after capturing the butterfly in the field. Testes were stored in the fixative for 3-36 months at +4 °C. Then the gonads were stained in 2% acetic orcein for 30-60 days at +18-20 °C. Different stages of male meiosis, including metaphase I (MI) and metaphase II (MII) were examined using an original two-phase method of chromosome analysis (Lukhtanov et al. 2006(Lukhtanov et al. , 2008. Abbreviation ca (circa) means that the count was made with an approximation due to an insufficient quality of preparation or overlapping of some chromosomes or bivalents.
Leica DM2500 light microscope equipped with HC PL APO 100×/1.44 Oil CORR CS lens and S1/1.4 oil condenser head was used for bright-field microscopy analysis. Leica DM2500 light microscope equipped with HC PL APO 100×/1.40 OIL PH3 lens was used for phase-contrast microscopy analysis.  (Forster, 1960) and four males of P. (A.) iphigenides zarmitanus subsp. nov. revealed the same (or almost the same) haploid chromosome number n = 66 or n = 67 in all studied taxa ( Table 2). The karyotype structure was also found to be identical in all studied individuals, with three large bivalents in the center of metaphase plates (Fig. 2). Bivalent 1 was 1.2-1.5 times larger than bivalent 2, and the latter was 1.2-1.5 times larger than bivalent 3.

Chromosomal stasis
It has been found that all taxa within P. (A.) damone species complex demonstrate chromosomal stasis and share the same or very similar haploid chromosomal number (n = 66 or n = 67). This result is unexpected since the karyotypes are known to be very diverse and species-specific in the subgenus Agrodiaetus.
It is believed that an unusual diversity of karyotypes is the most remarkable characteristic of Agrodiaetus. Species of this subgenus exhibit one of the highest ranges in chromosome numbers in the animal kingdom (Vershinina and Lukhtanov 2017). In Agrodiaetus haploid chromosome numbers (n) range from n = 10 in P. (A.) caeruleus (Staudinger, 1871) to n = 134 in P. (A.) shahrami (Skala, 2001) (Lukhtanov et al. 2005). The genus Polyommatus as a whole shows numbers from n = 10 to n = 226 (Lukhtanov 2015). Additionally, the subgenus Agrodiaetus demonstrates a high level of karyotypic differentiation with respect to chromosome size (Lukhtanov and Dantchenko 2002b) and variation in number of chromosomes bearing ribosomal DNA clusters (Vershinina et al. 2015). These differences provide reliable characters for species delimitation, description and identification (de Lesse 1960(de Lesse , 1963Lukhtanov and Dantchenko 2002a, b).
The P. (A.) damone species complex represents an exception. In this group divergence in several phylogenetic lineages was not accompanied by changes in karyotypes, and the chromosome number n = 66-67 is the synapomorphic character for the species of the group.

DNA-barcode clusters
The DNA-barcode clusters revealed in our study correspond well to traditionally recognized species and certain specific geographic areas (Figs 3, 4)    . Distribution areas of the COI clusters revealed in this study. Cluster 1 corresponds to P. damone. Cluster 2 corresponds to P. juldusus + P. mediator. Cluster 3 corresponds to P. iphigenides (including P. iphigenides melanius). Cluster 4 corresponds to P. zarmitanus. Cluster 5 corresponds to P. karatavicus.
Up to our knowledge there are no data on sympatry of P. (A.) mediator and P. (A.) damone in Mongolia as it was reported or supposed earlier (Bálint and Johnson 1987;Bálint 1989).

Cluster 3 (P. iphigenides iphigenides + P. iphigenides melanius)
Polyommatus (Agrodiaetus) iphigenides is highly polymorphic with regard to the black suffusion on the wing upperside and the marginal and submarginal part of the wing underside in males as well as the white streak on hindwings in both sexes. In extreme cases, the suffusion can be practically absent resembling the upperside in P. damone or may extend almost to the discal spot which is observed as a fixed feature in two other taxa, P. iphigenides melanius and P. juldusus kirgisorum. The white streak is also very variable from clear visibility to complete absence. The taxa P. (A.) samusi Korb, 2017 (syn. nov.) and P. (A.) melanius komarovi Korb, 2017 (syn. nov.) are mainly described on the base of such extreme forms of the same population. Therefore, we consider these taxa as junior subjective synonyms of P. (A.) iphigenides iphigenides.
Cluster 3 also includes the taxon described as Lycaena kindermanni var. Melania Staudinger, 1886. For a long time, due to lack of material it had been considered to be a melanized form of P. (A.) iphigenides iphigenides (e.g. Forster 1960). But in recent years it has been treated as a separate species P. (A.) melanius with a local, nearly dot-like distribution in the border area between southwestern Kyrgyzstan and eastern Tajikistan in the Kyzylsu/Surkhob River basin (Dantchenko 2000;Eckweiler and Bozano 2016). We found that DNA barcodes of P. (A.) iphigenides and P. (A.) melanius are identical or differ by non-fixed 1-2 nucleotide substitutions. The main feature of P. (A.) melanius, a wide dark marginal border on the fore-and hindwings, is quite stable for the diagnosis of the taxon; however, the tendency towards such a wide border is expressed in different populations of P. (A.) iphigenides, too. Therefore, this trait can be hardly considered a species-specific character. Here we argue that P.(A.) melanius is rather a subspecies P. (A.) iphigenides than a species. However, this is not a final conclusion. There is indirect evidence in favour of a possible species status of P. (A.) melanius, e.g. the distribution areas of P. (A.) iphigenides iphigenides and P. (A.) iphigenides melanius almost touch each other, and an intergradation zone would be expected between them. However, such a zone is still unknown, and specimens of P. (A.) iphigenides iphigenides and P. (A.) iphigenides melanius from very close localities are clearly differentiated. We suppose that genome-wide analysis may be useful to verify the taxonomic status of P. (A.) iphigenides melanius.

Cluster 4 (P. iphigenides zarmitanus)
Morphologically this group is close to P. ipigenides iphigenides, whereas with regard to mitochondrial DNA it is close to sympatric species P. phyllides which is morphologically very different. In our opinion, two alternative evolutionary scenarios can explain this pattern.

Scenario 1
The cluster 4 (P. iphigenides zarmitanus) and the lineage 6 (P. phyllides) are sister species which recently evolved from a common ancestor by means of sympatric speciation.

Scenario 2
Cluster 3 (P. iphigenides) and cluster 4 (P. iphigenides zarmitanus) are sister taxa evolved in allopatry; therefore, they share an ancestral type of the wing pattern and coloration, although differentiated with respect to DNA barcodes. The similarity between completely sympatric cluster 4 (P. iphigenides zarmitanus) and lineage 6 (P. phyllides) is a result of ancient mitochondrial introgression.
Analysis of multiple nuclear markers is required in order to distinguish between these two scenarios. Scenario 2 seems to be more probable since mitochondrial introgression is not a rare phenomenon in butterflies (e.g. Gompert 2008;Cong et al. 2017) and is also documented in the subgenus Polyommatus (Agrodiaetus) . Therefore, below we describe the new lineage discovered in West Hissar region as a subspecies of P. iphigenides.
Genitalia. The male genitalia have a structure typical for other species of the subgenus Agrodiaetus (Coutsis 1986, Eckweiler andBozano 2016).
Females. (Fig. 6a, b) Forewing length 15-17 mm. Upperside: Ground color brown with slightly darker veins, discal strokes present, submarginal and antemarginal marking almost absent on fore wings and strong and clear visible on hindwings, antemarginal black spots on hindwings bordered with orange lunules, fringe whitish.
Underside: ground color and general design as in males but darker, brownish grey, greenish blue basal suffusion near invisible, white streak on hindwings clear visible, enlarged distally, fringe light greyish.
Diagnosis. The new subspecies is distinguished phenotypically from the most similar P. iphigenides iphigenides (Figs 5c-f, 6c, d) by the underside of the hind wing, which has a paler and less contrasting coloration. The white streak is also dim and weakly stands out against the background of the wing, is often reduced or absent. The same can be said about the basal greenish-blue suffusion: it is dim and weakly stands out against the background of the wing; its size, on average, is much smaller than that in P. iphigenides iphigenides. As a rule, it is limited by black dots of the basal row, while in P. iphigenides iphigenides it usually extends further in the distal direction, sometimes to spots of the discal row. This suffusion itself has a more greenish tint than that in P. iphigenides iphigenides (in the latter, it is more blue). The new species always has black dots of the basal row (although they are small), while in another species they are reduced.
The main differences between the species are still in the molecular characters. Polyommatus iphigenides zarmitanus can be distinguished from P. iphigenides iphigenides by using molecular markers from the COI gene. These mitochondrial diagnostic characters are in the following positions in the COI barcode region: adenine (A) in position 22, cytosine (C) in position 132, guanine (G) in position 180, cytosine (C) in position 286, guanine (G) in position 468, guanine (G) in position 468, and guanine (G) in position 627.
The new subspecies differs from sympatric (syntopic and synchronous) P. phyllides by milky blue (not greenish blue) wing upperside and white pubescence of the costal area of the forewings in males and by light grey color of the wing underside (P. phyllides has specific warm pinkish grey color of the wing underside). It also differs from P. phyllides by diagnostic nucleotide guanine (G) in position 627 of the COI barcode region.
Distribution area (Fig. 7). Uzbekistan: West part of the Hissar Range, Zeravshan Mts, Nuratau Mts, Boysun (= Baisuntau) Mts. Tajikistan: west part of the Zeravshan valley and Zeravshansky Range, West Hisar Range. Etymology. The name zarmitanus is an adjective of the masculine gender. This name originates from Zarmitan, the village in Uzbekistan.

Taxonomic conclusion
The discovered topology (Fig. 1) can be considered as a signal to taxonomic rearrangement within the group. However, since the volume of the studied material of these taxa is small, we prefer to leave the existing taxonomic hypotheses. Additionally, we assume that the hypothesis of the existence of a species called P. (A.). altaiensis with subspecies  Dantchenko et Lukhtanov, 1997,