Chromosomal evidence of species status and evolutionary relationships of the black fly Prosimulium petrosum (Diptera, Simuliidae) in Armenia

Abstract The karyotype of Armenian populations of the black fly Prosimulium petrosum Rubtsov, 1955 was characterized and compared with that of all other chromosomally known Palearctic members of the Prosimulium hirtipes group. Analysis of the polytene chromosomes established that Prosimulium petrosum is most closely related to European populations of Prosimulium latimucro (Enderlein, 1925) with which it shares an identical fixed chromosomal banding sequence. Its validity as a species, separate from Prosimulium latimucro, is based on its unique sex-differential sections in the expanded centromere region of chromosome I, in agreement with the unique structural configuration of the hypostomal teeth of its larvae. Prosimulium petrosum and Prosimulium latimucro, therefore, are homosequential species, demonstrating the value of a combined chromosomal and morphological approach in determining species status.


Introduction
Chromosomal rearrangements have long been considered a driving force in speciation in certain groups of organisms, based on a wealth of evidence, much of it indirect (White 1978, Rothfels 1989, Nevo 2012. If the chromosomes have played an integral role in the speciation process, individual species might be expected to carry unique signatures in their karyotype. In the dipteran family Simuliidae, the vast majority of species are chromosomally distinct from one another, even when they cannot be distinguished reliably by morphological criteria (Adler et al. 2010). Detailed banding sequences of the polytene chromosomes in the larval salivary glands of Simuliidae, consequently, have facilitated the discovery of cryptic species, provided insights into population structure and evolutionary relationships, and positioned the Simuliidae at the forefront of knowledge about the genetics of natural populations of insects (Adler and Crosskey 2015a).
Given the lack of chromosomal information for bona fide material of P. petrosum, we conducted a comparative band-by-band analysis of P. petrosum to characterize its karyotype and illuminate its taxonomic status and evolutionary relationships. In particular, we were interested in determining if P. petrosum is a species distinct from the morphologically similar European species, P. latimucro (Enderlein, 1925), or if they are conspecific.

Methods
Larvae were collected from three streams, up to about 210 km apart, in April and May in northern and southern Armenia (Table 1). The material was fixed in a 3:1 mixture of ethanol and glacial acetic acid. Pupae and adults were not collected, but 13 mature larvae with well-developed gill histoblasts were obtained. Larvae were identified morphologically as P. petrosum, based on structural characters (Rubtsov 1956, Terteryan 1968)-the apex of the median hypostomal tooth of our material was posterior to the apices of the lateral teeth, and the 16 filaments of the pupal gill were arranged on three, widely splayed primary trunks, with a branching formula of (3+3+2)+(2+2)+(2+2) or (3+3+2)+(2+1+1)+(2+2).
Polytene chromosomes from the larval salivary glands were stained using the lactoaceto orcein method (Bedo 1975), which also stained gonadal tissue. Preparations were spread by squashing chromosomes on a microscope slide. Larval gender was determined by the form of the gonads: rounded in males and elongated in females. Representative chromosomal arms and selected rearrangements were photographed under oil immersion (Figs 1-4). Composite digital images from different focal planes were made with Helicon Focus 5.3 and further processed with Adobe Photoshop CS6. The banding sequences of all six chromosomal arms were compared with the standard maps of the P. hirtipes group (Basrur 1962) and with maps of various species in the P. hirtipes group (Basrur 1959, Adler and Belqat 2001, Adler and Şirin 2014.
Fixed inversions (i.e., homozygous in all larvae) are italicized in the text and underlined on our maps; floating inversions (i.e., polymorphisms) are not italicized or underlined. Inversions identical to those identified in previous studies (i.e., IIS-6, IIS-7, IIL-9, and IIL-10) were given the same numbers assigned by Basrur (1959). Newly discovered inversions were numbered to follow the last number assigned to inversions in other species of the P. hirtipes group currently under study and as yet unpublished. Chromosomal terminology, including terms for landmarks, follows that of Basrur (1959Basrur ( , 1962.
Three morphological preparations of mature larvae were deposited in the Zoological Institute of the Russian Academy of Sciences, St. Petersburg, Russia. Additional morphological preparations and chromosomal photographs were deposited in the Moscow State Regional University, Moscow, Russia.

Results
Karyotype. In total, 64 larvae were analyzed. One larva from Site 1 chromosomally matched the banding sequence of P. rachiliense cytoform 'A' (sensu Adler and Şirin 2014). The other 63 larvae (16 males, 47 females) were assigned to P. petrosum. All larvae had a diploid number of 2n = 6, with tightly paired homologues (Fig. 1). The chromosomes were submetacentric (Fig. 1). Chromosome I (sections 1-44) was the longest, with the two arms (IS and IL) subequal in length, followed by chromosome II (sections 45-74) with the long arm (IIL) slightly longer than the short arm (IIS). Chromosome III (sections 75-100) was the shortest, with the long arm (IIIL) approximately 35% longer than the short arm (IIIS). The centromere regions of chromosomes I and II were transformed (CI t , CII t ; sensu Basrur 1959), producing an expanded, flocculent area from section 19 through section 21 (CI t ) and from the middle of section 57 through section 58 (CII t ) (Figs 1-3). The centromere region of chromosome III was not expanded (Figs 1, 4).
A single, primary nucleolus organizer was in the standard position for the P. hirtipes group, that is, in the base of IL (Fig. 2). Landmarks that remained in the standard  positions for the P. hirtipes group included the single Balbiani ring in the base of IIS (Fig. 3), "blister" in IIIS (Fig. 4), and "shield" and "triad" in IIIL (Fig. 4). A chromocenter was lacking, and supernumerary (B) chromosomes were absent.
Fixed (interspecific) inversions. The banding sequence of chromosome arms IS, IL, IIIS, and IIIL was identical with the standard banding sequence established by Basrur (1959Basrur ( , 1962 for the P. hirtipes group. Chromosome II, however, had four fixed inversions relative to the standard sequence-two overlapping inversions in the short arm, IIS-6 and IIS-7, and two tandem inversions in the long arm, IIL-9 and IIL-10 (Fig. 3). The four inversions involved 73% and 58% of the sections of IIS and IIL, respectively. IIL-9 moved the "group of 5" marker more centrally and IIL-10 reversed the polarity of the parabalbiani.
Autosomal (intraspecific) polymorphisms. Fifteen autosomal polymorphisms were detected; all were present in the heterozygous state only. These autosomal rear- 0.460 † All three collection sites combined. * Frequencies of rearrangements were calculated on the basis of the number of homologues. ‡ Three larvae, which had IIIL-dif (= IIIL-34, IIIL-35+IIIL-36+IIIL-37), also carried heterozygous deletion 100dl T , the frequency of which is accounted for separately; two additional larvae had heterozygous deletion 100dl T in the absence of IIIL-dif. § IIIL-dif was treated as a single inversion for the purpose of presenting means. rangements included 12 inversions, one heteroband (90hb), a duplication of bands (87dp), and a telomeric deletion (100dlT) ( Table 2, Figs 2, 3, 4). Four inversions in sections 96-100 of IIIL formed a complex set of loops collectively referred to as IIILdif (Fig. 4, e). The proposed breakpoints of these four inversions are shown in Fig. 4, f. All autosomal polymorphisms were expressed in low frequency (< 0.065; Table 2).
Sex-differential region. All 16 males had a heterochromatic band (20/21hc) at the junction of sections 20 and 21 and lacked conjugation in the CI t region, typically from section 20 through the beginning of section 21 (Fig. 2), although one male was unpaired from the beginning of section 19 to the beginning of section 21; no inversion could be discerned in the unpaired region. Females lacked the heterochromatic band and exhibited complete pairing of homologues in the CI t region. Thus, the expanded centromere region of chromosome I was the sex-differential segment, with males X 0 Y 1 and females X 0 X 0 . In two males, ectopic pairing of CI t and CII t occurred in some nuclei.

Discussion
Our chromosomal analysis requires taxonomic context, especially a reasonable assignment of the correct species name. The larvae from our three sites in Armenia are chromosomally cohesive. Based on gill structure, they conform to previous Armenian collections (Terteryan 1968), rather than to Azerbaijanian material (Rubtsov 1956), of P. petrosum. Based on hypostomal structure, they precisely match the Azerbaijanian (holotype) (Rubtsov 1956). Our Armenian collections and the type locality of P. petrosum in Azerbaijan are 140-160 km apart, and all are in the same ecoregion-the Caucasus Mixed Forests Ecoregion (World Wildlife Federation 2015). Although a slight difference in the branching pattern of the gill between Armenian and Azerbaijanian samples possibly indicates the presence of cryptic species, we attribute the difference to intraspecific variation, which is common, especially on the dorsal trunk, in members of the P. hirtipes group (Stloukalova 2004). Similarly, although we found mature larvae 1.0-1.5 months earlier (end of April-beginning of May) than did Terteryan (1968), the seasonal difference could be attributable to altitude or perhaps climatic variation among years. Given the minimal geographic distance, dispersal ability of simuliids (Adler et al. 2005), identical ecoregion, and morphological similarity, we conclude that our Armenian populations are conspecific with the holotype.
Although our material corresponds with the type (Caucasican) concept of P. petrosum (Rubtsov 1955), a larger question is whether P. petrosum is a unique species or conspecific with P. latimucro, as suggested by Adler and Crosskey (2015b), based on morphological similarity. Yankovsky (2003) suggested that the projection of the median hypostomal tooth anterior to the lateral teeth and the second-order branching (i.e., 3+3+2) of the upper gill trunk distinguish the larva of P. latimucro from that of P. petrosum. Accordingly, our samples correspond with P. petrosum, based on the hypostomal teeth, and with P. latimucro, based on the branching of the gill.
Males and females of P. petrosum consistently differ in the expression of their CI t region, indicating the general location of the sex-determining locus. The sex chromosomes of the Simuliidae often are associated with rearrangements, such as inversions and heterobands, although the X and Y also can be microscopically undifferentiated (X 0 Y 0 ) (Rothfels 1980, Post 1982. Any of the three chromosomes (I, II, or III) can function as the sex chromosome. Identical, differentiated sex chromosomes are rarely shared between species Freeman 1983, Adler et al. 2015). Thus, the sex chromosomes can be useful in species discovery and identification (Rothfels 1989).
Lack of pairing of homologues in the CI t region, observed in males of P. petrosum, also is found in at least some populations of other Palearctic members of the P. hirtipes group, such as P. hirtipes (Fries, 1824), P. latimucro, and P. "aff. 3", and often serves as the basis for further elaboration of the Y chromosome, such as the addition of sex-linked inversions and heterobands (Basrur 1959, Adler unpublished). The heterochromatic band 20/21hc of P. petrosum also appears on the Y chromosome of various members of the P. hirtipes group, including Moroccan populations of P. rufipes and P. latimucro and some European populations of P. latimucro and P. "aff. 3", often with various repatternings of banding in the CI t region (Adler and Belqat 2001, Adler unpublished). No conspicuous repatterning was observed in the CI t region of P. petrosum. We do not know if the unpaired CI t condition and heterochromatic band 20/21hc are identical across populations and species, and if so, if their shared nature reflects common ancestry, introgression, or independent origins. Species differences in other members of the P. hirtipes group, such as those in eastern North America often are based on minor differences in the centromere region, especially of CIII (Rothfels and Freeman 1977).
A Y chromosome based on an unpaired CI t region, coupled with 20/21hc, without an associated inversion or band repatterning, uniquely characterizes P. petrosum. The allopatric nature of P. petrosum and P. latimucro, however, presents a challenge for evaluating reproductive isolation; the nearest chromosomally analyzed populations of P. petrosum and P. latimucro are more than 1,500 km apart. Our analysis of the photographs by Ralcheva (1974) of putative P. petrosum from Bulgaria suggests that IS and IL are standard, IIS-6,7 and IIL-9 are present, and IIL-10 is absent; the sex chromosomes and larval morphology were not mentioned. Based on available evidence, P. petrosum of Ralcheva (1974), therefore, is probably P. latimucro, and most closely resembles populations in the Swiss Alps (as P. inflatum "aff. 1" of Basrur 1959). The chromosomal characteristics of all other analyzed populations identified as P. petrosum and P. latimucro are entirely congruent with the respective configurations of the hypostomal teeth. Thus, we argue that P. petrosum is a distinct species on the basis of unique chromosomal features corroborated by distinct hypostomal features. Prosimulium petrosum and European P. latimucro, therefore, are homosequential species-they have the same fixed banding sequence but differ morphologically, a phenomenon first discovered in Drosophila (Carson et al. 1967). Although not common in the Simuliidae, previous examples of homosequential species include several members of the Simulium vernum group (Hunter 1987, Adler et al. 2004, Seitz and Adler 2015.

as a Basis for Biological Control of Vectors of Dangerous Infectious Diseases and
Precautions against the Acts of Biological Terrorism". We thank three reviewers for comments that improved the presentation of the paper.