A new obligate groundwater species of Asellus (Isopoda, Asellidae) from Iran

With only 43 described stygobionts and only two isopod species the obligate groundwater fauna of Iran, a vast country with over 10% of limestone surface, is inadequately known. Here, we report the discovery of Asellus ismailsezarii sp. nov. from Zagros mountains, the first eyeless and depigmented asellid isopod from Iran. The new species is morphologically similar to Asellus monticola Birstein, 1932, but it is eyeless and fully depigmented, has a slightly curved pereopod IV and does not bear any setae on proximal margins of exopodite of pleopods IV and V. Species phylogenetic relationships using original and datamined mitochondrial DNA and nuclear rDNA, and estimation of molecular divergences with other Asellus species, suggest that A. ismailsezarii sp. nov. is sister to a larger clade that also contains the European A. aquaticus species complex. Surface populations of Asellus have colonized groundwater at multiple occasions and localities, both in Europe and Asia, giving rise to species and subspecies that have evolved troglomorphisms, such as depigmentation and loss of eyes. Of the 37 formally described species and subspecies of Asellus, 15 are from groundwater, including A. ismailsezarii sp. nov. We predict that many more obligate groundwater Asellus taxa are yet to be discovered in Asia. Subterranean Biology 42: 97–124 (2022) doi: 10.3897/subtbiol.42.79447 https://subtbiol.pensoft.net Copyright Mohammad Javad Malek-Hosseini et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. RESEARCH ARTICLE Subterranean Biology Published by The International Society for Subterranean Biology A peer-reviewed open-access journal

). The genus Asellus shows a wide geographic range, being present in Alaska (1 species, Bowman and Holmquist 1975), Asia (15 species, Henry and Magniez 1995;Sidorov and Prevorčnik 2016) and Europe (2 species). The occurrences of Asellus in Europe were all attributed to the Asellus aquaticus sensu lato species complex (Verovnik et al. 2005), with the exception of occurrences from two nearby surface water sites in eastern Ukraine, which were attributed to Asellus sp. (River Kharkiv and Pionersky pond; see Verovnik et al. 2005). The taxonomic status of Asellus populations within the A. aquaticus complex is not yet fully resolved. That complex actually contains two species, A. aquaticus Linnaeus, 1758 and A. kosswigi Verovnik et al., 2009, and eleven formally described subspecies (Sket 1965;Turk-Prevorčnik and Blejec 1998). COI-based species delimitation methods indicate that there may be many more species (Sworobowicz et al. 2015(Sworobowicz et al. , 2020. Also, several Asellus populations belonging to that complex have colonized caves in Europe and have evolved traits such as loss of pigmentation and eye reduction (Balázs et al. 2021).
The eastern geographic boundaries of the A. aquaticus complex are not clearly established. The global biodiversity information facility does not report any records of A. aquaticus Linnaeus, 1758 in Iran (consultation date: 16/11/2021). However, the presence of that species was reported by Rémy (1941) at a pond located 6 km from Chahi, Mazanderan province, and more recently by Henry and Magniez (1995) at a surface freshwater site in the Gorgan province contiguous to Turkmenistan. Birstein (1945) described Asellus aquaticus messerianus from Turkmenistan Lakes Delili, Yashka and Karatogelek located near the northern border of Iran. Another asellid species, Asellus monticola, originally described by Birstein (1932) from freshwater surface sites in Armenia and Georgia was collected in Northern Iran at Bora Laan Spring, a locality located near the Turkish border (Henry and Magniez 1996).
Here, we report on the discovery of Asellus ismailsezarii sp. nov. (Asellidae, Asellota, Isopoda, Pancrustacea), the first eyeless and depigmented asellid isopod described from Iran. We also provide morphological comparisons with A. monticola, the species that morphologically resembles Asellus ismailsezarii sp. nov. as well as comparisons with European cave-species and subspecies belonging to the A. aquaticus complex. We then use molecular data to corroborate the species status of A. ismailsezarii sp. nov. and to document its phylogenetic relationship to the A. aquaticus complex.

Morphological description
When necessary, pereopods I, IV and VII from one side of the body were dissected and heat-treated in a KOH solution, dyed with Trypan Blue for better visibility of spinulation and then temporary mounted on slides in glycerine. After dissection, also antennae I, II, pleopods, uropods and pleotelson, were mounted on slides in glycerine. Specimens and their body parts were photographed and measured using a Sony DXC390P digital camera mounted on a stereomicroscope or microscope (depending on the size of the structure), and measured with Leica Application suite -LAS EZ. The remains of the dissected specimens were then transferred to 70% ethanol for storage. In one of the specimens from the type locality, all appendages were dissected and prepared for drawing. All pereopods, as well as the trunk were heat-treated and dyed as described above, and then temporarily mounted in glycerine, alongside the rest of dissected specimen (mouth appendages, antennae, pleopods, uropods). Vector drawings were made from microphotographs using a graphics tablet (Wacom, Cintiq 13HD Creative Pen Display) and the free software KRITA 4.1.1 (https://krita.ord/). In males, we measured 90 morphometric characters (cf. Prevorčnik et al. 2004) for describing body proportions (trunk, appendages) and characterizing cuticular and sensory structures (number and length of spines, setae, aesthetascs). In antenna II, only the last two articles of the antennal basis (fourth and fifth) were included in the total antennal length together with the length of the flagellum. In pereopods IV and VII, length of the coxa was not included in the total pereopod length. We did not report full length of antenna II and pereopods because in most cases the basal articles of both appendages remained attached to the body during dissection. Throughout the ensuing description, we provide a single measurement value for the holotype, together with a range of values (in parentheses) referring to the males (♂♂) and females (♀♀) from the type material.
For observation with scanning electron microscopy (SEM), Plp I and Plp II of the male holotype and paratype stored in 70% ethanol were air-dried and mounted on sample stubs using conductive double sided carbon tape. Mounted samples were sputter-coated with platinum and observed with a JEOL JSM-7500F field emission scanning electron microscope (Jeol. Japan) at the Department of Biology, Biotechnical Faculty, University of Ljubljana.

DNA extraction, amplification and gene sequencing
We used the last three pereopods of three specimens (specimens AS1, AS2 and AS3, see Material examined) of A. ismailsezarii sp. nov. to obtain sequences of the mitochondrial cytochrome oxidase subunit I (COI) gene and 28S nuclear rDNA gene. We performed molecular analyses at the Evolutionary Zoology Laboratory, Jovan Hadži Institute of Biology ZRC SAZU, Ljubljana (EZ LAB, Slovenia) and LEHNA laboratory, Villeurbanne (LEHNA, France). We crosschecked DNA extraction protocols for the COI gene using specimen AS1 and DNA extraction and PCR protocols for the 28S using specimens AS1 and AS2. Full detail of the molecular protocol for each gene and specimen is provided in Suppl. material 1: Table S1 and Suppl. material 2: Table S2.
We performed robotic DNA extraction at EZ LAB on the three specimens using Mag MAX™ Express magnetic particle processor Type 700 with DNA Multisample kit (Thermo Fisher Scientific kit) and modified protocols following Vidergar et al. (2014). We also extracted DNA from specimen AS1 at LEHNA using a modified Chelex protocol from Casquet et al. (2012). We incubated pereopods in a solution of 150 μl of 7% chelex and 10 μl of proteinase K at 15 mg / ml for 90 minutes at 56 °C, then 15 minutes at 90 °C.
We performed polymerase chain reactions (PCRs) for COI fragments at LEHNA using a previously optimized protocol (Calvignac et al. 2011). However, we used a Taq polymerase (Eurobiotaq) amount of 0.05 U instead of 0.15 U and a PCR volume of 25 μl instead of 35 μl. To detect amplification of nuclear paralogs of mtDNA, we used a mix of four half-nested PCR strategies applied on a dilution range of initial DNA template (pure DNA solution of 2 μl to 1/2500 dilution, primers in Suppl. material 1: Table S1 and Suppl. material 2: Table S2). We ran the first round PCRs with one step of 3 min at 95 °C; 37 cycles of 20 s at 95 °C, 30 s at 51 °C, 45 s at 72 °C; and one step of 5 min at 72 °C. The second round PCRs were run on 1 μl of the first round PCR product using the same parameters but 35 cycles. Service providers (BIOFIDAL, Vaulxen-Velin, France) performed Sanger sequencing using PCR amplification primers.

Molecular species delimitation
Species and higher taxa and their associated names represent hypotheses of independently evolving lineages that should ideally be supported by different lines of evidence (de Queiroz 2007;Padial et al. 2010;Fišer et al. 2018). In addition to morphological evidence, we used a COI-based molecular species delimitation approach to corroborate whether A. ismailsezarii sp. nov. was indeed a new species. In effect, we follow the diagnosable interpretation of the phylogenetic species concept (de Queiroz 2007).
COI sequences of A. ismailsezarii sp. nov. produced in this study were analysed with all COI sequences of taxa belonging to the "Asellus pattern" (sensu Henry and Magniez 1995) available in NCBI (Sequence extraction date: 25 August 2021). In addition to sequences of A. ismailsezarii sp. nov., we retrieved 581 sequences belonging to the A. aquaticus sensu lato species complex (including A. kosswigi, Verovnik et al. 2005Verovnik et al. , 2009Sworobowicz et al. 2015Sworobowicz et al. , 2020Lafuente et al. 2021), Asellus sp. from Ukraine, Asellus hilgendorfii from Japan, and two species from Lake Baikal, Russia -Mesoasellus dybowskii and Limnoasellus poberezhnii (Nomen nudum in Hidding et al. 2003) (Suppl. material 3: Table S3). We aligned the sequences with ClustalO in Seaview (Gouy et al. 2010) (Suppl. material 4: Table S4). We checked visually for the presence of anomalies, including stop codons and frameshifts, and removed ambiguous sites with the embedded Gblocks (with options for less stringed selection enforced) (Castresana 2000). COI maximumlikelihood phylogenies were reconstructed using PhyML ) under the GTR + G + I model of substitution with 4 substitution rate categories and a gamma shape parameter (α = 0.519) as well as a proportion of invariant sites (0.392) estimated by maximum likelihood in PhyML. This model of evolution was previously used in several studies for delimiting species and reconstructing phylogenetic relationships among asellids (Lefébure et al. 2006;Morvan et al. 2013;Eme et al. 2018). We used MEGA (Kumar et al. 2018) as well, to find the best nucleotide substitution model and the optimal partitioning scheme for each marker. Molecular operational taxonomic units (MOTUs) were delimited using the fixed threshold method implemented by Lefébure et al. (2006) for crustaceans. This method is based on the observation made from 1500 COI sequences belonging to 276 species of crustaceans that two clades diverging by more than 0.16 substitution per site, as measured by patristic distances, have a strong probability (ca. 0.99%) of belonging to different species. That method is conservative insofar as it identifies both fewer MOTUs and MOTUs that are more divergent than other delimitation methods, thereby limiting the risk of considering two populations of the same species as belonging to two distinct MOTUs. Several studies (Morvan et al. 2013;Sworobowicz et al. 2015;Eme et al. 2018) showed that evolutionary units of Asellids as delimited by distance-based methods, including that implemented by Lefébure et al. (2006), were further subdivided by tree-based methods such as the General Mixed Yule-Coalescent method (Pons et al. 2006) and the Poisson tree processes (Zhang et al. 2013). Our conservative estimate avoids potential oversplitting.

Taxonomy
Diagnosis of male. Species of A. ismailsezarii sp. nov. with depigmented body and complete loss of eyes (Fig. 2). Body elongated oval, more than three times (3.2-3.5) as long as wide. Head up to 1.4 times as wide as long; frontal margin bisinuate, medially concave, lateral margins rounded, each with posterolateral prominence with few weak setae. Pereonites I-III with slightly convex lateral margins, and pereonites IV-VII with straight lateral margins, with only few small simple short setae of about the same length. Pleomere I-II width about 60% of pereonite VII width. Pleotelson subrectangular, lateral margins densely fringed with simple short setae of about the same length, terminal edge rounded, with only weak median prominence. Antennae I and II, lengths about 15% and 70% of body length and with 8 and 55 flagellar articles, respectively. Propodus I slender ovoid, with weak apophysa on palmar margin. Carpus and propodus IV only slightly curved, superior margin and submarginal surface of propodus IV with few (6) short and weak simple setae. Length of longest spiniform . The photo shows a specimen exposed to air after the stone it was under had been turned over. Body length is 8.6 mm. robust seta on propodus VII inferior margin only about 3% pereopod VII length. Pleopod I exopodite without symmetric concavity on lateral margin, lateral margin being rather straight to slightly convex. Pleopod II architecture following typical ''Asellus pattern'' sensu Henry and Magniez (1995). Its endopodite with short (around 25% of protopodite length) horn-shaped basal spur (processus calcariformis). Pleopod IV and V with small respiratory areae, linea areae beginning and ending at the distal exopodite margin. Uropod short, up to about a quarter of body length.
Pereonites (Fig. 3B) I-III with straight to slightly convex lateral margins and weakly rounded antero-and posterolateral angles, slightly protruding forwards. Last three pereonites with antero-and posterolateral angles slightly protruding backwards. Pereonite VI widest. Anterolateral, lateral and posterolateral margins of pereonites almost bare, with only sparse short simple setae. Coxopods well developed, margins of all epimerae dorsally visible, last three the most prominent.
Pleomere I-II short but wide (Fig. 3B), their width about 62% of pereonite VII width, forming a stalk largely covered by posterior margin of pereonite VII. Pleotelson Mandibulae (Fig. 4MdbL) robust: Pars molaris (molar process) U-shaped, with toothed margin. Pars incisiva (incisor) formed by few blunt cusps arranged in semicircle. Left lacinia mobilis with few cups and spine row of about 15 biserrate setae, the distal ones being longest. Palp of three segments. First palpal segment widest, with few simple setae distally. Second palpal segment 1.6 times longer than first, without setation along external margin, with two simple setae distalo-mesially, and few simple setae along internal margin. Third palpal segment around half as long as second, with row of about 20 robust biserrate setae along external margin. Maxilla I (maxillule; Fig. 4MxI) lateral lobe with few smooth (the outer spines cone-shaped) and few weakly serrate robust spines. Distal part of outer margin with 1 long slender seta. Mesial lobe with 4 robust long plumose setae. Maxilla II (Fig. 4MxII) lateral and middle lobe with 19 slender and simple, and 13 curved pectinate robust setae, respectively, mesial lobe with about 10 biserrate setae and parallel row of about 13 long simple setae along inner margin. Maxilliped (Fig. 4MxlpR) endite distal margin with about 11 biserrate robust setae, subapically with several rows of short simple setae. Mesial margin curved dorsally, with row of about 8 long biserrate setae, distomesial margin with setulose fringe and around 5 coupling hooks. Palp of five articles. First article with 2 short setae apically on outer margin; second about 2.5 times as long as first, subtrapezoidal, with 5 long stiff setae on outer margin and row of about 15 longer medially directed simple setae on inner margin. Third article a bit shorter as second, less broad, with 4 long stiff setae on outer margin and row of about 11 slender simple setae on inner margin. Fourth article approximately twice as long as third, slender, distally wider, with a row of 6 and about 20 long slender setae along outer and inner margin, respectively. Fifth article as long as first, ovoid, fringed with around 10 long slender setae and 2 longest stiff simple apical setae. Epipodite subrectangular, lateral margin almost bare, with only few (3) short simple setae.
Pleopod III (Fig. 6PlpIII) exopodite rounded triangular, about 1.6 times as long as wide, with almost straight medial margin. Medio-distal, terminal and latero-distal margins with around 8 long plumose setae. No setation along medial and lateral margins. Endopodite length about 0.7 of exopodite length.
Pleopod IV (Fig. 6PlpIV) exopodite broadly ovoid, about 1.6 times as long as wide, its area equally shaped as in pleopod V. Without setation along margins. Endopodite subrectangular, its length about 0.8 of exopodite length.
Morphological comparison with other Asellus species. Among the 18 species of Asellus presently described, A. ismailsezarii sp. nov. resembles morphologically A. monticola Birstein, 1932 (Suppl. material 6: Table S6). Pleopods I and II appear to be almost identical between the two species, although Henry and Magniez (1996) did not provide SEM photos of these pleopods for A. monticola. In both species, the endopodite of pleopod II has a spoon-shaped Capitulum with no spines. However, A. ismailsezarii sp. nov. differs from A. monticola in that it is eyeless and fully depigmented, has a slightly curved pereopod IV and bears no setae on proximal margins of exopodite of pleopods IV and V. These morphological differences are sufficient to consider the studied specimens as belonging to a distinct species. We also provided in Suppl. material 6: Table S6, a morphological comparison with A. kosswigi, A. aquaticus aquaticus, A. aquaticus infernus and A. aquaticus cavernicolus showing that A. ismailsezarii sp. nov. shows many morphological differences with both surface and subterranean aquatic species and subspecies of the A. aquaticus complex. Contrary to A. ismailsezarii sp. nov., all members of the A. aquaticus complex show a helical Capitulum, suggesting the new species does not belong to that complex. Molecular species delimitation. We found eleven MOTUs within the "Asellus pattern", including one MOTU corresponding to A. ismailsezarii sp. nov. (Fig. 8A). Maximal patristic distances among COI sequences within the same MOTU (i.e., within-MOTU distances) ranged from 0 to 0.19 (0.01 for A. ismailsezarii sp. nov.) (Suppl. material 7: Table S7). Minimal patristic distances among sequences belonging to A. ismailsezarii sp. nov. and sequences belonging to the other MOTUs (i.e., between-MOTU distances) ranged from 0.76 to 1.00 (Suppl. material 7: Table S7). The considerable gap between within-and between-MOTU patristic distances supports our taxonomic assessment that A. ismailsezarii sp. nov. is a distinct species.
Phylogenetic relationships as inferred from the COI gene indicated that A. ismailsezarii sp. nov. does not belong to the A. aquaticus complex, which itself contained six MOTUs, including A. kosswigi (Fig. 8A). Within that complex, only MOTU E corresponding to A. aquaticus Linnaeus, 1758 sensu stricto was widely distributed across Europe (Fig. 9). The resulting tree of the nuclear 28S rDNA sequences agreed well with the COI phylogeny, again indicating that A. ismailsezarii sp. nov. was a sister species  Lefébure et al. (2006) were collapsed. In B, branches to the main clades were collapsed. Colors and symbols as in Fig. 9. Branch supports are approximate likelihood ratio test (aLRT) values (black) and bootstrap values (red). of the A. aquaticus complex (Fig. 8B). Patristic distances derived from the 28S loci also corroborated COI results by revealing a strong divergence between A. ismailsezarii sp. nov. and the other MOTUs.
Etymology. The name of the new species is a patronym for ''Mohammad Ismail Sezari'', the legendary guard of the train tracks from the area of the type locality. In year 1986, he sacrificed his life to save 750 others that were on a collision course train.

Discussion
The present study provides morphological and molecular evidence supporting the species status of Asellus ismailsezarii sp. nov. This newly discovered species shows diagnostic morphological features of the Asellus genus and Asellus subgenus (Magniez and Henry 1970;Henry andMagniez 1993, 1996). In addition to being eyeless and fully depigmented, A. ismailsezarii sp. nov. differs from the surface aquatic species A. monticola to which it is supposedly most closely related, by the shape of its pereopod IV and the lack of setation on pleopods IV and V. Most but not all groundwater species lack a grasping pereopod IV, suggesting adult males no longer engage in precopulatory mate guarding (Henry 1976). The occurrence of a still curved pereopod IV among some groundwater asellids could indicate that groundwater colonization has occurred recently. However, this hypothesis deserves to be tested using a large number of population and/or species (see for example Balázs et al. 2021). We also present morphological evidence that Asellus ismailsezarii sp. nov. is distantly related to species belonging to the A. aquaticus complex. All species of that complex are characterized by a helical Capitulum, whereas both Asellus ismailsezarii sp. nov. and A. monticola have a spoon-shaped Capitulum.
In addition to morphological evidence, we show that patristic distances for the COI gene between A. ismailsezarii sp. nov. and any other MOTUs delimited within the "Asellus pattern" largely exceed the 0.16 threshold value above which two clades can be considered to belong to distinct species (Lefébure et al. 2006 Verovnik et al. 2009) systematically shows a large distribution throughout Europe. According to records by Rémy (1941) and Henry and Magniez (1995), A. aquaticus Linnaeus, 1758 is also present in Iran. Further sampling and sequencing of asellids in the Caspian See region is necessary to ascertain the existence of that contact zone between species of the European A. aquaticus complex and Asian species.
The molecular systematics of the Asellus and more largely of the "Asellus pattern" is yet to be established since most of its species have not yet been sequenced. Hence, except from the fact that A. ismailsezarii sp. nov. does not belong to the A. aquaticus complex, its phylogenetic position within the Asellus remains to be more precisely determined. We expect that A. ismailsezarii sp. nov. is phylogenetically closely related to A. monticola because the two species share many morphological characters. This hypothesis can be tested when fresh material of A. monticola becomes available for DNA sequencing. Ancestral populations of A. monticola might have colonized groundwater on multiple occasions and localities. If so, many more obligate groundwater Asellus taxa could be discovered in Iran.
We have several reasons to expect many more groundwater species discoveries in Iran as sampling effort increases. First, the number of 43 obligate groundwater species presently known from Iran is exceedingly low considering the extent and diversity of groundwater habitats (Vardanjani et al. 2017). Of the 2000 registered caves and many more wells and springs in Iran, a tiny proportion has actually been sampled. Second, about 70% of the species were described in the past eight years when sampling and taxonomic efforts intensified. Third, the number of species is very unevenly distributed among taxa, reflecting a strong taxonomic bias in description effort. Of the 43 obligate groundwater species, 23 belong to the genus Niphargus Schiødte, 1847, due to a recent increase in species description activity among Iranian experts (Malek-Hosseini and Zamani 2017; Zamanpoore et al. 2020; Bargrizaneh et al. 2021). Yet, recent species descriptions of Stenasellidae, aquatic Isopoda (Khalaji-Pirbalouty et al. 2018), Moitessieriidae, aquatic snails (Fatemi et al. 2019), Agnaridae, terrestrial Isopoda (Kashani et al. 2013) and Carabidae, terrestrial Coleoptera (Malek-Hosseini et al. 2021) indicate that many animal groups are likely to have presently unknown subterranean representatives in Iran. Local but repeated sampling is also revealing the presence of potentially species-rich aquifers with high conservation value. The karst aquifer feeding the Ganow spring, the type locality of A. ismailsezarii sp. nov., also harbors three described species of cave fish (Mousavi-Sabet et al. 2016;Vatandoust et al. 2019) as well as yet-undescribed species of gastropods (in first author collection). This paper contributes a piece in the puzzle of the Iranian cave biodiversity. located in Biotechnical faculty, University of Ljubljana, as a member of MRIC UL network. We thank Rudi Verovnik, one anonymous reviewer and Cene Fišer for their constructive comments on an earlier draft of this manuscript.