﻿Redescription of Tylosmaindroni Giordani Soika, 1954 (Crustacea, Isopoda, Oniscidea) based on SEM and molecular data

﻿Abstract The woodlouse species Tylosmaindroni Giordani Soika, 1954 (Crustacea, Isopoda, Oniscidea) is redescribed from the Persian Gulf based on light and scanning electron microscopy. This species differs from the closely related T.exiguus Stebbing, 1910, from the Red Sea (coasts of Sudan and Eritrea), and Socotra Island, by pereopod 1 superior margin without a prominent projection and pleopod 2 endopod 2.3 times as long as exopod, vs. 3.6 in T.exiguus. A distribution map for T.maindroni is provided. In addition, we studied the molecular differentiation of five populations of T.maindroni from the Persian Gulf, based on partial cytochrome c oxidase subunit I (COI) gene sequences. The results revealed low levels of population structuring between the analyzed populations.


Introduction
The isopod genus Tylos Audouin, 1826 has a worldwide distribution, with 21 species currently considered as valid (Boyko et al. 2008 onwards). Species in this genus are found in the marine sandy supralittoral zone, where animals can feed on algae and other organic material washed up on the beach by the waves (Kensley 1974). To avoid excessive predation by daytime predators (birds, crabs), feeding occurs at night (Schmalfuss and Vergara 2000). These animals are able to roll up into a perfect ball, with the antennae remaining inside the ball. Endoantennal conglobation can be also observed in Armadilliidae, Eubelidae, and Scleropactidae. Rolling up is not only a response to predators, but it can also help reduce water loss by about 35% (Schmalfuss and Vergara 2000;Sfenthourakis et al. 2020). According to Schmalfuss (2003), only two species of Tylos have been recorded from the northwestern areas of the Indian Ocean: T. exiguus Stebbing, 1910 from the Red Sea (coasts of Sudan and Eritrea) and the coasts of Socotra Island (Schmalfuss and Vergara 2000;Taiti and Ferrara 2004;Taiti and Checcucci 2010); and T. maindroni Giordani Soika, 1954 from the Gulf of Oman and the Persian Gulf. The original description of T. maindroni is brief and based on a single female from Muscat, Oman. Taiti and Ferrara (1991) later reported this species from Kuwait, Oman, and Iran (Busher coast), with illustrations of specimens from Kuwait. Recently, a molecular phylogenetic study including various species of the genus Tylos clearly revealed the existence of two distinct Tylos species along the coastal zones of the Arabian Peninsula (northwestern Indian Ocean): T. maindroni and T. exiguus (Hurtado et al. 2014).
Herein, we redescribe T. maindroni based on material from the Persian Gulf and provide new COI mtDNA sequence data.

Morphological analyses
Specimens used in this study were collected from four coastal sites in the Persian Gulf, Iran during field expeditions from 2006 to 2021 ( Fig. 1; Table 1). All specimens are held in the isopod collection of the Zoological Museum of Shahrekord University (ZMSU).
Specimens prepared for SEM were washed in a chilled 1% sodium acetate solution for 10 minutes, then cleaned for 10-20 seconds in an ultrasound cleaner in a weak solution of jewelry soap and distilled water to remove sediment and debris adhering to the cuticle. Specimens were dehydrated in an ethanol series (70, 75, 80, 85, 90, 95, 100%; 20 minutes per treatment). Specimens were transferred through ethanol and hexamethyldisilazane (HMDS) solutions (ethanol: HMDS ratios were 2:1, 1:1, 1:2) and finally into 100% HMDS (20 minutes per treatment). All samples were transferred to fresh HMDS, which evaporated overnight. Specimens were mounted on stubs using double adhesive carbon spots before being coated with gold in a sputter coater to 40 nm thickness. Micrographs were taken using a Hitachi S-2460N SEM at Zoologisches Forschungsmuseum Alexander Koenig in Bonn, Germany. Color images were taken using a Zeiss AxioCam ERc5s camera mounted on a Zeiss Stereomicroscope (Stemi 508).

Molecular analyses
We extracted genomic DNA from the legs of seven specimens, 1-2 individuals per locality, using the Aron-Gene Tissue DNA Extraction kit (Aron-Gene, Iran) following the manufacturer's protocol. A 536 base pair fragment of the mitochondrial Cytochrome Oxidase I (COI) gene was PCR-amplified using the LCO-1490 and HCO-2198 primer pair under standard conditions (Folmer et al. 1994). The PCR solution consisted of a 10 µl PCR Master Mix (SinaClon BioScience, Iran), 2 µl of template DNA (~50 ng), 1 µl of each primer (concentration 10 pm/ml), and 6 µl of nuclease-free water for a total volume of 20 µl. PCR products were examined using gel electrophoresis on 1% agarose gels, with positive PCR amplifications sequenced on an ABI 3130XL automated sequencer. We assembled, inspected, and edited sequences using Bioedit v.7.0.5.3.  Once assembled and edited, sequences produced in this study were combined with previously published COI sequences of T. maindroni as well as other Tylos species, provided that these sequences were > 500-bp long. Information for publicly available sequences included in this study can be found in Table 2. Sequences were aligned using the online MAFFT server (Katoh et al. 2019) and default settings. The resulting alignment was trimmed to remove end gaps. No evidence suggestive of pseudo-genes was observed in the final alignment. Given the high levels of divergence amongst Tylos species and differences in sequence lengths across studies, we re-aligned the COI sequences for T. maindroni individuals separately.
We used ASAP (Puillandre et al. 2021), a distance-based species delimitation approach, to determine if all T. maindroni sequences were assigned to a single species cluster. This analysis was carried out on the ASAP web portal (https://bioinfo.mnhn.fr/abi/ public/asap/) under the Kimura (K80 or K2P) model (Kimura 1980) and a ts/tv ratio of 2.0. All other settings were as default. We estimated pairwise genetic distances with the Kimura-2-Parameter (K2P) correction in MEGA v11.0.10 (Tamura et al. 2021).
Lastly, we visualized relationships between T. maindroni COI haplotypes by reconstructing branch connections using the TCS network option (Clement et al. 2002) of PopArt v1.7 (Leigh and Bryant 2015), with a 95% connection limit. Table 2. GenBank Accession information for sequences used in this study. Accession numbers of sequences produced in this study are in bold.  Fig. 2A-D), about 2.5 times as long as greatest width. Cephalon with a weak domed process on each side between eyes. Epistome triangular with narrowly rounded apex, labrum with rows of small tubercles, as figured (Fig. 3B). Eyes composed of 36-38 ommatidia in adults (Fig. 3E). Coxal plates 2-5 with rounded margin, coxal plates 6-7 rectangular with strait margin. Pleotelson framed by pleonite 5 laterally, distal margin with small setae, length about 0.55 times width (Fig 3F, G). Antennula (Fig. 3C). Small, disolateral and apical margins straight, medial margin concave, covered with cuticular scales, about 1.3 times as long as greatest width.

Number of individuals
Antenna (Fig. 3D). Extending to posterior margin of pereonite 1, basal peduncular articles 2-5 increasing in length; article 5 about 1.3 times as long as article 4; flagellum with 4 articles, distal article smallest, apex with cone-like tuft of setae.

Genetic differentiation
We obtained seven 534-bp long COI sequences from T. maindroni individuals from four locations across the Persian Gulf coastline of Iran. These sequences were combined with a previously published COI sequence of T. maindroni from Kuwait (GenBank Acc. KJ468116; BIN: BOLD: ACQ3230). We identified four highly similar COI haplotypes as indicated by K2P divergences (0.0-0.4%, 1-3 nucleotide differences, Fig. 7). These haplotypes, however, were highly divergent from those found in other Tylos species (16.2-33.9% COI K2P divergences, Table 3).   Combining the T. maindroni sequences with other previously published sequences of the genus Tylos resulted in a 517-bp long alignment containing 410 sequences. ASAP analyses of this dataset identified two partitioning schemes with nearly similar numbers of hypothetical species groups (23 and 24), threshold distances (0.068107 and 0.051440), and low ASAP scores (7 and 8). This last measure reflects both the p-value and the relative barcode gap width rank for a given partitioning scheme, with lower values reflecting stronger support for a given partitioning scheme. All COI haplotypes from T. maindroni individuals were placed in a single cluster that included no sequences from other Tylos species, regardless of the partitioning scheme.

Discussion
Tylos maindroni was first described by Giordani Soika in 1954; however, the original description was brief and did not include a discussion or illustration of characters used in the taxonomy of this genus. A later work by Taiti and Ferrara (1991) suggested that T. maindroni's geographic range extends into the Persian Gulf, including locations on the coasts of Kuwait and Iran, but additional work remains necessary to clarify the status of this species and its geographic range. Additionally, considering the high levels of genetic divergence reported in several coastal isopod taxa (Hurtado et al. 2013;Raupach 2014, 2016;Raupach et al. 2014;Hurtado et al. 2017;Santamaria et al. 2017;Greenan et al. 2018;Hurtado et al. 2018;Santamaria 2019), it would be important to determine if T. maindroni harbors cryptic diversity in its native range.
Our Persian Gulf specimens correspond morphologically quite well to the brief description and illustrations of T. maindroni from Oman by Giordani Soika (1954) and from Kuwait by Taiti and Ferrara (1991). Nevertheless, there is a slight difference in the number of lung pores on the exopod of the pleopods: the exopod of pleopod 2 has 8 pores rather than 7 and the exopod of pleopod 3 has 9 pores rather than 8. Tylos maindroni is morphologically most similar to T. exiguus Stebbing, 1910, a Red Sea species shown by Hurtado et al. (2014) to be a sister taxon to T. maindroni based on several mitochondrial markers. The former species differs from T. maindroni by having pereonite 1 posterior margin with a distinctly deeper concavity at the lateral side, pereopod 1 superior margin with a prominent projection, and pleopod 2 endopod 3.6 times as long as exopod (vs. 2.3 times in T. maindroni).
Molecular data are in concordance with the above findings. All Tylos specimens that were morphologically identified as T. maindroni have highly similar COI haplotypes differing by a maximum of three positions (K2P distances amongst haplotypes < 0.5%). Furthermore, sequences recovered from T. maindroni individuals were highly divergent from all other COI sequences recovered from other Tylos species including T. exiguus (16.2-33.9% COI K2P). Not surprisingly, all T. maindroni haplotypes were assigned to a single species cluster in species delimitation analyses carried out in ASAP, regardless of the partitioning scheme.
The low level of diversification herein reported between individuals of T. maindroni collected at Persian Gulf locations stands in contrast with those reported for other coastal oniscid taxa (Khalaji-Pirbalouty and Raupach , 2016Raupach et al. 2014;Hurtado et al. 2017;Santamaria et al. 2017;Greenan et al. 2018;Hurtado et al. 2018;Santamaria 2019), including other Tylos species (Hurtado et al. 2013). For instance, the molecular characterizations of Tylos populations from the Gulf of California showed genetic differentiation in COI sequences ranging from 3.6 to 17.3%, indicating long-standing isolation of the populations in the region as well as the possible presence of cryptic species (Hurtado et al. 2013). Similarly, T. granulatus populations in South Africa have shown to harbor two highly divergent mitochondrial lineages (Mbongwa et al. 2019). In contrast to this, the COI K2P divergences observed in T. maindroni were less than 0.5%.
The low levels of genetic divergence within T. maindroni in the Persian Gulf is likely a reflection of the young age of this marine waterbody. Although there is disagreement on the extent of the Persian Gulf coastline during the Holocene and Late Pleistocene (Sissakian et al. 2020), the Gulf Basin is thought to have been free of marine influence up until the last glacial maximum ~18,000 ya., with marine flooding due to rising sea levels and glacial displacement starting ~14,000 ya (Lambeck 1996). Thus, the geology of the region suggests that the ancestor to T. maindroni populations in the Persian Gulf area invaded the Gulf in the past ~14,000 years. Alternatively, the low divergence levels between COI sequences reported herein may be the result of infection with Wolbachia. Infection with this endosymbiotic bacterium has been proposed to reduce mitochondrial polymorphisms in arthropods, including isopods (Marcadé et al. 2009;Xiao et al. 2012;Delhoumi et al. 2019; but see Tang et al. 2019). We cannot determine whether Wolbachia have reduced mitochondrial diversity in T. maindroni in the Persian Gulf as we did not test for the presence of Wolbachia in our specimens. However, given the recent geological and hydrological history of the Persian Gulf, we propose that the low levels of divergence in T. maindroni reported herein are likely the result of the young age of the modern Persian Gulf. Nevertheless, future work remains needed to conclusively discern between these explanations. Future studies also remain needed to clarify the origins and evolution of T. maindroni in the region. The closest extant relative of T. maindroni in the Persian Gulf is T. exiguus (Hurtado et al. 2014), suggesting that the Persian Gulf populations of T. maindroni likely originated from an ancestor inhabiting coastal habitats in the Indian Ocean basin. As our sampling did not include T. maindroni populations from the Indian Ocean, future work would be best served by incorporating these populations.