Morphological and molecular evidence reveals a new species of chewing louse Pancola ailurus n. sp. (Phthiraptera: Trichodectidae) from the endangered Chinese red panda Ailurus styani

Lice are six-legged, wingless, insect parasites of mammals and birds, and include two main functional groups: blood-sucking lice and chewing lice. However, it is still not clear whether the Chinese red panda Ailurus styani is infested with the parasitic louse. In the present study, we describe a new genus and a species of chewing louse, Pancola ailurus (Phthiraptera: Trichodectidae) based on morphological and molecular datasets. The morphological features showed that Pancola is closer to Paratrichodectes. The genetic divergence of cox1 and 12S rRNA among the Pancola ailurus n. sp. and other Trichodectidae lice was 29.7 – 34.6% and 38.9 – 43.6%, respectively. Phylogenetic analyses based on the available mitochondrial gene sequences showed that P. ailurus n. sp. is more closely related to Trichodectes canis and Geomydoecus aurei than to Felicola subrostratus and together nested within the family Trichodectidae. This study is the first record of parasitic lice from the endangered Chinese red panda A. styani and highlights the importance of integrating morphological and molecular datasets for the identification and discrimination of new louse species.


Introduction
Parasitic lice (Psocodea: Phthiraptera) are permanent, obligate, and host-specific ectoparasites commonly found on the surface of birds and mammals (Kim, 1988;Price et al., 2003). The entire life of the lice is completed on their hosts (including egg, three nymphal, and adult stages). To date, five taxonomic groups of lice are recognized in the world: Anoplura, Amblycera, Ischnocera, Rhynchophthirina, and Trichodectoidea (De Moya et al., 2021). The lice in the groups of Amblycera, Ischnocera, Rhynchophthirina, and Trichodectidea are all chewing lice, which mainly feed on the feathers and dermal debris of their hosts. The Anoplura lice are all sucking lice, which only feed on the host blood (Light et al., 2010). Although chewing lice are relatively benign parasites, the infestation of lice can cause public health impacts. Heavy infestations can cause their hosts' severe skin irritation, scratching and rubbing, hair loss, fur damage, fleece and wool damage, restlessness, weight loss, and severe anemia (Mullen and Durden 2009;Clayton et al., 2015;Bush and Clayton, 2018). Approximately 400 species of chewing lice in 19 genera are recognized in the family Trichodectidae. The members of Trichodectidae only parasitize eutherian mammals and can be morphologically distinguished by only one tarsal claw on each leg, and the antennae in at least one sex have either a reduced number of segments or the reduced size of the last two segments (Ewing, 1936). Trichodectidae poses severe threats to livestock husbandry worldwide (Price et al., 2003). For instance, the dog-biting louse (Trichodectes canis) and the cat-biting louse (Felicola subrostratus) are intermediary hosts of the dog tapeworm Dipylidium caninum, which not only infect canids but also humans (Dantas-Torres, 2008;Rousseau et al., 2022).
Red pandas (Carnivora: Ailuridae), which are strictly endemic to the Himalaya-Hengduan Mountains mainly live in high altitudes (2200 to 4800 m) (Roberts and Gittleman, 1984;Glatston et al., 2015), are endangered mammals to the world. Based on morphology, biogeography, and population genetic evidence, red pandas are classified into two distinct species within the geographic boundary of the Yalu Zangbu River: Ailurus fulgens (Himalayan) and Ailurus styani (Chinese) (Hu et al., 2020). Populations of Chinese red pandas have been on the decline by 50% over the past 30 years (Glatston et al., 2015;Karki et al., 2021). The main factors linked to this decline have been hunting and fragmentation, degradation, and destruction of habitats (Pradhan et al., 2001;Thapa et al., 2018;Karki et al., 2021). Another likely threat to the red panda is parasite infections (Sharma and Achhami, 2021). Chinese red pandas have been reported with several parasites including protozoans, cestodes, nematodes, and trematodes (Bertelsen et al., 2010;Bista et al., 2017;Sharma and Achhami, 2021). Interestingly, ticks and mites were also reported (Bista et al., 2017), but no louse species has previously been recorded from the red pandas.
The traditional methods for the identification and differentiation of louse species are based on the morphological features and the relevant information of hosts or geographical origin (Madrid et al., 2020), and the morphological method helped identify and distinguish over 5000 louse species based on the last worldwide checklist (Price et al., 2003). To date, morphology is still considered the mainstream for identifying new species in a lot of studies (Lei et al., 2020;Wang et al., 2020Wang et al., , 2021Sychra and Palma, 2021;Jie et al., 2022). As a supplementary method, the molecular method to identify and distinguish louse species seems to be helpful for non-experts, and it could expand the understanding of the possible divergence of new species. With the assistants of molecular methods, it is easier to distinguish body lice and head lice that are similar in shape (Fu et al., 2022). Additionally, the molecular method was also developed as a rapid detection of lice infestations (Tran et al., 2022). Moreover, a large literature on both sucking and chewing lice added molecular information as supplementary data when describing new species (Valim and Weckstein, 2013;Najer et al., 2014;Kolencik et al., 2018;Durden et al., 2019;Madrid et al., 2020). Herein, we aim to combine morphological characters with molecular data for better identifying louse species.

Specimens collection
Louse specimens were collected from Chinese red pandas A. styani in the Sichuan province of China in November 2020. The collection procedures followed animal ethics and have no damage to both hosts or parasites. Using dandruff combs, the chewing lice on the surface of red pandas were combed down and collected (Fig. 1). The collected lice were washed with physiological saline solution. Some of them were used for morphological examination, specimen collection, and DNA extractions, and the rest were stored at -80 • C for further use.

Morphological examination
For further morphological examination, clearing, staining, and mounting of intact specimens with minimum content and stretched legs on the slides as Palma detailed (Palma, 1978). With paratergal plates pre-pierced, lice were dipped into 20% potassium hydroxide solution (KOH) for 24 -48 h until soft, and transparent. After gently squeezing the abdomens with tweezers to expel the digested tissues, lice were transferred into ultrapure water for 30 min. Then lice were transferred into a 10% acetic acid solution for 30 min. After that, the specimens were subsequently stained in 1% acid fuchsin for 4 h, then gradually dehydrated with 40%, 80%, and absolute ethyl alcohol for 30 min, respectively. After finishing the steps above, specimens were immersed in clove oil to purify dying for 1 -2 weeks. The specimens were then applied on slides with a drop of Canada balsam at room temperature (25 • C) drying for 2 -3 weeks. The slides were examined under a digital photomicroscope (OLYMPUS BX51). A key to Trichodectidae proposed by Lyal (1985) has been used for identification and comparison. All measurements were taken in micrometers (range followed by mean). Descriptive format and abbreviations of morphological features, with names of setae spelled out in full at first mention, follow Price and Hellenthal (1981), Lyal (1985), Gustafsson and Bush (2017), and Mey (2021).

DNA isolation, amplification, and sequencing
Whole-genome including mitochondrial (mt) and nuclear DNA was extracted from 38 single lice using Wizard® SV Genomic DNA Purification System (Promega, USA) according to the manufacturer introduction. The preliminary morphological identification inferred red panda lice similar to the family Trichodectidae species, and we then designed pairs of primers based on available trichodectids sequences to amplify the nearly complete mt genes (cox1 and 12S rRNA) and nuclear gene (18S rRNA) of these lice (Barker et al., 2003;Song et al., 2019), since most of the sequences available online are partial sequences and may not be the same fragments. The primers were listed in Table 1. PCR amplifications used a 2 μl DNA template, 1.5 μl of each primer (forward and reverse), 25 μl of Master Mix (Takara, Shiga, Japan), and added ddH 2 O to a 50 μl system. The amplification conditions were followed as denaturation at 94 • C for 1 min, followed by 37 cycles of denaturation at 98 • C for 10 s, then 30 s of annealing at 45 • C for cox1 and 47 • C for 12S and 18S, 1 min of extension at 72 • C, and 2 min of the final extension at 72 • C. PCR products were then verified with 1.5% agarose gel electrophoresis. Purified products were sequenced by the Sanger method in BGI Tech Solutions Co., Ltd. (Shenzhen, China).

Phylogenetic analyses and divergence dating
Phylogenetic relationships among species were inferred based on combined mt sequences (cox1 and 12S rRNA) within Trichodectidae and Bovicoliidae (Table 2) using the Maximum likelihood (ML) algorithm, with Liposcelis bostrichophila (KY656897) as the outgroup. All nucleotide sequences of families Trichodectidae and Bovicoliidae species were then aligned into a single alignment dataset with ClustalX 1.83 (Aiyar, 2000), and excluded ambiguous regions based on Gblocks 0.91b webserver with the selection of "less stringent" (Castresana, 2000;Dereeper et al., 2008). The most suitable model "GTR + I + G" was selected by jMo-delTest v2.1.5 with Akaike information criterion (AIC) (Darriba et al., 2012). ML analysis was conducted based on concatenated cox1 and 12S rRNA gene sequences in PhyML v3.1 with a nucleotide substitution model. To verify the reliabilities of each inferred phylogeny, 100 bootstraps were applied.
The divergence time of chewing lice from families Trichodectidae and Bovicoliidae was determined using BEAST v1.10.4 . Considering the divergence time of the recent common ancestor (Pediculus humanus and P. schaeffi) used in the present study was inferred mainly based on the cox1 gene (Light et al., 2010), phrased partial mt gene (cox1) implemented in BEAST v1.10.4 to estimate the BEAST coalescent species. The most suitable model was determined based on ModelFinder with "mtREV" used as the gamma site model. Divergence time analysis was set as followed: a random starting tree, a strict clock model, and a birth-death tree prior with an uncorrelated lognormal clock rate (Drummond et al., 2006). Considering the uncertainty of default calibrations, we introduced members whose divergence dates had been published before: (i) the calibration of booklouse (L. bostrychophila), which is seen as the closest free-living relative of parasitic lice (Yoshizawa and Johnson, 2010), was considered a potential ancient divergence time in 100 million years ago (Mya) (Grimaldi and Engel, 2006); (ii) a calibration for human lice (P. humanus) and chimpanzee lice (P. schaeffi) of 5 -7 Mya was used to represent the potential split time between both lice (Light et al., 2010;Ashfaq et al., 2015); (iii) P. schaeffi was estimated that the divergence time was 12 Mya for cox1 gene (Ashfaq et al., 2015). The L. bostrychophila was then set as the outgroup of whole taxa to calibrate the height of the BI tree and the mean age was set as 52 Mya. The minimum age was set up at 5 Mya using an exponential distribution. Each analysis was run for 40 million generations with sampling every 1000 generations. Tracer v1.7.2 was applied to monitor the convergence between runs the effective sample size (ESS) should be more than "200" . TreeAnnotator v2.6.2 was used to access the posterior probabilities of the maximum clade credibility (MCC) tree with the first 10% tree burn-in (Dellicour et al., 2021). The phylogenetic relationship and divergence time were visualized in FigTree v1.4.3.

Taxonomy
Superfamily Trichodectoidea Kéler, 1938 Lyal (1985) is used for the identification of Pancola with other genera in Trichodectidae. Based on the presence of exactly five abdominal spiracle openings, the abdominal chaetotaxy of both sexes, and the fact that the parameres are not fused to the lateral struts of the basal apodeme, Pancola is the closest to Paratrichodectes. However, these two genera can be separated by the following combination of characteristics: 1) the anterior head margin of the Pancola is broad, medium slightly convex, or non-convex, whereas the anterior of the head with osculum present in Paratrichodectes (Lyal, 1985); 2) the head marginal temporal carina with rounded edges in Pancola, differentiate it from Paratrichodectes, which has a convex or rectangular temple margin (Lyal, 1985); 3) the 6 pairs meso-metasternal setae present on Pancola, which is different from Paratrichodectes. There are no meso-metasternal setae on Paratrichodectes (Lyal, 1985); 4) the male genitalia of Pancola has narrow mesomeres, which is absent in Paratrichodectes (Lyal, 1985).
Wider than thorax, oval to rounded in shape. Dorsally, 5 pairs of spiracles in small sizes are present on segments III-VII. Abdominal chaetotaxy as in Fig. 2A and B. Tergal posterior setae (tps), and sternite setae (sts) are present, all setae long in size. Tergopleurite plates with paratergal seta (ps) and post-spiracular seta (pss) present.
Female. As in Fig. 2E. Ventral terminalia with small tapered gonapophyses VIII in females, the inner margin of gonapophyses VIII gently convex with marginal setae present, the outer margin of gonapophysis VIII within the abdomen boundary.
Etymology: The genus name is a noun referring to the common name of the host.
Head: As in Fig. 2A. Roughly round in shape, slightly wider than long. Pre-antennal Region: Anterior head margin broad, medium slightly convex or non-convex; frons sclerotized; mc along the head margin and partially pigmented; prmc continuous with pomc; pran large, blunt; conus blunt, smaller than the posterior end of scape in male and female. Dorsally, 1 pair of as1, as2, as3, pcs, ads1, ads2, and pas present on each side.; Ventrally, 1 pair of avs1, avs2, avs3, and mds on each side. Antennal Region: Antennae sexual dimorphic. Male antennae scape stout and larger than females. Antennae with 3 segments, antennal groove slightly deep; pedicel and flagellomere unfused, pedicel smaller than scape; flagellomere with the large process, slightly curved. Postantennal Region: Gular plate absent; the margin of temple convex; marginal temporal carina with rounded edges. One pair of pns, os, pos, and pts on each side. Five pairs of mts on each side, vary in size, mts 4 large.
Abdomen: As in Fig. 2A. Wider than the thorax, oval to rounded in shape. Dorsally, 5 pairs of spiracles in small sizes are present on segments III-VII. One tergite per segment, except segments I and VIII without tergites and segment II with 2 tergites. Tergite 1 with 4 pairs of long tps, medium pair smaller. Tergites 2 and 3 each with 9 pairs of long tps. Tergites 4 and 5 each with 10 pairs of tps, medium pair very small. Tergite 6 with 9 pairs of long tps, medium pair very small. Tergite 7 with 9 pairs of tps, lateral pair large but medium pair very small. Segment 8 with 2 rows of setae, 1st row with 11 pairs of very small setae, 2nd row with 8 pairs of very small setae. Ventrally, six sternites with sts present from segment II to VII, all setae long in size. Sternite 1 with 17 long sts. Sternite 2 with 9 pairs of long sts. Sternite 3 and 4 with 8 pairs of long sts. Sternite 5 with 15 long sts. Sternite 6 with 5 pairs of long sts. Tergopleurite II with 2 pairs of ps and 3 pairs of pss on each side. Tergopleurites III to VI with a small spiracle, 2 pairs of ps, and 4 pairs of pss on each side. Tergopleurite VII with a very small spiracle, 2 pairs of ps, and 3 pairs of pss on each side. Tergopleurite VIII has no spiracles, with 2 small pairs of ps and 2 pairs of pss on each side.
Head: As in Fig. 2B. Much as in males, except for the antennae parts, female antennae scape much slimmer and shorter than in males. Antennae scape in female stout and elongated with the lateral process; pedicel and flagellomere unfused, pedicel smaller than scape; flagellomere with the process, no curve.
Thorax: As in Fig. 2B. As much as in males. Abdomen: As in Fig. 2B. Wider than the thorax, oval to rounded in shape. Dorsally, 5 pairs of spiracles in small sizes are present on segments III-VII. One tergite per segment, except segments I and VIII without tergites and segment II with 2 tergites. Tergite 1 with 9 long tps, equal in size. Tergites 2 with 25 long tps. Tergite 3 with 13 pairs of long tps. Tergite 4 with 15 pairs of long tps. Tergite 5 with 14 pairs of long tps. Tergite 6 with 13 pairs of long tps. Tergite 7 with 12 pairs of tps, varies in size. Ventrally, six sternites with sts present from segment II to VII, all setae long in size. Sternites 1 and 2 each with 27 sts. Sternites 3 -5 each with 12 pairs of sts. Sternite 6 with 8 pairs of sts. Tergopleurite II with 2 pairs of ps and 4 pairs of pss on each side. Tergopleurite III with a small spiracle, 2 pairs of ps, and 4 pairs of pss on each side. Tergopleurites IV and V each with a small spiracle, 2 pairs of ps, and 3 pairs of pss on each side. Tergopleurite VI with a very small spiracle, 2 pairs of ps, and 4 pairs of pss on each side. Tergopleurite VII with a very small spiracle, 2 pairs of ps, and 3 pairs of pss on each side. Tergopleurite VIII with no spiracles, 2 small pairs of ps, and 2 pairs of pss on each side.
Female genitalia: As in Fig. 2E. The subgenital plate is broad and subtriangular-shaped with three pairs of setae, apical pair setae shortest, other two pairs equal in size; the terminal portion of the subgenital plate without apical setae; two projections bordering the concavity at the apex with 4 pairs of setae on each projection; ventral terminalia with small tapered gonapophyses VIII, the inner margin of gonapophyses VIII gently convex with 9 marginal setae each side; the outer margin of gonapophysis VIII within abdomen boundary.
Etymology: The species epithet is a noun in apposition referring to the species scientific name of the host.

Phylogenetic analyses and divergence time
We combined the mt cox1 and 12S rRNA genes for further phylogenetic analyses. All cox1 and 12S rRNA genes of specimens were individually aligned by ClustalX 1.83 (Aiyar, 2000), respectively. The mt genes phylogenetic reconstructions revealed similar topologies (Figs. 3 and 4), and strongly supported the monophyly of the family Trichodectidae lineage. Among phylogenetic analyses, the family Trichodectidae clade was comprised of four genera Trichodectes, Geomydoecus, Felicola, and Pancola; the other obvious clade was the family Bovicoliidae, which mainly consisted of two genera Bovicola and Damalinia. The P. ailurus grouped the consistent topology with species Felicola subrostratus, Geomydoecus aurei, and Trichodectes canis with moderate support (Fig. 3), and had a closer relationship with Trichodectes and Geomydoecus.
Considering genetic divergence at the species level in various mammal groups was from 2 -11% (Baker and Bradley, 2006), the distance for uncorrected cox1 and 12S rRNA genes among different species from families Trichodectidae and Bovicoliidae was 29.7 -34.6% and 38.9 -43.6%, respectively (Table 3), indirectly representing P. ailurus was the distinct species from genera Felicola, Geomydoecus, and Trichodectes.
Divergence analysis of cox1 alignment supported a clear separation with Pancola, Felicola, and Trichodectes within the family Trichodectidae, and dated the split time within them at ~24.52 Mya (with 95% highest posterior density, HPD, Mya) (Fig. 4). The topology of divergence analysis was the same as ML analysis. In all analyses, disregarding other unknown molecular data, the cox1 gene sequences from families Trichodectidae and Bovicoliidae were grouped into two branches separated from primate lice, Pediculus humanus and P. schaeffi. The divergence tree further indicated that associated divergence time within the family Trichodectidae (Felicola, Trichodectes, and Pancola) from 24.52 Mya (95% HPD, 15.93-35.52 Mya) to 31.91 Mya (95% HPD, 21.40-45.98 Mya), and the origin of Pancola was closer to Trichodectes.

Discussion
We describe one new species, Pancola ailurus n. sp., from Chinese red pandas. This description increases the number of known genera of Trichodectoidea from 19 to 20. First, the morphological features that distinguished those species have been identified as above-described. Second, results showed a phylogenetic relationship within the family Trichodectiidae with a divergence time of 24.52 -31.91 Mya (Fig. 4) and genetic distances greater than 29.0% (Table 3), which is higher than species-level (Baker and Bradley, 2006), supporting that Pancola was a distinct genus, and P. ailurus was a validate species. Consistent with hosts divergence and phylogenetic analyses that canids are closer to red pandas (Agnarsson et al., 2010;Nyakatura and Bininda-Emonds, 2012), relationships and the origin between Trichodectes and Pancola species were more related. However, Lyal (1983) proposed Trichodectes was more related to Felicola, but a relationship based on mt genes showed Trichodectes may be closer to Geomydcecus.
As L. bostrychophila was added as the outgroup with cox1 for 100 Mya and the estimated split time for human lice (P. humanus) and chimpanzee lice (P. schaeffi) was 5-7 Mya for cox1, the divergence time was estimated for Bovicola, Damalinia, Trichodectes, Felicola, and Pancola approximately ranging from 5.01 Mya to 36.27 Mya for cox1 gene. Similar to Light et al. (2010) analyses that the divergence time of the same genus was equal to/less than 15 Mya, the divergence time among Bovicola species (B. ovis, B. caprae, and B. bovis) was from 5.01 to 9.98 Mya, and Pediculus species diverged to P. schaeffi and P. humanus about 5.75 Mya.
Based on the morphological keys of Lyal (1985), the genus Pancola is probably closely related to Paratrichodectes. However, available sequences are absent from a large number of trichodectid genera, including Paratrichodectes, and no direct comparisons could be made. The correct placement of Pancola within Trichodectidae will therefore have to be evaluated when more sequences of different trichodectid genera are available. Additionally, overall phylogenetic relationships of Trichodectidae using molecular information were similar to Lyal (1983) and Kéler (1938) observations, suggesting that molecular markers could be a useful supplemented method to identify lice.   manuscript.

Data availability statement
The sequence data is uploaded to the NCBI GenBank and the raw sequences are available under the accession of ON964542, ON973802, ON964865, ON964866, and ON964867.

Table 3
Uncorrected pairwise genetic distance of the mitochondrial genes (cox1 and 12S rRNA) sequences of chewing lice from families Trichodectidae and Bovicoliidae used in this study. The lower left represents the genetic distance of cox1, and the upper right represents the genetic distance of 12S rRNA.