A molecular phylogeny of the Chinese Sinopoda spiders (Sparassidae, Heteropodinae): implications for taxonomy

Sinopoda spiders are a diverse group with limited dispersal ability. They are remarkably sympatric among related species, which often results in misidentification and incorrect matching of sexes. In order to understand the evolutionary relationships and revise the taxonomy problems in this genus, we offer the first molecular phylogeny of Sinopoda. Our results strongly support the monophyly of Sinopoda and its sister relationship with Spariolenus and reject the monophyly of the S. okinawana species group. We establish three new species groups based on both molecular and morphological data. Our phylogeny also illuminates some taxonomic issues and clarifies some species limits: (1) Supporting the newly revised matching of sexes in S. longiducta and S. yaanensis by Zhong et al. (2019). (2) The original description of S. campanacea was based on mismatched sexes. S. changde is proposed as a junior synonymy of S. campanacea, while the original female ‘S. campanacea’ is here described as a new species: S. papilionaceous Liu sp. nov. (3) The type series of S. serpentembolus contains mismatched sexes. The female is considered as S. campanacea, while we here report the correctly matched females of S. serpentembolus. (4) We describe one additional new species: S. wuyiensis Liu sp. nov. Our first molecular phylogeny of Sinopoda provides a tool for comparative analyses and a solid base for the future biodiversity and taxonomic work on the genus.


INTRODUCTION
Taxonomy can be challenging in groups that have undergone relatively little morphological change through the speciation processes. This is especially true when there is extensive sympatry among related species, a rather rare phenomenon (Agnarsson et al., 2016). While the majority of spider taxonomy is still purely based on morphological data, integrative approaches are critical to address taxonomy in such challenging groups (Godwin & Bond, Most of these individuals were collected by the members of our laboratory and others were provided by the colleagues from Southwest University. A total of 856 individuals from 12 Provinces (Fujian, Gansu, Guizhou, Hainan, Henan, Hubei, Hunan, Jiangxi, Liaoning, Shanxi, Sichuan and Yunnan), one Municipality (Chongqing) and 1 Autonomous Regions (Xizang Autonomous Region) were collected from the field. Every specimen was given a unique identification number ('S' number). Species were initially sorted by morphological characteristics and stored in 70% ethanol for morphological work and in 100% ethanol for molecular analyses. In total, we included 70 specimens's sequences of the genus Sinopoda to molecular analyses including three individuals from Genbank. Individual data (including species name, sample locations and GenBank Accession Numbers) are provided in Supplement 1.

Molecular protocols
One or two legs of each individual (depending on the size of specimens) were used to extract total genomic DNA. DNA extraction was achieved with the Universal Genomic DNA Kit (CWBIO, Beijing, China). We used a target gene approach including both mitochondrial and nuclear genes. Six loci were targeted with different degrees of variability. Two mitochondrial genes (two regions including 16S ribosomal RNA gene (16S) and cytochrome c oxidase subunit 1 (COI)) and four nuclear genes (protein-coding histone H3 (H3), 18S ribosomal RNA gene (18S), 28S ribosomal RNA gene (28S) and Internal Transcribed Spacer 2 (ITS2)) were used in this research. Primers (Folmer et al., 1994;Simon et al., 1994;White et al., 1990) and PCR conditions are shown in Table 1. Multiple primers were employed in the amplification of a large region of COI (approximately 1.2 kb). These primers include the pairs LCOI1490 and HCOI2198, and Jerry and C1-N-2776. Fragments were sequenced by the companies of Tsingke Biological Technology (Wuhan, China) and Sunny Biotechnology Company Limited (Shanghai, China) in both directions. Sequences were assembled and edited using the Chromaseq module in Mesquite (Maddison & Maddison, 2011a;Maddison &Maddison, 2011b) employing Phred andPhrap (Green, 1999;Green & Ewing, 2002). After assembly, to all sequences were blasted against Genbank (National Center for Biotechnology Information (NCBI)) to verify they all belonged to the family Sparassidae.

Phylogenetic analyses
All sequences were aligned with MAFFT (Katoh, 2013) on XSEDE in parallel on the Cyberinfrastructure for Phylogenetic Research Project (CIPRES Science Gateway) at the Table 1 Molecular markers and primers used for amplification. The amplification was performed in 50 µl final volume containing 18 µl of ultra-pure water (dd H 2 O), 25 µl of I-5 TM 2X High-Fidelity Master Mix, 2 µl of each primer (100 pmol/ µl), 3 µl of the genomic spider DNA templates. PCR settings list Initial Denaturation (ind), followed by /n cycles (Denaturation: de, Primer Annealing: pra, Primer Elongation: pre), and one Terminal Elongation (tee). (Temperature in • C following by time in seconds).

Marker
Primer name Premier sequence ( UC San Diego Supercomputing Center (Miller, Pfeiffer & Schwartz, 2010). Other large analyses were performed also using this platform. Considering the lack of gaps, we used the L-INS-i method to align the protein-coding genes H3 and COI. We verified absence of stop codons by translating sequences to amino acids. In virtue of the highly variable structure of ribosomal RNA genes, the ambiguously aligned regions were excluded by using the E-INS-i method to align the following four genes: 16S, 18S, 28S, ITS2 (Wheeler et al., 2016). We concatenated these six aligned genes in Mesquite.
Two analytical methods (Maximun Likehood and Bayesian) were used to estimate the phylogenetic relationships. In all analyses, we treated the gaps and ambiguous as missing data. Trees for all target genes were also reconstructed. Bayesian inference analyses were performed via the parallel MrBayes 3.2.6 (Ronquist et al., 2012) on XSEDE. Due to the highly substitution rates of the third position, protein-coding genes (COI and H3) were implemented three different partition schemes, namely as COI-1st, COI-2nd, COI-3rd, H3-1st, H3-2nd and H3-3rd. For sensitivity analyses of the multilocus dataset, six genes were divided into ten data partitions, the jModelTest2 on XSEDE (2.1.6) (Darriba et al., 2012) were used to choose the most suitable and best-fit models for mtDNA and nuDNA, according to the Akaike information criterion (AIC) (Posada & Buckley, 2004). The model parameters were estimated during the analyses and the choice by the jModelTest2 on XSEDE (2.1.6). For 16S, 28S, COI-2nd and ITS2, we used the model of GTR + I + G. The best model for 18S is GTR. GTR + G for COI-3rd and H3-1st. HKY + I + G model were used for the partitions of COI-1st. HKY + G for H3-3rd. For H3-2nd, we used the model of HKY. For every analysis, 5*10 7 generations were run for two simultaneous independent analyses with four Markov Chains (one cold and three heated) and every 1000th states were saved for the current tree file. Based on the TRACER v1.7.1 (Rambaut & Drummond, 2007), all the results for the posterior distributions of the parameters had an Effective Sample Size (ESS) ≥ 200. The first 25% trees (1.25 ×10 7 generations) of every run were discarded as burn in. Maximum likelihood (ML) analyses were performed using IQ-Tree on XSEDE (Nguyen et al., 2015) on the focal matrix with same partitions as implemented in the Bayesian analysis. Node support was estimated using ultrafast bootstrapping with 1,000 replicates (Hoang et al., 2018) and the Shimodaira-Hase-gawa approximate likelihood-ratio test (SH-aLRT) with 1,000 replicates (Guindon et al., 2011).

Taxonomy
Specimens were examined with an Olympus SZX16 stereomicroscope; details were further investigated with an Olympus BX51 compound microscope. Epigyne were cleared in proteinase K at 56 • C to dissolve non-chitinous tissues. Photos were taken with Leica M205C stereomicroscope and Olympus BX51 equipped with a Micropublisher 3.3 RTV camera (QImaging, Surrey, BC, Canada). The digital images depicting the habitus and genital morphology were a composite of multiple images taken at different focal planes along the Z axis and assembled using the software package Helicon Focus 3.10.
Leg measurements are shown as: total length (femur, patella, tibia, metatarsus, tarsus). Numbers of spines are listed for each segment in the following order: prolateral, dorsal, retrolateral, ventral (in femora and patellae ventral spines are absent and fourth digit is omitted in the spination formula). All measurements are in millimeters.

Nomenclatural acts
According to the International Commission on Zoological Nomenclature (ICZN), the electronic version of this article in portable document format (PDF) will represent a published work. The new species names contained in the electronic version are effectively published under that Code from the electronic edition alone. This article and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix http://zoobank.org/. The LSID for this publication is: urn:lsid:zoobank.org:pub: DE32C06B-FB70-497D-A690-74DED52939DB.

PHYLOGENETIC RESULTS
Our full DNA matrix contains 83 individuals, 70 of which belongs to 38 Sinopoda species including about one-third (29%) of the known species and one new species. For outgroups we included 13 individuals belonging to 10 species in five genera of the subfamily Heteropodinae. The aligned sequences amounted to 460 bp for 16S (65 individuals), 818bp for 18S (42 individuals), 698bp for 28S (79 individuals), 1155bp for COI (79 individuals), 330bp for H3 (74 individuals), 386bp for ITS2 (66 individuals). The phylogenetic trees from the two phylogenetic methods (Bayesian inference and Maximum Likelihood) were highly consistent with relatively high posterior probabilities (PP) and bootstrap values (BS). Hence, we showed the nodal supports with these two analyses together (Fig. 2) on the BI topology. In general, Bayesian posterior probabilities were slightly higher than ML bootstrap supports. The monophyly of Sinopoda was robustly supported (PP 1.00, BS 100%). Spariolenus was supported as the sister group of Sinopoda (PP 0.93, BS 64%), which is not consistent with former study (Moradmand, Schönhofer & Jäger, 2014). The phylogeny suggested the polyphyly of the S. okinawana-group, as S. nuda is far removed from the remaining group members. We established three new species groups, according to the phylogeny, all supported by morphological and molecular characters: S. anguina-group (Fig. 3, PP 1.00, BS 97%), S. globosa-group (Fig. 4, PP 0.97, BS lower than 50%) and S. tumefacta-group (Fig. 5, PP 1.00, BS 100%). These are diagnosed, described and illustrated in detail in the following taxonomy part. Male and female individuals from the Tianping Mountain belonging to S. campanacea and S. changde were analysed. Male S. campanacea form a monophylum with the female S. changde (PP 0.98, BS 94%). The originally described female of S. campanacea is considered as a new species: S. papilionaceous Liu sp. nov. which is redescribed in current paper according to the original illustration (Wang, 1990: 7, Figs. 4-5). For S. serpentembolus, the male of S. serpentembolus from Shanxi Province and female individuals from Henan Province are monophyletic (PP 1.00, BS 100%). However, females differ significantly from the originally matched female of S. serpentembolus (Zhang et al., 2007: 251, Figs. 5-6)

DISCUSSION
We provide the first phylogenetic analysis of Sinopoda, focusing on the Chinese species. Our analysis strongly supports the monophyly of Sinopoda. According to a previous study (Moradmand, Schönhofer & Jäger, 2014), Sinopoda was hypothesized to group with Heteropoda and Spariolenus. We find support to the sister relationship between Sinopoda and Spariolenus, however, further sampling of Heteropodinae genera will be necessary to further clarify the placement of Sinopoda. Since the taxonomy of Sinopoda is poorly known and hitherto little genetic data has been available, species identification and matching of sexes has been challenging and has relied chiefly on field data. However, because many species share very similar morphology and due to extensive species sympatry, misidentifications are common and mismatches of sexes are common in Sinopoda. The sex matching of Sinopoda spiders is guided by field data and morphology but should in general be tested using molecular evidence. Such molecular testing of hypothesis is absolutely critical in cases of highly similar sympatric species. Our phylogeny and DNA evidence clarify the taxonomy and species level classification of Sinopoda spiders, and tests and clarifies recent taxonomic rematching of sexes (Zhang, Zhang & Zhang, 2015;Zhong et al., 2019). Our study also provides strong evidence for further rematching of sexes and recircumscription of some species, and the results in the description of two new species. The S. campanacea holotype is male, we therefore propose S. changde as a new synonym of S. campanacea. The originally described female of S. campanacea is considered as a new species: S. papilionaceous Liu sp. nov. which is redescribed in current paper according to the original illustration (Wang, 1990: 7, Figs. 4-5). For S. serpentembolus, the male of S. serpentembolus from Shanxi Province and female individuals from Henan Province are monophyletic (PP 1.00, BS 100%). However, females differ significantly from the originally matched female of S. serpentembolus (Zhang et al., 2007: 251, Figs. 5-6) based on morphological data. Therefore, we revised the S. serpentembolus as following: (1) The correct female of S. serpentembolus is reported for the first time.
(2) The originally mismatched female of S. serpentembolus is found to be similar to S. campanacea where we tentatively place it. We note, however, that it shows some morphological differences with S. campanacea and further molecular data is needed to clarify its placement. In addition to providing crucial matching information on species our phylogeny aids in the organization of species into species groups that can be useful units for evolutionary and biogeographical questions. On the basis of morphological resemblance, many Sinopoda species have been classified into different groups. We established three new groups: S. anguina-group, S. globosa-group and S. tumefacta-group basing on both molecular and morphological data (Figs. 2-5). These groups have their unique biogeography, although some have partially overlapping geographical distribution. Members of S. anguina-group, which includes 12 species (only four are included in our analysis) distributed in southern area of the Ailao Shan-Red River Fault zone (Southeast Asia, Hengduan Moutains of Yunnan). Members of S. globosa-group are distributed in the mountains surrounding the Sichuan Basin.
Our analysis indicates that the morphologically-based ''S. okinawana-group'' (Fig. 6) is polyphyletic. Because diagnostic characters for the females are weak (Jäger & Ono, 2002;Liu, Li & Jäger, 2008;Zhong et al., 2018), this group was original diagnosed mainly based on male genital characters, especially the reduced embolic apophysis and ventral RTA. However, the reduced embolic apophysis is homoplastic and also occurs in other Sinopoda species such as S. tumefacta and S. yaanensis. In addition, the embolic apophysis is totally absent in S. longshan Yin et al., 2000 andS. nuda Liu, Li &Jäger, 2008. These evidences suggest that the embolic apophysis evolves rapidly enough in this genus to not be reliable for group diagnostics. Our observations of the RTA, and prior studies, suggest that it may also evolve rapidly. In general, the characters of male palp may not be the best evidence to guide classification at the higher level due to rapid evolution (Eberhard, 2004). This conclusion is consistent with the finding in Australian huntsman spiders (Agnarsson & Rayor, 2013). A further revision of the ''S. okinawana-group'' is necessary including more species and other putative synapomorphies. It seems likely that at least a portion of the current group will form a clade as almost all members of ''S. okinawana-group'' are distributed in similar low-altitude areas, such as Northeast China Plain (Zhang et al., submitted).
In conclusion, we argue that integrative taxonomy-approaching species delimitation and description using multiple lines of evidence (Godwin & Bond, 2021)-is critical in Sinopoda, a diverse genus of morphologically similar spiders. Our study provides a solid base for understanding the biodiversity and taxonomy of Sinopoda, and a tool for comparative studies, such as analyses of biogeographical patterns in this genus. Nevertheless, we highlight that our only contains about a third of all the known species and that some of the deeper nodes, and several of the shallower clades in this phylogeny, are weakly supported. Further assessment of Sinopoda biodiversity and phylogeny further sampling of species, and updated genetic approaches through next generation sequencing techniques will be necessary.

Family Sparassidae Bertkau, 1872 Subfamily Heteropodinae Thorell, 1873
Genus Sinopoda Jäger, 1999 Type species: Sarotes forcipatus (Karsch, 1881) Diagnosis: Small to large spiders with laterigrade legs. Male palp with bifurcated RTA (retrolateral tibial apophysis), membranous conductor which is arising from distal part of tegulum, with embolic apophysis in most species. Female epigynum with modified epigynal rims and distinct lobal septum. Female vulva with internal duct system fused along median line which is divided into a basal part and a head, situated laterally from the entrance of copulatory ducts into internal duct system (Grall & Jäger, 2020;Jäger, 1999;Liu, Li & Jäger, 2008;Zhang, Zhang & Zhang, 2015). Diagnosis. S. campanacea is similar to S. serpentembolus in having strongly curved and sheet-shaped embolic apophysis, developed RTA with short and broad vRTA, longer dRTA in male, the epigyne with anterio-lateral margins of lateral lobes almost parallel with posterior margin of epigyne in female, but can be distinguished from the latter by the following characters: (1) The sperm duct of S. campanacea is almost straight, but significantly curved in S. serpentembolus. (2) The tegular apophysis is absent in S. campanacea, but present and located posteriorly in S. serpentembolus. (3) The glandular appendages are widely separated from posterior part of internal duct system in S. campanacea, but distinctly close with each other in S. serpentembolus. Description. For details see S. changde Zhong et al. (2019). Remarks. S. changde is proposed as the new synonym of S. campanacea based on the following evidences: (1) S. changde was also collected in Tianpingshan Scenic Area where the holotype of S. campanacea was collected (Fig. 2). (2) The same RTA, the curved and sheet-shaped embolic apophysis, the short and slender embolic tip indicate that the male of S. changde belongs to S. campanacea. (3) Zhong et al. (2019) indicated that the main difference in the male palp between S. changde and S. campanacea was the palpal tegulum covering proximal part of embolus in S. changde but not in S. campanacea, we found it was due to the photos taken at different angle. (4) Zhong et al. (2019) proposed S. changde as a new species mainly based on the difference between S. changde and S. campanacea is in the female genitalia, while the matched individuals of S. changde (including two females and one male collected from different localities) are strongly monophyletic in the molecular phylogeny (Fig. 2). Therefore, we consider that the originally assigned female of S. campanacea should be another new species which is described as a new species in the following part.

Diagnosis.
This new species can be distinguished from other Sinopoda species by the papilionaceous shape of internal duct systems based on the original illustrations from Wang (1990: 7, Fig. 5). . Mediaum sized Heteropodinae. PL 5.8, PW 4.8; OL 6.1, OW 4.0. Cheliceral furrow with 3 anterior and 4 posterior teeth. Dorsal prosoma deep yellowish-brown, with a yellow spot in the middle part. Fovea and redial furrows distinctly dark brown. Dorsal opisthosoma brownish black, and a yellow triangular macula in posterior part. Ventral opisthosoma yellow. Female genitalia: Epigynal field wider than long, without anterior bands. Lobal septum narrow. Anterior and posterior margins of lateral lobes almost parallel. Internal duct system anteriorly touching each other at the median line but posteriorly and widely separated. Posterior part of internal duct system slightly wider than anterior part. Fertilization ducts arising posterio-laterally (Wang, 1990: 7, Figs. 4-5). Distribution. Hunan Province, China. Remarks. We didn't examine the holotype specimen of S. papilionacea, because the type specimens may be lost. Female of S. papilionacea is easily identified as a new species according to its special internal duct system based on the original illustrations. by brown hairs. Ventral opisthosoma uniformly yellowish-brown with some irregular. Legs yellowish-brown, with dark setae. Female genitalia: Epigynal field wider than long, without anterior bands and slit sensilla. Lobal septum anteriorly around 4/5 of epigyne width, anterior part wider than middle part. Lateral lobes fused. Anterior part of internal ducts system diverging. Glandular appendages long and distinctly curved. Posterior part of spermathecae, bulging laterally, fertilization ducts arising posterio-laterally (Figs. 7A-7B). Abbreviations ALE anterior lateral eyes AME anterior median eyes AB anterior bands AW anterior width of prosoma C conductor CH clypeus height dRTA dorsal retrolateral tibial apophysis E embolus