New Insights into the Origin and Evolution of Mysmenid Spiders (Araneae, Mysmenidae) Based on the First Four Complete Mitochondrial Genomes

Simple Summary The complete mitochondrial genome has been widely applied in the phylogenetics, population genetics and ecological research of animals due to its characteristics such as strict maternal inheritance and comparatively conserved genomic structure. Currently, the complete mitochondrial genomes of the Mysmenidae are not available. In this study, we reported the first four complete mitochondrial genomes of the Mysmenidae, including one aboveground species (Trogloneta yuensis) and three cave-dwelling species (T. yunnanense, Yamaneta kehen and Y. paquini). T. yunnanense was more similar to Yamaneta in mitogenome size than T. yuensis, possibly showing the convergent evolution of cave spiders. High variability was detected between the genera Trogloneta and Yamaneta. The phylogenetic analysis supports that Mysmenidae is a sister clade to the family Tetragnathidae. Our data and findings could enrich the gene database and contribute to the better understanding of the molecular characteristics of the family Mysmenidae, which will provide help for further studies related to the population genetics, molecular biology and phylogenetics of these spiders. Abstract The mitochondrial genome (mitogenome) is recognized as an effective molecular marker for studying molecular evolution and phylogeny. The family Mysmenidae is a group of widely distributed and covert-living spiders, of which the mitogenomic information is largely unclear. In this study, we obtained the first four complete mitogenomes of mysmenid spiders (one aboveground species: Trogloneta yuensis, and three cave-dwelling species: T. yunnanense, Yamaneta kehen and Y. paquini). Comparative analyses revealed that their lengths ranged from 13,771 bp (T. yuensis) to 14,223 bp (Y. kehen), containing a standard set of 37 genes and an A + T-rich region with the same gene orientation as other spider species. The mitogenomic size of T. yunnanense was more similar to that of Yamaneta mitogenomes than that of T. yuensis, which might indicate the convergent evolution of cave spiders. High variability was detected between the genera Trogloneta and Yamaneta. The A + T content, the amino acid frequency of protein-coding genes (PCGs) and the secondary structures of tRNAs showed large differences. Yamaneta kehen and Y. paquini contained almost identical truncated tRNAs, and their intergenic spacers and overlaps exhibited high uniformity. The two Yamaneta species also possessed a higher similarity of start/stop codons for PCGs than the two Trogloneta species. In selective pressure analysis, compared to Yamaneta, Trogloneta had much higher Ka/Ks values, which implies that selection pressure may be affected by habitat changes. In our study, the phylogenetic analysis based on the combination of 13 PCGs and two rRNAs showed that Mysmenidae is a sister clade to the family Tetragnathidae. Our data and findings will contribute to the better understanding of the origin and evolution of mysmenid spiders.


Introduction
Spiders are among the most diverse groups dating back to the Late Carboniferous. They are found in almost all terrestrial ecosystems on the planet and considered one of the most successful animal groups due to their outstanding evolutionary radiation and ecological plasticity [1,2]. The order Araneae currently contains more than 50,000 described species in 132 families [3]. The World Spider Catalog has recorded a total of 158 species in 14 genera of Mysmenidae. Although the family Mysmenidae is distributed worldwide, it is one of the least-studied family-level groups among orb-weaving spiders, and its diversity is grossly undersampled due to their small size and cryptic life style [4]. Most previous studies were restricted to a relatively small dataset for phylogenetic analyses with limited morphological and/or behavioral characteristics, or a few mitochondrial/nuclear gene sequences such as cytochrome oxidase subunit 1 (cox1), rRNAs (16S and 12S) and tRNAs, and nuclear genes of rDNA (18S and 28S) and histone (H3) [4][5][6]. These relatively short and rapidly evolving genetic makers are generally inadequate for resolving deeper-level relationships [7].
Generally, a typical mitochondrial genome (mitogenome) of a metazoan is a closedcircular, double-stranded molecule of 11-20 kb in size, which contains 13 protein-coding genes (PCGs), 22 transfer RNAs (tRNAs), two ribosomal RNAs (16S and 12S rRNA) and a non-coding sequence known as the control region for replication and transcription [1]. Since the 1980s, the mitogenome has given a new impetus to evolutionary genetics [8], which is characterized by a rapid evolutionary rate, a low recombination rate, maternal inheritance, a neutral evolution pattern and haploidy [9]. For example, birds and crocodiles are closely related in evolutionary history [10], and hippopotamus and whales began to diverge from about 55 million years ago [11].
Mysmenidae is widely distributed in tropical and temperate regions, even in some extreme environment, such as caves [4]. A cave is a special ecosystem with no sunlight at all, no plant growth, a high CO 2 concentration, a constant temperature that is close to the mean annual region temperature and little food [12]. To adapt to these hard habitats, the cavedwelling species exhibit diverse morphological characteristics such as pigment reduction, eye regression and appendage elongation [13]. Trogloneta yunnanense [14], Yamaneta kehen and Y. paquini [15] are karst cave spiders that live in the mountains of southwest China. The unavailability of complete mitogenomes of these cave-dwelling spiders could definitely confine the understanding of the evolution and genetic adaptation of these troglobionts. In this study, we sequenced, annotated and characterized the first four spider mitogenomes of the Mysmenidae (the three cave species aforementioned and Trogloneta yuensis, an aboveground species), to (1) explore the general characterizations of Mysmenidae mitogenomes, (2) assess whether the selection pressure between different habitats exhibits some differences and (3) further conduct phylogenetic analysis to explore the place of Mysmenidae in the evolutionary history of Araneae.

Sampling
The samples of four spider species were collected from Yunnan, Guizhou and Hunan Provinces, China, and their sampling locality information is provided in Table 1. The field collections did not involve endangered or protected species, and no specific permits were required for our collecting. Each specimen was preserved in the field in 95% ethanol and taken back to the laboratory stored at −20 • C. We identified species by the morphology of male palp and female epigyne.

Mitogenome Assembly, Annotation, Visualization and Comparative Analysis
The overall quality of the sequences was assessed from their Phred scores using fastp software [16]. Ambiguous nucleotides and raw sequence reads with lower than Q20 Phred score were trimmed and removed. NOVOPlasty [17] was used to de novo assemble the mitogenome. The assembled mitogenomes were submitted to MITOS WebServer [18] for initial gene annotation. The regions of protein-coding genes (PCGs), transfer RNAs (tRNAs) and ribosomal RNAs (rRNAs) were further validated using nucleotide-nucleotide BLAST (BLASTn) [19]. The maps of the mitogenomes were drawn using OGDraw v1.2 software (http://ogdraw.mpimp-golm.mpg.de/) (accessed on 20 March 2022) [20].
The secondary structures of 22 tRNA genes were predicted by MITOS WebServer [18], tRNAscan-SE 2.0 WebServer [21] and ARWEN [22]. For those tRNA genes failed to be identified, we determined them by comparing the mitogenomes with published spider mitogenomes in GenBank and proofread tRNA secondary structure features.
Statistical analyses of base content, amino acid composition and codon usage in the mitogenome sequences of four spider species were performed using MEGA X [23]. Overlapping regions and spacer regions between genes were detected manually. The mitochondrial genome skew values were calculated as follows: AT skew = (A − T)/(A + T) and GC skew = (G − C)/(G + C) [24]. The non-synonymous substitution rate (Ka), synonymous substitution rate (Ks) and Ka/Ks ratio were inferred with DnaSP v6.0 (the Ka/Ks ratio >1, =1, and <1 indicate positive, neutral and purifying selection, respectively) [25].

Phylogenetic Analysis
A total of 55 mitogenomes available in the Genebank and four mitogenomes obtained in this study were used for phylogenetic analysis, including 58 mitogenomes of spider species and a mitogenome of Limulus polyphemus ( Table 2). Limulus polyphemus is an ancient and slow morphological-evolving group of species, and is regarded as a keystone group for studies of evolution and arthropod phylogeny [26]. It was used as outgroup in this analysis. We aligned each gene of 13 PCGs and two rRNAs respectively, and Gblock [27] was used to analyze them, and then concatenated them all. The final aligned sequences were 10,506 sites for 13 PCGs and 1378 sites for two rRNAs. We used DAMBE to test whether the sequence is suitable for constructing a phylogenetic tree [28]. ModelFinder was used to find the most suitable model for both Maximum Likelihood (ML) analysis and Bayesian inference (BI) under the Akaike information criterion (AIC) [29]. The BI analysis was performed by using Mrbayes 3.2.7a [30] with two simultaneous runs and four Monte Carlo Markov chains (10,000,000 generations, sampling every 1000 generations and the first 25% of sampled trees were burn-in) until the average standard deviation of split frequencies was less than 0.01. The ML phylogenetic analysis was conducted in IQ-TREE v1.6.12 [31] with 1000 standard bootstrap replicates. In this study, ML and BI trees were constructed by a set of software integrated in the PhyloSuite program [32].  Figure 1). They were relatively smaller than other spider mitogenomes [33], especially T. yuensis, owing to its tiny A + T-rich region and extremely truncated tRNAs. A previous study on Habronattus oregonensis of the Salticidae indicated an overall trend toward minimization of the spider mitogenomes [33]. Interestingly, the mitogenome size of T. yunnanense was more similar to that of Yamaneta than T. yuensis. Each mitogenome shared the same 37 typical metazoan genes (13 PCGs, 22 tRNAs and two rRNA genes,) and a non-coding control region (Table 3). Meanwhile, the gene order of these four mitogenomes was conserved and identical to many other spiders. Circular maps of all four mitogenomes are shown in Figure 1. Similar to other published spider mitogenomes, the nucleotide composition of the mitochondrial genomes of the four spiders clearly favored A/T of the J-strands ( Table 4). The A + T content of Trogloneta was significantly higher than Yamaneta in each mitogenome region (Table 4). Additionally, all four mitogenomes showed negative AT-skews (-0.174 to -0.054) and positive GC-skews (0.288 to 0.428) ( Table 4). The AT-skews of the genus Trogloneta were lower than those of the genus Yamaneta, while the GC-skews of T. yunnanense, Y. kehen and Y. paquini were similar but not for T. yuensis (Table 3).
Each mitogenome had a large number of intergenic sequences (spacers and overlaps) ( Table 2). Mitogenomes of these four spiders were characterized with more intergenic overlaps than spacers. The largest spacer of Yamaneta was located between trnN and trnA (89 bp in Y. kehen and 87 bp in Y. paquini), which were much longer than those of Trogloneta (both 10 bp between trnP and nad6) ( Table 2). The higher content of intergenic overlaps than spacers was also evident in the mitogenomes of other spider species: such as Tetragnatha maxillosa, Tetragnatha nitens, Neoscona Scylla and Lyrognathus crotalus [34][35][36]. Moreover, the spacers and overlaps of Yamaneta showed higher uniformity than those of Trogloneta.

Protein-Coding Genes and Codon Usage
Among the 13 PCGs, only four genes (nad4, nad4L, nad5 and nad1) were encoded on the minority strand (N-strand), while the others were encoded on the majority strand (J-strand). In Trogloneta, the start codons of PCGs were characterized with five types: ATA, ATT, ATG, TTA and TTG, while there were six codon types in Yamaneta (additionally including ATC) ( Table 2). Most PCGs terminated with the TAA or TAG stop codon, while nad4 and nad4L had an incomplete stop codon T−.
The mitogenomes of Trogloneta only possessed five pairs of identical start/stop codons, while the Yamaneta had cox1 (TTA/TAA), atp6 (ATG/TAA), nad4 (ATA/T), nad4L (ATT/T) and nad5 (ATA/TAA) (Table 3). Overall, the presence of ATN as start codons was common in most PCGs except for cox1, cox2, cox3 and nad6. The start codons for cox1 and cox3 were TAA and TTG in these four spider species, respectively. The start codon for cox2 was TTG in Trogloneta and ATT in Yamaneta. The start codon for nad6 was TTA in Y. paquini and ATT in the other three spider species. Trogloneta had almost the same start/stop codons of PCGs except for the start codons of nad2 (ATA in T. yunnanense and ATT in T. yuensis), the stop codons of cytb (TAA in T. yunnanense and TAG in T. yuensis) and nad1 (TAA in T. yunnanense and TAG in T. yuensis) ( Table 3). Each pair of stop codons between Y. kehen and Y. paquini were the same, while they shared different stop codons in nad6 (ATT in Y. kehen and TTA in Y. paquini) and nad1 (ATT in Y. kehen and ATC in Y. paquini) ( Table 3).
In this study, a truncated stop codon (T) was detected as existing in nad4 and nad4L in all four spider species, which was similar to the posttranscriptional animals as numerous studies have reported [36,37], Ebrechtella tricuspidate [38], Tetragnatha maxillosa [34], Tetragnatha nitens [34] and Argiope perforate [39]. It was assumed that these incomplete stop codons were complemented by posttranscriptional polyadenylation [40]. Additionally, more than 10 bp overlaps were detected in the junctions between atp8 and atp6 in each spider mitogenome. Generally, hairpin structures at the 3' end of the upstream protein's mRNA may act as a signal for the cleavage of the polycistronic primary transcript. Both truncated stop codon and overlaps between genes indicated selective pressure power to reduce mitochondrial gene size.   The total numbers of non-stop codons in T. yunnanense, T. yuensis, Y. kehen and Y. paquini were 3593, 3584, 3585 and 3634, respectively. Amino acid frequencies varied significantly between genera but not within genus. The most frequently used amino acids occurred on the leucine (the UUR codon) (mean value = 12.17%) in Trogloneta and the phenylalanine (the UUY codon) (mean value = 9.21%) in Yamaneta (Figure 2). Cysteine was the least used amino acid in all four spider mitogenomes (mean value = 0.77%). Trogloneta contained more codons for Phe, Leu2, Ile, Met, Tyr, Asn and Lys than Yamaneta, while Leu1, Val, Ser2, Pro, Thr, Ala, Asp, Glu, Cys and Gly encoded fewer (Figure 2). Analysis of the relative synonymous codon usage (RSCU) revealed the biased usage of A/T rather than G/C at the third codon position. Yamaneta used C/G as the third codon more frequently than Trogloneta (Figure 3). In this study, a truncated stop codon (T) was detected as existing in nad4 and nad4L in all four spider species, which was similar to the posttranscriptional animals as numerous studies have reported [36,37], Ebrechtella tricuspidate [38], Tetragnatha maxillosa [34], Tetragnatha nitens [34] and Argiope perforate [39]. It was assumed that these incomplete stop codons were complemented by posttranscriptional polyadenylation [40]. Additionally, more than 10 bp overlaps were detected in the junctions between atp8 and atp6 in each spider mitogenome. Generally, hairpin structures at the 3' end of the upstream protein's mRNA may act as a signal for the cleavage of the polycistronic primary transcript. Both truncated stop codon and overlaps between genes indicated selective pressure power to reduce mitochondrial gene size.
The total numbers of non-stop codons in T. yunnanense, T. yuensis, Y. kehen and Y. paquini were 3593, 3584, 3585 and 3634, respectively. Amino acid frequencies varied significantly between genera but not within genus. The most frequently used amino acids occurred on the leucine (the UUR codon) (mean value = 12.17%) in Trogloneta and the phenylalanine (the UUY codon) (mean value = 9.21%) in Yamaneta (Figure 2). Cysteine was the least used amino acid in all four spider mitogenomes (mean value = 0.77%). Trogloneta contained more codons for Phe, Leu2, Ile, Met, Tyr, Asn and Lys than Yamaneta, while Leu1, Val, Ser2, Pro, Thr, Ala, Asp, Glu, Cys and Gly encoded fewer (Figure 2). Analysis of the relative synonymous codon usage (RSCU) revealed the biased usage of A/T rather than G/C at the third codon position. Yamaneta used C/G as the third codon more frequently than Trogloneta (Figure 3).    To explore the selection pressure between species with different habits, we calculated Ka, Ks and Ka/Ks values within genera. The Ka values varied between 0.017 and 0.107 in Yamaneta. Compared to Yamaneta, the Ka values displayed relatively higher ranges (0.043-0.346) in Trogloneta, Among the 13 mitochondrial PCGs, atp8 had the biggest Ka/Ks values (0.167 in Yamaneta and 0.835 in Trogloneta). By contrast, the lowest was cox1 (0.054 in Yamaneta and 0.116 in Trogloneta) ( Table 5). Mitochondria play a key role in energy metabolism. Previous studies have found that cave spiders and ground spiders have different metabolic rates [41]. In this study, the Ka/Ks values of all 13 PCGs were less than 1, which means they are all under purifying selection. However, compared to Yamaneta, Trogloneta has much higher Ka/Ks values, which may imply that the effective population size is small in the cave population affected by habitats changes.

Transfer RNA and Ribosomal RNA Gene
It has been widely accepted that the cloverleaf secondary structure of transfer RNA (tRNA) is one of the most conserved features for the mitogenome since the first extremely truncated tRNAs were found in the jumping spider H. oregonensis [33]. However, almost all reported spider species contain atypical tRNA secondary structures. A lost-arm tRNA in Arachnids compared to other metazoans or arthropods has been reported more often [42]. It was indicated that the genome-wide propensity to lose sequences that encode canonical cloverleaf structures likely evolved multiple times within arachnids [43]. Given the otherwise extreme conservation of tRNA structure across all of life, one hypothesis was that it was the result of parallel evolution under the pressure of selection pressure [44]. Meanwhile, a posttranscriptional editing mechanism likely edited spider mitogenome tRNA acceptor stems to enable them to function [45]. Thus, these tRNAs with truncated arms could be valuable markers for deep-level phylogenetic inference.
For these 22 typical animal tRNA genes in each Mysmenidae mitogenome, 14 tRNAs were encoded by the J-strand and the remaining eight were located on the N-strand, ranging from 37 bp (trnV in T. yuensis) to 70 bp (trnI in Y. kehen and Y. paquini). Most of the tRNAs in Trogloneta and Yamaneta mitogenomes had aberrant cloverleaf secondary structures, including a truncated aminoacyl acceptor stem and mismatched (lacking wellpaired) aminoacyl acceptor stem (Figure 4). The tRNAs' secondary structure exhibited a conservative type in Yamaneta. All Yamaneta tRNAs possess a anticodon arm, and 11 pairs of the same tRNAs (trnC, trnD, trnG, trnK, trnL1, trnN, trnI, trnV, trnA, trnS1 and trnW) lack a TΨC (pseudouracil) arm. However, in Trogloneta, six of the twenty-two tRNAs share different secondary structures (trnD, trnG, trnF, trnI, trnL1 and trnV), and a DHU (dihydrouracil) arm was absent in three tRNAs (trnA, trnS1 and trnS2) of both T. yunnanense and T. yuensis (Figure 4). A TΨC arm was absent in nine tRNAs of T. yunnanense and twelve tRNAs of T. yuensis. Our results demonstrated a large divergence in the secondary structure of tRNAs at the genus level, and only four tRNAs (trnG, trnK, trnN and trnW) shared the common absence of a TΨC arm among these four spider species. (Figure 4), suggesting a high degree of variability between the two genera. Previous studies found that trnS2 (UCN) and trnS1 (AGN) lacked the DHU arm in many arthropod mitochondrial genomes [46]. Wolstenholme [47] once described that the gene coding for the DHU arm of trnS1 had been absent prior to metazoan diversification. However, we found that the trnS1 in Yamaneta has an intact secondary structure ( Figure  4). The typical cloverleaf structure was also found in Adoxophyes honmai and Pseudocellus pearsei [48,49]. The absence of the TψC arm or DHU arm resulted in shorter tRNA gene lengths and more compact gene structures. Lavrov [45] suggested that RNA editing mechanisms may play a key role in posttranscriptional editing to modify these atypical tRNAs. Previous studies found that trnS2 (UCN) and trnS1 (AGN) lacked the DHU arm in many arthropod mitochondrial genomes [46]. Wolstenholme [47] once described that the gene coding for the DHU arm of trnS1 had been absent prior to metazoan diversification. However, we found that the trnS1 in Yamaneta has an intact secondary structure (Figure 4). The typical cloverleaf structure was also found in Adoxophyes honmai and Pseudocellus pearsei [48,49]. The absence of the TψC arm or DHU arm resulted in shorter tRNA gene lengths and more compact gene structures. Lavrov [45] suggested that RNA editing mechanisms may play a key role in posttranscriptional editing to modify these atypical tRNAs.
Large and small subunit rRNAs (rrnL and rrnS) were adjacent on the N-strand and spaced by a single tRNA (trnV). Yet it was hard to accurately predict the ends of rRNAs using DNA sequencing alone. We assumed that the ends of the rRNAs extended to the boundaries of the flanking genes [34]. The length of the predicted rRNAs of these four spider species did not differ much. The length of rrnL ranged from 1001 bp in T. yunnanense to 1043 bp in Y. kehen, and the length of rrnS ranged from 686 bp in T. yunnanense to 692 bp in T. yuensis and Y. kehen. The A + T contents of rRNAs were distinct between the two genera. In Trogloneta, the A + T contents were 83.8% in T. yunnanense and 83.5% in T. yuensis; however, in Yamaneta, they were 66.9% in Y. kehen and 66.5% in Y. paquini.

Control Region
The putative control region located between rrnS and rrnL was the longest non-coding region in the whole mitogenome. It played a role in initiating and regulating replication and transcription in mitochondria. The full lengths of the CR in the four mitogenomes were 631 bp (T. yunnanense), 317 bp (T. yuensis), 591 bp (Y. kehen) and 581 bp (Y. paquini), respectively. The A + T content of the control region of the Trogloneta and Yamaneta mitogenomes was AT-rich (Table 2), with negative AT skewness value in T. yunnanense and positive values in T. yuensis, Y. kehen and Y. paquini. The GC skewness value was positive in all four spider mitogenomes.

Phylogenetic Analysis
The results of DAMBE and the best-fit models of partitioned analyses can be seen in Figure S1 and Table S1. Almost identical topologies of phylogenetic trees were obtained by the BI and ML methods ( Figure 5 and Figure S2). It was clearly observed that the two genera (Trogloneta and Yamaneta) were clustered into two adjacent clades. The Mysmenidae is a sister group to the Tetragnathidae with bootstrap = 87 and posterior probabilities = 1.00, indicating that Mysmenidae is a member of the superfamily Araneoidea. This evolutionary structure is consistent with earlier studies that used morphological and gene sequence data [50]. In addition, the superfamily Araneoidea was confirmed to be monophyly according to the phylogenetic analysis. The Hypochilidae, which is regarded as an ancient and special taxon in spiders, was also isolated from the families Dysdeidae, Sicariidae and Pholcidae. The RTA clade, containing nine families in this analysis, showed high supporting values for monophyly. However, there were some branches in the RTA clade that were not well supported, such as Telamonia vlijmi and Plexippus paykulli (bootstrap = 57 and posterior probabilities = 0.736). In this study, two infraorder Mygalomorphae and Araneomopha were clearly split, as inferred from both BI and ML analyses. Moreover, the results strongly supported the monophyletic characteristic of two suborders (Opisthothelae and Mesothelae) in Araneae, with bootstrap = 100 and posterior probabilities = 1.00.

Conclusions
In the present study, the complete mitogenomes of T. yunnanense, T. yuensis, Y. kehen and Y. paquini were determined and characterized. Each mitogenome contained an identical composition, gene order and relatively shorter length compared with most spider mitogenomes. However, the mitogenomes of Trogloneta and Yamaneta showed high variability. The overall lengths of the Yamaneta mitogenomes were relatively longer. They possessed an extremely large spacer about 90 bp (between trnN and trnA), which had never been found in previous studies. The content of A + T in Trogloneta was significantly higher than that in Yamaneta among each gene. In addition, the amino acid frequencies of the PCGs and secondary structures of tRNAs between the two genera were significantly different. We found that the mitogenomes of Yamaneta were more conserved, especially for the secondary structures of tRNAs that were regarded as markers for deep-level phylogenetic inference. Y. kehen and Y. paquini contained almost identical truncated arms Figure 5. Phylogenetic tree from Araneae species based on nucleotide sequence of 13 PCGs and 2 rRNAs using Bayesian inference (BI). Limulus polyphemus was used as an outgroup. Numbers below the nodes refer to Bayesian posterior probabilities in percentages. The specimen used in our experiment are marked in red.

Conclusions
In the present study, the complete mitogenomes of T. yunnanense, T. yuensis, Y. kehen and Y. paquini were determined and characterized. Each mitogenome contained an identical composition, gene order and relatively shorter length compared with most spider mitogenomes. However, the mitogenomes of Trogloneta and Yamaneta showed high variability. The overall lengths of the Yamaneta mitogenomes were relatively longer. They possessed an extremely large spacer about 90 bp (between trnN and trnA), which had never been found in previous studies. The content of A + T in Trogloneta was significantly higher than that in Yamaneta among each gene. In addition, the amino acid frequencies of the PCGs and secondary structures of tRNAs between the two genera were significantly different. We found that the mitogenomes of Yamaneta were more conserved, especially for the secondary structures of tRNAs that were regarded as markers for deep-level phylogenetic inference. Y. kehen and Y. paquini contained almost identical truncated arms of tRNAs. The trnS1 of Y. kehen and Y. paquini had an intact secondary structure, which was a very rare phenomenon in other species. In Trogloneta, on the other hand, six of the twenty-two tRNAs shared different secondary structures. Moreover, the spacers and overlaps were more even between Y. kehen and Y. paquini, and they also possessed more similar start/stop codons for PCGs than T. yunnanense and T. yuensis. All of the above evidence supports a higher degree of genetic divergence between T. yunnanense and T. yuensis. For phylogenetic analysis, the mitogenome revealed the lineage of Araneae with high quality. Our studies indicate that the mitogenomes possess the potential for better exploration of Araneae, and to obtain the sites that are hard to reach with morphological or other molecular methods.
Supplementary Materials: The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/ani13030497/s1, Figure S1: The result of substitution saturation test by using DAMBE; Figure S2: The ML phylogenetic tree from Araneae species based on nucleotide sequence of 13 PCGs and 2 rRNAs; Table S1: The best-fit models of each partition Author Contributions: Conceptualization and writing-review and editing, C.Z. and Y.L.; writingoriginal draft preparation, S.L.; data analysis, S.L., S.W. and Q.C. All authors have read and agreed to the published version of the manuscript.
Funding: This study was supported by the National Natural Science Foundation of China by Yucheng Lin (NSFC-31772410, 31972870).

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The complete mitochondrial genomes generated in this study have been deposited in GenBank with accession numbers listed in Table 2.