Comparative Mitogenomics of Jumping Spiders with First Complete Mitochondrial Genomes of Euophryini (Araneae: Salticidae)

Simple Summary Salticidae, a highly diverse lineage of spiders with 671 genera and 6497 species reported worldwide, are ideal model organisms for ecological, behavioral and evolutionary studies. Euophryini, one of the largest tribes of jumping spiders, lacks information on the features of mitochondrial genomes based on completely annotated data. Here, we sequenced and assembled the mitochondrial genomes of two euophryine species, Corythalia opima (G. W. Peckham & E. G. Peckham, 1885) and Parabathippus shelfordi (G. W. Peckham & E. G. Peckham, 1907). By comparing all available well-characterized mitogenomes of 13 salticid species, the common features and differences in the mitochondrial genomes of jumping spiders were explored. The implications of mitogenomic features for the evolution and taxonomy of jumping spiders were investigated. Abstract Salticidae is the most species-rich family of spiders with diverse morphology, ecology and behavior. However, the characteristics of the mitogenomes within this group are poorly understood with relatively few well-characterized complete mitochondrial genomes. In this study, we provide completely annotated mitogenomes for Corythalia opima and Parabathippus shelfordi, which represent the first complete mitogenomes of the tribe Euophryini of Salticidae. The features and characteristics of the mitochondrial genomes are elucidated for Salticidae by thoroughly comparing the known well-characterized mitogenomes. The gene rearrangement between trnL2 and trnN was found in two jumping spider species, Corythalia opima and Heliophanus lineiventris Simon, 1868. Additionally, the rearrangement of nad1 to between trnE and trnF found in Asemonea sichuanensis Song & Chai, 1992 is the first protein-coding gene rearrangement in Salticidae, which may have an important phylogenetic implication for the family. Tandem repeats of various copy numbers and lengths were discovered in three jumping spider species. The codon usage analyses showed that the evolution of codon usage bias in salticid mitogenomes was affected by both selection and mutational pressure, but selection may have played a more important role. The phylogenetic analyses provided insight into the taxonomy of Colopsus longipalpis (Żabka, 1985). The data presented in this study will improve our understanding of the evolution of mitochondrial genomes within Salticidae.


Introduction
Mitochondrial genomes have been widely applied in studies, for example on the molecular evolution and the phylogeny of various animal lineages, because of their highly conserved structure, low recombination frequency and rapid evolutionary rate [1][2][3]. The mitochondrial genome of Metazoa is a circular, double-stranded molecule that typically consists of 13 protein-coding genes (PCG), 22 transfer RNA genes (tRNA), 2 ribosomal genes (rRNA) and a large non-coding control region (CR, also known as an AT-rich region) [4]. Certain

Taxon Sampling
The complete mitochondrial genomes were obtained for two species of the tribe Euophryini. In addition, 11 of the 18 available salticid mitochondrial genomes that are complete or only lack part of control region, as well as 8 mitochondrial genomes of outgroups (other spider families), were downloaded from the Genbank (see Table 1 for detailed information). Specimens of the two euophryine species, Corythalia opima and Parabathippus shelfordi, were collected from Estación de Biología Los Tuxtlas, Veracruz, Mexico (18.

DNA Extraction, Library Preparation and Sequencing
Genomic DNA was extracted from the cephalothorax and legs using QIAGEN DNeasy Blood & Tissue Kit (Hilden, Germany), and RNA was removed with QIAGEN RNase A (Hilden, Germany). The quantity of DNA was checked using a Qubit TM fluorometer (Singapore). The genomic DNA was sent to Novogene Co. Ltd. (Tianjin, China) for library preparation using the Truseq Nano DNA HT sample preparation kit (Illumina, San Diego, CA, USA), and then sequenced on the Illumina NovaSeq 6000 platform with 150 bp pairedend reads and insert size around 350 bp. About 5 Gb of raw data was obtained for each species to assemble the mitochondrial genome.

Mitochondrial Genome Assembly, Annotation and Sequence Analysis
After performing the quality control of raw reads to remove the adapters and lowquality reads (with ≤10% unidentified nucleotides, or with >50% bases having Phred quality <5, or with >10 nt aligned to the adapter, or read 1 and read 2 of the paired-end reads being completely identical), we proceeded with mitogenome assembly with the cleaned reads using MitoZ v3.4 [32,33] with the SPAdes assembler [34] and multi-kmer strategy (clade = Arthropoda; genetic_code = 5). The assembled complete mitogenomes were annotated with the online tool MITOchondrial genome annotation Server (MITOS; http://mitos.bioinf.uni-leipzig.de/index.py/ 10 February 2023) [35], and some tRNA loci were further annotated by ARWEN [36] and manual inspection. The secondary structures of tRNAs were predicted by MITOS and ARWEN, as well as manual inspection after determining the ends of adjacent protein-coding genes. The open reading frames (ORFs) of PCGs were examined in Geneious Prime ® 2023.0.1. The fully annotated circular maps of the two newly sequenced euophryine mitogenomes were then visualized in CHLOROPLO [37]. The above annotation procedures were also conducted on the two mitochondrial genomes, Dendryphantes sp. (MW832855) and Heliophanus lineiventris (MW832849), which were downloaded as "unverified" from the Genbank. In addition, the mitogenome of Carrhotus xanthogramma (Latreille, 1819) (KP402247) was re-annotated since Li et al. [17] suggested a different gene order from that of NCBI and the original publication [29]. The annotations for the other downloaded mitogenomes available in the Genbank were used in this study.
The  [39]. The 13 PCGs were aligned using MACSE v2 [40], and the aligned sequences were used to calculate the nonsynonymous substitution rates (Ka), synonymous substitution rates (Ks) and Ka/Ks of each PCG with the software KaKs_Calculator v3.0 with Model Averaging (MA) method [41]. The nucleotide diversity (Pi) of PCGs was calculated by a sliding window analysis using DnaSP v5.10 [42] with a window length of 100 bp and a step size of 25 bp. The pairwise genetic distances of each PCG were calculated by MEGA v11 [43] with the maximum composite likelihood method. The program Tandem Repeat Finder [44] (http://tandem.bu.edu/trf/trf.html 25 February 2023) was used to predict tandem repeats in the control region. The software IBM SPSS Statistics v26.0 was used to test the correlations between the effective codon (ENC) and average GC content in the third positions of PCG codons (GC3), and the average GC content in the first and second positions of PCG codons (GC12) and GC3, respectively.

Phylogenetic Analyses
The mitochondrial genome sequences of 21 spider species were included in the phylogenetic analyses, of which 13 were from the ingroup jumping spider species, and 8 were from the outgroup families (1 each of Cheiracanthiidae, Ctenidae, Lycosidae, Oxyopidae, Philodromidae, Pisauridae, Selenopidae and Trochanteriidae; Table 1). Both the nucleotide and amino acid sequences of the 13 PCGs were used for phylogenetic reconstruction.
The 13 PCG sequences were aligned, respectively, using MAFFT v7.505 [45] with the L-INS-i strategy, and the gaps and misaligned sites were trimmed in trimAl v1.2rev57 [46] with the "automated1" mode. The trimmed alignments were concatenated in PhyloSuite v1.2.3, and the ModelFinder v2.2.0 [47] was used to select the best partition and model with Bayesian information criterion (BlC). The maximum likelihood (ML) analyses were performed in IQ-TREE v2.2.0 [48] using the optimized model and partition scheme, and an ultrafast bootstrap [49] analysis with 1000 replicates was conducted to assess the node support. The Bayesian inference (BI) was completed using MrBayes v3.2.7 [50] with two independent runs (500 million generations and four chains in each run, sampling every 1000 generations). The resulting log files were imported into Tracer v1.7.2 [51] to check the convergence, and the trees sampled during the first 20% of generations were discarded as burn-in. The resulting trees were visualized via Figtree v1.4.4 (http://tree.bio.ed.ac.uk/ software/figtree/ 20 March 2023).

General Features of Mitogenomes of Euophryini
This study presents the first report on the complete characterization of mitogenomes for the jumping spider tribe Euophryini. The assembled and annotated complete mitochondrial genomes of the two euophryine species were uploaded to Genbank with the accession numbers OQ281589 and OQ429315 (  Table S1). Consistent with other jumping spiders, the mitogenomes of Euophryini are double-stranded circular, comprising 22 tRNAs, 13 PCGs, 2 rRNAs and a non-coding control region (also known as an AT-rich region), of which 9 PCGs, 13 tRNAs and the control region are in the majority strand (J-strand), and 4 PCGs, 9 tRNAs and 2 rRNAs are in the minority strand (N-strand) ( Figure 1). The base composition of the full mitogenome is A = 29%, T = 43.1%, C = 8.4%, G = 19.4% for Corythalia opima, and A = 32.9%, T = 41.3%, C = 8.6%, G = 17.2% for Parabathippus shelfordi, and both species have high A + T content (72.1% and 74.2%, respectively). All jumping spider species with complete mitogenomes have the highest A + T content in the third codon positions, except Corythalia opima, which has the highest A + T content in the control region ( Figure 2A). Similar to other jumping rRNAs are in the minority strand (N-strand) ( Figure 1). The base composition of the full mitogenome is A = 29%, T = 43.1%, C = 8.4%, G = 19.4% for Corythalia opima, and A = 32.9%, T = 41.3%, C = 8.6%, G = 17.2% for Parabathippus shelfordi, and both species have high A + T content (72.1% and 74.2%, respectively). All jumping spider species with complete mitogenomes have the highest A + T content in the third codon positions, except Corythalia opima, which has the highest A + T content in the control region ( Figure 2A). Similar to other jumping spiders, the two euophryine species show typical negative AT skews (−0.196 and −0.113, respectively) and positive GC skews (0.394 and 0.334, respectively) in the full mitogenome.
The PCGs in jumping spider mitogenomes show a typical AT preference with the A + T content of all PCGs ranging from 71.0% to 78.4% (Supplementary Table S2), and the third codon positions having much higher A + T content (78.5-94.0%) than the first and second codon positions (66.6-73.0%; Figure 2A). Among the 13 PCGs, atp8 has higher A + T content than other PCGs in all jumping spider species except Corythalia opima and Habronattus oregonensis, in which nad3 has the highest A + T content, and cox1 has the lowest A + T content ( Figure 2B; Supplementary Table S3). The nucleotide skewness analysis shows that the PCGs of all jumping spider species prefer T and G over A and C with AT skewness from −0.166 to −0.121 and GC skewness from 0.036 to 0.108 (Supplementary Table S2). Insects 2023, 14, x FOR PEER REVIEW 6 of 18
The PCGs in jumping spider mitogenomes show a typical AT preference with the A + T content of all PCGs ranging from 71.0% to 78.4% (Supplementary Table S2), and the third codon positions having much higher A + T content (78.5-94.0%) than the first and A B Figure 2. The A + T contents of (A) transfer RNAs (tRNAs), ribosomal RNAs (rRNAs), the whole, first, second, third positions of protein-coding genes (PCGs), control region and full mitogenome; and (B) the 13 PCGs in all jumping spider species (the control regions of Dendryphantes sp. and Heliophanus lineiventris are not complete, so their A + T contents are not provided).
The sliding window analysis shows that the nucleotide diversity (Pi) of the 13 PCGs in Salticidae is highly variable, with the highest Pi for atp8 (0.393) followed by nad2 (0.320) and nad5 (0.279), and the lowest Pi for cox1 (0.148) ( Figure 3A). The analysis of pairwise genetic distances among the jumping spider species shows that among the 13 PCGs, atp8 (0.671) and nad2 (0.466) have evolved relatively quickly, while cox1 (0.167) relatively slowly ( Figure 3B).
To estimate the evolutionary rate, the Ka/Ks values were calculated for the 13 PCGs of Salticidae. The results show that atp8 has the highest mean Ka/Ks value, which indicates it may have evolved more rapidly than the other PCGs in Salticidae, whereas cox1 has the lowest mean value of Ka/Ks, implying a slower rate of evolution ( Figure 4A Table S2).
The sliding window analysis shows that the nucleotide diversity (Pi) of the 13 PCGs in Salticidae is highly variable, with the highest Pi for atp8 (0.393) followed by nad2 (0.320) and nad5 (0.279), and the lowest Pi for cox1 (0.148) ( Figure 3A). The analysis of pairwise genetic distances among the jumping spider species shows that among the 13 PCGs, atp8 (0.671) and nad2 (0.466) have evolved relatively quickly, while cox1 (0.167) relatively slowly ( Figure 3B). To estimate the evolutionary rate, the Ka/Ks values were calculated for the 13 PCGs of Salticidae. The results show that atp8 has the highest mean Ka/Ks value, which indicates it may have evolved more rapidly than the other PCGs in Salticidae, whereas cox1 has the lowest mean value of Ka/Ks, implying a slower rate of evolution ( Figure 4A)

Transfer and Ribosomal RNA Genes
The mitochondrial genomes of jumping spiders contain 22 tRNA genes, 1 for each amino acid, with an additional isoform for each of serine and leucine. Of the 22 tRNA genes, 13 are on the J-strand and the remaining nine are on the N-strand (Supplementary Table S1). The length of the 22 tRNAs in the jumping spiders ranges from 1144 bp to 1342 bp, and their AT content ranges from 73.4% to 79.2%. The tRNAs in jumping spiders do not show strong A/T bias with AT-skew values close to 0, but they clearly prefer G over C with GC-skew values ranging from 0.129 to 0.271 (Supplementary Table S2). The gene order in jumping spider mitogenomes is conserved, but a rearrangement between trnL2 and trnN is found in Corythalia opima and Heliophanus lineiventris, and a rearrangement of nad1 to between trnE and trnF is found in Asemonea sichuanensis.
The predicted secondary structures of the 22 tRNAs for the two Euophryini species are shown in Supplementary Figures S2 and S3, respectively. Some tRNAs fail to fold into the typical cloverleaf-shaped secondary structures, as observed in the mitochondrial genomes of many arachnids [13,22]. The trnS1 of both Euophryini species lacks the dihydrouridine (DHU) arm, which is a common feature of most metazoans [3]. There are

Transfer and Ribosomal RNA Genes
The mitochondrial genomes of jumping spiders contain 22 tRNA genes, 1 for each amino acid, with an additional isoform for each of serine and leucine. Of the 22 tRNA genes, 13 are on the J-strand and the remaining nine are on the N-strand (Supplementary Table  S1). The length of the 22 tRNAs in the jumping spiders ranges from 1144 bp to 1342 bp, and their AT content ranges from 73.4% to 79.2%. The tRNAs in jumping spiders do not show strong A/T bias with AT-skew values close to 0, but they clearly prefer G over C with GC-skew values ranging from 0.129 to 0.271 (Supplementary Table S2). The gene order in jumping spider mitogenomes is conserved, but a rearrangement between trnL2 and trnN is found in Corythalia opima and Heliophanus lineiventris, and a rearrangement of nad1 to between trnE and trnF is found in Asemonea sichuanensis.
The predicted secondary structures of the 22 tRNAs for the two Euophryini species are shown in Supplementary Figures S2 and S3, respectively. Some tRNAs fail to fold into the typical cloverleaf-shaped secondary structures, as observed in the mitochondrial genomes of many arachnids [13,22]. The trnS1 of both Euophryini species lacks the dihydrouridine (DHU) arm, which is a common feature of most metazoans [3]. There are mismatched base pairs in the secondary structures of tRNAs, including U-U, U-C, A-C, T-G, A-A, A-G, G-G, A-C and C-C, with the mismatches occurring mainly in the amino acid (AA) arms and the TΨC arms.

Control Region
The control region (CR), also known as the AT-rich region with A + T content ranging from 74.4% to 82.7%, is located between trnQ and trnM in the jumping spider mitogenomes, and has typical negative AT skew and positive GC skew, indicating a preference for T and G. The control region of Asemonea sichuanensis (1793 bp) is significantly longer than that of other jumping spider species (657 bp to 968 bp) (Supplementary Table S2). The tandem repeats are found in the control regions of three jumping spider species with variation in the copy number and length (Figure 7): the CR of Corythalia opima has one tandem repeat of 24 bp with two partial sequences of 22 bp and 9 bp, Epeus alboguttatus has one tandem repeat of 323 bp with two partial sequences of 319 bp and 10 bp, and Telamonia vlijmi Prószyński, 1984, presents one tandem repeat of 326 bp with a partial sequence of 308 bp. No tandem repeats are observed in the other well-characterized mitogenomes of Salticidae. from 74.4% to 82.7%, is located between trnQ and trnM in the jumping spider mitogenomes, and has typical negative AT skew and positive GC skew, indicating a preference for T and G. The control region of Asemonea sichuanensis (1793 bp) is significantly longer than that of other jumping spider species (657 bp to 968 bp) (Supplementary Table S2). The tandem repeats are found in the control regions of three jumping spider species with variation in the copy number and length (Figure 7): the CR of Corythalia opima has one tandem repeat of 24 bp with two partial sequences of 22 bp and 9 bp, Epeus alboguttatus has one tandem repeat of 323 bp with two partial sequences of 319 bp and 10 bp, and Telamonia vlijmi Prószyński, 1984, presents one tandem repeat of 326 bp with a partial sequence of 308 bp. No tandem repeats are observed in the other well-characterized mitogenomes of Salticidae.

Phylogenetic Analyses
The results of phylogenetic analyses from the nucleotide and amino acid sequences of the 13 PCGs are shown in Figure 8 and Supplementary Figures S4-S6. The BI and ML analyses on the nucleotide dataset recovered the same topology, in which the monophyly of Salticidae (posterior probability, pp = 1; bootstrap, bs = 96%) and Salticinae (pp = 1; bs = 100%) are supported. Both analyses suggested the sister relationship of Salticidae with Cheiracanthiidae + Philodromidae, but this is not well supported (pp = 0.6; bs = 78%). Within Salticinae, the tribe Plexippini is rendered paraphyletic with Colopsus longipalpis.

Phylogenetic Analyses
The results of phylogenetic analyses from the nucleotide and amino acid sequences of the 13 PCGs are shown in Figure 8 and Supplementary Figures S4-S6. The BI and ML analyses on the nucleotide dataset recovered the same topology, in which the monophyly of Salticidae (posterior probability, pp = 1; bootstrap, bs = 96%) and Salticinae (pp = 1; bs = 100%) are supported. Both analyses suggested the sister relationship of Salticidae with Cheiracanthiidae + Philodromidae, but this is not well supported (pp = 0.6; bs = 78%). Within Salticinae, the tribe Plexippini is rendered paraphyletic with Colopsus longipalpis.

Discussion
The size and content of mitochondrial genomes are conserved in Salticidae, including Euophryini, as shown in the other groups of spiders [17]. The nucleotide skewness analysis is commonly used to reveal the nucleotide composition dynamics of mitogenomes [39,52,53], which shows that the mitogenomes as a whole in jumping spiders are clearly 0.05

Selen o p s _ b u r s a r i u s_NC_024878 Cheir a c a nth i u m_triviale_MN334527
Asemone a _ sichuanensis_MN651970

Discussion
The size and content of mitochondrial genomes are conserved in Salticidae, including Euophryini, as shown in the other groups of spiders [17]. The nucleotide skewness analysis is commonly used to reveal the nucleotide composition dynamics of mitogenomes [39,52,53], which shows that the mitogenomes as a whole in jumping spiders are clearly AT-biased with much higher A + T content than G + C content, and often prefer T over A and G over C with negative AT skews and positive GC skews. This has been found repeatedly in other spiders, such as Leucauge celebesiana (Walckenaer, 1841) [54], Argiope perforata Schenkel, 1963 [55] and Argiope ocula Fox, 1938 [56]. The apt8 has the highest A + T content among the mitochondrial PCGs in almost all salticid species examined except Corythalia opima and Habronattus oregonensis, but its significance still needs further investigation with extended sampling in a comparative mitogenomic study.
Genomic rearrangements, which have been considered useful markers for deep phylogenetic inference in some lineages, are relatively common in arachnids [55,56]. Two types of gene arrangements have been found in well-characterized jumping spider mitogenomes ( Figure 9). One type of rearrangement occurs between the trnL2 and trnN in Corythalia opima (Euophryini) and Heliophanus lineiventris (Chrysillini), which has also been discovered in other spider families such as Desidae (Desis jiaxiangi Lin, Li & Chen, 2020 [15]). This type of rearrangement can be classified as shuffling because the genes did not cross PCGs but moved from their original positions to the adjacent positions. As reported in previous studies on spider mitogenomic gene arrangement [57,58], the tRNA rearrangements in jumping spiders can be explained by the tandem duplication and random loss model (TDRL), which assumes that the rearrangement of mitochondrial genes occurs through the tandem duplication of specific genes, followed by the random loss of one copy of each gene [59,60]. This mechanism would change the position of the gene but not its orientation. The other type represents the first PCG rearrangement of the mitochondrial genomes in spiders, which involves rearranging nad1 to between trnE and trnF. So far, this rearrangement has only been observed in Asemonea sichuanensis (Asemoneinae), the only species of the basal lineages of jumping spiders with a well-characterized mitogenome. The basal lineages of jumping spiders comprise six subfamilies (Asemoneinae, Eupoinae, Hisponinae, Lyssomaneinae, Onomastinae and Spartaeinae) of Salticidae, all except the subfamily Salticinae that possesses the bulk species diversity of jumping spiders [19]. Resolving the relationships of these basal lineages is essential to clarify the jumping spider phylogeny. However, it is hard to determine if the rearrangement of nad1 found in Asemonea sichuanensis is ancestral in jumping spiders and has phylogenetic implications due to the lack of knowledge on the mitogenomic features of other basal salticid lineages.
tRNAs are key in translation, serving as adapter molecules between mRNA codons and amino acids [61]. All the examined jumping spider mitogenomes have 22 tRNAs. Although the loss of tRNA genes in animal mitochondrial genomes is rare, it has been reported in a variety of animal groups [62], including some arachnids such as scorpions [63], mites [64] and wolf spiders (Pirata subpiraticus (Bösenberg & Strand, 1906), lacking trnG [17]). Mismatches of base pairs in the AA arm are common in most tRNAs of spider mitogenomes, which might be a typical feature of spider mitochondrial tRNAs [17]. Previous studies have proposed that a fully paired AA arm in spiders may be formed by tRNA editing [13,22]. Lacking a DHU or TΨC arm in tRNAs is pretty common in spider mitogenomes [17], and truncated mitochondrial tRNAs may cause difficulty in the accurate annotation of certain tRNA genes [17]. However, the truncated secondary structure in certain tRNAs may not influence the translation process. For instance, a previous study has shown that in nematodes, mitochondrial tRNAs with extremely short structures, can still be recognized by synthesizing enzymes and can be aminoacylated [65]. The control region is an essential element involved in the replication and transcription initiation of the mitochondrial genome [66]. As the largest non-coding region, the length of the control region in jumping spiders varies, with that of Asemonea sichuanensis being almost twice as long as the other examined salticid species. However, this variation Epeus_alboguttatus_MH922026 A S1 R E -F Figure 9. Genomic rearrangements of Salticidae. The left phylogenetic tree is from the ML analysis of the nucleotide dataset, and the red box lines represent the rearrangements of tRNAs and PCG within jumping spiders.
The other type represents the first PCG rearrangement of the mitochondrial genomes in spiders, which involves rearranging nad1 to between trnE and trnF. So far, this rearrangement has only been observed in Asemonea sichuanensis (Asemoneinae), the only species of the basal lineages of jumping spiders with a well-characterized mitogenome. The basal lineages of jumping spiders comprise six subfamilies (Asemoneinae, Eupoinae, Hisponi-nae, Lyssomaneinae, Onomastinae and Spartaeinae) of Salticidae, all except the subfamily Salticinae that possesses the bulk species diversity of jumping spiders [19]. Resolving the relationships of these basal lineages is essential to clarify the jumping spider phylogeny. However, it is hard to determine if the rearrangement of nad1 found in Asemonea sichuanensis is ancestral in jumping spiders and has phylogenetic implications due to the lack of knowledge on the mitogenomic features of other basal salticid lineages.
tRNAs are key in translation, serving as adapter molecules between mRNA codons and amino acids [61]. All the examined jumping spider mitogenomes have 22 tRNAs. Although the loss of tRNA genes in animal mitochondrial genomes is rare, it has been reported in a variety of animal groups [62], including some arachnids such as scorpions [63], mites [64] and wolf spiders (Pirata subpiraticus (Bösenberg & Strand, 1906), lacking trnG [17]). Mismatches of base pairs in the AA arm are common in most tRNAs of spider mitogenomes, which might be a typical feature of spider mitochondrial tRNAs [17]. Previous studies have proposed that a fully paired AA arm in spiders may be formed by tRNA editing [13,22]. Lacking a DHU or TΨC arm in tRNAs is pretty common in spider mitogenomes [17], and truncated mitochondrial tRNAs may cause difficulty in the accurate annotation of certain tRNA genes [17]. However, the truncated secondary structure in certain tRNAs may not influence the translation process. For instance, a previous study has shown that in nematodes, mitochondrial tRNAs with extremely short structures, can still be recognized by synthesizing enzymes and can be aminoacylated [65].
The control region is an essential element involved in the replication and transcription initiation of the mitochondrial genome [66]. As the largest non-coding region, the length of the control region in jumping spiders varies, with that of Asemonea sichuanensis being almost twice as long as the other examined salticid species. However, this variation seems to not be completely due to the variation in the length and copy number of tandem repeats (TRs), since no TR is found in Asemonea sichuanensis. TRs have been found in other spider species, such as Argiope amoena L. Koch, 1878 [67], Argiope bruennichi [68] and Trichonephila clavata (L. Koch, 1878) [69]. However, among the 11 examined jumping spider mitogenomes with complete CR regions, TRs are present in only three species (Corythalia opima, Epeus alboguttatus and Telamonia vlijmi), indicating that the presence of TRs may not be a conserved feature, at least within Salticidae. The highly repetitive sequences (often A-T-rich) in the control region may inhibit DNA polymerases and lead to sequencing failures [70,71], which may account for the incomplete control regions in some registered salticid mitogenomes, e.g., Heliophanus lineiventris (MW832849) and Dendryphantes sp. (MW832855).
The PCGs of the jumping spider species share the same AT and GC skewness pattern as the full mitogenome. The mean Ka/Ks value for the PCGs is less than 1 in all the examined jumping spider species, indicating that the mitogenomes of the jumping spiders are generally subject to purifying selection, as documented in other organisms such as spider mites [72] and insects [73]. Among all the PCGs, the atp8 and nad2 exhibit relatively high nucleotide diversity (Pi), pairwise genetic distances and Ka/Ks values, with Ka/Ks values for atp8 sometimes around or higher than 1 (Figure 4). This suggests these two genes may have evolved under a relatively relaxed purifying selection that may be related to an adaptation to new environments [74]. For instance, positive selection on genes such as nad4, cytb and atp8 was believed to have acted to meet the enormous changes in energy demand and have played a pivotal role in the evolution of bat flight [75]. The cox1 gene has the lowest Ka/Ks value and less amino acid variation, so it has been widely applied as a barcode marker in spiders and other animals [15]. Codon usage patterns are often influenced by selection and mutational pressures [76]. ENC and neutral maps are commonly used to detect the proportion of evolutionary pressure [77,78]. The ENC-GC3 plot ( Figure 5) showed a distribution of the salticid PCGs below the standard curve, which indicates the codon usage bias in jumping spider mitogenomes is affected by natural selection [79]. The neutrality plot analysis showed a positive correlation between GC12 and GC3, indicating that mutational pressure also affects codon usage in the evolution of jumping spider mitogenomes. The slope of the regression line (0.249) suggested that natural selection (75.1%) plays a more important role in the evolution of amino acid composition in jumping spiders than mutational pressure (24.9%) [76].
Although the phylogenetic analyses in this study have few exemplar taxa due to the limited number of well-characterized mitogenomes, especially for jumping spiders, they show the potential of mitochondrial sequences for resolving the phylogenetic relationships of spiders and providing insight into the placement of certain problematic taxa. Both analyses on the nucleotide and amino acid datasets recovered the same relationships among the Salticidae and the outgroup families, which are largely congruent with recent transcriptome-or UCE (ultra-conserved element)-based phylogenomic studies [80][81][82][83]. Within the Dionycha clade, Salticidae was suggested to be a sister to the clade of Cheiracanthiidae + Philodromidae (Figure 8) through mitogenomic PCGs, but the phylogenomic study on Dionycha combining genomic, morphological and Sanger data indicated a sister relationship between Salticidae and Philodromidae [83]. This discrepancy is likely due to the limited taxon sampling in the mitogenomic dataset. Recent progress in molecular phylogenetic studies of jumping spiders, especially the application of the phylogenomic approach, has strongly promoted an understanding of the relationships of major salticid lineages [19,84]. However, only two (Asemoneinae and Salticinae) of the seven salticid subfamilies were included in the mitogenomic phylogenetic analyses of this study, in which the monophyly of the subfamily Salticinae was strongly supported. Thus, the performance of mitogenome sequences in reconstructing the phylogeny of salticids needs to be further tested with a more extended sampling of jumping spider mitogenomes.
An interesting finding through the phylogenetic analyses with the mitogenomic PCGs is the placement of Colopsus longipalpis, which was not grouped with the other members of the tribe Plexippini in the resulting phylogeny ( Figure 8). Colopsus longipalpis was originally classified in the genus CheliceroidesŻabka, 1985, and was designated as the type species [85]. Additionally, Cheliceroides was considered a member of the salticid tribe Hasariini [19]. However, Logunov synonymized Cheliceroides with Colopsus based on morphological characteristics, which resulted in the current placement of Colopsus longipalpis within the tribe Plexippini [86]. The mitogenomic phylogeny challenged Logunov's taxonomic treatment and suggested that Colopsus longipalpis might not be a member of Plexippini, but further evidence is needed to clarify its taxonomic status and phylogenetic position.
Supplementary Materials: The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/insects14060517/s1, Figure S1: The (A) relative synonymous codon usage (RSCU) and (B) number of amino acids in the 13 protein coding genes of salticid species. Codon families are labeled below the X-axis; Figure S2: Predicted secondary structures of tRNAs in the mitochondrial genome of Corythalia opima; Figure S3: Predicted secondary structures of tRNAs in the mitochondrial genome of Parabathippus shelfordi; Figure S4: Phylogenetic tree from the ML analysis of the amino acid dataset, with the numbers on the branches indicating bootstrap supports; Figure S5: Phylogenetic tree from the BI analysis of the amino acid dataset, with the numbers on the branches indicating posterior probabilities; Figure S6: Phylogenetic tree from the BI analysis of the nucleotide dataset, with the numbers on the branches indicating posterior probabilities; Table S1: Annotation of the mitochondrial genomes of the jumping spider species; Table S2: Nucleotide composition of mitogenomes of Salticidae; Table S3: The A + T contents of the 13 PCGs of Salticidae; Table S4: The GC12 content, GC3 content and ENC of PCGs of Salticidae.   Table 1.