Complete mitochondrial genome sequence of Labriocimbex sinicus, a new genus and new species of Cimbicidae (Hymenoptera) from China

Description of new species

Specimens were examined with a Leica S8APO dissection microscope. Adult images were taken with a Nikon D700 digital camera and a series of images edited using Helicon Focus (HeliconSoft), while detailed images were taken with Leica Z16 APO/DFC550. A cylinder of semitransparent plastic was placed around the specimen to disperse the light, that methods follows Vilhelmsen (2019). The specimen must be sufficiently relaxed in a moist chamber before dissection. Dissected ovipositor valves, gonoforcep and penis valves were permanently mounted on slides in gum Arabic and images produced and composited automatically with a Nikon Ci-L/DS-Fi3. We used Adobe Photoshop CS 6.0 for further image processing. The terminology of sawfly genitalia follows Ross (1945), and that of general morphology follows Viitasaari (2002). For a few terms (e.g., middle fovea and lateral fovea), we followed Takeuchi (1952). Abbreviations used were: OOL = distance between the eye and outer edge of lateral ocelli; POL = distance between the mesal edges of the lateral ocelli; OCL = distance between a lateral ocellus and the occipital carina or hind margin of the head.

The holotype and some paratypes of the new species are deposited in the Asian Sawfly Collection, Nanchang, China (ASCN). The most remaining paratypes are deposited in the Insect Collection of Central South University of Forestry and Technology, Changsha, Hunan, China (CSCS). A few paratypes are kept in Lishui Academy of Forestry (LSAF).

The electronic version of this article in Portable Document Format (PDF) will represent a published work according to the International Commission on Zoological Nomenclature (ICZN), and hence the new names contained in the electronic version are effectively published under that Code from the electronic edition alone. This published work 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: EE7F5193-78B2-42CE-87C1-B3FE947CB70F. The online version of this work is archived and available from the following digital repositories: PeerJ, PubMed Central and CLOCKSS.

DNA library construction and sequencing

Total DNA was extracted from L. sinicus using an E.Z.N.A.^® Tissue DNA Kit (Omega, Norcross, GA) and was stored at −20 °C, in accordance with the manufacturer’s instructions. Sequencing libraries with approximately 250-bp insertions were constructed using a NEXT flex™ Rapid DNA-Seq Kit (Illumina, San Diego, CA, USA) in accordance with the manufacturer’s protocol. Each library was sequenced using an Illumina Hiseq 4,000 to generate 150-bp paired end reads at BGI-Shenzhen, China. The sequencing reads have been deposited in NCBI SRA database under accession number: PRJNA507477.

Mitochondrial genome assembly

Next generation sequencing and bioinformatic analyses were performed by Shanghai Majorbio Bio-pharm Technology Co., Ltd. Reconstruction of the mitochondrial genome from Illumina reads was carried out using three different approaches to ensure the accuracy of the assemblies: SOAPdenovo v2.0 (Luo et al., 2012), MITObim v1.8 (Hahn, Bachmann & Chevreux, 2013) and NOVOPlasty v2.7.1 (Dierckxsens, Mardulyn & Smits, 2017). The assembled mitochondrial fragments were identified using BlastX and T. anthracinum (NC029733) mitochondrial genes as queries. Prediction and annotation of protein-coding, tRNA and rRNA genes were performed using DOGMA (http://dogma.ccbb.utexas.edu/) or MITOS (http://mitos.bioinf.uni-leipzig.de/index.py) with annotation from a reference mitochondrial genome. Queries were then corrected manually.

Mitochondrial genome annotation and secondary structure prediction

All RNA genes were identified by employing the online MITOS tool (http://mitos.bioinf.uni-leipzig.de/index.py) (Bernt et al., 2013) with the invertebrate mitochondrial genetic code. The initiation and termination codons of PCGs were determined using Geneious v11.0.3 (http://www.geneious.com) with reference sequences from other symphytan species with subsequent manual adjustment. The A + T content of nucleotide sequences and relative synonymous codon usage (RSCU) were calculated using MEGA v7.0 (Kumar, Stecher & Tamura, 2016). Strand asymmetry was calculated using the formulae (Perna & Kocher, 1995): GC–skew = (G − C)/(G + C) and AT–skew = (A − T)/(A + T), for the strand encoding the majority of the PCGs.

The secondary structures of the rrnS and rrnL were partitioned into four areas and six areas, respectively. The secondary structures of rRNAs were inferred using alignment to models predicted for T. anthracinum. First, the primary sequence and the secondary structure of this species were aligned in MARNA (Siebert & Backofen, 2005) to identify a consensus sequence as well as a consensus structure in the output files. Secondly, the secondary structures of the rrnS and rrnL in L. sinicus were predicted by specific structure models in SSU-ALIGN (Nawrocki, 2009). Finally, the structures were artificially transformed to their relative secondary structure with micro changes.

The predicted secondary structures of RNAs were drawn using VARNA v3-93 (Darty, Denise & Ponty, 2009) and RNAviz 2.0.3 (De Rijk, Wuyts & De Wachter, 2003). Helix numbering was performed following the Apis mellifera rRNA secondary structure (Gillespie et al., 2006) including minor modifications.

Taxon sampling

We sampled all known mitochondrial genomes representatives from the symphytan of Hymenoptera (34 species of Symphyta and two representatives of Apocrita) and four Non-hymenopteran outgroups (Mecoptera, Diptera, Megaloptera, Coleoptera) including some mitochondrial genomes downloaded from GenBank which had previously been sequenced, and the newly sequenced mitochondrial genome in this study (Table 1).

To investigate the phylogenetic relationship of Labriocimbex within Cimbicidae, we used 40 species (43 samples) of seven genera belonging to three subfamilies according the classification system of Abe & Smith (1991), by using partial cytochrome oxidase subunit I (COI) gene of mitochondrial genome. Composite of ingroup and outgroup taxon, as indicated in Table 2, were developed by sequences from different taxon either from our own sequences or those deposited in GenBank.

DNA extraction, polymerase chain reaction (PCR) and sequencing of COI gene

Total genomic DNA was extracted from muscles or single leg of adult specimens stored in ethanol at −20 °C using the DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany). PCRs (50 µl) contained 25µl Taq MasterMix (CW0682M) reaction buffer, 2 µl of each primer, 2 µl DNA template and 19 µl PCR grade H2O. The PCR programme consisted of an initial denaturing step at 95 °C for one min, followed by 42 cycles of 20 s denaturing at 95 °C, 30 s annealing at 55 °C and 80 s extension at 68 °C; the last cycle was followed by a final 7 min extension step at 68 °C. The primers of COI referenced Nyman et al. (2006). Purified PCR products were sequenced in both directions with the BigDye v3.1 Mix Sequencing Kit (Applied Biosystems) and an ABI 3730 XL automated sequencer (Applied Biosystems). Sequence trace files were read and edited using Sequencher v4.8 (BGI-Shenzhen cop, China).

COI gene sequences were checked and assembled in Geneious v11.0.3 (http://www.geneious.com), prior to submission to GenBank (accession numbers MN076590 –MN076605).

Phylogenetic analysis

We used the Maximum Likelihood (ML) and Bayesian Inference (BI) methods to construct phylogenetic trees. A total of 13 PCGs were aligned by MUSCLE in MEGA v7.0 individually, two rRNAs and COI gene were aligned by MAFFT (https://www.ebi.ac.uk/Tools/msa/mafft/) (Katoh & Standley, 2013). Then, these (13 PCGs and two rRNAs) aligned nucleotide sequences were concatenated using SequenceMatrix v1.7.8 (Vaidya, Lohman & Meier, 2011) and partitioned into several data blocks.

The partitioned data block file was used to infer both partition schemes and substitution models in PartitionFinder v1.1.1 (Lanfear et al., 2012), with “unlinked” branch lengths under the “greedy” search algorithm. The standard partitioning schemes “bic” and “aicc” were selected for BI and ML analyses, respectively. BI analyses were conducted with the GTR+I+G model and HKY+G model using MrBayes v3.2.2 (Ronquist et al., 2012). Four simultaneous Markov chains (three cold, one heated) were run for five million generations in two independent runs, with sampling every 1,000 generations and the first 25% of generations were discarded as burn-in.

The best partitioning scheme were chosen for 13 PCGs (Table 3). The best-fit model of nucleotide substitution and phylogenetic construction based on ML were created using the IQ-TREE web server (http://iqtree.cibiv.univie.ac.at/). The previous data block file was used as well as the original parameters. In addition, 0.1 was employed as the disturbance intensity and 1,000 as the IQ-TREE stopping rule.

The COI gene phylogenetic construction also based on ML and BI analyses. All related files have been uploaded to Figshare (https://doi.org/10.6084/m9.figshare.7339334.v1).

Description

Type species: Labriocimbex sinicus Yan & Wei, sp. nov.

Description. Body middle to large-sized; black, without metallic luster (Fig. 1); head and thorax with dense and long yellowish brown hairs; clypeus distinctly broader than distance between lower margin of eyes, anterior margin with arcuate notch, furrow between clypeus and supraclypeal area deep (Fig. 2A); base of labrum much broader than apex, lateral margin of labrum distinctly narrowed upward (Fig. 2A); mandibles elongate, with three teeth in total, basal one truncate at apex (Figs. 3A, 3B); maxillary palp with 6 palpomeres, apex 1–2 combined distinctly shorter than palpomere 4; labial palp with 4 palpomeres, short (Fig. 2G); malar space 2.3 times the diameter of lateral ocellus, about as long as scape and pedicel combined; eyes moderately large, inner margins parallel, distance between eyes slightly longer than longest axis of eye (Figs. 2A, 2B); lateral part of head distinctly dilated behind eyes in lateral view (Fig. 2B) and dorsal view (Fig. 2D); postocellar area with median and lateral furrows distinct, frontal carina indistinct (Fig. 2D). Antenna longer than breadth of head, club of antenna strongly enlarged with obscure annular suture, with 5 antennomeres before club, antennomere 3 slender and distinctly longer than antennomeres 4 and 5 combined (Fig. 2H). Propleuron and sternum merged; median suture of mesonotum shallow, notaulus distinct; mesoscutellum flat, anterior margin subtruncate, posterior margin roundly triangular (Fig. 2E); distance between inner margin of cenchri 3.3 times the longest axis of a cenchrus, distance between outer margin of cenchri longer than breadth of mesoscutellum (Fig. 2E). Coxae and femur of leg with long hairs; ventral side of middle and hind femur without tooth near apex (Figs. 2F, 3C), hind coxae close to each other; inner spur of hind tibia as long as apical breadth of tibia, apex blunt and membranous (Fig. 3J), about 0.4 times length of metabasitarsus; metabasitarsus slightly shorter than tarsomeres 2 and 3 combined, base of hind tibia narrower than apex (Fig. 2F); 1st and 2nd tarsal pulvilli long, nearly contiguous (Fig. 2F); claw simple, roundly bent (Fig. 3K).Fore wing with crossvein 2r present, base of vein Rs absent (Fig. 1A); vein 2r-m and 2m-cu almost interstitial, pterostigma long and narrow; anal cell strongly narrowed in basal 1/3 with a short anal crossvein, apical anal cell about 2 times the length of basal anal cell; cell Rs and M closed in hind wing, apex of anal cell quadrate, petiole of anal cell longer than length of vein cu-a, jugum region only with 1 longitudinal vein, without crossvein (Fig. 1A). Sternites and basal abdominal terga with long hairs, posterior margin of abdominal tergum 1 shallowly incised, without middle carina and lateral carina (Figs. 2E, 2F). Genital plate of female developed with wide incision and arcuate in middle (Fig. 3L); apical ovipositor sheath short and roundish in lateral view (Fig. 3D), tapering toward apes in dorsal view (Fig. 3F); apex of lancet and lance curved upwards (Figs. 3M, 3N), each annulus with 1 pore, serrulae sub-truncate at apex, lateral teeth sharp (Fig. 3G). Each sternite of male incised in middle, both sides roundish (Fig. 3E); penis valve broad, with apical lobe bulge, ventral hook small, lateral ridge distinct (Fig. 3H); harpe small, longer than broad (Fig. 3I).

Etymology. The generic name is composed of “labrio-” and “-cimbex”, emphasizing the shape of labrum differs from other genera of the family. Gender masculine.

Distribution. China.

Host plant: Prunus pseudocerasus of Rosaceae (Female adult were observed laying eggs on it).

Remarks.

The name Labriocimbex was mentioned in two papers before (Li & Wu, 2010; Vilhelmsen, 2019). The name is a nomina nuda in these papers as it was not accompanied by a proper description and type designation. The name was originally proposed by the senior author of this paper (MW) for the genus hear described as new to science. The present paper constitutes the proper establishment of the name.

This new genus is similar to Praia Andre and Trichiosoma Leach. It is similar to Praia having the head and thorax with dense and long hairs; the antenna with 5 antennomeres before the club; the coxae and femur of leg with long hairs; the ventral side of hind femur without a denticle near apex; the base of hind tibia narrower than apex; the 1st and 2nd tarsal pulvilli long. It differs from Praia by having the triangular labrum, the base of labrum much wider than its apex and the basal breadth about half the breadth of clypeus; the inner margins of eyes parallel; the anal cell strongly narrowed in basal 1/3 with a short crossvein and apical anal cell about 2 times the length of basal anal cell; the abdomen all black, without transverse band. It differs from Trichiosoma by the ventral side of the hind femur without a subapical denticle; the 1st and 2nd tarsal pulvilli in the male very long and nearly contacting to each other, and the different pattern of the male genitalia.

Trichiosomini includes 3 genera: Trichiosoma, Pseudoclavellaria and Leptocimbex (Abe & Smith, 1991; Deng, 2000). Most of the characteristics of the new genus suggest placing it in the tribe Trichiosomini. The most important characteristics include: the labrum large with the basal breadth about half the breadth of clypeus, the jugum region in hind wing without crossvein, the clypeus very short and much broader than lower distance between eyes and not merging with supraclypeal area. See the key below for the relationship of the new genus and other genera of the family.

Key to extant Holarctic genera of Cimbicinae

Female. (Holotype) Body length 21 mm (Fig. 1A). Black; apical 1/2 of mandible reddish brown (Figs. 3A, 3B); inner and ventral side of club of antennae largely brown, outer side dark brown (Fig. 2H); cenchri pale yellowish white; posterior half of mesepimeron, metapleuron largely, metanotum except for a small macula behind cenchri and most of metapostnotum, median triangular macula and narrow posterior margin of abdominal tergum 1 (Fig. 2E) yellowish white; apex of each tibia, tarsus and claw reddish brown, tarsal pulvillus grayish white (Fig. 2F). Wings brownish hyaline, stigma black, basal 3/5 of vein C in fore wing, basal 3/7 of vein Sc+R and entire vein M+Cu pale yellow, vein A pale brown, other veins largely black, vein J and basal parts of all other veins in hind wing pale yellow (Fig. 1A). Hairs on face and gena black at base and yellowish white at apex (Fig. 2A); hairs on vertex of head and mesonotum black (Fig. 2D); hairs on pronotum and scutellum yellowish white largely except for black basal 0.2; hairs on mesopleuron, coxae and femora yellowish brown largely with less than basal 0.3 black (Figs. 2C, 2E, 2F); inner hairs of fore tibia reddish yellow; abdominal terga 1–2 and posterior margins of terga 3–4 with yellowish white hairs; hairs on ventral side of terga and sternites 1–3 black in basal 0.4 and reddish yellow in apical 0.6.

Body densely microsculptured, matt; lower margin of orbit, small fovea lateral to lateral ocellus, apical half of mandible, narrow lateral side of mesoscutal lateral lobe, ventral part of trochanters and of femora distinctly shiny, ventral half of mesepisternum feebly microsculptured mixed with some minute punctures, shiny; venter of abdomen feebly shiny.

Apex of labrum thickened with middle notch (Fig. 2A); median fovea round and deep, lateral foveae obscure (Fig. 2D); middle of frons concave, lateral furrow of frons shallow; postocellar furrow distinct, interocellar furrow long and deep; postocellar area quadrate, middle furrow very shallow, indistinct; lateral furrows shallow, weakly divergent backwards (Fig. 2D). Long hairs on gena clearly shorter than 1/3 head width in dorsal view. Club of antennae as long as length of antennomeres 4 and 5 combined, with obscure annular suture (Fig. 2H). Mesopleuron without middle oblique ridge (Fig. 2C); cenchrus 2.1 times broader than long, reniform (Fig. 2E). Coxae and femora with dense hairs longer than breadth of femora (Fig. 2F); inner hairs of tibia dense and short (Fig. 2F). Vein 2r in fore wing joining cell 2Rs at basal 0.4; cu-a joining cell 1M close to vein 1M (Fig. 1A). Abdominal terga 1–2 and posterior margin of terga 3–4 with dense and long hairs, other terga with sparse and short hairs. Sternites and ventral side of abdominal terga with spare and long hairs (Fig. 2F). Ovipositor sheath 0.8 times as long as metatarsomere 1 and 2 combined, apical margin roundish in lateral view (Fig. 3D), acute at apex in dorsal view (Fig. 3F). Lancet with 45 serrulae (Fig. 3N), middle serrulae as Fig. 3G, annular spine bands narrow, membranous area between serrulae roundly protruding, middle serrulae subtruncate at apex, with 5–6 proximal and 4–5 distal subbasal teeth (Fig. 3G).

Male: Body length 21.5 mm (Fig. 1B); body color and structure similar to female except for following parts: labrum broad and large; anterior margin of clypeus arc-shape, without incision; metathorax and abdominal tergum 1 entire black; subgenital plate slightly broader than long (Fig. 3E), apical margin round; apex of each sternite with clear middle incision, both sides roundly arcuate. Penis valve shown in Fig. 3H, gonoforcep shown in Fig. 3I.

Holotype. Female (CSCS13010_Lab001). China: Hunan Province, Wugang County, Mt. Yun, Yunfengge alt. 1,380 m, 26°38.630′N, 110°37.299′E, April 13, 2013, Zejian Li leg.

Paratypes: 17 Females, 15 Males (CSCS13010_Lab002–033). Collecting information as the holotype. 18 Females, 10 Males (CSCS13015_Lab034–061), locality and collector as the holotype, April 15, 2013. 45 Females, 17 Males (CSCS13014_Lab062–123), locality and collecting time as the holotype, Liwei Qi, Biao Chu leg. 36 Females, 51 Males (CSCS11009_Lab124–210), China: Hunan Province, Wugang County, Mt. Yun, Shengli Temple, alt. 1,145 m, 26°38.859′N, 110°37.026′E, April 18–22, 2011, Zejian Li, Liwei Qi leg. 17 Females, 22 Males (CSCS05001_Lab211–249), China: Hunan Province, Wugang County, Mt. Yun, alt. 800–1,100 m, April 24–26, 2005, Meicai Wei, Shaobing Zhang, Wei Xiao leg. One Male (CSCS1999001_Lab250), China: Hunan Province, Wugang County, Mt. Yun, alt. 1,300 m, April 3, 1999, Wei Xiao leg. Two Females, six Males (LSAF18029_Lab251–258), China: Zhejiang Province, Lin’an City, Mt. Tianmu, alt. 1,506, 30.349° N, 119.424° E, April 19, 2018, Zejian Li, Mengmeng Liu leg. One Females (LSAF17053_Lab259), locality and collector as the former, April 16, 2017. One Females, 26 males (LSAF17054_Lab259–285), locality as the former, April 17, 2017, Tingting Ji leg. Four Females, two Males (CSCS18006_Lab286–291), China: Hunan Province, Wugang County, Mt. Yun, alt. 1,124 m, 26°38.059′N, 110°37.017′E, April 03, 2018, Meicai Wei, Gengyun Niu, Hannan Wang leg. Seven Females, one Males (CSCS18007_Lab292–299), locality as the former, alt. 1,129 m, 26°39.003′N, 110°37.027′E, April 04, 2018, Meicai Wei, Hannan Wang leg.

Variation. Body length 18–24 mm in female, 19–24 mm in male; club of antennae color brown to pale yellowish brown; hairs color on pronotum and scutellum yellowish white to yellowish brown.

Distribution. China (Hunan, Zhejiang).

Etymology: The specific name of the new species refers to the distribution area, China.

Remarks.

Labriocimbex pilosus sp. nov. (Li & Wu, 2010) and Labriocimbex sinicus (Vilhelmsen, 2019) were two nomina nuda and have never been properly described before this paper. The two manuscript names were originally proposed by the senior author of this paper (MW) for the two undescribed species found in China. The former species represents only by a few specimens from different localities and so it is not described here, and specimen collection record: one female, China, Gansu Province, Mt. Xiaolong, Maiji forest farm, Sun hill; alt. 1,620, 34°25′11.0″N, 105°46′30.1″E, April 17, 2009, Wu XingYu leg.

Labriocimbex sinicus Yan & Wei, sp. nov. is similar to L. zaraeoides (Malaise, 1939) comb. nov. (Fig. 4A), but differs from the latter in the following characters: the clypeal notch deep, depth about 1/2 length of clypeus; between the clypeus and supraclypeal area with a distinct transverse furrow; the long hairs on gena 3.5 times as long as diameter of lateral ocellus, longer than the shortest axis of an eye; the long hairs on mesopleuron about 4.5 times as long as diameter of lateral ocellus; the abdominal tergum 1 largely black.

Distribution. Northern Myanmar.

Remarks. This species is similar to L. sinicus Yan & Wei sp. nov., the majority of the characters place it in the new genus Labriocimbex. The most important of these characters are: the broadly emarginated clypeus, the triangular labrum (Fig. 4C), the form of the antennae (antennae with 5 antennomeres before the rigid club; joints of the club very indistinct [sic!] Malaise, 1939); the slender hind legs (the coxae and femur of leg with long hairs; the ventral side of hind femur without a denticle near apex; the base of hind tibia narrower than apex; the 1st and 2nd tarsal pulvilli long, Fig. 4D) the venation (the anal crossvein punctiform and the anal cell strongly narrowed, Fig. 4A); the color of body (the posterior half of mesepimeron, metapleuron, metanotum and abdominal tergum 1 yellowish white, Figs. 4B, 4D and 4G). The characters that separate this species from all known Trichiosoma are the yellowish white color of the metanotum and the base of the abdomen, and the ventral side of the hind femur without a large denticle near apex. L. zaraeoides differs from L. sinicus in the following characters: the clypeal notch shallow, depth about 1/4 length of clypeus; the transverse furrow between clypeus and supraclypeal area absent (Fig. 4C); the long hairs on gena 2.5 times the diameter of lateral ocellus, shorter than the shortest axis of an eye (Fig. 4E); the long hairs on mesopleuron about 3.5 times as long as diameter of the lateral ocellus (Fig. 4B); the abdominal tergum1 largely yellow brown (Fig. 4G).

General features of the L. sinicus mitochondrial genome

We sequenced the complete mitochondrial genome of L. sinicus (GenBank accession no. MH136623), a typical set of 37 genes, including 13 PCGs, 22 tRNAs and two rRNAs. Most of the genes were located on the J- strand except for four PCGs, two rRNAs and seven tRNAs (Table 4).

A total of 14 pairs of genes were directly adjacent, without overlapping or intergenic nucleotides. The total length of the intergenic regions was 268 bp in 18 locations with a size ranging from 1 to 50 bp (Table 4). The longest was located between trnH and ND4, while the second longest was 45 bp located between rrnS and trnM. In comparison with the mitochondrial genome of T. anthracinum and C. lateralis, there were differences in the length of intergenic spacers and locations. The longest (414 bp) was located at the start of the mitochondrial genome before trnY in T. anthracinum. The longest length of the intergenic spacers was 345 bp located between the ND6 and CYTB genes in C. lateralis. We found that homologous searches on the longest intergenic region of L. sinicus revealed substantial differences from any identified Symphyta sequence.

There were in total 32 overlapping nucleotides between neighboring genes in six locations, and the range of length of the overlapping sequence is from 3 to 14 bp: trnM and trnQ, ATP8 and ATP6, ND4 and ND4L, trnN and trnS2, and ATP6 and COIII; and the longest was 14 bp between ATP6 and COIII (Table 4). The common motifs such as: ATGATAA between ATP8 and ATP6, and ATGTTAA between ND4 and ND4L, which also exist in T. anthracinum, and are not found in C. lateralis, are common features of many other insect mitochondrial genomes (Song et al., 2016; Doğan & Korkmaz, 2017).

Protein-coding genes and codon usage

The mitochondrial genome of L. sinicus contains 13 PCGs, and its length is 12,456 bp, accounting for 80.86% of the total length (Table 5). All PCGs were initiated by ATN codons. All PCGs were ended with TAA as stop codon except for ND5 which ended with T (Table 4).

The codon usage of L. sinicus also shows a significant bias towards A/T Leu, Ile, Phe and Ser, were found as the most frequently used amino acids. TTA-Leu showed the highest RSCU of 5.04 (Table 6). Comparisons of the RSCU with those of C. lateralis and T. anthracinum showed a similar pattern for codon usage bias and reflected a significant correlation between codon preference and nucleotide composition, that is similar to other symphytan species (Dowton et al., 2009a; Wei et al., 2010; Wei, Wu & Liu, 2015; Korkmaz et al., 2015; Korkmaz et al., 2016; Korkmaz et al., 2017; Song et al., 2015; Song et al., 2016; Niu et al., 2019a; Niu et al., 2019b; Du et al., 2018; Ma et al., 2019; Tang et al., 2019). Codons rich in C and G, CGC-Arg and CTC-Leu was absent, CGG-Arg, GGC-Gly, AGC-Ser, ACG-Thr, CTG-Leu, GTC-Val, GTG-Val and TGC-Cys, were used once, AGG-Ser, TCG-Ser, TCC-Ser, CCG-Pro and GCG-Ala were rarely used, which is similar to both cimbicid mitochondrial genomes (Table 6). The ratio can be calculated by rate of G + C rich codons (Pro, Ala, Arg, and Gly) and A + T rich codons (Phe, Ile, Met, Tyr, Asn, and Lys), and it is 0.28 in L. sinicus, which is similar to those of other symphytan species (0.28–0.31) (Korkmaz et al., 2015). The translation, initiation, and termination signals as well as the codon usage of the L. sinicus mitochondrial genome do not display any unusual characteristics (Table 6).

Gene rearrangement and nucleotide composition

The mitochondrial genome of symphytan species appears to be more conserved than that of Apocrita (Song et al., 2016; Wei, Niu & Du, 2014). However, compared with the putative ancestral mitochondrial genome of insects, we detected several rearrangement events in three tRNA gene clusters in L. sinicus (Fig. 5), The first rearrangement event is found in the clusters of trnI-trnQ-trnM, where trnM and trnQ was founding swapped positions, in addition, trnM-trnQ was translocated from the trnI-trnQ-trnM cluster to a downstream position of rrnS; which have not been reported for any symphytan mitogenome to date. The second event is corresponding to the remote inversion of trnY and the translocation of trnC from a location between trnW and COX1 to upstream of trnI, which has great similarity to the gene order and rearrangement events observed in T. anthracinum. The arrangement of cluster of trnW-trnC-trnY appears to be mostly conserved in almost all known symphytan mitogenomes, except for representative cimbicid species. The last event is only found in the TP cluster of L. sinicus, and here trnT is inverted. The gene order from COI to rrnS is conserved in all sequenced species of Cimbicidae.

Similar to previously reported symphytan mitochondrial genomes (Ma et al., 2019; Doğan & Korkmaz, 2017; Song et al., 2016), the nucleotide compositions of L. sinicus (43.5% A, 37.7% T, 7.7% G and 11.1% C) were biased towards A and T, with an average 81.2% A+T content; a stronger AT bias was found in the N strand (81.4% A+T content) than in the J- strand (78.7%) (Table 5).

Further analysis of the PCGs indicated that the third codon position demonstrates the highest A + T content (93.5%), in agreement with symphytan mitochondrial genomes (Ma et al., 2019; Doğan & Korkmaz, 2017; Song et al., 2016). The gene with the highest A + T content was ATP8 with 88.3% (Table 5). Here we observed that the AT-skew was slightly positive (0.0714), and the GC-skew was negative (−0.1809) when considering the whole genome (Table 5). This indicates that the occurrence of A is higher than that of T, and the occurrence of C is higher than that of G, which is a general phenomenon observed in all reported symphytan mitochondrial genomes, except for those of Tremex columba and Xiphydria sp. (Ma et al., 2019; Doğan & Korkmaz, 2017; Song et al., 2016; Wei, Wu & Liu, 2015; Castro & Dowton, 2005; Dowton et al., 2009a). However, a deviation was found in the PCGs of L. sinicus, in terms of AT-skew (−0.1389) and GC-skew (0.0348), which also occurred in both C. lateralis and T. anthracinum. This deviation can exert influences on the selection forces acting on the PCG codon positions, in accordance with study by Korkmaz et al. (2015).

Transfer RNA genes

In the mitochondrial genome of L. sinicus 15 tRNAs were encoded by the J- strand, while the remaining tRNAs were encoded by the opposite N-strand. All tRNAs folded into a common clover-leaf structure, except trns1-AGN, where the dihydrouridine (DHU) arm was missing (Fig. 6). The size of the tRNAs ranged from 64 bp (trnG) to 71 bp (trnC, trnK), and this usually depends on the length of the variable loop, T ΨC loop and D-loops (Clary & Wolstenholme, 1985). The DHU arm was 3–4 bp, the AC arm was 4–5 bp, and the T ΨC arm varied from 4–5 bp, while the amino acid acceptor (AA) stem and anticodon (AC) loops were conserved at 7 bp in all of the tRNA genes.

In the mitochondrial tRNA secondary structures, mismatches mainly occur in the DHU arm, AA arm and AC arm, and sometimes in the T ΨC arm. A total of 16 unmatched base pairs were scattered among the following tRNA genes, including 12 G-U mismatched pairs occurring in trnA, trnD, trnQ, trnG, trnH, trnL1, trnP, trnF, and trnY, and four U-U mismatches occurring in trnR, trnT and trnL1. The number of mismatches were 24 (12 G–U pairs, five U–U pairs, three A–A pairs, two A–C pairs, one A–G pair and one C–U pair) in C. lateralis (Fig. 7), and 18 (15 G–U pairs, two U–U pairs and 1 A–C pair) in T. anthracinum (Fig. 6), which is typical for Hymenoptera (Ma et al., 2019; Castro & Dowton, 2005; Dowton et al., 2009b). The phenomenon of aberrant mismatches, loops, or extremely short arms for tRNAs has been shown to be common in metazoan mitochondrial genomes (Wolstenholme, 1992).

In addition, there were some tRNA structural differences between L. sinicus and T. anthracinum (Fig. 6). The identified anticodons were almost identical to those of the cimbicid species, with the exception of the anticodon of trnS1 (AGn), which is UCU in L. sinicus and T. anthracinum, as well as this is true of all previously reported of Symphyta (Ma et al., 2019; Doğan & Korkmaz, 2017; Song et al., 2016; Castro & Dowton, 2005; Dowton et al., 2009b).

Ribosomal RNA genes

The rrnL gene of L. sinicus was 1,341 bp in length with an 84.2% A+T content, while rrnS was 791 bp in length with an 84.1% A+T content (Table 5). This was in a comparable range to homologous genes in T. anthracinum (1,351 bp; 800 bp) and C. lateralis (1,359 bp; 493 bp rrnS partial gene), and also identical to all reported hymenopteran species (Gillespie et al., 2006; Wei et al., 2010; Doğan & Korkmaz, 2017; Song et al., 2016; Korkmaz et al., 2015). Both genes were encoded on the N-strand (Table 4).

Similar to the known symphytan mitochondrial genomes, the rrnL gene is positioned between trnL1 and trnV in three species of Cimbicidae (Fig. 5). The predicted structure of rrnL in L. sinicus is consistent with the observed pattern in C. lateralis and T. anthracinum, whereby 45 helices belonging to five domains were identified in those species (Figs. 8, 9). Domain III is absent as in other arthropods (Korkmaz et al., 2015), and domain II is variable in base composition, forming a long stem with a big loop structure in the area II terminal. Domains IV and V are more conserved within the Tenthredinidae than domains I, II and VI. Eight helices (H563, H579, H777, H822, H2023, H2043, H2455 and H2547) of rrnL are highly conserved. The H183, H991, H1057, H1196 and H2077 helices display helical length and loop size/structure variability within three cimbicid rrnL genes (Figs. 8, 9).

The rrnS secondary structure of L. sinicus is between trnV and an AT-rich region, and contains four domains and 26 helices. Compared with T. anthracinum, it is significantly different in terms of base composition in domain II (Fig. 10). Specifically, H47 is variable among the different hymenopteran species, having a large loop. The loop size is variable and determined by overall rrnS length, except for in the cephid species (Gillespie et al., 2006; Wei et al., 2010; Doğan & Korkmaz, 2017; Song et al., 2016; Korkmaz et al., 2015). The structures of domains I and II of C. lateralis are missing, so they cannot be compared with those of L. sinicus, but the structures are similar in domains III and IV (Fig. 11). In rrnS, domain III and domain VI were more conserved within Tenthredinidae than domains I and II (Figs. 10 and 11).

Phylogenetic relationships

Phylogenetic relationships within the “Symphyta” were reconstructed using both BI and ML analyses. The topologies of the two phylogenetic trees were almost identical, thus we combined the two phylogenetic trees (Fig. 12). The clade consisting of (Tenthredinidae + Cimbicidae) + (Argidae + Pergidae), was very stable with the highest nodal supports. The recovered trees supported a relationship consisting of Xyelidae + (Tenthredinoidea + (Pamphiliidae + ((Megalodontesidae + Xiphydriidae) + (Cephidae + (Orussidae + (Siricidae + Apocrita)))))) in Hymenoptera. It seems unusual that Siricidae is the sister group to Apocrita, although, the sister relation of Tremex columba (MH422968) and Tremex columba (AY206795) according to the phylogenetic inference based on IV and V areas of rrnL (YY Zhang, 2019, unpublished data), suggests the former is valid. It showed that more mitochondrial genomes should be involved to evaluate the correctness of these unusual sister groups.

To investigate the phylogenetic relationships of Labriocimbex within Cimbicidae, we analyzed 43 cimbicid and two outgroups sequences, an approximately 850-bp piece of the COI gene were obtained in this study, combined with the other species sequences from GenBank aligned by MAFFT. The phylogenetic trees were reconstructed using both BI and ML analyses (Fig. 13). We found that the generic relationships of Cimbicidae revealed by the phylogenetic analyses based on COI genes agree quite closely with the systematic arrangement of the genera based on the morphological characters.

The subfamily classification scheme suggested by Abe & Smith (1991) is confirmed by the results of the present analyses. The monotypic Corynidinae are always retrieved with strong support, and Abiinae + Cimbicinae is best-supported internal node in the present analyses of Cimbicidae, which was also supported by Vilhelmsen (2019).

The clade consisting of ((((Labriocimbex + Praia) + T. anthracinum) + Trichiosoma) + Leptocimbex) + Cimbex, was highly supported in both trees (posterior probability >0.9511 and bootstrap support >87). Labriocimbex is always retrieved as monophyletic (two samples were collected in Hunan, one was collected in Zhejiang), and it is the sister group of Praia. Praia + Labriocimbex is the sister group of T. anthracinum. Malaise (1939) placed T. anthracinum and other two species, T. hymalayensis and T. sikkimensis, in the subgenus Asitrichiosoma within Trichiosoma. According to our analyses (Fig. 13), T. anthracinum renders the genus Trichiosoma paraphyletic, and it is necessary to reestablish the genus Asitrichiosoma.

Additionally, we demonstrated that COI gene sequences can be used to solve phylogenetic relationships between genera of Cimbicidae. However, to reconstruct the generic phylogeny of Cimbicidae the further studies of the mitochondrial genomes and external morphology of more taxa of the family are needed.