The adults of Meloe proscarabaeus were collected from a hill slope in Dongsheng, Inner Mongolia, China (39°45.33'N, 110°01.83'E). The fresh samples were immediately preserved in 100% ethanol and stored in a −20 °C refrigerator. We identified the specimens according to the morphological characters that descried by Pan and Bologna (2021). Total genomic DNA was extracted from a frozen adult using Tianamp Genomic DNA kit following the manufacturer’s protocol. The quality of DNA was determined using 1% agarose gel electrophoresis.

The library was constructed using an Illumina TruSeq Library Preparation kit with an insert size of 250 bp and sequenced using the paired-end strategy on an Illumina HiSeq 2500 platform. A total of 4.4 Gb raw data was yielded with an average read length of 150 bp. The raw data were trimmed and filtered using fastp with parameters of phred quality ≥ 30 and unqualified percent < 20 to remove adapters and low-quality reads (Chen et al. 2018). Then the clean data were used to assemble the mitochondrial genome of M. proscarabaeus by MITObim v. 1.9.1 with the default settings (Hahn et al. 2013). The mitochondrial genome of Lytta caraganae (GenBank accession number NC_033339.1; Du et al. 2017) was employed as a reference sequence. The assembled sequence was aligned with other meloid mitochondrial genomes described by Du et al. (2017) to ensure the assembling quality.

The complete mitochondrial genome of M. proscarabaeus was automatically annotated by the software MitoZ (Meng et al. 2019) and manually compared with other meloid mitochondrial genomes. Of these, PCGs were checked by the identification of open reading frames and aligning with mitochondrial PCGs of other meloids. The annotated mitochondrial genome was analysed its genome characteristics, including nucleotide composition, the composition of skewness, codon usages, and relative synonymous codon usage (RSCU), using MEGA6 (Tamura et al. 2013). All available mitochondrial genomes of 15 meloid species were used to calculate the genetic distances within and between genera in Meloidae. P-distances were calculated using MEGA6 with the bootstrap method of 1,000 replications for the variance estimation.

To infer the phylogenetic relationships among the family Meloidae, all available mitochondrial genomes of 15 meloid species, including the M. proscarabaeus, were employed to reconstruct the maximum-likelihood (ML) trees with the Tribolium castaneum (GenBank accession number NC_003081.2; Friedrich and Muqim 2003) as the outgroup. According to GenBank annotations, the nucleotide and amino acid sequences of 13 mitochondrial PCGs from each species were extracted and stop codons removed. Orderly combined sequences were aligned using MAFFT with the default settings (Katoh and Standley 2013) and the gaps and ambiguous sites removed to concatenate into consensus sequences including a 10,934 bp nucleotide and a 3,633 amino acid dataset, respectively. IQ-Tree was employed to estimate the best-fit models of partitioning schemes (Nguyen et al. 2015). The nucleotide and the corresponding amino acid datasets were partitioned for 13 genes individually. The best-fitting models of partitioning schemes were selected with the greedy search algorithm, under the Bayesian information criterion. The ML trees of both datasets were also reconstructed using IQ-Tree with 1000 bootstraps to assess the node support.

The 5.4 Gb raw data was yielded by the next-generation sequencing with 36,048,728 reads, and the 4.9 Gb (95.06%) clean data was obtained after filtering. The complete mitochondrial genome of M. proscarabaeus was assembled using 1,160,002 reads, and the average depth of coverage was assessed at 7,635.95 X. The mitochondrial genome was annotated and then submitted to the GenBank under the accession number OL840851. The complete mitochondrial genome of M. proscarabaeus is 15,653 bp in length, which is of moderate length among mitochondrial genomes within Meloidae (Du et al. 2016, 2017; Zhou et al. 2021).

The complete mitochondrial genome of M. proscarabaeus is a circular DNA molecule consisting of 13 protein-coding genes, 22 transport RNAs, two ribosomal RNAs, and a control region (Fig. 1). The length of rrnL and rrnS were determined to be 1,329 bp and 816 bp, respectively, and the control region was 1,013 bp in length. The nucleotide base composition of the mitochondrial genome was 37.5% A, 31.6% T, 19.1% C, and 11.9% G. The total A + T content was 69.1%, and the AT skew was 0.0854.

Figure 1. 

The circular structure of the mitochondrial genome of Meloe proscarabaeus. The inner circle represents GC content (the ratio of guanine to cytosine), the outer circle represents genetic characteristics, orange represents ribosomal RNA, red represents transport RNA, and green represents protein-coding regions. Genes in the inner circle (on the J chain) are transcribed clockwise, while those outside the circle (on the N chain) are transcribed counter-clockwise.

The total length of the 13 mitochondrial protein-coding genes in M. proscarabaeus is 11,127 bp, accounting for 71.10% of the total length of the genome, encoding 3,711 codons in total. All 13 protein-coding genes start using regular initiation codons, including ATT and ATG, and ATA (Table 1), which were commonly used as start codons in insect mitochondrial genomes. Most protein-coding genes terminated with conventional stop codons (such as TAA or TAG), except cox3, nad5, and nad4 stopped with T (Table 1).

The A + T bias is also manifested in the codon usage of protein-coding genes (Fig. 2). Relatively synonymous codon usages, excluding stop codons, showed that the third position of synonymous codons always has more frequency with A or T than G or C. Additionally, the first three frequently used codons UUA (Leu2), UCU (Ser2), GUU (Val), and some other frequently used codons, including AUU (Ile), UUU (Phe) AAU (Asn), etc. are comprised of two or three A and/or T nucleotides.

Figure 2. 

Relative synonymous codon usage (RSCU) in the Meloe proscarabaeus mitochondrial genome. An average of 3,711 codons were analysed, excluding stop codons. Codon families are provided on the x-axis. Leu, Leu2, Ser, and Ser2 indicate trnL1 (CUN), trnL2 (UUR), trnS1 (AGN), and trnS2 (UCN), respectively.

All 22 transfer RNAs were annotated in the mitochondrial genome of M. proscarabaeus. Their length ranged from the shortest trnS with 59 bp to the longest trnK with 71 bp (Table 1). The maldistribution of transfer RNAs was also found in the mitochondrial genome. For example, there are two clusters comprising six transfer RNAs (trnA-trnR-trnN-trnS-trnE-trnF) between nad3 and nad5 genes and three transfer RNAs (trnW-trnC-trnY) between nad2 and cox1. The remaining RNAs are dispersed among other genes in a single or double way (Fig. 1). Two ribosomal RNAs mitochondrial genome of M. proscarabaeus were aligned with these of other meloid and assigned to the blanks between neighbouring genes. The rrnL gene was located between trnL and trnV with a length of 1,328 bp, the rrnS was located between tnrV and the control region with a length of 815 bp (Table 1). The control region was located between rrnS and trnI with 1,015 bp in length (Table 1; Fig. 1).

The genetic distances within and between genera were calculated using the nucleotide and amino acid data of 13 mitochondrial PCGs among 15 meloid taxa. The result from the nucleotide data showed that the p-distances within genera for Hycleus, Epicauta, and Mylabris are 0.167, 0.173, and 0.232, respectively, with an average of 0.191, while the p-distances between genera ranged from 0.234 to 0.281 with an average of 0.258 (Fig. 3). The result from the amino acid data showed that the p-distances within genus of these three genera are 0.124, 0.107, and 0.163, respectively, with the mean of 0.115, and the p-distance between genera ranging from 0.173 to 0.226 with the mean of 0.187 (Fig. 3). The p-distances within genus are significantly lower than between genera from both datasets (p < 0.01). The p-distance between M. proscarabaeus and M. poggii was 0.116 and 0.064 from the nucleotide and the amino acid data, and far less than the corresponding distances between genera.

Figure 3. 

Genetic distance within and between genera. Each boxplot represents the p-distance based on the nucleotide and the amino acid datasets from 13 mitochondrial PCGs. The lower horizontal bar represents the smallest observation, the lower edge of the rectangle represents the 25 percentile, the central bar within the rectangle represents the median, the upper edge of the rectangle represents 75 percentile, and the upper horizontal bar represents the largest observation.

The phylogenetic relationships within the family Meloidae were inferred by using maximum likelihood methods, based on the nucleotide and amino acid data from mitogenomes of 15 meloid taxa. The best-fit partitioning schemes and corresponding substitution models are shown in Table 2. Log likelihoods of consensus trees constructed from 1000 bootstrap trees are −88,585.1106 and −32,785.8864 for the nucleotide and amino acid data, respectively.

The ML trees resulting from both datasets strongly support the monophyly of Meloe and Epicauta, whereas the genera Hycleus and Mylabris are not monophyletic (Fig. 4). The monophyletic Meloe, including M. proscarabaeus and M. poggii, sisters with Lytta into a branch in both phylogenetic trees, but the branch clusters with Epicauta or Hycleus in the ML tree from nucleotide or amino acid data, respectively (Fig. 4).

Figure 4. 

Phylogenetic trees of 14 meloid species based on the nucleotide (a) and amino acid (b) dataset inferred from the maximum likelihood. The numbers abutting branches refer to the bootstrap supports. Tribolium castaneum (Tenebrionidae) was used to root the trees as an outgroup.