The complete mitogenome of S. inferens was found to be a 15,413 bp long circular molecule, and its sequence was deposited into GenBank (Accession number: JN039362) (Figure 1 and Table 1). The size was shorter than that of H. cunea (15,481 bp), L. dispar (15,569 bp), and O. lunifer (15,593 bp) [4,5,7]. All of the 37 typical animal mitochondrial genes (13 PCGs, 22 tRNA and 2 rRNA genes) were present, and the gene order was the same as in other sequenced noctuid moths, in an order of trnM, trnI, and trnQ. The position of the trnM gene was usually translocated to the 5′ upstream position of trnI in the lepidopteran species, which was different from those of other order species [14]. This indicated that the mitochondrial gene arrangement in lepidopteran species evolved independently after splitting from its stem lineage [15]. In addition, a total of 257 bp of intergenic spacer sequences were present in 18 locations with the exception of the A+T-rich region, and a total of 25 bp of overlapping nucleotides was scattered over six locations in the S. inferens mitogenome.

The nucleotide composition of the S. inferens mitogenome was also biased toward AT nucleotides (80.3%), which was slightly higher than those of L. dispar (79.9%) and O. lunifer (77.8%), but was slightly lower than those of H. armigera (81.0%) and H. cunea (80.4%) [4–7] (Table 2). The AT skew in the S. inferens mitogenome was slightly negative (−0.001), indicating the occurrence of more Ts than As. This value was different from those of other Noctuidae species, such as H. armigera (0.002), H. cunea (0.010), L. dispar (0.016), O. lunifer (0.030) [4–7]. Meanwhile, the GC skew in the S. inferens mitogenome was negative (−0.228), demonstrating the occurrence of more Cs than Gs. This was also true for H. armigera (−0.184), H. cunea (−0.230), L. dispar (−0.247), O. lunifer (−0.318) [4–7].

The mitochondrial sequence of S. inferens, as in most insects, contained 13 PCGs (Table 2). Most of the PCGs utilized typical ATN start codons (ATG, ATT, and ATC), among which six genes started with codon ATG (atp6, cox3, nad4, nad4L, cob, and nad1), five with ATT (nad2, cox2, atp8, nad3, and nad5), and the nad6 gene with ATC. However, the cox1 gene in the S. inferens mitogenome had the start codon CGA as observed in most other lepidopteran species sequenced to date [16]. Notably, the start codon of the cox1 gene is usually found to use nonstandard putative codons, as previously reported in arthropod species [17–20].

The conventional termination codon TAA, likely resulting from post-transcriptional polyadenylation [21], was observed in 10 PCGs. However, the cox1, cox2, and nad4 genes utilized truncated termination codons, which are commonly observed in lepidopteran species.

The average AT content of the 13 PCGs in the S. inferens mitogenome was 78.6%, with a negative AT skew of −0.148 and a slightly positive GC skew of 0.037 (Table 2). The AT and GC skews were similar to those of H. armigera (−0.139, 0.029), H. cunea (−0.146, 0.023), and L. dispar (−0.148, 0.013), indicating that the contents of T and G were higher than those of A and C, respectively [5–7]. It is to be noted that both the AT skew (−0.141) and the GC skew (−0.004) were slightly negative in the 13 PCGs of O. lunifer [4]. Also, the AT contents at different codon positions was analyzed. In the S. inferens mitogenome, the AT contents was 87.7%, 72.9% and 75.2% at the first, second and third codon positions, respectively, with the highest AT contents at the first codon position, which was similar to the result from L. dispar mitogenome (90.2%, 72.9%, 70.1%), but was different from that in H. armigera (70.2%, 94.1%, 73.8%), H. cunea (73.2%, 70.3%, 92.0%) and O. lunifer mitogenomes (72.0%, 70.0%, 85.1%) [4–7] (Table 2).

Excluding the start and termination codons, the 13 PCGs in the S. inferens mitogenome consisted of 3702 codons in total, consistent with the observations in four other noctuid moths, which ranged from 3687 in the H. armigera to 3718 codons in the O. lunifer mitogenome [4–7] (Figure 2). The codon families exhibited a very similar behavior among the five species (Figures 2 and 3). There were at least six codon families with at least 50 codons (CDs) (Asn, Ile, Leu2, Met, Phe, Tyr), and two families with at least 100 CDs per thousand CDs (Leu2 and Phe) as observed in five insects. In addition, the AT-rich CDs favor synonymous CDs with a lower AT content, as revealed by the Relative Synonymous Codon Usage (RSCU) results, and exemplified by the Leu2 family, in which the TTA codon accounted for the large majority of CDs (Figure 3). The codon families with high CDs had A and T predominantly in the third position, which might reflect selection for optimal tRNA usage, genome bias, and the speed and efficacy of genome/DNA repair [22].

The mitogenome of S. inferens had the characteristic 22 tRNAs sets interspersed by rRNAs or PCGs, with an AT content of 81.5% and a total of 1478 bp in size ranging from 64 to 73 bp/tRNA. All tRNA genes had the typical cloverleaf secondary structures except for the trnS1(AGN) gene, in which a stable stem-loop structure of the dihydrouridine (DHU) arm was missing (Figure 4), as observed in many metazoan mitogenomes [23–25]. Eight tRNAs were encoded by the L-strand and the remaining 14 by the H-strand, which was identical in all the species. All tRNA genes usually contained a 7-bp amino acid acceptor (AA) stem, where most nucleotide substitutions were compensatory. The anticodon (AC) stem and the loop (7 bp) were both conserved in all tRNAs, whereas two U–U pairs were usually located at the second and third couplets in the anticodon stem of trnS2(UCN). The length of DHU was 3–4 bp, except for trnS1(AGN). The TΨC arm was usually 3–6 bp in length. Except for trnD, trnE, and trnY, a total of 29 unmatched base pairs were detected in the S. inferens tRNAs, but 18 of them were G-U pairs, which form a weak bond in the tRNAs, and are well-known non-canonical pairs in tRNA secondary structures [26]. The remaining 11 were mismatched base pairs including seven U-U, one C-U, and three A-C pairs. Notably, the numbers of mismatches varied in different noctuid moths, with 4, 7, 23 and 10 mismatches in H. armigera, H. cunea, L. dispar, and O. lunifer, respectively [4–7].

As in other mitochondrial sequences from Insecta species, there were two rRNAs in S. inferens with a total length of 2169 bp and an AT content of 84.1%. The large ribosomal gene (rrnL) had a length of 1385 bp, located between trnL1(CUN) and trnV, whereas the small one (rrnS) had a length of 784 bp between trnV and the A+T-rich region. The gene sizes and locations were typical as in other noctuid moths’ mitogenomes. Both tRNA and rRNA genes of Noctuidae species predominantly contained A and T [4–7] (Table 2). Furthermore, both the AT and GC skews were slightly positive in the tRNA and rRNA genes, consistent with the results in H. armigera and H. cunea mitogenomes [5,6].

Both the rrnL and rrnS secondary structures were predicted according to models proposed for these genes in other insects [27,28]. Although the length of individual helices varied in insect species, the secondary structures of both rrnL and rrnS in S. inferens were quite similar to their counterparts in Apis mellifera, Manduca sexta, and other insect species [27,28]. In S. inferens, the rrnL gene contained five domains (labeled I, II, IV, V, and VI) with 49 helices, as observed in other noctuid moths (Figure 5). Each domain was separated by a single-stranded region, and domain III was absent in arthropod mitogenomes. Also, S. inferens had the rrnS with 33 helices in three domains (labeled I, II, III) (Figure 6).

In the S. inferens mitogenome, there was a total of 257 bp of intergenic spacer sequences, which resided at 18 locations (except for the A+T-rich region) and varied in size from 1 to 68 bp. It is to be noted that there were four major intergenic spacers (s1–s4) spanning at least 18 bp, all of which were rich in As and Ts (Figure 7). The s1 spacer (68 bp) located between trnQ and nad2 had a higher AT content compared with the counterpart in the O. lunifer mitogenome. And so did the other three spacers. The s3 spacer (44 bp), inserted between nad6 and cob, is not conserved in five noctuid moths, which was mostly a 4 bp repeat in the S. inferens mitogenome while the other had a stretch of TA dinucleotides in the L. dispar mitogenome. However, the s2 spacer (44 bp), placed between nad4 and nad4L, was longer than its counterparts in H. armigera and L. dispar mitogenomes, but it was twice the length of the counterpart in the L. dispar mitogenome (44 vs. 22) and about the same size as the counterpart in the H. armigera mitogenome (44 vs. 43). In addition, the s4 spacer (18 bp), located between trnS2(UCN) and nad1, did not contain the motif “ATACTAA” that was conserved across four other noctuid moths and other lepidopteran species mitogenomes [28,29].

Also, a total of 25 bp overlapping nucleotides was scattered over six locations and ranged in sizes from 1 to 11 bp in the S. inferens mitogenome, with the longest one (11 bp) located between atp8 and atp6. Interestingly, an 8-bp motif “AAGCCTTA” conserved in the four other noctuid moths’ mitogenomes [4–7] was also detected between trnW and trnC in the present study.

In mitochondrial genomes, the A+T-rich region had been reported to possess essential elements involved in the initiation of replication and transcription of mitogenome [30]. The A+T-rich region of the S. inferens mitogenome was 311 bp in length with an AT content of 95.8%. Similarly, the sizes and AT contents in the A+T-rich regions of four other noctuid moths were 328 bp and 95.1% AT content (H. armigera), 357 bp and 95.0% AT content (H. cunea), 369 bp and 96.1% AT content (L. dispar), and 319 bp and 93.4% AT content (O. lunifer) [4–7]. Notably, the S. inferens A+T-rich region displayed a higher AT content (95.8%) than all the other locations in the mitogenome (Table 2).

There was a conserved structure that combined the motif “ATAGA” and a 19-bp poly-T stretch in S. inferens A+T-rich region (Figure 8). A very similar pattern occurred in four other noctuid moths, and it was widely conserved in lepidopteran mitogenomes and might be the origin of light-strand replication [4–7,30]. There was also a stem-and-loop structure in S. inferens A+T-rich region containing a 5′ flanking “TATA” and 3′ flanking “G(A)_nT” sequence, which were considered to be the site of initiation of secondary strand synthesis [31,32] (Figures 8 and 9). The stem-and loop structure might be the characteristic of insect mitogenomes, and it had been found in several insect orders including Orthoptera, Lepidoptera, Diptera, Plecoptera, and Hymenoptera, but it was not observed in four other noctuid moths [30,33–36]. Also, the conserved 3′ flanking “G(A)_nT” sequence might not be present in all insects or the sequence might take on different forms [35,37].

The presence of varying copy numbers of tandemly repeated elements was one of the characteristics in insect mitogenomes [30]. The S. inferens A+T-rich region contained a duplicated 17-bp repeat element with minor variations, and a decuplicated highly divergent 8-bp segment “ATATTAAT” (Figure 8), which resembled the octuplicated 8-bp repeat element “TATATATT” in the L. dispar A+T-rich region [7]. Additionally, there was a microsatellite “(AT)₇” in S. inferens A+T-rich region (Figure 8), and similar patterns were observed in H. cunea “(AT)₈” and O. lunifer “(AT)₇” mitogenomes [4,5].

A conserved 9-bp poly-A element was observed upstream of trnM in H. cunea, L. dispar, and O. lunifer mitogenomes, which might be involved in controlling transcription and replication initiation [4,5,7,38]. However, this structure was not detected in S. inferens and H. armigera mitogenomes [6].

In this study, the amino acid sequences of the 13 PCGs were concatenated, rather than analyzed separately, to construct the phylogenetic relationships, which may result in a more complete analysis [39]. Based on morphological and genomic analyses, the relationship among six superfamilies in lepidopteran species has been reported in other articles [29,40,41].

In the present study, the NJ and MP trees showed similar topologies except for some node confidence values (Figure 10). The S. inferens mitogenome together with mitogenomes of H. armigera, H. cunea, L. dispar, and O. lunifer clustered with other superfamilies, which was consistent with the morphological classification [40]. Geometroidea (Phthonandria atrilineata) was closely related to Noctuoidea and Bombycoidea (Bombyx mandarina, Bombyx mori, Eriogyna pyretorum, M. sexta) [28,42–44], but Papilionoidea (Calinaga davidis and Coreana raphaelis) [15,45] was the sister to Pyraloidea (C. suppressalis and Ostrinia furnacalis) [29,46], Tortricoidea (Adoxophyes honmai and Grapholita molesta) [16,38] and other superfamilies, in disagreement with the tree topology [29,40,41]. The results indicated that more mitogenomes could elaborate on phylogenetic relationships among lepidopteran superfamilies in greater detail to some extent.

Adult S. inferens were collected from rice paddies surrounding Yangzhou (32°40.025N, 119°44.017E), Jiangsu Province, China. Samples were preserved in 100% ethanol and stored at −70 °C until DNA extraction was performed. Whole genomic DNA was extracted from a single sample using the protocol of DNAVzol (Bioteke, Beijing, China) and then used for the PCR amplification.

Two pairs of LA-PCR primers were used for amplification of two overlapping fragments of S. inferens mitogenome, which were 16SAA (5′-ATGCTWCCTTTGCACRGTCAAGATACYGCGGC-3′), 16SBB (5′-CTTATCGAYAAAAAAGWTTGCGACCTCRATGTTG-3′), LP01 (5′-TGATTAGCTCCACAAATTTCTGAACATTGACC-3′), and LP02 (5′-WACACCAGTTCATATTDAACCAGAATGATATT-3′) [48]. Sub-PCR primers were designed from comparisons of 35 lepidopteran sequences available in GenBank and referenced to the universal primers of insect mitogenomes [49].

The conditions for amplification of two long fragments were as follows: An initial denaturation for 2 min at 96 °C followed by 35 cycles of 10 s at 98 °C, 10 min at 58 °C, and a subsequent 10 min final extension at 72 °C. PCR conditions for amplification of other fragments were as follows: An initial denaturation for 5 min at 95 °C, followed by 35 cycles of denaturation for 1 min at 94 °C, annealing for 1 min at 45–50 °C, elongation for 1–3 min (depending on putative length of the fragments) at 68 °C, and a final extension step of 72 °C for 10 min. For most fragments, LA Taq polymerase (TaKaRa, Kyoto, Japan) was used for the PCR amplification, but for fragments less than 1.3 kb, LA Taq polymerase was replaced by Taq polymerase (TaKaRa, Kyoto, Japan) in the PCR reaction. All PCR reactions were performed in an ABI thermal cycler (PE Applied Biosystems, CA, USA).

PCR products were separated by electrophoresis in a 1.0% agarose gel and purified using a DNA Gel Extraction Kit (Bioteke, Beijing, China). Purified PCR products were ligated into the T-vector (TaKaRa, Kyoto, Japan) and then transformed into XL-1 blue competent bacteria, according to the method of Wei et al. [50]. The positive recombinant clone was sequenced using upstream and downstream primers from bi-directions on ABI 3730XL Automated DNA Sequencer (PE Applied Biosystems, CA, USA) at least three times.

The Staden package was used for sequence assembly and annotation [51]. The PCGs and rRNA genes in S. inferens mitogenome were identified by sequence alignment with other lepidopteran species, especially the Noctuidae by using Clustal X version 2.0 [4–7,52]. The PCG nucleotide sequences without start and termination codons were translated on the basis of the Invertebrate Mitochondrial Genetic Code. The AT content was calculated using MEGA version 5.0 [35]. Composition skew analysis was carried out with formulas AT skew = [A − T]/[A + T] and GC skew = [G − C]/[G + C] [53].

Transfer RNA genes were identified using tRNAscan-SE software available online [54], and the software of XRNA 1.2.0b was used to draw the secondary structure of RNA genes. The secondary structures of the rrnL and rrnS genes were inferred based on models developed for other insect species [27,55]. To infer the rRNA gene secondary structures, we used a commonly accepted comparative approach to correct for unusual pairings with RNA-editing mechanisms that are well known in arthropod mitogenomes [56,57]. The entire A+T-rich region was subjected to a search for tandem repeats using the Tandem Repeats Finder program [58].

To illustrate the phylogenetic relationship of Lepidoptera, the other sixteen complete lepidopteran mitogenomes were downloaded from Genbank, including A. honmai (DQ073916), B. mori (AY048187), B. mandarina (FJ384796), C. davidis (HQ658143), C. raphaelis (DQ102703), C. suppressalis (JF339041), E. pyretorum (FJ685653), G. molesta (HQ392511), H. armigera (GU188273), H. cunea (GU592049), L. dispar (FJ617240), M. sexta (EU286785), O. furnacalis (AF467260), O. lunifer (AM946601), P. atrilineata (EU569764), Drosophila melanogaster (U37541) and Locusta migratoria manilensis (GU344101) were used as outgroup.

The alignment of the concatenated amino acid sequences of each 13 mitochondrial PCGs was aligned with Clustal X [52] using default settings. Then, based on the concatenated amino acid data set from the 13 PCGs, phylogenetic analyses were constructed using MEGA version 5.0 software with Neighbor Joining (NJ) and Maximum Parsimony (MP) methods [35]. Bootstrap analysis was done with 1000 replications, and bootstrap values were calculated using the 50% majority rule.