The Complete Mitochondrial Genome of the Damsel Bug Alloeorhynchus bakeri (Hemiptera: Nabidae)

Abstract

The complete sequence of the mitochondrial DNA (mtDNA) of the damsel bug, Alloeorhynchus bakeri, has been completed and annotated in this study. It represents the first sequenced mitochondrial genome of heteropteran family Nabidae. The circular genome is 15, 851 bp in length with an A+T content of 73.5%, contains the typical 37 genes that are arranged in the same order as that of the putative ancestor of hexapods. Nucleotide composition and codon usage are similar to other known heteropteran mitochondrial genomes. All protein-coding genes (PCGs) use standard initiation codons (methionine and isoleucine), except COI, which started with TTG. Canonical TAA and TAG termination codons are found in eight protein-coding genes, the remaining five (COI, COII, COIII, ND5, ND1) have incomplete termination codons (T or TA). PCGs of two strands present opposite CG skew which is also reflected by the nucleotide composition and codon usage. All tRNAs have the typical clover-leaf structure, except the dihydrouridine (DHU) arm of tRNA^{Ser (AGN)}which forms a simple loop as known in many other metazoa. Secondary structure models of the ribosomal RNA genes of A. bakeri are presented, similar to those proposed for other insect orders. There are six domains and 45 helices and three domains and 27 helices in the secondary structures of rrnL and rrnS, respectively. The major non-coding region (also called control region) between the small ribosomal subunit and the tRNA^Ile gene includes two special regions. The first region includes four 133 bp tandem repeat units plus a partial copy of the repeat (28 bp of the beginning), and the second region at the end of control region contains 4 potential stem-loop structures. Finally, PCGs sequences were used to perform a phylogenetic study. Both maximum likelihood and Bayesian inference analyses highly support Nabidae as the sister group to Anthocoridae and Miridae.

Introduction

Mitochondrial (mt) genome sequence and structure is widely used to provide information on comparative and evolutionary genomics, on molecular evolution and patterns of gene flow, on phylogenetics and population genetics ^{1, 2}. Several analyses have demonstrated recently that complete mt genomes provide higher levels of support than those based on individual or partial mt genes ^3-5. Mt genome of insect is typically a double-stranded, circular molecule of 14-20 kb in length, which usually encodes 13 protein-coding genes (PCGs), two ribosomal RNA (rRNA) genes, and 22 transfer RNA (tRNA) genes ^{6, 7}. Additionally, insect mt genome contains a major non-coding region known as the A+T-rich region or the control region (CR) that plays a role in initiation of transcription and replication ⁶. The CRs of different insect taxa have turned out to be very divergent, showing differences in primary sequence, organization, as well as in their location relative to flanking genes, raising the question of whether CRs are homologous across different taxa ⁷. Moreover, the length of this region is also highly variable due to its high rates of nucleotide substitution, insertions/deletions, and the presence of varying copy numbers of tandem repeats ^{8, 9}.

The reconstruction of the phylogeny of insects has been a focus of studies for more than a century ^{10, 11}. The growing interest in phylogenetic reconstruction of the mt genome has triggered a rapid increase in the number of published complete mt genome sequence ¹². To date, the complete or nearly complete mt genomes of 32 species of true bugs are available at NCBI (status April 25, 2011).

Nabidae is a relatively small family of Heteroptera with 20 genera and approximately 500 species ¹³. The members of this family are important natural enemies of pests and are distributed throughout the world. Nabidae is proposed to be one of the most primitive families in the infraorder Cimicomorpha and hence it is of major importance for the classification and phylogeny of this infraorder ¹⁴. No complete mt genome has been sequenced from members of this family prior to this study. Here, we present the complete mt genome of Alloeorhynchus bakeri, a representative of Prostemmatinae, and provide analyses of the nucleotide composition, codon usage, compositional biases, RNA secondary structure, and evaluate the phylogenetic position of Nabidae in Heteroptera based on the sequences of PCGs.

Materials and Methods

Samples and DNA extraction

Adult specimens of A. bakeri were collected from Mengla (21°43.474N, 101°32.635E), Yunnan Province, China in April 2007. All specimens were preserved in 95% ethanol in the field. After being transported to the laboratory, they were stored at -20℃ until DNA extraction. Total genomic DNA was extracted from thorax muscle tissue using a CTAB-based method ¹⁵. Voucher specimens (Nos. VHem-00101), preserved in alcohol, are deposited at the Entomological Museum of China Agricultural University (Beijing).

PCR amplification, cloning and sequencing

The genome was amplified in overlapping PCR fragments (Supplementary Material: Table S1). Initially, 13 fragments were amplified using the universal primers from previous work ¹⁶ (Fig. 1). Seven perfectly matching primers were designed on the basis of these short fragments for secondary PCRs.

Fig 1.

Map of the mt genome of A. bakeri. The tRNAs are denoted by the color blocks and are labeled according to the IUPACIUB single-letter amino acid codes. Gene name without underline indicates the direction of transcription from left to right, and with underline indicates right to left. PCGs are denoted by the grey blocks indicate the direction of transcription from right to left, and the sky-blue indicate the direction of transcription from left to right. Overlapping lines within the circle denote PCR fragments used for cloning and sequencing.

Short PCRs were conducted using Qiagen Taq DNA polymerase (Qiagen, Beijing, China) with the following cycling conditions: 5 min at 94℃, followed by 35 cycles of 50 s at 94℃, 50 s at 48-55℃, and 1-2 min at 72℃. The final elongation step was continued for 10 min at 72℃. Long PCRs were performed using NEB Long Taq DNA polymerase (New England Biolabs) under the following cycling conditions: 30 s at 95℃, followed by 45 cycles of 10 s at 95℃, 50 s at 48-55℃, and 3-6 min at 65℃. The final elongation was continued for 10 min at 65℃. These PCR products were analyzed by 1.0% agarose gel electrophoresis.

The fragments were ligated into pGEM-T Easy Vector (Promega) and the resultant plasmid DNA was isolated using the TIANprp Midi Plasmid Kit (Qiagen). All fragments were sequenced in both directions using the BigDye Terminator Sequencing Kit (Applied Bio Systems) and the ABI 3730XL Genetic Analyzer (PE Applied Biosystems, San Francisco, CA, USA) with two vector-specific primers and internal primers for primer walking.

Sequence analysis and inferences of secondary structures

Raw sequence files were proof-read and aligned into contigs in BioEdit version 7.0.5.3 ¹⁷. Protein-coding regions and ribosomal RNA genes were identified by sequence comparison with published insect mt sequences.

The tRNAs were identified by tRNAscan-SE Search Server v.1.21 ¹⁸ with default setting. Some tRNA genes that could not be found by tRNAscan-SE were identified by comparing to other hemipterans. Secondary structures of the small and large ribosomal RNAs were inferred using alignment to the models predicted for Drosophila melanogaster and D. virilis ¹⁹, Apis mellifera ²⁰, Manduca sexta ²¹ and Ruspolia dubia ²². Stem-loops were named using both the conventions of A. mellifera ²⁰ and M. sexta ²¹.

Protein-coding gene sequences were aligned using Clustal X ²³. The aligned data were further analyzed by MEGA version 4.0 ²⁴ for the codon usage. The putative control region was examined for regions of potential inverted repeats or palindromes with the aid of the mfold web server (http://www.bioinfo.rpi.edu/applications/mfold/) ²⁵. Strand asymmetry was calculated using the formulae: AT skew= [A−T]/ [A+T] and GC skew= [G−C]/ [G+C] ²⁶, for the strand encoding the majority of the protein-coding genes.

Phylogenetic analysis

Phylogenetic analysis was carried out based on the 32 complete or nearly complete mt genomes of true bugs from GenBank. Four species from Sternorrhyncha and Auchenorrhyncha were selected as outgroups (Table 1). Based on an analysis of mt genomes of nine Nepomorpha and five other hemipterans, Pleidae were suggested to be raised from a superfamily to the infraorder Plemorpha ²⁷. Since we didn't add samples to solve this problem, Paraplea frontalis was treated as incertae sedis, and was not included in the phylogenetic analysis to ensure the stability of the topology.

Table 1.

Summary of sample information used in present study

Order/suborder	Infraorder/superfamily	Family	Species	Accession Number	Reference
Sternorrhyncha
	Psylloidea	Psyllidae	Pachypsylla venusta	NC_006157	34
	Aphidoidea	Aphididae	Acyrthosiphon pisum	NC_011594	35
Auchenorrhyncha
	Fulgoroidea	Fulgoridae	Lycorma delicatula	NC_012835	27
		Issidae	Sivaloka damnosa	NC_014286	36
Heteroptera
	Gerromorpha
	Hydrometroidea	Hydrometridae	Hydrometra sp.	NC_012842	27
	Gerroidea	Gerridae	Gerris sp.	NC_012841	27
	Nepomorpha
	Corixoidea	Corixidae	Sigara septemlineata	FJ456941	27
	Ochteroidea	Gelastocoridae	Nerthra sp.	NC_012838	27
		Ochteridae	Ochterus marginatus	NC_012820*	27
	Notonectoidea	Notonectidae	Enithares tibialis	NC_012819	27
		Pleidae	Paraplea frontalis	NC_012822	27
	Nepoidea	Nepidae	Laccotrephes robustus	NC_012817	27
		Belostomatidae	Diplonychus rusticus	FJ456939*	27
	Naucoroidea	Naucoridae	Ilyocoris cimicoides	NC_012845	27
		Aphelocheiridae	Aphelocheirus ellipsoideus	FJ456940*	27
	Leptopodomorpha
	Saldoidea	Saldidae	Saldula arsenjevi	NC_012463	49
	Leptopodoidea	Leptopodidae	Leptopus sp.	FJ456946	27
	Cimicomorpha
	Naboidea	Nabidae	Alloeorhynchus bakeri	HM 235722
	Cimicoidea	Anthocoridae	Orius niger	NC_012429*	49
	Reduvioidea	Reduviidae	Triatoma dimidiata	NC_002609	33
			Valentia hoffmanni	NC_012823	27
	Miroidea	Miridae	Lygus lineolaris	EU401991*	Roehrdanz, unpublished
	Pentatomomorpha
	Aradoidea	Aradidae	Neuroctenus parus	NC_012459	49
	Pentatomoidea	Pentatomidae	Nezara viridula	NC_011755	49
			Halyomorpha halys	NC_013272	37
		Cydnidae	Macroscytus subaeneus	NC_012457*	49
		Plataspidae	Coptosoma bifaria	NC_012449	49
	Lygaeoidea	Berytidae	Yemmalysus parallelus	NC_012464	49
		Colobathristidae	Phaenacantha marcida	NC_012460*	49
		Malcidae	Malcus inconspicuus	NC_012458	49
		Geocoridae	Geocoris pallidipennis	NC_012424*	49
	Pyrrhocoroidea	Largidae	Physopelta gutta	NC_012432	49
		Pyrrhocoridae	Dysdercus cingulatus	NC_012421	49
	Coreoidea	Alydidae	Riptortus pedestris	NC_012462	49
		Coreidae	Hydaropsis longirostris	NC_012456	49
		Rhopalidae	Aeschyntelus notatus	NC_012446*	49
			Stictopleurus subviridis	NC_012888	49

^* Mt genome sequence was incomplete.

A DNA alignment was inferred from the amino acid alignment of the 13 protein-coding genes using Clustal X ²³. Alignments of individual genes were then concatenated excluding the stop codon.

Model selection was done with MrModeltest 2.3 ²⁸ and Modeltest 3.7 ²⁹ for Bayesian inference and ML analysis, respectively. According to the Akaike information criterion, the GTR+I+G model was optimal for analysis with nucleotide alignments. MrBayes Version 3.1.1 ³⁰ and a PHYML online web server ³¹ were employed to analyze this data set under the GTR+I+G model. In Bayesian inference, two simultaneous runs of 3, 000, 000 generations were conducted for the matrix. Each set was sampled every 200 generations with a burnin of 25%. Trees inferred prior to stationarity were discarded as burnin, and the remaining trees were used to construct a 50% majority-rule consensus tree. In ML analysis, the parameters were estimated during analysis and the node support values were assessed by bootstrap resampling (BP) ³² calculated using 100 replicates.

Results

Genome organization and structure

The mt genome of A. bakeri was a double-stranded circular molecule of 15, 851 bp in length (GenBank: HM 235722; Fig.1), and it contained the entire set of 37 genes usually present in most insect mtDNAs (13 PCGs, 22 tRNA genes, and two rRNA genes), and a large non-coding region (control region) (Table 2).

Table 2.

Organization of the A. bakeri mt genome

Gene	Direction	Location	Size	Anticodon	Codon		Intergenic nucleotides a
					Start	Stop
tRNAIle	F	1-63	63	30-32 GAT
tRNAGln	R	67-133	67	102-104 TTG			3
tRNAMet	F	133-198	66	164-166 CAT			-1
ND2	F	199-1197	999		ATT	TAA	0
tRNATrp	F	1196-1258	63	1227-1229 TCA			-2
tRNACys	R	1251-1316	66	1281-1283 GCA			-8
tRNATyr	R	1319-1381	63	1347-1349 GTA			2
COI	F	1383-2916	1534		TTG	T-	1
tRNALeu(UUR)	F	2917-2981	65	2946-2948 TAA			0
COII	F	2982-3660	679		ATT	T-	0
tRNALys	F	3661-3730	70	3691-3693 CTT			0
tRNAAsp	F	3730-3794	65	3761-3763 GTC			-1
ATP8	F	3795-3953	159		ATA	TAA	0
ATP6	F	3947-4630	684		ATG	TAA	-7
COIII	F	4617-5404	788		ATG	TA-	-14
tRNAGly	F	5404-5463	60	5433-5435 TCC			-1
ND3	F	5464-5817	354		ATA	TAA	0
tRNAAla	F	5821-5880	60	5850-5852 TGC			3
tRNAArg	F	5884-5946	63	5914-5916 TCG			3
tRNAAsn	F	5945-6010	66	5976-5978 GTT			-2
tRNASer(AGN)	F	6010-6078	69	6037-6039 GCT			-1
tRNAGlu	F	6078-6141	64	6109-6111 TTC			-1
tRNAPhe	R	6140-6202	63	6167-6169 GAA			-2
ND5	R	6202-7907	1706		ATT	TA-	-1
tRNAHis	R	7905-7966	62	7933-7935 GTG			-3
ND4	R	7966-9294	1329		ATG	TAA	-1
ND4L	R	9288-9581	294		ATT	TAG	-7
tRNAThr	F	9593-9655	63	9624-9626 TGT			11
tRNAPro	R	9656-9718	63	9687-9689TGG			0
ND6	F	9721-10218	498		ATA	TAA	2
CytB	F	10218-11354	1137		ATG	TAG	-1
tRNASer(UCN)	F	11353-11420	68	11384-11386TGA			-2
ND1	R	11441-12362	922		ATA	T-	20
tRNALeu(CUN)	R	12363-12428	66	12397-12399TAG			0
lrRNA	R	12429-13680	1252				0
tRNAVal	R	13681-13749	69	13716-13718 TAC			0
srRNA	R	13750-14539	790				0
Control region		14540-15851	1312

^a Negative numbers indicate that adjacent genes overlap.

Twenty-three genes were transcribed on the majority strand (J-strand), whereas the others were oriented on the minority strand (N-strand). Gene overlaps were found at 17 gene junctions and involved a total of 54 bp; the longest overlap (14 bp) existed between ATP6 and COIII. In addition to the control region, there were 45 nucleotides dispersed in 8 intergenic spacers, ranging in size from 1 to 20 bp. The longest spacer sequence was located between tRNA^{Ser (UCN)} and ND1.

Transfer RNAs

The entire complement of 22 tRNAs was found in A. bakeri, and 20 of them were determined using tRNAscane-SE ¹⁸. The tRNA^Arg and tRNA^{Ser (AGN)} genes were not detected by software, and were determined through comparison with previously published hemipteran mt genomes ^{27, 33}. All tRNAs could fold into the typical clover-leaf structure except for tRNA^{Ser (AGN)}, in which its dihydrouridine (DHU) arm simply formed a loop (Fig. 2).

Fig 2.

Inferred secondary structure of 22 tRNAs of the A. bakeri mt genome. The tRNAs are labeled with the abbreviations of their corresponding amino acids. Dashed (-) indicate Watson-Crick base pairing and (+) indicate G-U base pairing.

The length of tRNAs ranged from 60 to 70 bp. The aminoacyl (AA) stem (7 bp) and the AC loop (7 nucleotides) were invariable, and most of the size variation was the DHU and TΨC (T) arms, within which the loop size (3-9 bp) was more variable than the stem size (2-5 bp). The size of the anticodon stems was conservative, with the exception of tRNA^{Ser (AGN)} which possessed a long optimal base pairing (9 bp in contrast to the normal 5) and a bulged nucleotide in the middle for the AC stem.

Based on the secondary structure, a total of 28 unmatched base pairs were found in the A. bakeri tRNAs. Twenty-three of them were G-U pairs, which form a weak bond, located in the AA stem (8 bp), the DHU stem (9 bp), the AC stem (2 bp), the T stem (4 bp), the remaining 5 included C-U (2 bp) mismatches in the AA stem and the T stem of tRNA^Arg, respectively; A-A (2 bp) mismatches in the AA stem of tRNA^Arg; U-U mismatches (1 bp) in the AA stem of tRNA^Ala.

Ribosomal RNAs

The boundaries of rRNA genes were determined by sequence alignment with that of Triatoma dimidiata ³³ and Valentia hoffmanni ²⁷. As in most other insect mt genomes, the large and small ribosomal RNAs (rrnL and rrnS) genes in A. bakeri were located between tRNA^Leu(CUN) and tRNA^Val and between tRNA^Valand the control region, respectively (Fig. 1; Table 2). The length rrnL and rrnS were determined to be 1, 252 bp and 790 bp, respectively. The secondary structure of rrnL consisted of six structural domains (domain III is absent in arthropods) and 45 helices (Fig. 3), and the rrnS consisted of three structural domains and 27 helices (Fig. 4).

Fig 3.

Predicted secondary structure of the rrnL gene in the A. bakeri mt genome. Roman numerals denote the conserved domain structure. The numbering system follows ²⁰. Dashed (-) indicate Watson-Crick base pairing and dot (•) indicate G-U base pairing.

Fig 4.

Predicted secondary structure of the rrnS gene in the A. bakeri mt genome. Roman numerals denote the conserved domain structure. Dashed (-) indicate Watson-Crick base pairing and dot (•) indicate G-U base pairing. Structural annotations follow Fig. 3.

Protein-coding genes: Translation initiation and termination signals

All but one PCGs of A. bakeri initiated with ATN as the start codon (four with ATG, four with ATT and four with ATA) (Table 2). The only exception was the COI gene, which used TTG as a start codon.

The majority of the PCGs of A. bakeri had the complete termination codons TAA (ND2, ATP8, ATP6, ND3, ND4 and ND6) or TAG (ND4L and CytB), and the remaining five had incomplete termination codons, TA (COIII and ND5) or T (COI, COII and ND1) (Table 2).

Nucleotide composition and codon usage

The nucleotide composition of the A. bakeri mtDNA was significantly biased toward A and T. The A+T content was 73.5% (A = 40.1%, T = 33.4%, C = 16.3 %, G = 10.2%). The A+T content of isolated PCGs, tRNAs, rRNAs and the CR is 72.6%, 75.4%, 75.7% and 75.7%. The skew statistics of the total PCGs demonstrated that the J-strand PCGs were CG-skewed and consisted of nearly equal A and T while the N-strand PCGs were GC-skewed and much more TA-skewed, and the N-strand tRNAs had also higher GC-skewed than the J-strand tRNAs.

The nucleotide bias was also reflected in the codon usage. Analysis of base composition at each codon position of the concatenated 13 PCGs showed that the third codon position (81.2%) was higher in A+T content than the first (68.5%) and second (66.3%) codon positions (Table 3). There were different nucleotide frequencies in all codon position between the two strands in A. bakeri. If the J-strand alone was inspected, the third codon position sites showed a preponderance of A nucleotides, whereas for N-strand, the third codon position sites biased toward T (Table 3).

Table 3.

Nucleotide composition of the A. bakeri mt genome

	Proportion of nucleotides
Feature	%T	%C	%A	%G	%A+T	AT Skew	GC Skew	No. of nucleotides
Whole genome	33.4	16.3	40.1	10.2	73.5	0.09	-0.23	15851
Protein-coding genes	40.6	14.0	32.0	13.3	72.6	-0.12	-0.03	11085
First codon position	33.7	13.1	34.8	18.4	68.5	0.02	0.17	3695
Second codon position	46.7	18.5	19.6	15.2	66.3	-0.41	-0.10	3695
Third codon position	41.5	10.4	39.7	8.3	81.2	-0.02	-0.11	3695
Protein-coding genes-J	35.9	17.0	35.5	11.6	71.4	-0.01	-0.19	6819
First codon position	28.6	15.8	38.0	17.6	66.6	0.14	0.05	2273
Second codon position	44.3	20.9	20.6	14.2	64.9	-0.37	-0.19	2273
Third codon position	34.6	14.4	47.9	3.0	82.5	0.16	-0.66	2273
Protein-coding genes-N	48.3	9.2	26.4	16.1	74.7	-0.29	0.27	4267
First codon position	41.9	8.8	29.7	19.5	71.6	-0.17	0.38	1422
Second codon position	50.4	14.7	18.0	16.9	68.4	-0.47	0.07	1422
Third codon position	52.5	4.1	31.6	11.9	84.1	-0.25	0.49	1422
tRNA genes	36.7	10.6	38.7	14.1	75.4	0.03	0.14	1427
tRNA genes-J	35.2	12.3	39.9	12.6	75.1	0.06	0.01	908
tRNA genes-N	39.1	7.5	36.6	16.8	75.7	-0.03	0.38	519
rRNA genes	41.2	8.8	34.5	15.6	75.7	-0.09	0.28	2042
Control region	39.4	16.5	36.3	7.9	75.7	-0.04	-0.35	1312

Four most frequently used codon, TTA (leucine), ATT (isoleucine), TTT (phenylalanine) and ATA(methionine), were all composed wholly of A and/or T, and NNA and NNC codons were more frequent than NNU and NNG in PCGs encoded on the J-strand, whereas the N-strand genes showed exactly the opposite trend (Fig. 5).

Fig 5.

Relative synonymous codon usage (RSCU) in the A. bakeri mt genome. Codon families are provided on the x-axis.

The control region

The 1, 312 bp long control region of A. bakeri mt genome was located at the conserved position between rrnS and tRNA^Ile- tRNA^Gln- tRNA^Met gene cluster (Fig. 1), and was composed of 75.7% A+T content, which was the most A+T-rich region (Table 3).

The control region of A. bakeri can be divided into four parts (Fig. 6A): (1) a 533 bp region that was bordered by rrnS, of which the G+C content (33.2%) is higher than the whole genome, and at the beginning of this region contained two 21 bp C-rich repetitive sequences (TCCCCCCTCCGGTGGTCGCTA); (2) a 39 bp region heavily biased toward A+T (89.7%); (3) a region composed of five tandem repeats; (4) a region at the end of control region containing 4 potential stem-loop structures, the largest one with a stem of 20 bp and 21 bp loop (Fig. 6B).

Fig 6.

Control region of the A. bakeri mt genome. (A) Structure elements found in the control region of A. bakeri. The control region flanking genes rrnS, trnI (I), trnQ (Q), and trnM (M) are represented in grey boxes; the blue and azury boxes with roman numerals indicate the tandem repeat region; “G+C” (yellow) indicates high G+C content region; repeats (pink) indicate two repeat regions at the beginning of the high G+C content region; “A+T” (green) indicates high A+T content region. (B) The putative stem-loops structure found in the control region. The grey boxes indicate highly conserved flanking sequence.­

Phylogenetic relationships

We performed phylogenetic analysis using nucleotide sequences of 13 mt PCGs from 32 heteropteran species and 4 outgroup hemipteran insect species ^{27, 34-37}. BI and ML analyses generated identical tree topologies (Fig. 7).

Fig 7.

Phylogenetic tree inferred from the mt genomes of 32 heteropterans. Phylogenetic analysis was based on all 13 protein-coding genes. The tree was rooted with four outgroup taxa (P. venusta, A. pisum, S. damnosa and L. delicatula). Cycles indicate bootstrap support; percentages of Bayesian posterior probabilities (upper) and ML bootstrap support values (underside).

In the present study, the sister-relationship within the infraorders were supported for the Pentatomomorpha (14 taxa), Nepomorpha (8 taxa), Leptopodomorpha (2 taxa) and Gerromorpha (2 taxa) by BI and ML analysis. Two Gerromorpha superfamilies were monophyletic in the basal position of these five infraorders. Within Cimicomorpha, Reduviidae was paraphyletic with respect to the Nabidae, Anthocoridae and Miridae. The sister-relationship of Nabidae, Anthocoridae and Miridae was confirmed. The infraordinal relationships tended to be poorly resolved with low support.

Discussion

The mt genome of A. bakeri is a double-stranded circular molecule, with the same gene content (37 genes and 1 control region) and gene order as that in D. yakuba ³⁸. The overall organization of the A. bakeri mt genome is very compact, and the overlaps between ATP8/ATP6 (7 bp) and ND4/ND4L (7 bp) are often been found across the Metazoa ^{39, 40}.

Dihydrouridine (DHU) arm of A. bakeri tRNA^{Ser (AGN)} simply forms a loop. This phenomenon is common in sequenced true bug mt genomes, and has been considered as a typical feature of metazoan mtDNA ⁴¹. Some A. bakeri tRNA genes possessed non-Watson-crick matches, aberrant loops, or even extremely short arms. It is not known whether the aberrant tRNAs lose their function in every case, but a post-transcriptional RNA editing mechanism has been proposed to maintain function of these tRNA genes ^{42, 43}.

The secondary structures of the A. bakeri mt rrnL and rrnS are drawn following the previously published models for M. sexta ²¹. In rrnL, H837 forms a long stem structure with a small loop in the terminal as observed in other insects ^{20, 21, 44, 45}. In many insect mtDNA, the helix H2077 is absent as the bases do not form complementary pairs ^{21, 46}, whereas it includes a 23 paired bases stem and 12 bp loop in A. bakeri. The helix H2347 is also highly variable in insect, and in A. bakeri this region consisting of 5 paired bases is similar to that proposed for M. sexta ²¹. H2735, the last stem-loop of rrnL, only forms a 4 bp stem and 6 bp loop in A. bakeri which is difference in size from M. sexta, 7 bp stem and 22 bp loop ²¹.

Domains I and II are alterable regions in terms of sequence and structure, whereas domain III is highly conserved part of the rrnS of A. bakeri. Helix 47 is variable among different insects, but the terminal portion of this stem is conserved ^{21, 45}, and in A. bakeri two loops are formed similar to those in Evania appendigaster ⁴⁵. The sequence between H577 and H673 can't be folded, similar to that in M. sexta ²¹. H1047 and associated stems H1068, H1074 and H1113 may yield multiple possible secondary structures due to its high AT bias and several non-canonical base pairs, as discussed in other insects ^{20, 21, 44, 47, 48}.

An unconventional TTG start codon was detected only for the COI gene in A. bakeri, which is consistent with some other true bugs ^{27, 49, 50}, and other insects (mainly in Diptera) ^{38, 51-54}. The presence of an incomplete stop codon is a common phenomenon found in mt genomes of insects and it has been proposed that the complete termination codon TAA could be generated by the posttranscriptional polyadenylation ^{55, 56}.

The A+T content of A. bakeri corresponds well to the AT bias generally observed in hexapod mt genomes, which range from 64.8% in Japyx solifugus ⁵⁷ to 87.4% in Diadegma semiclausum ⁴⁴.

Metazoan mt genomes usually present a clear strand bias in nucleotide composition ^{58, 59}, and the strand bias can be measured as AT- and GC-skews ²⁶. AT- and GC-skews of A. bakeri mt genomes is consistent compared to the usual strand biases of metazoan mtDNA (positive AT-skew and negative GC-skew for the J-strand, and whereas the reverse is observed in the N-strand). The underlying mechanism that leads to the strand bias has been generally related to replication, because this process has long been assumed to be asymmetric in the mtDNA and could therefore affect the occurrence of mutations between the two strands ⁵⁸. It is possible that the overall genome A-bias is driven by mutational pressure on the N-strand and the GC-skew may be correlated with the asymmetric replication process of the mtDNA ⁶⁰.

The nucleotide bias is also reflected in the codon usage. As reported for other metazoan mtDNAs, the most commonly used codon in degenerate codon families often does not match the anticodon ⁶. All codons are present in A. bakeri mtDNA PCGs, but GCG codon is not represented on the J-strand, and CAC and CGC codons for the N-strand, reflecting the influence of a strong biased codon usage ⁵³. Codon usage may be influenced by other molecular processes such as translational selection efficiency and accuracy, which apparently have a stronger influence in organisms with rapid growth rates ^{57, 61, 62}.

The largest non-coding region (1, 312 bp) was flanked by rrnS and the tRNA^Ile gene in the A. bakeri mt genome. It was highly enriched in AT (75.7%) and could form stable stem-loop secondary structures. Repeated sequences are common in the control region for most insects, and length variations due to the various numbers of repeats are not without precedent ⁶³. In the case of A. bakeri, the control region includes four 133 bp tandem repeat units plus a partial copy of the repeat (28 bp of the beginning).

The stem-loop structure in the control region is suggested as the site of the initiation of secondary strand synthesis in Drosophila ⁶⁴. The flanking sequence of the structure is suggested to be highly conserved among some insects, possessing the consensus 'TATA' sequence at the 5' end and 'G(A)nT' at the 3' end ^{65, 66}. However, in the A. bakeri control region, no highly conserved flanking 'TATA' sequence existed at the 5' end, but we found 'G(A)nT' at the 3' end (Fig. 6B).

The poly-thymine stretch is relatively conserved across insects ⁶³. In A. bakeri this stretch locates in the beginning of the fourth part of control region and spans 12 thymine nucleotides with one adenine. It has been speculated that this poly-thymine stretch may be involved in transcriptional control or may be the site for initiation of replication ⁶⁴.

The topology of infraordinal relationships of Heteroptera is similar to previous work ⁶⁷, and future analyses should focus on phylogeny studies including Dipsocoromorpha and Enicocephalomorpha mt genome data and additional representatives for some poorly sampled clades. The sister-relationship of Nabidae, Anthocoridae and Miridae is confirmed in the present study. But the position of Reduviidae is not improved, although the mt genome of A. bakeri is added. Cimicomorpha comprise over 20, 000 species currently placed in 17 families ⁶⁸, but only 4 families have the mt genome data, and it is too limited to resolve the phylogeny of Cimicomorpha, and increased taxon sampling may be the best way to resolve this problem.

Supplementary Material

Table S1: Primer sequences used in this study.