First mitogenome for the tribe Saccharosydnini ( Hemiptera : Delphacidae : Delphacinae ) and the phylogeny of three predominant rice planthoppers

The mitochondrial genome of Saccharosydne procerus (Matsumura) is the fi rst sequenced in the tribe Saccharosydnini (Hemiptera: Delphacidae: Delphacinae). In addition, the mitogenome sequence of Sogatella vibix (Haupt) (in Delphacini) is also sequenced. The Sa. procerus mitochondrial genome is 16,031 bp (GenBank accession no. MG515237) in length, and So. vibix is 16,554 bp (GenBank accession no. MG515238). The existence of purifying selection was indicated by the rate of nonsynonymous and synonymous substitutions. Three species of Delphacini, Laodelphax striatellus (Fallén), Sogatella furcifera (Horváth) and Nilaparvata lugens (Stål), are important pests of rice. The phylogeny of these three rice planthoppers based on the mitochondrial genome sequence was (L. striatellus + (So. vibix + So. furcifera)) + (N. muiri + N. lugens). * Corresponding author; e-mail: qindaozh0426@aliyun.com INTRODUCTION The planthopper subfamily Delphacinae is the most speciose and economically important group in the family Delphacidae. It comprises three tribes (Delphacini, Tropidocephalini and Saccharosydnini) and contains over 80% of all delphacid species (Asche, 1985; Bourgoin, 2017). Some members in this subfamily are pests of crops or vectors of plant pathogens, causing economic losses widely reported around the world, for example, three species of Delphacini, Laodelphax striatellus (Fallén), Sogatella furcifera (Horváth) and Nilaparvata lugens (Stål) as important pests of rice (e.g. Cai et al., 2003; Wilson, 2005; Grilli, 2006; Grimshaw & Donaldson, 2007; Wang et al., 2008; Heong et al., 2014; Zhang et al., 2014). Despite several recent studies on the phylogeny of this group (Asche, 1985, 1990; Yang et al., 1987; Emeljanov, 1996; Dijkstra et al., 2003, 2006; Hamilton, 2006; Urban et al., 2010; Huang et al., 2017), more data (including mitochondrial genomes evidence) are still needed to better understand the evolution of Delphacinae. Insect mitochondrial genomes (mitogenomes) are small, double stranded, circular DNA molecules, ranging in size from 14 to 19 kb. They are composed of 37 genes (13 protein-coding, 22 transfer RNA, and 2 ribosomal RNA genes), and a control region (A + T-rich region) that is thought to Eur. J. Entomol. 115: 242–248, 2018 doi: 10.14411/eje.2018.023

Insect mitochondrial genomes (mitogenomes) are small, double stranded, circular DNA molecules, ranging in size from 14 to 19 kb.They are composed of 37 genes (13 protein-coding, 22 transfer RNA, and 2 ribosomal RNA genes), and a control region (A + T-rich region) that is thought to nin et al. ( 2005): AT-skew = ( A -T) / (A + T) and GC-skew = (G -C) / (G + C).The number of synonymous substitutions per synonymous site (Ks) and the number of nonsynonymous substitutions per nonsynonymous site (Ka) for each concatenated 13 PCGs of Delphacini mitogenome were calculated by DnaSP 5 (Rozas et al., 2003), with stop codons and codons with alignment gaps excluded, using the sequence of Sa. procerus from Sac charosydnini as a reference sequence.

Phylogenetic analysis
Two newly generated mitogenomes and 12 from GenBank (Table 1) were analyzed in this study, with Sa. procerus selected as an outgroup.Alignment of PCGs was conducted by using MAFFT 7.3.1 using G-INS-I algorithms (Katoh & Tandley, 2016).Two rRNA segments were aligned by the R-Coffee web server (Moretti, 2008).Subsequently, all alignments were concatenated in a single matrix using DAMBE (Xia, 2013).We used PartitionFinder 1.1.1(Lanfear et al., 2012) to infer the optimal partitioning strategy; the best-fi tting model was selected for each partition using the BIC (Bayesian Information Criterion).
Both ML (Maximum likelihood) and BI (Bayesian inference) analyses were conducted on the concatenated dataset for phylogeny reconstruction.Maximum likelihood analysis was conducted in IQtree v1.4.1 (Lam-Tung et al., 2015) using the best-fi t substitution model.An ultrafast bootstrap (UFB) (Bui et al., 2013) of 1000 replications and the SH-aLRT test were used in this analysis to assess branch supports.

Sequencing and assembly
A whole genome shotgun (WGS) strategy was used with sequencing on the Illumina Miseq platform.The quality of data was checked by FastQC (Andrews, 2016).The adapters of raw data were removed by AdapterRemoval version 2 (Schubert et al., 2016).SOAPec version 2.01 was used for quality correction, setting K-mer to 17. Reads with a length of less than 50 bp were excluded.Assembly of the mitochondrial (mt) genome was done using A5-miseq version 2.0 (Coil et al., 2014).

Mitochondrial genome annotation
The tRNA genes were identifi ed and secondary structures of tRNAs were predicted using MITOS WebSever, setting the parameters with the Invertebrate Mito genetic code (Bernt et al., 2013).Every sequence of tRNA genes was checked separately by eye.Protein-coding genes (PCGs) were identifi ed as open reading frames corresponding to the 13 PCGs in the metazoan mt genomes.The rRNA genes and control region were identifi ed by the boundary of the tRNA genes and by alignment with other Delphacidae mitogenomes.The mitogenome map was produced using CGView (Grant & Stothard, 2008).

Comparative analysis
Base composition and relative synonymous codon usage (RSCU) were analyzed using MEGA 6.0 (Tamura et al., 2013).GC and AT asymmetry were measured in terms of GC and AT skews using the following formulae suggested by Hassa-  Bayesian inference analysis was conducted using BEAST 1.8.0 (Drummond et al., 2012).Chains were run for 20 million generations, with sampling every 2000 generations.Tracer 1.6.0(Rambaut et al., 2014) was used to verify the posterior distribution and to ensure effective sample sizes (ESSs) > 200 from the Markov Chain Monte Carlo (MCMC) output.TreeAnnotator in the BEAST package was used to summarize tree data with "median height".The fi rst 25% of samples were discarded as burn-in and the remaining samples were used to generate a 50% majority rule consensus tree.FigTree v.1.3.1 (Rambaut, 2009) was used to view the resulting trees.

RESULTS AND DISCUSSION
The Sa. procerus mitochondrial genome (GenBank accession no.MG515237) is 16,031 bp in length (Fig. 1), and the overall nucleotide composition exhibits a high A + T   content of 80.5% (Table 2).The mitogenome of So. vibix (GenBank accession no.MG515238) is 16,554 bp long with an A + T content of 76.0% (Table 3), likewise heavily biased toward the A and T nucleotides (Fig. 2).The mitogenomes of bot h species encode a complete set of 37 genes (Tables 4-5 ) which are usually found in animal mitogenomes, consisting of 13 protein-coding genes (PCG), 2 ribosomal RNA (rRNA) genes and 22 transfer RNA (tRNA) genes (Cameron, 2014).The gene arrangements in the mitochondrial genomes of Sa. procerus and So.vibix are conserved, similar to other mitogenomes of Delphacidae, with the exception of Nila parvata lugens (Stål).Zhang et al. (2013) found three trnC genes in N. lugens, but only one trnC gene was found by Lv et al. (2015) which corresponds to most hemipteran insects sequenced so far (Wang et al., 2015).Most PCGs share the start codon ATT or ATG, with nad6 of Sa. procerus starting with ATA.Four genes of So. vibix (cox1,atp6,cox3,nad5) and fi ves genes of Sa. procerus (cox1, atp6 , cox3, nad5, nad1) use the incomplete stop codon T. Four genes of So. vibix (cox2,nad4l,cytb,nad1) and nad4l of Sa. procerus use TAG.The remaining PCGs  use the stop codon TAA.The stop codon of nad1 in Sa. procerus (T) is different from those in Delphacini (TAA or TAG).This suggests that during evolution the nad1 gene in Sa. procerus acquired a different mechanism for transcription termination.Further genome sequencing is needed to fi nd out whether this feature exists only in Sa. procerus or in the tribe Saccharosydnini.The use of anti-codons for 22 tRNAs are all the same between So. vibix and Sa.procerus.
The relative synonymous codon usage (RSCU) of Sa. procerus and So.vibix are shown in Figs 3-4.The codon usage in these mitogenomes shows a high AT content.The most frequently used amino acids were Phe, Leu and Ile, while TTT (Phe), TTA (Leu) and ATT (Ile) were the most frequently utilized c odons.All three of these most frequently utilized codons are composed of A and T. Additionally, it is obvious that the preferred codon usage is A or T in the third position rather than G and C .Almost all of the frequently used codons ended with A or T, which may contribute to the signifi cant bias towards A and T.
The rate of nonsynonymous substitutions (Ka), synonymous substitutions (Ks), and the ratio of Ka/Ks were calculated for PCGs of each delphacine mitogenome with Sa. procerus as the reference sequence (Fig. 5).All of the Ka, Ks and the ratios of Ka/Ks values were less than 1, indicating the existence of purifying selection in these species.
Saccharosydne procerus (tribe Saccarosydnini) was selected as the outgroup based on results of previous analyses that placed this tribe (plus Tropidocephalini) as sister to Delphacini (Asche, 1985(Asche, , 1990;;Urban et al., 2010;Huang et al., 2017).Peregrinus maidis was also included to test t he polarity of the phylogeny.The result placed P. maidis as sister to the remaining Delphacini, which is concordant with our previous study (Huang et al., 2017).We therefore think the use of Sa. procerus as the outgroup taxon is appropriate.
The phylogenetic analyses of ML and BI based on mitogenome datasets yield two identical topologies (Fig. 6) when rooted with Sa. procerus (of Saccharosydnini); remaining species form the tribe Delphacini, with P. maidis being sister to the remaining species.The conformation of the clade containing the three rice planthoppers (L.striatellus, So. furcifera and N. lugens) was (L.striatellus + (So.vibix + So. furcifera)) + (N.muiri + N. lu gens).Moreover, the relationships among biotypes of N. lugens were recovered.
This study documents the fi rst mitogenome of Saccharosydnini and the mitogenome of So. vibix, which both contain 37 typical metazoan mitochondrial genes and retain the organization of the most other Delphacidae mitogenomes.The phylogeny based on more taxa is needed to better understand the evolution of Delphacidae.Therefore, more mitogenomes need to be sequenced in further studies.

Fig. 5 .
Fig.5.Evolutionary rates of Delphacini mitochondrial genomes.The number of nonsynonymous substitutions per nonsynonymous site (Ka), the number of synonymous substitutions per synonymous site (Ks), and the ratio of Ka/Ks for each Delphacini mitochondrial genome are given, using that of Saccharosydne procerus as a reference sequence.

Fig. 6 .
Fig. 6.Phylogenetic tree of three predominant rice planthoppers obtained from ML analysis based on concatenated data of 13 PCGs and two rRNA genes.The numbers at nodes indicate ML bootstrap values/Bayesian posterior probabilities, respectively.Accession numbers are given for species obtained from GenBank.

Table 1 .
Taxa included in the phylogenetic analyses in this study.

Table 2 .
Nucleotide composition of the Saccharosydne procerus mitochondrial genome.

Table 3 .
Nucleotide composition of the Sogatella vibix mitochondrial genome.

Table 4 .
Organization of the mitogenome of Saccharosydne procerus.

Table 5 .
Organization of the mitogenome of Sogatella vibix.