Sequence Analysis of Insecticide Action and Detoxification-Related Genes in the Insect Pest Natural Enemy Pardosa pseudoannulata

The pond wolf spider Pardosa pseudoannulata, an important natural predatory enemy of rice planthoppers, is found widely distributed in paddy fields. However, data on the genes involved in insecticide action, detoxification, and response are very limited for P. pseudoannulata, which inhibits the development and appropriate use of selective insecticides to control insect pests on rice. We used transcriptome construction from adult spider cephalothoraxes to analyze and manually identify genes enconding metabolic enzymes and target receptors related to insecticide action and detoxification, including 90 cytochrome P450s, 14 glutathione S-transferases (GSTs), 17 acetylcholinesterases (AChEs), 17 nicotinic acetylcholine receptors (nAChRs), and 17 gamma-aminobutyric acid (GABA) receptors, as well as 12 glutamate-gated chloride channel (GluCl) unigenes. Sequence alignment and phylogenetic analysis revealed the different subclassifications of P450s and GSTs, some important sequence diversities in nAChRs and GABA receptors, polymorphism in AChEs, and high similarities in GluCls. For P450s in P. pseudoannulata, the number of unigenes belonging to the CYP2 clade was much higher than that in CYP3 and CYP4 clades. The results differed from insects in which most P450 genes were in CYP3 and CYP4 clades. For GSTs, most unigenes belonged to the delta and sigma classes, and no epsilon GST class gene was found, which differed from the findings for insects and acarina. Our results will be useful for studies on insecticide action, selectivity, and detoxification in the spider and other related animals, and the sequence differences in target genes between the spider and insects will provide important information for the design of selective insecticides.


Introduction
Biological control is an important way for pest management. Currently, increasing interest in safe, effective, and sustainable strategies for insect pest control is encouraging the use of biological control strategies, but the primary way for insect pests control in China remains the application of chemical insecticides, which not only cause environmental contamination but also reduce the populations of natural enemies [1]. Spiders are recognized as important natural enemies to reduce pest populations. The pond wolf spider Pardosa pseudoannulata, an important predatory enemy of rice planthoppers and leafhoppers, is one of the most abundant spiders in paddy fields and able to effectively reduce rice pest populations. Nevertheless, spiders have been variably affected by the extensive use of insecticides, especially neurotoxic substances [2]. Neonicotinoid insecticides, acting selectively as neurotoxins on insect nicotinic acetylcholine (ACh) receptors (nAChRs) and extensively used to control rice insect pests, demonstrated relative safety to P. pseudoannulata [3]. In previous studies, the chemical structure and differences in the nAChRs between insects and P. pseudoannulata were found to contribute directly to neonicotinoid sensitivities [4][5][6].
To develop approaches to protect the natural enemy P. pseudoannulata and improve insecticide selectivity, pharmacological studies of insects and P. pseudoannulata are necessary. Molecular mechanisms of insecticide resistance in insects have been studied in depth, but such studies on spiders are rare, with only 3% of the toxicology papers on natural enemies devoted to spiders [2]. Two principal regulatory molecular mechanisms of insecticide resistance exist in insects: enhanced insecticide metabolism and reduced sensitivity of targets to insecticides [7]. Enhanced insecticide metabolism is mainly due to the increased metabolism enzyme activities or overexpression of enzyme genes, such as cytochrome P450s and glutathione S-transferases (GSTs) [8]. Reduced sensitivity of targets to insecticides mainly occurs through target mutation by reducing binding of the insecticide to its target [9]. The most frequently studied targets are acetylcholinesterases (AChEs), nicotinic acetylcholine receptors (nAChRs), gamma-aminobutyric acid (GABA) receptors, and glutamate-gated chloride channels (GluCls).
The lack of genetic information on P. pseudoannulata limits the studies of differences between insects and P. pseudoannulata, especially the genes related to insecticide action and detoxification. Prior to this study, only 11 amino acid sequences of P. pseudoannulata were found in NCBI (National Center for Biotechnology Information) database (http://www.ncbi.nlm.nih. gov), including four nAChR subunits and two AChE genes. Fortunately, high-throughput sequencing technology can provide the rapid high-quality genetic information. In the present study, we used the high-throughput sequencing Illumina HiSeq 2000 instrument to complete sequencing of the entire transcriptome and acquired a large volume of genetic information on P. pseudoannulata. Genes related to insecticide action and detoxification were manually identified and analyzed in sequences.

Results and Discussion
De novo assembly, sequence homology distribution, and function annotations Using Illumina sequencing, P. pseudoannulata transcriptome was generated from male and female adult spider cephalothoraxes. Basic information on the transcriptome is summarized in S1 Table. Length distribution and homology analysis of unigenes that hit in the NCBI nonredundant (Nr) protein database are shown in S1 and S2 Figs To acquire accurate annotation information, the Gene Ontology (GO), Clusters of Orthologous Groups (COG), and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases were also used to annotate the unigenes (S3 and S4 Figs, and S2 Table). The transcriptome data have been deposited in NCBI (accession number: GCKE00000000).

Identification and analysis of genes related to insecticide action and detoxification
Insecticides such as neonicotinoids show high toxicity to insect pests, but relative hypotoxicity to mammal and natural enemies. One of the main reasons for this is the differences in their insecticide targets and metabolism enzymes [6,10]. As an important predatory enemy of rice insect pests, genetic information on P. pseudoannulata is highly lacking. In this study, the P. pseudoannulata transcriptome supplied significantly more genetic information than that from the literatures. The genes related to insecticide action and detoxification were manually identified, including the metabolism enzymes cytochrome P450s and GSTs, and insecticide targets AChEs, nAChRs, GABA receptors, and GluCls (Table 1).

Cytochrome P450 family
The cytochrome P450 family, one of the largest superfamilies of insect species, has functions in xenobiotic metabolism and detoxification. The P450 superfamily includes four main clades: CYP2, CYP3, CYP4, and the mitochondrial CYP clade (CYP M) [11]. The multitudinous and highly active nature of P450s allows insects to metabolize nearly all classes of insecticides and other xenobiotics [12]. Changes in P450s were found to result in high insecticide resistance in many insect species [13][14][15][16][17][18][19][20][21]. In our dataset, 90 P450 unigenes with an average length of 583 bp were hit in the NCBI Nr database, and 79 putative P450 genes were identified (S3 Table). Most P450 unigenes belonged to CYP2 clade (34/90), followed by CYP4 and CYP3 clades (19/90 and 17/90), and nine unigenes were annotated to CYP M (9/90). However, due to the existence of unigenes with a short open reading frame and fragmentary distribution, only 24 P450 unigenes were well classified in the sequence alignment and phylogenetic analysis with model insect P450s (Fig 1).
In insects, the CYP3 clade contains the most numerous P450 genes and functions in xenobiotic metabolism and insecticide resistance. Studies have found that changes in allele or overexpression of genes in CYP3 clade resulted in high insecticide resistance in insects, such as CYP337B3 and CYP6AY1 [17,21]. The CYP4 clade P450 genes in insects were also found to be involved in both pesticide metabolism and chemical communication [22]. Overexpression of genes in the CYP4 clade also could lead to high insecticide resistance, e.g., CYP4C27 and CYP4G19 [23,24]. The mitochondrial P450s are all involved in steroid or vitamin D metabolism in vertebrates and xenobiotic metabolism in insects [25]. CYP2 clade is a relatively small clade when compared to CYP3 and CYP4 clades in insects, with the primary functions involved in juvenile hormone biosynthesis and ecdysone metabolism. In the present P. pseudoannulata transcriptome, CYP2 clade was superior in P450 unigene number and was close to the sum of CYP3 and CYP4 clade. This result differed markedly from that for insects, in which more P450 genes were found in CYP3 and CYP4 clades [26][27][28][29][30]. These findings indicated that the significant differences existed in xenophobic metabolism and detoxification between arachnoidea and insects. Glutathione S-transferase GST (E.C. 2.5.1.18) belongs to the multifunctional detoxification enzymes involved in the detoxification of xenobiotics. GSTs catalyze the conjugation of reduced glutathione (GSH) with exogenous and endogenous toxic compounds or their metabolites, rendering them more water-soluble, less toxic, and easier to excrete [31]. In insects, increased expression and/or activity of GSTs can enhance insecticide resistances [32,33]. GSTs consist of seven main classes: delta, epsilon, omega, sigma, theta, zeta, and microsomal. The delta and epsilon classes have been described as the insect-specific GST classes and implicated in xenobiotic metabolism, especially in insecticide detoxification. An epsilon GST was able to detoxify DDT and result in DDT-resistance in Anopheles gambiae [8]. Sigma GSTs show low activity with the typical GST substrates, but have high affinity with the lipid peroxidation product, 4-hydroxynonenal, and are localized in metabolically active tissues in flies [22]. Like the cytosolic enzymes, the microsomal GSTs play a similar role in general detoxification reactions and protection against oxidative stress [22]. In the present study, 14 GST unigenes were found in the P. pseudoannulata transcriptome belonging to three classes: delta (6/14), sigma (5/14), microsomal (2/14), and one unknown (S4 Table). Phylogenetic analysis with model insects produced 12 GST unigenes (Fig 2), most of which belonged to the delta and sigma classes, and no epsilon GST gene was found. The results differed from those for insects, in which the main GST classes were delta and epsilon, with no microsomal class. Moreover, the findings were also different from the main GST classes in acarina, which were delta and microsomal classes, but without sigma class [34]. In the multiped spider P. pseudoannulata, which has no wings, the sigma class GSTs possibly mainly function in the muscles related to jumping, but not in flight muscles.

Acetylcholinesterase
AChE (EC 3.1.1.7) is an important neurotransmitter hydrolase in insect nervous system that terminates transmission at the cholinergic synapses by hydrolyzing ACh [35]. Organophosphate and carbamate insecticides act as inhibitors on AChEs, blocking normal nerve conduction and finally resulting in the death of insects. Studies have found that changes in insect AChEs could lead to high resistance to many insecticides. In several insects, such as Drosophila melanogaster, Leptinotarsa decemlineata, Nilaparvata lugens, and Plutella xylostella, a single amino acid mutation could decrease the sensitivity of AChEs to a single insecticide alone or insecticides used in combination. [36][37][38][39]. In our previous study, two AChE genes cloned from P. pseudoannulata exhibited significantly different sensitivities to insecticides, and some key amino acid differences between P. pseudoannulata and insect AChEs might play important roles in insecticide selectivity [35]. In the present study, 17 AChE unigenes were identified from the P. pseudoannulata transcriptome (S5 Table). Among the unigenes, two were perfectly hit to previously reported AChE1 and AChE2, and one unigene showed high similarity to Metaseiulus occidentalis AChE4 (Fig 3). However, several unigenes also showed high similarity to P. pseudoannulata AChE2, such as contig2866, contig4230, contig2350, and unigene15709 (Fig 3). The results indicated that P. pseudoannulata may have more than four AChEs. The number of ace genes in invertebrates varies, e.g., four in nematodes, one or two in different insect species, and one in the spider mite Tetranychus urticae [35]. Further studies are required to determine the number and function of AChEs in P. pseudoannulata.

Nicotinic acetylcholine receptors
nAChRs are ligand-gated ion channels and abundant within the nervous systems of both invertebrates and vertebrates, mediating fast cholinergic synaptic transmission. They are targets of many insecticides, such as neonicotinoids, nereistoxin, and spinosad [4]. In studies of nAChRs,~10-12 nAChR subunits were found in insects. Previous studies showed that residue mutations in a key region of nAChR subunits influenced insecticide sensitivity, and in many cases represented important mechanisms for insecticide resistance [40][41][42][43][44]. Neonicotinoids selectively act on insect nAChRs, but show relative safety for mammals and P. pseudoannulata, partly because of the differences in nAChRs [5]. In our previous study on nAChRs, key amino acid differences between P. pseudoannulata and insect β1 influenced neonicotinoid sensitivity [45]. In the present transcriptome, 17 nAChR unigenes were identified, including six alpha subunits (α2, α3, α6, α7, α8, α9) and two beta subunits (β2, β3) (S6 Table, S5 Fig). The amino acid differences were found in key regions in specific subunits between P. pseudoannulata and insects. For example, differences were found in loops A, B, and C of α2 and α6 subunits between P. pseudoannulata and insects (Fig 4 and S5 Fig).

Ionotropic γ-aminobutyric acid receptor
Insect ionotropic γ-aminobutyric acid (GABA A ) receptor is a chloride channel and an important target for insecticides such as fipronil and cyclodienes, which act as antagonists and bind with higher affinities to insect than to vertebrate ionotropic GABA A receptors [46]. In insects, the GABA A receptor has three subunits: Rdl (resistance to dieldrin), Lcch3 (ligand-gated chloride channel homolog 3), and Grd (the GABA and glycine-like receptor of Drosophila). Studies have found that amino acid substitution in Rdl could confer insecticide resistance; e.g., A302S and A302G resulted in dieldrin resistance in D. melanogaster and D. simulans, A296S substitution was associated with dieldrin resistance in A. gambiae, Anopheles arabiensis, and Aedes aegypti, and an A302N mutation conferred resistance to fipronil in Sogatella furcifera [47][48][49].
In the present transcriptome, 17 GABA receptor unigenes were found, including nine GABA A receptor unigenes and eight GABA B receptor unigenes (S7 Table). Four Rdl genes, three Lcch3 genes, and one Grd gene were identified using phylogenetic analysis and NCBI Nr annotation in GABA A receptor unigenes (Fig 5). Some of the annotated Rdl, Lcch3, and Grd unigenes were manually aligned with GABA A receptor genes from other species and showed very high similarities (S6 Fig). The Rdl gene is attracting significant research interest in the field of insecticide resistance, but it has not been reported in the literature for the natural enemy spider. In the present study, we acquired almost the complete sequence of Rdl for P. pseudoannulata, which could aid in analyzing the differences between insects and natural enemies.

Glutamate-gated chloride channel
GluCls, as members of the cysteine loop ligand-gated ion channel superfamily, are important targets for various insecticides, such as avermectin, ivermectin, and fipronil. To date, GluCls are only found in invertebrates, such as insects, which make them the ideal targets for insecticides [50]. However, insecticide resistance is a major problem in insecticides targeting GluCls. Studies have shown that mutations or alternative splicing in GluCl subunits can reduce insecticide sensitivities [51,52]. In the present study, 12 GluCls unigenes were found with an average length of 955 bp (S8 Table). Phylogenetic analysis and sequence alignment were also employed in related unigenes, and several unigenes showed high similarities with insect, nematode, and acarid GluCls (Fig 6). Notably, the transmembrane regions of unigene42266 were mostly identical to insect GluCls (S7 Fig). The high similarity in GluCls between P. pseudoannulata and insects suggests that P. pseudoannulata may also show high sensitivity to insecticides targeting insect GluCls.

Conclusions
The transcriptome provided significant genetic information on the natural predator P. pseudoannulata. The genes related to insecticide action and detoxification were manually identified, including 90 P450s, 14 GSTs, 17 AChEs, 17 nAChRs, 17 GABA receptors, and 12 GluCl unigenes. The main categories and differences in these genes between insects and P. pseudoannulata were analyzed, which provide useful information for the development of selective insecticides between targeted insect pests and this natural enemy.

Materials and Methods
Sample preparation P. pseudoannulata was collected from a field of hybrid paddy rice in Nanjing (Jiangsu, China). Spiders were narcotized with CO 2 and the legs, chelicerae, and abdomen were manually removed. Samples contained an equal ratio of male and female adult spiders. We confirmed that the location was not privately owned or protected in any way and that the species collections did not involve endangered or protected species.

RNA isolation, library construction, and Illumina sequencing
Total RNA was isolated using TRIzol reagent (Invitrogen/Life Technologies, Paisley, UK). RNA purity and integrity were assessed by the absorbance ratio and agarose gel electrophoresis, and the qualified RNA sample was used for cDNA library construction. mRNA was enriched, fragmented, and used as the template to synthesize the first strand cDNA, after which the second strand cDNA was synthesized. These cDNA fragments were purified and resolved with buffer EB (Elution Buffer) for end reparation and addition of poly(A), then ligated to adaptors. Suitable fragments were selected as templates for PCR amplification to create the cDNA

De novo assembly and transcript annotation
Before assembly, low-quality reads were removed from the raw reads to generate clean reads. De novo transcriptome analysis of the clean reads were assembled using the short-read assembly program Trinity [53].
Finally, all assembled unigenes with a significant cutoff E-value <10 -5 were determined using BLASTx against the NCBI Nr protein, Swiss-Prot, KEGG, and the COG databases. The best matches were used to identify coding regions and to determine the sequence direction. If the alignment results of different databases conflicted with each other, a priority order of NCBI Nr, Swiss-Prot, KEGG, and COG was used. The NCBI Nr database was used to predict the functional annotations of sequences, the GO annotation of unigenes using Blast2GO and WEGO to perform GO functional classification, and to determine the distribution of the gene functions of the species [54]. The KEGG database was used to analyze gene products during metabolism processes and related gene functions in cellular processes [34]. The unigene sequences were also aligned to the COG database to predict and classify functions [55].

Identification and analysis of genes related to insecticide action and detoxification
Genes related to insecticide action and detoxification were identified from the BLAST results generated with the NCBI Nr database with a cutoff E-value of <10 -5 . The unigenes found in the same BLAST results or with high homology to one another were eliminated selectively as allelic variants or as different parts of the same gene [26]. The complete coding region was determined using the open reading frame finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.html), and protein BLAST results were generated. Genes of other species were downloaded from the NCBI database and used as references. MEGA 5.05 software was used to analyze the phylogenetic relationships between genes of interest with the related genes of other species. Phylogenetic trees were generated using the neighbor-joining method and bootstrapped with 1000 iterations to evaluate the branch strength of the tree [56,57].