P. interpunctella were reared on crushed wheat seeds in the laboratory of the Plant Protection Institute, Hebei Academy of Agricultural and Forestry Sciences, at 28 ± 1°C, 60 ± 5% RH and 14:10 L:D photoperiod (Jia et al., 2018). The last-instar larvae were separated and individually reared in glass vials (diameter 2 cm, height 4.5 cm) until their eclosion. The samples from tissues (antennae, thoraces, abdomens, legs, and wings) were prepared following our previous method (Liu et al., 2019). All samples were immediately frozen in liquid nitrogen and stored at −80°C until further RNA extraction. Total RNA extraction, purity evaluation, and concentration determination were performed as previously reported (Jia et al., 2018).

The identification of antennal GSTs from P. interpunctella was mainly based on previously reported transcriptome datasets (accession number: SRR6002827 and SRR6002828) (Jia et al., 2018). The putative GSTs were preliminarily retrieved from annotations based on the latest database, including non-redundant protein (NR), Gene Ontology (GO), Swiss-Prot, and the Kyoto Encyclopedia of Genes and Genomes (KEGG). Subsequently, all candidates were manually validated using the NCBI BLASTx (http://blast.ncbi.nlm.nih.gov/) with an E-value of < 10⁻⁵.

Quantitative real-time PCR tests were conducted on an ABI 7500 (Thermo Fisher Scientific, United States) using Bestar^® SybrGreen qPCR mastermix kit (DBI^® Bioscience) and using the β-actin gene, which was identified from the antennal transcriptome of P. interpunctella, as the reference gene (paired primers: 5′-GTATCAACGGATTTGGTCG-3′ and 5′-CACCTTCCAAGTGAGCAGAT-3′) (Liu et al., 2019). Each reaction was completed in a 20 μL system blend, comprising 10 μL of Bestar SyBr Green qPCR mastermix, 0.2 μM of each primer, 0.4 μL of 50x ROX Reference Dye, 2 μL of cDNA template, and 6.8 μL of RNase-free water at conditions of 1 cycle of 95°C for 2 min, followed by 40 cycles of 95°C for 10 s, 55°C for 34 s, and 72°C for 30 s. Each sample had three independent biological replicates, and each replicate was tested in three technical repeats. All primers are available in Supplementary Table 1. The amplification efficiency for each primer pair ranged from 91.6% to 100.3% based on the standard curve analysis. Relative expression of all GST genes was determined using the comparative 2^−ΔΔCt method (Livak and Schmittgen, 2001). The heatmaps were created by Heatmapper (http://www.heatmapper.ca/) based on the transformed data of log₂ (2^−ΔΔCt + 1) values (Babicki et al., 2016).

The GST sequences were characterized by corresponding bioinformatics tools. GST-ORFs were identified using ORF Finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.html). The sequence lengths, molecular weights (MWs), and isoelectric points (pI) were predicted by using ExPASy tools (https://web.expasy.org/compute_pi/) (Gasteiger et al., 2005). The conserved domains were predicted by using the hmmsearch tool from the pfam website (http://pfam.xfam.org/) (Mistry et al., 2021). Identification of conserved motifs of GSTs was conducted with the MEME online program for protein sequence (http://meme.nbcr.net/meme/intro.html) (Bailey et al., 2009) with the optimized parameters being any number of repetitions, a maximum number of 10 motifs, and optimum 6–50 residue length per motif.

Deduced amino acid sequences of GST genes from different insects were aligned with the GST sequence identified from the antennae of P. interpunctella using ClustalW with default parameters (https://www.genome.jp/tools-bin/clustalw). After sequence alignments, the phylogenetic tree was constructed by MEGA5.0 software using the neighbor-joining method with the following parameters: Poisson model, pairwise deletion, and 1,000 bootstrap replications (Tamura et al., 2011). The dendrogram was further decorated using Evolview software (https://www.evolgenius.info/evolview/). The homologous GST sequences were used to reconstruct a phylogenetic tree from eight species, including Plutella xylostella (You et al., 2015), Cydia pomonella (Huang et al., 2017), Bombyx mori (Yu et al., 2008), Chilo suppressalis (Liu et al., 2015), Acyrthosiphon pisum (Francis et al., 2001), Drosophila melanogaster (Younus et al., 2014), Anopheles gambiae (Ding et al., 2003), and Tribolium castaneum (Shi et al., 2012). All sequences were obtained from NCBI (https://www.ncbi.nlm.nih.gov/).

The homology model was constructed by the SWISS-MODEL server (https://swissmodel.expasy.org/interactive). The models of PiGSTd1 were built based on the target-template alignment using ProMod3 (Guex et al., 2009). The QMEAN scoring function was used to assess the global and per-residue model quality (Studer et al., 2020). Then, an automated model BmGSTd1 (PDB ID: 4e8e.1) was selected as the template from PDB database. Pictures of three-dimensional structures were generated with PyMOL (DeLano, 2002). Multisequence alignments were performed using ClustalX 2.1, and the results were presented by GeneDoc software (http://nrbsc.org/gfx/genedoc). The secondary structure was predicted with PSIPRED software (McGuffin et al., 2000).

The PiGSTd1 sequence without signal peptide was amplified by PCR using TransStart^® FastPfu PCR SuperMix (TransGen Biotech, China). The paired primers were forward 5′-ATGCCGGCTCAAGCCATCAA-3′ and reverse 5′-CTAATCTTTCTTCAGAAATGATGC-3′. The amplification was carried out under the conditions of denaturation at 95°C for 1 min followed by 35 cycles of 95°C for 20 s, 55°C for 20 s, and 72 °C for 1 min, and a final extension at 72°C for 5 min. The PCR products were ligated into a pEASY-Blunt E1 Expression vector (TransGen Biotech, China) and transformed into Escherichia coli Trans-T1 (Liu et al., 2017). After sequence confirmation by Sangon Biotechnology (Shanghai, China), the positive recombinant plasmids were designated as pEASY-Blunt E1-PiGSTd1.

PiGSTd1 expression and purification were conducted as previously described with a slight modification (Song et al., 2020). Briefly, the recombinant vector (pEASY-Blunt E1-PiGSTd1) was transformed into E. coli BL21 (DE3), and the positive clones were isolated for expression. Cultures were started from single colonies, in LB broth with 50 μg/mL ampicillin in a 37°C shaker (220 rpm). When OD of 600 nm reached 0.6, isopropyl β-D-1-thiogalactopyranoside (IPTG) was added to 1 mM. After cultured for 6 h at 25°C and 220 rpm, cells were harvested by centrifugation at 8,000 g at 4°C and suspended in 20 ml of PBS buffer (pH = 7.0).

After ultrasonic cell disintegration, the collected bacteria were centrifuged at 14,000 rpm at 4°C for 20 min. After confirming the expression by 12% SDS–PAGE, the supernatants were loaded on a Ni-chelating affinity column (GE, United States), which had been equilibrated with 20 mM Tris–HCl (pH = 7.9) supplemented with 100 mM NaCl, and eluted with imidazole (50, 100, 150, and 200 mM) in an ascending series. The recombinant PiGSTd1 purity was assayed by SDS–PAGE. Its concentration was determined using Bradford's method with BCA Protein Assay Kits (Legend biotech, Beijing, China). Proteins were stored at −20°C before use.

The kinetic parameter of PiGSTd1 was determined based on the CDNB (1-chloro-2,4-dinitrobenzene) method (Li et al., 2018). Briefly, 0.4 μg of the purified PiGSTd1 was added into 200 μL acetate–phosphate buffer (pH 7.5) containing 50 mM GSH and a series of CDNB (0, 0.2, 0.4, 0.6, 0.8, 1.0, 1.2, and 1.4 mM) at 35°C in a transparent 96-well plates, and the absorbance at 340 nm in 0–1 min was recorded in a Multiskan Spectrum Microplate Spectrophotometer (BioTek, Shoreline, WA). Heat-inactivated PiGSTd1 was used as the negative control. The Km and Vmax were calculated by the linear regression of a double reciprocal plot (Balakrishnan et al., 2018). To optimize the reaction pH and temperature of PiGSTd1, the assays were conducted at fixed concentrations of GSH (1 mM) and CDNB (0.5 mM) with varying acetate–phosphate buffer (pH = 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5 and 9.0) and reaction temperature (20, 25, 30, 35, 40, 50, 55, and 60°C for 30 min). All determinations were performed three times.

A GC-MS (7890A-5975C; Agilent, United States) with a DB-WAX column (30m × 0.25mm × 0.25μm, Agilent) was used to evaluate the degradation activities of PiGSTd1 on the main sex pheromone and environmental volatiles (Supplementary Table 2). The degradation assays were conducted in 1 mL acetate–phosphate buffer (pH = 7.0) containing 2.5 μg purified PiGSTd1, 10 mM GSH, and 20 μg substrates. After reacting for 1 h at 35°C, the reaction mixture was extracted with 1 mL hexane immediately. Subsequently, substrates in the organic phase were qualitatively and quantitatively analyzed on the GC-MS with the chromatographic conditions setting as helium carrier gas at 1 ml·min⁻¹; oven temperature initiated at 50°C (hold 1 min), increased to 120°C at 5°C·min⁻¹ (hold 2 min), and subsequently increased to 230°C at 10°C·min⁻¹ (hold 5 min). The ionization current and ionization voltage were 100 μA and 70 eV, respectively. All assays were repeated three times with the heat-inactivated PiGSTd1 as the negative control. Degradation data were analyzed by one-way ANOVA (SPSS 19.0 for Windows) with Tukey's test. The least significant difference was set at P < 0.05.

From the antennal transcriptome of P. interpunctella, we identified a total of 17 sequences encoding putative GSTs, which were designated as PiGSTd1-PiGSTm3. Sequence characteristics (ORFs, MW, and pI) and Blastx results are listed in Table 1. Among all PiGSTs, 15 sequences were intact ORFs, while PiGSTo4 and PiGSTd2 were incomplete with truncated 3′-regions. The sequence lengths of the PiGSTs ranged from 149 to 290 amino acid (aa), and their calculated MWs ranged from 16.35 to 33.15 kDa. BLAST_X search of the best hits showed that all PiGST sequences shared relatively high sequence identities (62−97%) with their respective orthologs from other lepidopteran species (Table 1).

The phylogenetic tree was reconstructed with 169 GST sequences from nine species, including model insects (e.g., B. mori and D. melanogaster), typical species in varying families, as well as congeneric Pyraloid moths. Although these GSTs were derived from diverse species, they showed relative conservation in classification. According to their sequence similarities, 17 PiGSTs were distributed into eight branches of the phylogenetic tree: delta (PiGSTd1 to PiGSTd3), epsilon (PiGSTe1 and PiGSTe2), omega (PiGSTo2 to PiGSTo4), sigma (PiGSTs1 and PiGSTs2), theta (PiGSTt1), zeta (PiGSTz1), and unclassified class (PiGSTm1) (Figure 1). PiGSTd1 was clustered with CpGSTd2, a well-characterized enzyme involved in odorant degradation for chemosensory perception in C. pomonella (Huang et al., 2017), indicating it could potentially degrade odorants.

The analyses of conserved domains among the PiGSTs revealed two domains of the protein sequences: a fairly conserved N-terminal domain and a more variable C-terminal domain among different subclasses (Supplementary Figure 1). Besides, a member of conserved membrane-associated proteins was identified in eicosanoid and glutathione metabolism (MAPGE) from microsomal GSTs (Supplementary Figure 1). A schematic representing the structure of all complete PiGSTs sequences was constructed from the MEME motif analysis results. PiGSTs in the same subclass usually shared a similar motif composition and showed highly similar motif distributions, e.g., the clustered PiGST pairs, PiGSTs1-2 and PiGSTe12 (Figure 2). Among all motifs, motif 3 and motif 4 were found in all cytosolic GST proteins, while motif 2, motif 6, and motif 9 were exclusively expressed in microsomal GSTs (PiGSTm1-3).

Based on qRT-PCR determination, only PiGSTd1 expression was antennae-specific, and its expression level was significantly higher in males than in females (Figure 3), indicating that it has a close association with odorant recognition. In contrast, PiGSTe2 was almost equally expressed in female and male antennae and was also found in the abdomens, but it was not antennae-enriched. PiGSTo2, PiGSTo3, PiGSTm1, PiGSTm2, PiGSTm3, PiGSTs1, PiGSTs2, PiGSTz1, PiGSTe1, PiGSTd3, and PiGSTt1 were abundantly expressed in the abdomens of both sexes (Figure 3). Other GST genes were ubiquitously expressed in all tested tissues.

According to a multiple alignment of PiGSTd1 with delta GSTs from other moths, PiGSTd1 showed relatively high identities (63.11–68.83%) with HvGSTd1 (AWX68884.1), OfGSTd1 (QIC35737.1), CsGSTd1 (AKS40338.1), SeGSTd1 (ASN63930.1), BmGSTd1 (NP_001037183.1), and PrGSTd1 (APW77568.1) (Supplementary Figure 3), indicating high conservation between PiGSTd1 and moth delta GSTs. Additionally, the multiple alignments and homology modeling on the basis of BmGSTd1 suggested that PiGSTd1 adopted the classic GST fold and was composed of an N-terminal domain, a C-terminal domain, and a linker in between (Figure 4). In the conserved N-terminal domain, a three α-helices and four β-strands motif (βαβαββα) of thioredoxin fold-served as the glutathione binding site (G-site). Ser40 in PiGSTd1 appeared to be responsible for enzyme catalysis.

The entire coding sequence of PiGSTd1 was successfully expressed in E. coli strain BL21 through pEASY-Blunt E1 vector. SDS–PAGE showed that Ni⁺-column-purified PiGSTd1 displayed a single band with a MW of ~27kDa (Figure 5A). Using CDNB and reduced GSH as substrates, the optimized catalytic conditions for PiGSTd1 were 35 °C and pH=7.0 (Figures 5B,C). Under these conditions, K_m and V_max of recombinant PiGSTd1 were determined as 0.2292 ± 0.01805mM and 14.02±0.2545 μmol·mg⁻¹·min⁻¹, respectively (Figure 5D).

The ability of recombinant PiGSTd1 to degrade odorants was evaluated by GC-MS. The results showed that PiGSTd1 more efficiently degraded the sex pheromone component Z9-12:Ac (75.63 ± 5.52%) as compared with the pheromone analog Z8-12:Ac (58.47 ± 1.64%), despite only a differently positioned double bond. Besides sex pheromones, PiGSTd1 also displayed high efficiency in degrading various host odorants and environmental volatiles (Supplementary Table 2), e.g., α-pinene (68.83 ± 2.37%), hexanal (89.10 ± 2.21%), heptanal (63.19 ± 5.36%), (E)-2-octenal (73.58 ± 3.92%), (E)-2-nonenal (75.81 ± 1.90%), and (E)-2-decenal (61.13 ± 5.24%). The results indicated that PiGSTd1 highly expressed in P. interpunctella antennae was involved in degrading sex pheromones and host volatiles.