Cantharidin Impedes Activity of Glutathione S-Transferase in the Midgut of Helicoverpa armigera Hübner

Previous investigations have implicated glutathione S-transferases (GSTs) as one of the major reasons for insecticide resistance. Therefore, effectiveness of new candidate compounds depends on their ability to inhibit GSTs to prevent metabolic detoxification by insects. Cantharidin, a terpenoid compound of insect origin, has been developed as a bio-pesticide in China, and proves highly toxic to a wide range of insects, especially lepidopteran. In the present study, we test cantharidin as a model compound for its toxicity, effects on the mRNA transcription of a model Helicoverpa armigera glutathione S-transferase gene (HaGST) and also for its putative inhibitory effect on the catalytic activity of GSTs, both in vivo and in vitro in Helicoverpa armigera, employing molecular and biochemical methods. Bioassay results showed that cantharidin was highly toxic to H. armigera. Real-time qPCR showed down-regulation of the HaGST at the mRNA transcript ranging from 2.5 to 12.5 folds while biochemical assays showed in vivo inhibition of GSTs in midgut and in vitro inhibition of rHaGST. Binding of cantharidin to HaGST was rationalized by homology and molecular docking simulations using a model GST (1PN9) as a template structure. Molecular docking simulations also confirmed accurate docking of the cantharidin molecule to the active site of HaGST impeding its catalytic activity.

protein, rHaGST. Homology modeling and molecular docking simulation techniques were employed to rationalize our experimental results.

Bioassay
Bioassay results showed an increase in levels of H. armigera mortality with time after feeding on cantharidin-treated artificial diet. The artificial diet incorporated cantharidin of 250 μg g −1 at 12, 24, 48, 72 h after treatment caused significant larval mortality of 23%, 41% and 69% and 98%, respectively ( Figure 1).

SDS-PAGE Analysis
The positive clones were transformed in BL-21 (DE-3) and protein expression induced by the addition of IPTG was detected initially by SDS-PAGE using standard protein molecular weight marker. The expected band was detected at 27 kDa as the MW of the HaGST is about 24 kDa and the pET-28a tag is about 3 kDa (Figure 2). The expression of recombinant protein was detected by 6× His mouse monoclonal primary antibody on a PVDF membrane.

Specific Activity of HaGST
The inhibitory effects of cantharidin in vivo on GSTs enzyme within midguts dissected from larvae treated with sub-lethal dose of cantharidin is shown in Figure 3. A sub-lethal dose of 25 μg g −1 exerted inhibitory effects on the GSTs in midgut as compared to the untreated control. Data showed that the specific activity of the GSTs tended to decrease in treatment from 24 to 96 h, whereas specific activity tended to increase in untreated controls. The inhibitory effect on the activity at 96 h after treatment was 22.8 compared to 48.32 μM min −1 mg −1 of control.  . Specific activity of the GSTs in larval midgut of H. armigera subjected to sub-lethal dose of cantharidin using glutathione (GSH) and 1-chloro-2,4-dinitrobenzene (CDNB) secondary substrate. The activity of the GSTs was measured at 340 nm both in treated and control samples.

Kinetic Properties of GSTs
The inhibitory effects of cantharidin on the GSTs enzyme extract activity using GSH as substrate is shown in Figure 4. To find out the kind of cantharidin inhibition of the GSTs enzyme extract by cantharidin, the GSTs activity was calculated with variable concentrations of GSH. A Lineweaver-Burk plot with GSH as the variable substrate and for the type of inhibition by cantharidin is shown in Figure 4A. The Lineweaver-Burk plot revealed that cantharidin inhibited GSTs non-competitively as V max lowered down, whereas Km remained unchanged with respect to GSH. . Lineweaver-Burk plot of the GSTs activity in crude enzyme extract. (A) Specific activity of GSTs with and without cantharidin. IC50 value of cantharidin for the GSTs using glutathione (GSH) and 1-chloro-2,4-dinitrobenzene (CDNB) secondary substrate; (B) IC50 value was obtained using a plot of percent activities vs. varying concentrations of cantharidin. S-Hexylglutathione (GTX) was used as positive control.

Kinetic Properties of Purified rHaGST
The inhibitory effects of the cantharidin on the affinity of the column purified soluble rHaGST activity using GSH as substrate is shown in Figure 5. To determine the type of inhibition of the rHaGST by cantharidin, the rHaGST activity was calculated under different concentrations of GSH. The Lineweaver-Burk plot for GSH as the variable substrate and type of inhibition is shown in Figure 5A. The Lineweaver-Burk plot revealed that cantharidin inhibited rHaGST non-competitively as V max lowered, whereas Km remained unchanged with respect to GSH.

IC50 of Cantharidin
To calculate IC50 of cantharidin for the GSTs and the soluble purified rHaGST, variable concentrations of cantharidin were used using GSH as substrate. Results showed that cantharidin inhibited activity of the GSTs and the purified rHaGST in a dose-dependent manner with 50% inhibitory concentration at 9.77 μM ( Figure 4B) and 12.5 μM ( Figure 5B), respectively.

Homology Modeling of HaGST
In order to study the binding of cantharidin with HaGST, a better crystallographic R-factor (20.9%) and higher overall sequence identity (57%) was considered ( Figure 7). We finally selected the complete crystal structure of an insect delta-class glutathione S-transferase from a DDT-resistant strain of the malaria vector Anopheles gambiae in complex with its inhibitor named S-hexylglutathione (GTX) as the template (PDB:1PN9) [22] at a resolution of 2.0 Å. The best model was selected with the lowest value of DOPE assessment score (−27,127.846). The analysis of the Ramachandran plot showed that 97.2% (212/218) of all residues were in favored (98%) regions, and 99.1% (216/218) of all residues were in allowed (>99.8%) regions. Only two amino acids GLU66 and ALA212 were found in the disallowed region of the Ramachandran plot. The resultant 3D structure of the HaGST is shown in Figure 8. The active site is located in a deep cleft formed at the interface of the two domains.  Blue and red colors represent a chain trace from the N-terminus to C-terminus, respectively. α helices and β sheets are represented by H and B, respectively. The active site is located in the cleft formed at the interface of H8, H3 and between loops of H2, B3.

Molecular Docking Simulations
As shown in Figure 9 the final binding mode of the GTX-HaGST was obtained by molecular docking and superimposition (RMSD of only 0.210Å) of the conformations of the GTX in the 1PN9 and HaGST. The O26 atom in the GTX and the NH atom of the HIS52 side chain formed hydrogen bonds having a distance of 2.2 Å between atoms. The NH and the O atoms, which are derived from the VAL54 backbone, interact with the O13 atom and the H8 atom in the GTX by hydrogen bonds respectively, having distances of 1.8 Å and 2.0 Å between them. The NH atom of SER67 backbone also interacts with the O5 atom of GTX with a hydrogen bond distance of 1.9 Å between atoms. The hydrogen bonds are formed between the H2 atom of the GTX and the O atom from GLU66 side chain whose distance is 2.4 Å ( Figure 10). While comparing the schematic representation of residual-ligand hydrogen bond interactions, only one interaction site in insect delta class GST, ILE52 in Anopheles gambiae may have mutated to VAL54 in Helicoverpa armigera. Although the VAL54 and the ILE52 are non-polar hydrophobic amino acids, the NH atom derived from the backbone of ILE52 only interacts with the O13 atom of the GTX. Based on our findings, it could be assumed that the hydrogen bond network is more stable in the GTX-HaGST 3D model complex than those in the crystal structure named 1PN9. Based on the docking work mentioned above, the accurate binding mode of the HaGST 3D model with the cantharidin was obtained as shown in Figure 11.   . Binding mode of the cantharidin-HaGST complex. Red dotted lines show hydrogen bonding between the amino acid residues of the active site and atoms of cantharidin. The O atom of GLY10 and NH atom of ALA12 interact with the cantharidin O3 atom, simultaneously by hydrogen bonding. The OH atom of TRY116 also interacts with the O3 atom of cantharidin. A OH atom of TRY108 forming a hydrogen bond with O1 of cantharidin plays an important role in binding affinity based on its lowest inter-atomic distance. The schematic representation of residue-ligand interactions for cantharidin is shown in Figure 12. The cantharidin docks inside the active site formed by amino acid residues TYR116, PHE120, SER11, ALA8, GLY10, ALA12, ARG15, PRO13, TYR108, LEU35, VAL54. The O atom of GLY10 and the NH atom of ALA12 which are both derived from the backbone of the HaGST 3D model simultaneously interact with the O3 atom of the cantharidin by hydrogen bonds, whose distances are 3.4 Å and 2.1 Å, respectively. The OH atom from the side chain of TYR116 also interacts with the O3 atom of the cantharidin by hydrogen bonds. The hydrogen bonds are formed between the OH atom of TYR108 and the O1 atom of the cantharidin which are only1.7 Å apart. The smaller distance of 1.7 Å between the OH atom of TYR108 and the O1 atom of the cantharidin may have a profound impact on their binding affinity. The hydrogen bond network occupying the active site of HaGST allows the formation of the stable complex with cantharidin.

Binding Energy Calculations
As shown in Table 1, the calculated binding energy delta G (DG) indicates that cantharidin has a higher binding affinity than the GTX. In our current research we have used GST as a molecular target of cantharidin since GSTs plays an important role in defense against plant allelochemicals [6] as well as insecticides [7].
Earlier studies have suggested the increased level of GSTs as one of the major reasons for the development of resistance using metabolic detoxification of insecticides. Many researchers have documented the development of insecticide resistance caused by increased levels of GSTs especially in diamondback moth (DBM) and H. armigera which are phylogenetically related. In one report, indiscriminate use of insecticides, multiple generation of DBM per annum and year round availability of host crops have been mentioned as reasons for development of resistance in this pest to all kinds of insecticides [23,24]. There is a clear correlation between resistance and level of GSTs in insects. Likewise, there is higher activity of GST in the DBM-R population by a 3 to 4 fold increase in enzyme activity over susceptible strains of Plutella xylostella [25]. Similarly, a 1.5 to 2 fold increase in parathione selected DBM was observed [26]. Higher activity of GSTs in pyrethroid resistant strain has established that GST plays an important role as a detoxifying mechanism in DBM [27]. There are earlier reports of resistance in H. armigera to synthetic and nonsynthetic pyrethroid insecticides in China, with resistance to deltamethrin, cyhalothrin, fenpropathrin, esfenvalerate, cyfluthrin and methomyl ranging between 10 and 50-fold [28]. GST could provide passive protection by binding pyrethroid molecules in a sequestration mechanism in resistant field strains of H. armigera [29]. Detoxifying GSTs are a family of enzymes that catalyze the conjugation of glutathione with electrophilic substrates including insecticides [30]. The GSTs are involved in O-dealkylation or dearylation of OPs [31]. High frequencies of profenofos resistance were moderately correlated with GST activity toward 1-chloro-2,4-dinitrobenzene in larvae of H. virescens that were collected in Louisiana cotton fields during the 1995 cotton growing season [32]. A recent study suggests that GSTs act as an antioxidant-defense agent and confer pyrethroid resistance in Nilaparvata lugens and possibly in other insects [33]. Enhanced activities of GSTs that confer insecticide resistance result from both quantitative and qualitative alterations in gene expression. First, there is evidence for over-expression of one or more GST isoforms in resistant insects. For example, the highest activity found in an insecticide-resistant strain of M. domestica is correlated with a high level of GST1 transcript [12].
Our investigations revealed that cantharidin is an effective inhibitor of the GSTs and the HaGST in the midgut of H. armigera with IC50 of 9.77 and 12.5 μM, respectively. The low value of IC50 in vivo suggests that it may be a general inhibitor of GSTs. Enzyme kinetics studies were carried out to detect the inhibitory mechanism of GSTs in enzyme extract and soluble purified rHaGST by cantharidin with respect to GSH. Results revealed that cantharidin was non-competitive inhibitor of the GSTs and the HaGST with respect to GSH, suggesting that cantharidin binding to GSTs may have caused conformational changes consequently leading to the enzyme inactivation. Furthermore, cantharidin has the potential to bind to GSTs through hydrogen a bond causing steric obstruction leading to inactivation of catalytic activity.
Homology modeling and molecular docking simulations were employed to rationalize our results more precisely. Our docking results confirmed putative binding of cantharidin to the catalytic active site residues, ALA12 and TRY108, resultantly causing conformational changes that lead to the inactivation of enzyme catalytic activity.

Insects
Helicoverpa armigera larvae were procured from Henan Jiyuan Baiyun Industry Co., Ltd. China and reared on artificial diet [34] until F1 were available for use in the experiments. Groups of 24 larvae were placed into 24 chamber plastic boxes obtained from the company. The boxes were placed in an incubator at 27 + 1 ○ C and 40% to 50% RH with a 12 h photoperiod.

Bioassay
Diet incorporation bioassay was used to determine the toxicity of cantharidin. Batches of healthy homogeneous third-instar larvae were selected for bioassay. A homogeneous group comprising 24 larvae per replication was subjected to bioassay. The bioassay experiment was replicated thrice. Cantharidin dissolved in acetone was added to the semisolid artificial diet at the rate of 250 μg g −1 and mixed well. Acetone was allowed to evaporate for 1 h before allowing insects to feed. The larvae were starved for 8 h before their introduction to treated diet. One larva of third-instar was introduced to each cell of the 24-cell plastic bioassay tray. The mortality data were recorded at 12, 24, 48 and 72 h.

Larval Treatment
Cantharidin used in the experiment was extracted in the laboratory. The third-instar larvae were selected for this study. The insects were starved for 8 h before their introduction to the cantharidin-treated artificial diet containing 25 μg g −1 cantharidin. Afterward, larvae were collected at an interval of 24, 48, 72 and 96 h. The collected larvae were flash frozen in liquid nitrogen for storage at −80 °C and subsequently used for extraction of enzyme extract and total RNA for synthesis of cDNAs.

Total RNA Preparation and Synthesis of cDNAs Template for Real-Time qPCR
A total of 10 × 3 larvae per time interval (10 larvae per replication), treated as mentioned above were used for extraction of total RNA. At first midguts were dissected from the larvae stored at −80 °C and homogenized using liquid nitrogen before addition of RNAiso Plus (TaKaRa, Dalian, China). RNA was extracted under the guidelines of the manufacturer's instructions. The quality of the RNA samples was examined by running on 1% (w/v) agarose gel. DNA contamination was removed by DNaseI (Fermentas, Beijing, China). The cDNAs were synthesized by reverse transcription using RevertAid™ Reverse Transcriptase (Fermentas, Beijing, China) in a 20 μL reaction containing 5 μL total RNA having 1 μg RNA, oligo (dT) 18

Cloning of Glutathione S-Transferase Gene from H. armigera
Glutathione S-transferase (GenBank accession no. EF033109) was amplified from cDNA by polymerase chain reaction (PCR) using a pair of sense and antisense primers, respectively ( Table 2). The BamHI restriction site was incorporated to sense primer, whereas HindIII restriction site was incorporated to antisense primer for double restriction digestion reaction. The amplification reaction was performed by the PCR program: first step denaturation for 3 min at 95 °C followed by 34 cycles of 95 °C for 30 s, 55 °C 30 s, 72 °C for 1 min and final extension of 5 min at 72 °C. The PCR product was run on 1% (w/v) agarose gel and visualized by ethidium bromide using the BioRad imaging system. Target gene amplified product was gel purified by a gel extraction kit (Biomiga, San Diego, CA, USA). Gel purified PCR product was then ligated to pMD-19T vector (TaKaRa, Dalian, China) and transformed into Escherichia coli DH5α. The transformants were selected on LB agar plates containing 50 μg mL −1 kanamycin after overnight incubation at 37 °C. The resultant PCR clones were sequenced by Shanghai Sunny Biotech, Shanghai, China.

Construction of Recombinant Expression Plasmid
The pMD19T-HaGST was subjected to double restriction digestion by BamHI and HindIII. The digested fragments were gel purified and ligated into the prokaryotic expression vector, pET-28a (Novagen, Darmstadt, Germany) using TaKaRa quick ligation kit to secure the recombinant plasmid, rHaGST. The ligation reaction was transformed into BL-21 (DE-3) competent cells. The transformed Bl-21 cells were cultured in LB media containing kanamycin (50 μg mL −1 ) at 37 °C with 220 rpm shaking until the absorbance at OD 600 reached 0.5 nm, then isopropyl-β-D-thiogalactopyranoside (IPTG) was added to the culture in final concentration of 1 mM at 30 °C for expression of recombinant protein.

SDS-PAGE Analysis of Recombinant Protein and Immunoblotting
The expressed recombinant protein of 29.81 kDa was confirmed and visualized by 12% SDS-PAGE using a standard protein marker (TaKaRa, Dalian, China). After running the recombinant protein on 12% SDS-PAGE, the gel was subjected to Coomassie brilliant blue R250 staining and proteins were transferred to a polyvinylidene fluoride membrane (PVDF). Immunoblotting was done [35] with 6-His monoclonal primary antibody and peroxidase-conjugated goat anti-mouse IgG secondary antibody.

Prokaryotic Expression and Purification of the Soluble Recombinant Protein, rHaGST
The E. coli strain Bl-21 (DE-3) cells with the rHaGST were grown at 37 °C in 50 mL LB media containing 50 μg mL −1 kanamycin until the OD 600 reached 1. The media was then added 0.5 mM IPTG and cells were grown at 28 °C for 6 h with shaking at 200 rpm. The cells were harvested at 10,000 rpm for one min. The resultant pellet was washed with sterile water three times and finally mixed with the binding buffer containing 20 mM imidazol in equilibration buffer. Lysozyme was added to the mixture at the rate of 1 mg mL −1 . After incubation for 30 min on ice, cells were lysed by gentle vortexing. The lysate was subjected to centrifugation at 12,000 rpm for 10 min to remove the cellular debris. The supernatant was passed through a 0.45 nM syringe filter for removal of any remaining debris. The protein extract obtained was passed through a Ni 2+ -nitrilotriacetate (NTA) chromatographic column. The column was washed with equilibration buffer containing, 300 mM NaCl, 50 mM sodium phosphate buffer, 10 mM imidazol and 0.01 M Tris-Cl (pH 8.0). Protein was eluted with a linear imidazole gradient of 50, 100, 150 and 200 mM. The eluted protein was desalted using a dialysis membrane against 50 mM sodium phosphate buffer, pH 7.2 for 24 h at 4 °C.

Enzyme Extract Preparation
A total of 10 × 3 larvae per time interval (10 larvae per replication), treated as mentioned above were used for enzyme extraction. The midguts dissected from the larvae were washed with 1× phosphate buffer and homogenized in 0.1 M potassium phosphate buffer, pH 6.5 on ice using a glass homogenizer. The homogenates were centrifuged at 10,000g for 15 min at 4 ○ C. The supernatants were used as an enzyme extract solution.

GST Activity Determination Assay
Glutathion S-transferase activity was determined by an earlier method with little modifications [36]. Ten microliters of ten-times diluted enzyme extract solution was added to a total volume of 200 μL in microplate wells. The reaction mixture was incubated with 10 mM glutathione at 25 ○ C for 10 min and added 10 μL 10 mM 1-chloro-2,4-dinitrobenzene (CDNB). Enzyme activity was measured spectrophotometrically at 340 nm using a TECAN Infinite ® 200 PRO multimode micro-plate reader. The enzyme activity was determined using the extinction coefficient of 9.6 mM −1 cm −1 for CDNB.

Kinetic Properties of GST
The enzyme extract solution was prepared as mentioned above. Ten microliters of 10 μM cantharidin dissolved in acetone was added to the reaction mixture as an inhibitor. S-Hexylglutathione (GTX) in a concentration of 10 μM was used as a positive control. The reaction was started by the addition of 10 mM glutathione and the reaction mixture was incubated at 25 °C for 10 min and finally added 10 μL of 10 mM CDNB. The enzyme activity was measured spectrophotometrically at 340 nm using a TECAN Infinite ® 200 PRO multimode micro-plate reader. The enzyme activity was determined using the extinction coefficient of 9.6 mM −1 cm −1 for CDNB.

Kinetic Properties of Purified Soluble rHaGST
In this assay the purified rHaGST as mentioned above was used as the enzyme source. Ten microliters of 10× diluted purified GST enzyme solution was added to the reaction mixture. Ten microliters of 10 μM cantharidin dissolved in acetone was added to the reaction mixture as an inhibitor. S-Hexylglutathione (GTX) in a concentration of 10 μM was used as a positive control. The reaction was started by the addition of 10 mM glutathione and the reaction mixture was incubated at 25 °C for 10 min and added 10 μL 10 mM CDNB. Enzyme activity was measured spectrophotometrically at 340 nm using the TECAN Infinite ® 200 PRO multimode micro-plate reader. The enzyme activity was determined using the extinction coefficient of 9.6 mM −1 cm −1 for CDNB. The enzyme kinetics module of the SigmaPlot computer package was used to analyze enzyme kinetics data (SigmaPlot, Systat Software, San Jose, CA. USA).

Determination of IC50
To calculate the concentration of inhibitor that inhibits 50% activity of enzyme (IC50), variable concentration of cantharidin dissolved in acetone were added to the mixture containing 170 μL GST buffer, 10 μL enzyme extract and 10 μL of 10 mM GSH solution at 25 °C. No inhibitor was added to the control. The inhibition reaction was carried out at 25 °C. The value of IC50 was calculated by percent inhibitory activity vs. concentration of inhibitor using Microsoft Excel 2003.

Real-Time qPCR Analysis of Gene Expression
The real-time qPCR was performed by a BioRad iQ 5 cycler. The real-time qPCR reaction was carried out in PCR strips. SYBR Green was used to detect an amplification signal. The reaction mixture consisted of 1 μL 10× diluted cDNA templates, 0.5 μL of 10 μM forward/reverse primers and Maxima SYBR Green/ROX qPCR Master Mix (Fermentas, Beijing, China) in a final volume of 25 μL. Forward and reverse primers were used as mentioned above. The real time PCR condition used: initial denaturing at 95 °C for 30 s., forty cycles of 95 °C for 10 s, 60 °C for 30 s and 72 °C for 30 s. The real time data were acquired at 72 °C. Three replicate for each sample were used for real time PCR analysis. The quantification of the relative transcript levels was performed using the comparative CT method. The expression ratio (R) was calculated as recommended by the manufacturer and corresponds to 2 −ΔΔCT , where: Relative quantification relies on the comparison between expression of a target gene versus a reference gene (Helicoverpa armigera β-actin) and the expression of the same gene in the target sample versus the reference sample [37].

Homology Modeling of HaGST
In this study, we used glutathione S-transferase of Helicoverpa armigera (HaGST) (ABK40535) obtained from the GenBank database [21] having 220 amino acid residues as a model sequence. All Homology Modeling computations were performed using the Modeller 9.10 [38] on a Linux High performance workstation based on 2 Intel Xeon 5680 processors (Super Micro Computer Inc., San Jose, CA, USA).
Using our model HaGST sequence as a probe, the PDB95 database was searched for non-redundant PDB sequences to select the most appropriate template for our query sequence. In order to visualize differences among 12 similar structures selected at 95% sequence identity as the candidate template, we compared all of the structures in the dendrogram, which can generate the weighted pair-group average clustering based on a distance matrix. The best template was picked according to the crystallographic R-factor and overall sequence identity. The sequence alignment between the HaGST sequence and the best template was generated by Align2D and also identified by ClustalX with the Blosum scoring function [39]. The best alignment was selected according to both the alignment score and the reciprocal positions of the conserved residues, especially those in or close to the GTX-binding sites of the template.
Once the target template alignment was constructed, it calculated 100 3D models of the target automatically by the automodel of Modeller using the optimization and refinement protocol. Each model was first optimized with the variable target function method (VTFM) with conjugate gradients (CG), and was then refined using molecular dynamics (MD) with simulated annealing (SA). We also used the loop model class in Modeller to refine the conformation of the loop between residues. To measure the relative stability of the protein conformation, GA341 and Discrete Optimized Protein Energy (DOPE) scores were employed and credible structure of HaGST was selected based on the lowest DOPE energy. MolProbity was used to generate the Protein Main-Chain dihedral Ramachandran map to identify the rationality of the stereochemical for the structure [40].