Lepidopteran HMG-CoA reductase is a potential selective target for pest control

Insects

Larvae of the tobacco hornworm, M. sexta, were raised from eggs provided by Carolina Biological Supply Company (Burlington, NC, USA) and reared on an artificial diet (Bio-Serv, NJ, USA) at 25 °C under a long-day (16 h light/8 h dark) photoperiod (Bell & Joachim, 1976). Pharate 5th instar larvae were set aside 4–7 h before lights off. The larvae molted within a few hours and were designated day 0. At the start of wandering, indicated by the appearance of a prominent dorsal vessel, the larvae were transferred to plastic vials containing vermiculite until pupation. Freshly ecdysed pupae were transferred to a chamber containing a tobacco plant and 10% sucrose under a long-day photoperiod into which the adult moths emerged (Lee, Chamberlin & Horodyski, 2002).

Newly emerged mated female D. punctata (day 0) were isolated from stock cultures. Mating was confirmed by the presence of a spermatophore. Stocks and isolated females were fed Lab Chow and water ad libitum, and were kept at 27 ± 1 °C and 50 ± 5% relative humidity with a 12 h light/12 h dark cycle (Kai et al., 2009).

Worker larvae of A. mellifera were collected from apiaries in Shanghai, China, and placed in an incubator at 34 °C and 80% relative humidity, fed a diet that was prepared with 40% pollen collected from combs and 60% honey. Fourth instar worker larvae were distinguished by the differences in maximum width of their head capsules (Rachinsky, Tobe & Feldlaufer, 2000).

Chemicals

L[¹⁴C-S-methyl] methionine was purchased from Amersham Biosciences (Piscataway, NJ, USA). HMGR inhibitors fluvastatin, pitavastatin and lovastatin and high-performance liquid chromatography (HPLC)-grade isooctane was purchased from Sigma-Aldrich (St. Louis, MO, USA).

Bioassays

Sequence alignment of HMGR

A sequence database of all known HMGR was collected from the literature and GenBank by using a combination of BLAST and keyword searches. Amino acid multiple sequence alignments for HMGR were constructed with ClustalW (Thompson, Higgins & Gibson, 1994) and adjusted by eye to ensure structural motifs were maintained. Poorly aligned regions and major gaps were deleted.

Homology modeling

Because there was no crystal structure of insect HMGR, the homology models of HMGR of M. sexta, A. mellifera, and D. punctata were prepared respectively, to explore the three-dimensional structural differences of the HMGR from different organisms, especially the differences at their active site. A crystal structure of human HMGR in complex with Fluvastatin (PDB ID: 1HWI) was used as the 3D coordinate template for the homology modeling (Istvan & Deisenhofer, 2001). The homology models for HMGRs were generated by using the FUGUE and ORCHESTRAR modules in Sybyl. The initial model was optimized energetically by using the minimize program with steepest descent algorithm, AMBER7 FF99 as the force field and Gasteiger-Huckel as the atomic point charges. The minimization was terminated when the RMS gradient convergence criterion of 0.05 kcal/(mol Å) was reached. The qualities of these models were analyzed by PROCHECK (Laskowski et al., 1993).

Docking calculations

A ligand lovastatin used for the docking studies with HMGRs from different organisms was constructed by using the 2D sketcher module in Sybyl. Minimum energy conformations of all structures were calculated with the Minimize module of Sybyl. The force field was MMFF94 with an 8 Å cutoff for nonbonded interactions, and the atomic point charges were also calculated with MMFF94 (Halgren, 1999). Minimizations were achieved with the steepest descent method for the first 100 steps, followed by the Broyden–Fletcher–Goldfarb–Shanno (BFGS) method until the Root-Mean-Square (RMS) of the gradient became less than 0.005 kcal/(mol Å) (Head & Zerner, 1985; Kai et al., 2006).

The Surflex-Dock (Spitzer & Jain, 2012) module implemented in the Sybyl program was used for the docking studies. The 3-D structures of M. sexta, A. mellifera and D. punctata were performed with homology modeling. Each inhibitor was docked into the binding site of the corresponding protein by an empirical scoring function and a patented search engine in Surflex-Dock applied with the automatic docking. Other parameters were established by default in the software.

Molecular dynamics simulations

Docking calculations as described above were performed for each inhibitor in complex with HMGR in which energy was minimized and used the minimized program in Sybyl-X 2.0 with the optimization algorithm BFGS (Head & Zerner, 1985). The force field was AMBER7 FF99 and the atomic point charge was Gasteiger-Huckel for 500 steps to remove bad contacts (Kai et al., 2006). The system was equilibrated at 400 K for 0.1 ns followed by data collection, at regular intervals, for 10 ns. Each structure collected was subjected to 0.1 ns of simulated annealing to 300 K. The final 100 structures were energy-minimized and clustered using cut-off distance of <0.2 nm. AMBER7 FF99 was used for force field and Gasteiger-Huckel for charges in molecular dynamics simulation using the dynamics program of Sybyl.

Sequence analysis

As the aim of our study was to find an eco-friendly insecticide target, HMGR sequences from different species were aligned relative to that of M. sexta, and the identity values were recorded (Fig. 1). The HMGR sequences of Blattaria, Isoptera, Coleoptera, Hymenoptera, Homoptera and Diptera were significantly different from Lepidoptera, (Dunnett’s multiple comparison test of identity values) (Fig. 1A). The identity values of more distantly related organisms, i.e., Malacostraca, Actinopterygii, Amphibia, Aves, Mammalia, Monocotyledoneae, and Dicotyledoneae were below 60% (Fig. 1B), in comparison with the identity value of the Lepidoptera (approximately 90%). This suggests that the HMGR sequences of Lepidopteran insects are significantly different from other organisms.

Homology modeling, docking and molecular dynamics optimization

With the goal of discovery of an eco-friendly insecticide target, the three-dimensional structures (especially the active site structures) of different species were analyzed. The HMGR of M. sexta, A. mellifera and D. punctata were selected for the structural comparison. The homology models of M. sexta, A. mellifera and D. punctata were generated using the crystal structure of H. sapiens (PDB ID: 1HWI) as the template. To select the best model, we checked the structural validity by PROCHECK (http://services.mbi.ucla.edu/SAVES). The geometry of the final refined models were evaluated with Ramachandran plot calculations computed using the PROCHECK program. The torsion angles of φ and ψ (the two torsion angles of the polypeptide chain, also called Ramachandran angles, describe the rotations of the polypeptide backbone around the bonds between N-Cα called φ and Cα-C called ψ.) in the generated model was represented in the Ramachandran plot as shown in Fig. S1. The Ramachandran plot showed 89.7% of the residues of M. sexta, 88.2% residues of A. mellifera and 90.5% residues of D. punctata existed in the most favored regions. The percentages of residues in disallowed regions of M. sexta, A. mellifera and D. punctata are 0.0%, 0.1% and 0.3%, respectively. This indicated that the backbone dihedral angles, phi and psi, of the three homology models were reasonably accurate.

One HMGR structure was assembled with the same four subunits. Each two adjacent subunits constituted a ligand binding pocket, which means one HMGR contained the same four binding pockets. Lovastatin, a commercial HMGR inhibitor, was used to identify the binding pocket of the aforementioned HMGR structures with docking calculations. Molecular dynamics (MD) simulations of the three homology models complexed with lovastatin as the ligand were performed for 10 ns to obtain the stable and low energy conformations. By reporting the root mean square deviation (RMSD) of the protein structure from the starting model, the receptor changes in structure and reaches a relatively stable conformational minimum after approximately 3 ns. The conformations with the lowest energy of the final 100 structures from the MD simulation were selected as the final structures.

Structure comparison

Both subunits that constituted the binding pocket of A. mellifera and D. punctata formed hydrogen bonds with lovastatin, whereas only one subunit of M. sexta can form hydrogen bonds with the ligand (Arg 579 and Lys 680 of chain A). This suggested that the binding pocket of M. sexta is more flexible.

The surface properties of the binding pocket of the above three structures were defined using MOLCAD calculations (an interactive visualization of molecular scenarios) in Sybyl to analyse these binding pockets. Figure 2 sketches the molecular surfaces of the pockets of these four HMGRs. The cavity of M. sexta was smooth and did not penetrate deep into the structure compared with the cavities of A. mellifera and D. punctata (Fig. 2). This result was in accordance with the results of hydrogen bond interaction. The electrostatic potential of the binding pockets of M. sexta was positively-charged, because the whole surface of the pocket was colored in yellowish green, whereas that of A. mellifera was electroneutral (Fig. 2). The front of the binding pocket of D. punctata was colored with blue, which suggests that some part of this pocket is electronegative. The lipophilic potential of the binding pocket of M. sexta was lipophilic, colored with brown, whereas that pocket of D. punctata was more hydrophilic, colored with blue. The green color suggested that the pocket of A. mellifera is neutral (Fig. 2). These results show that the active pockets of the three species are different, suggesting that it is possible to design eco-friendly insecticides using differences in surface properties. Thus, the insect HMGR may represent a potential eco-friendly insecticide target.

The surface properties of these three pockets suggests that increasing the molecular volume, electronegativity and lipophilicity of the ligand can strengthen the binding affinity between ligand and HMGR of M. sexta, whereas it also can weaken the binding affinities with other species.

Effects of HMGR inhibitors on JH biosynthesis in vitro and in vivo

To validate the results of sequence alignment and structure comparison, three commercial human HMGR inhibitors (statins) were used for the assay of JH biosynthesis in vitro and in vivo in M. sexta, A. mellifera and D. punctata. Lovastatin is a type I statin. Fluvastatin and pitavastatin are type II statins. The IC₅₀ value of each compound is shown in Table 1. For the lepidopteran pest M. sexta, all the compounds have potent inhibitory activity on JH biosynthesis. However, these compounds have little or no effect on A. mellifera, which suggests that HMGR inhibitors tested in the present work are safe for honeybees. Similarly, the inhibitory effects on D. punctata were much lower than on M. sexta. The above results suggest that insect HMGR (in particular, lepidopteran HMGR) might be an eco-friendly insecticide target.

Table 1:

The IC₅₀ values of HMGR inhibition of JH biosynthesis in vitro.

Compound	Structure	M. sexta, IC50 value (nM)	A. mellifera, IC50 value (nM)	D. punctata, IC50 value (nM)
Fluvastatin		5.11	18100	150.0
Lovastatin		99.45	No effect	884.7
Pitavastatin		5.23	157500	395.2

DOI: 10.7717/peerj.2881/table-1

Type II statins have a greater effect than type I for the inhibition of JH biosynthesis in M. sexta. The IC₅₀ value of lovastatin was 99.4 nM, which is much higher than that of fluvastatin or pitavastatin (their IC₅₀ values are 5.1 nM and 5.2 nM, respectively). These results suggest that type II statins should be good lead compounds for new insecticide design. In addition to their effects in vitro, the statins also showed significant effects on JH production by M. sexta following treatment in vivo.

Effects on JH biosynthesis following oral administration

In addition to the effects in the injection bioassays, the statins also showed a significant effect on JH production following oral administration (Fig. 3A). In bioassays at 1 µM, inhibition of JH biosynthesis by fluvastatin, pitavastatin and lovastatin was 58.9 ± 8.9%, 54.4 ± 3.1% and 62.6 ± 3.4%, respectively. This suggests that HMGR inhibitors can also inhibit JH biosynthesis following oral administration.

Pest control application

Larvicidal effects following oral administration of statins

The experiment in ‘Effects on JH biosynthesis following oral administration’ demonstrated that the statins have a significant effect on JH biosynthesis following oral administration. Accordingly, a stomach toxicity test was performed. We first determined which instars of M. sexta were most sensitive to the statins following feeding at high concentration (1,000 µM) in 1st, 2nd and 3rd instars, respectively, and recorded the mortality in Table 3. All larvae that commenced feeding from the 1st stadium died prior to pupation, and most died in the 1st stadium. Larvae fed from the 2nd stadium were also sensitive to the statins with a high mortality (above 85%). However, when the treatment commenced from the 3rd stadium, the mortality before the next molt was less than 10%, and mortality just prior to pupation was less than 40%. For all the inhibitors, the earlier instars were more sensitive than the later instars.

Table 3:

Mortality following feeding with statins at 1,000 µM during the first three stadia.

Compound	Feeding treatment	Mortality before next molt (%)	Mortality before pupation (%)
Fluvastatin	From 1st instar	100	100
	From 2nd instar	51	92
	From 3rd instar	6	13
Pitavastatin	From 1st instar	82	100
	From 2nd instar	80	85
	From 3rd instar	10	40
Lovastatin	From 1st instar	100	100
	From 2nd instar	90	100
	From 3rd instar	20	90

DOI: 10.7717/peerj.2881/table-3

We then topically treated larvae with different concentrations of the statins, commencing with 1st instars, and recorded larval mortality (Table 4). The statins showed significant larvicidal activity at 100 µM. The IC₅₀ values of fluvastatin, pitavastatin and lovastatin were 2,101 µM, 63.0 µM and 298.3 µM, respectively.

Table 4:

Larval mortality following treatment of 1st instars.

Compound	Concentration (µM)	Mortality before 3rd instar (%)	Larval mortality (%)
Fluvastatin	1,000	100	100
	100	25	25
	10	8.3	8.3
	1	16.7	16.7
Pitavastatin	1,000	100	100
	100	100	100
	10	41.7	41.7
	1	25	41.7
Lovastatin	1,000	100	100
	100	33.3	33.3
	10	0	8.3
	1	0	0

DOI: 10.7717/peerj.2881/table-4

In the dead larvae, the most striking characteristic was the darkening of the cuticle in some animals as well as molting disturbances (Fig. 4). This is consistent with the phenomenon of the inhibition of JH biosynthesis (Monger et al., 1982).

Growth regulation

A long-term feeding study with low concentrations of inhibitors was performed to identify their effects on growth. We fed larvae with inhibitors starting with 1st instars, and recorded the number of days from hatching to the 5th stadium (larvae which died before the 5th stadium were not recorded). For fluvastatin, the number of days from newly hatched larvae to the 5th stadium at 100 µM, 10 µM and 1 µM was 17.3, 13.6 and 13.5 days, respectively. In the control group (not fed inhibitors), the interval was 11.9 days. It appears that the HMGR inhibitors significantly slowed the growth rate of M. sexta. Figure 5A shows the difference in growth more clearly. Three larval groups hatched on the same day; one was fed normal food as control, one was treated with 100 µM of fluvastatin, and the other was treated with 1 µM of fluvastatin. The difference in size was readily apparent by comparison with the control 4th instars. The other statins also showed the same growth effect as fluvastatin (Fig.  S2).

Following treatment with a low concentration (1µM) of the statins, we observed that pupation of these larvae was not normal. Most larvae died in the process of pupation as a result of malformation (Fig. 5B). This suggests that the statins can be potent insect growth  regulators.