Structure, Catalysis, and Inhibition of OfChi-h, the Lepidoptera-exclusive Insect Chitinase*

Chitinase-h (Chi-h) is of special interest among insect chitinases due to its exclusive distribution in lepidopteran insects and high sequence identity with bacterial and baculovirus homologs. Here OfChi-h, a Chi-h from Ostrinia furnacalis, was investigated. Crystal structures of both OfChi-h and its complex with chitoheptaose ((GlcN)7) reveal that OfChi-h possesses a long and asymmetric substrate binding cleft, which is a typical characteristics of a processive exo-chitinase. The structural comparison between OfChi-h and its bacterial homolog SmChiA uncovered two phenylalanine-to-tryptophan site variants in OfChi-h at subsites +2 and possibly −7. The F232W/F396W double mutant endowed SmChiA with higher hydrolytic activities toward insoluble substrates, such as insect cuticle, α-chitin, and chitin nanowhisker. An enzymatic assay demonstrated that OfChi-h outperformed OfChtI, an insect endo-chitinase, toward the insoluble substrates, but showed lower activity toward the soluble substrate ethylene glycol chitin. Furthermore, OfChi-h was found to be inhibited by N,N′,N″-trimethylglucosamine-N,N′,N″,N″′-tetraacetylchitotetraose (TMG-(GlcNAc)4), a substrate analog which can be degraded into TMG-(GlcNAc)1–2. Injection of TMG-(GlcNAc)4 into 5th-instar O. furnacalis larvae led to severe defects in pupation. This work provides insights into a molting-indispensable insect chitinase that is phylogenetically closer to bacterial chitinases than insect chitinases.

Insect chitinases belong to glycoside hydrolase family 18 (GH18) 3 and can be classified into 11 groups based on sequence similarity and domain architecture (1,2). Among them, chitinase-h (Chi-h) is noteworthy because its members are only found in lepidopteran insects, one of the most destructive crop pest (3,4). Chi-hs and their bacterial homologs share more than 70% sequence identity, suggesting that a gene horizontal transfer occurred between these two phylogenetic-distant species (4,5).
The physiological role of Chi-h in lepidopteran insects is mostly related to cuticle chitin degradation. During molting and metamorphosis, lepidopteran insects secrete molting fluid, which contains three chitinases (EC 3.2.1.14, group I chitinase (ChtI), group II chitinase (ChtII) and Chi-h), one N-acetyl-Dhexosaminidase (EC 3.2.1.52, Hex), and several kinds of proteases to degrade and shed the old cuticle (6). Chitinases degrade polymeric chitin into chitobiose and chitotriose, which are then further degraded into N-acetyl-D-glucosamine (GlcNAc) by Hex (7). Compared with the extensively studied ChtI (8 -19), there is limited information about the function of ChtII and Chi-h. RNAi of SeChi-h from Spodoptera exigua led to molting deficiency and death indicating that Chi-h is indispensable for molting (17). The spatial and temporal expression patterns of Chi-hs from Bombyx mori (4,5) and S. exigua (17) are similar to that of ChtI but different from ChtII. This suggests that Chi-h and ChtI may work synergistically throughout insect development.
Several crystal structures of GH18 chitinases have been determined from archaea (20), bacteria (21)(22)(23)(24)(25)(26)(27)(28)(29), fungi (30 -34), plants (35)(36)(37)(38)(39)(40)(41), and mammals (42,43). These structures show that although all of the GH18 chitinases use the same catalytic mechanism, they have large discrepancies in the shape of the substrate binding cleft. The crystal structure of OfChtI gave structural evidence that ChtI has a long and open-ended substrate binding cleft with symmetrically distributed subsites that is believed to be a structural characteristic of an endo-acting chitinase (44). According to a structure-based sequence alignment, we found that Chi-h does not contain such a substrate binding cleft but contains a long substrate binding cleft with asymmetrically distributed subsites, a structural characteristic of the processive exo-acting chitinase SmChiA from Serratia marcescens (45). Thus, it is unlikely that Chi-h would be able to act through the same mode of action as ChtI. Revealing the structure of Chi-h will increase our knowledge of why and how lepidopteran insects acquired Chi-h for old cuticle shedding.
In this study, OfChi-h, a Chi-h from the pest Ostrinia furnacalis, was investigated. The crystal structures of OfChi-h and OfChi-h in complex with a substrate analog (GlcN) 7 were obtained and resolved. Through structure-based comparison as well as biochemical characterization, we demonstrate that Chi-h acts synergistically with ChtI to degrade cuticle chitin. Moreover, N,NЈ,NЉ-trimethylglucosamine-N,NЈ,NЉ,NЉЈ-tetraacetylchitotetraose (TMG-(GlcNAc) 4 ), an inhibitor against chitinolytic Hexs (46), inhibits OfChi-h, providing a valuable clue for designing efficient inhibitors. Because Chi-h is absent in most beneficial insects including parasitic wasps and bees, this work will also help develop novel and eco-friendly agrochemicals to protect plants and defend economical loss.

Results
Overall Structure of OfChi-h-The structure of OfChi-h was determined by molecular replacement using the bacterial SmChiA as a search model and was refined to a resolution of 3.2 Å (Table 1). OfChi-h adopts a compact and elongated structure with two domains: domain I (residues 18 -125) and domain II (residues 151-553) (Fig. 1A). According to SCOP (Structural Classification of Proteins) classification (47), domain I is different from domain II. Domain I (fibronectin III domain) is an immunoglobulin-like ␤-sandwich domain comprised of eight ␤-strands. And domain II (catalytic domain) is a (␤/␣) 8 -barrel composed of eight ␤-strands and eight ␣-helices. A chitinase insertion domain (residues 437-509), which consists of five antiparallel ␤-strands flanked by two ␣-helices, is observed in the domain II (48). Domain I and domain II are connected via a 25-amino acid linker (residues 126 -150) and interact with each other via a motif consisting of two antiparallel ␤-strands and one short ␣-helix (residues 34 -51). Two N-glycosylation sites (Asn 391 and Asn 456 ) were observed (Fig. 1A).
One of the most striking features of OfChi-h is a number of aromatic residues lining the groove starting from the far end of domain I and ending at the far end of the substrate binding cleft of domain II (Fig. 1B). They are nine in total, including Trp 27 , Trp 63 , Trp 238 , Trp 225 , Tyr 163 , Trp 160 , Trp 532 , Trp 268 , and Trp 389 . Seven of these aromatic residues are in domain II, but the first two come from domain I. According to the catalytic mechanism (23), the crucial catalytic residues, Asp 304 , Asp 306 , and Glu 308 , are located in the middle of the substrate binding cleft.
Substrate Binding Cleft of OfChi-h-Although our attempts to obtain the structure of OfChi-h complexed to its substrate hexa-N-acetylchitohexaose ((GlcNAc) 6 ) failed, the structure of OfChi-h complexed to chitoheptaose ((GlcN) 7 ), a substrate analog, was obtained by soaking OfChi-h crystals with (GlcN) 7 . The structure was determined by molecular replacement using the unliganded form of OfChi-h as a searching model. The final structure was refined to a resolution of 2.7 Å ( Table 1). The sugar binding subsites were named according to Davies et al. (49), where subsite Ϫn represents the non-reducing end, subsite ϩn represents the reducing end, and the enzymatic cleavage happens between the Ϫ1 and the ϩ1 subsites.
The overall structure of OfChi-h-(GlcN) 7 is very similar to unliganded OfChi-h, with a root mean square deviation (r.m.s.d.) of 0.3 Å. The electron density map supports (GlcN) 7 binds along the substrate binding cleft and occupies the sub- sites from Ϫ5 to ϩ2 (Fig. 2). It is worth noting that the electron density signals from GlcN residues at Ϫ5, Ϫ4, Ϫ3, ϩ1, and ϩ2 are stronger than those at Ϫ2 and Ϫ1. This is different from the electron density map of (GlcN) 5 complexed with OfChtI in that the electron density signals of GlcN residues are strong at subsite Ϫ1 and Ϫ2 but weaker at Ϫ3, Ϫ4, to Ϫ5 (50). (GlcN) 7 binds the substrate binding pocket of OfChi-h in a bent conformation. According to the Cremer-Pople parameter calculation (51), the conformation of the GlcN residues at subsites Ϫ1 and Ϫ2 are 1 S 5 and 4 H 5 , respectively, whereas the conformation of the GlcN residues at the other subsites is 4 C 1 (Fig. 2). The (GlcN) 7 binds OfChi-h mainly via stacking interactions between sugar rings and aromatic residues, specifically Ϫ5 GlcN with Tyr 163 , Ϫ3 GlcN with Trp 160 , Ϫ1 GlcN with Trp 532 , ϩ1 GlcN with Trp 268 , and ϩ2 GlcN with Trp 389 . In addition, polar interactions were also observed between OfChi-h and (GlcN) 7 , including C3-hydroxyl group of the Ϫ1 GlcN with Glu 308 and C6-hydroxyl group of the Ϫ1 GlcN with Arg 439 , respectively (Fig. 2). Enzymatic Activities of OfChi-h-The substrate spectrum of OfChi-h was determined using various insoluble substrates including insect cuticle, ␣-chitin, chitin nanowhisker (CNW) as well as soluble substrate ethylene glycol chitin (EGC). Two chitinases, OfChtI and SmChiA, were chosen to compare with OfChi-h. Because the two tryptophans along the substrate binding cleft of OfChi-h were phenylalanines in SmChiA, the mutant SmChiA-F232W/F396W was thus constructed to test the effects of these site mutations ( Fig. 1B and see Fig. 6A). Among the four substrates, OfChi-h and the other enzymes exhibited the highest hydrolytic activity toward the soluble EGC but lower activities toward insoluble substrates ( Table 2). Table 2, OfChi-h exhibited higher activities toward insoluble substrates than OfChtI but showed lower activities toward EGC than OfChtI. SmChiA-F232W/F396W outperformed SmChiA in hydrolyzing insect cuticle, ␣-chitin, and CNW but showed lower activities toward EGC.

As shown in
The hydrolytic mode of OfChi-h was investigated using (GlcNAc) 6 as the substrate. In addition, the hydrolytic modes of OfChtI, SmChiA, SmChiB, and SmChiC were also investigated for comparison. Because chitin is a ␤-1,4-linked polymer of GlcNAc and GH18 chitinases hydrolyze chitin via a retaining mechanism, ␤-anomeric products will be left after cleavage. SmChiA has been experimentally determined with transmission electron microscopy (52) and high speed atomic force microscopy (45) to be an exo-chitinase that attacks chitin from the reducing end. Supplemental Fig. S2 showed HPLC analysis of ␣and ␤-anomeric hydrolytic products of (GlcNAc) 6 in the presence of OfChi-h, OfChtI, SmChiA, SmChiB, and SmChiC. As shown in Table 3, the percentage of hydrolytic ␤-anomeric products in the presence of OfChi-h was very similar to that by SmChiA but different from those in the presence of any of OfChtI, SmChiB, or SmChiC. Therefore, we deduce OfChi-h perhaps acted in a similar way as did SmChiA.
Because both OfChi-h and OfChtI are present in the insect molting fluid and may act through different hydrolytic modes, the synergistic effect was investigated using insect cuticle chitin as substrate (Fig. 3). The results clearly indicated a synergistic effect between OfChi-h and OfChtI because the activity of the combination of OfChi-h and OfChtI was significantly higher than the specific activity calculated from individual activities during the reaction. The synergistic coefficient between

TABLE 3
Percentage of ␤-anomers after partial hydrolysis of (GlcNAc) 6  OfChi-h and OfChtI was calculated by using the following function (53).

by insect and bacterial chitinases
It is worthy to note that the synergistic coefficient at different time points increased with the reaction time from 1.24 at 2 h to 1.98 at 6 h.
To test the in vivo activity, 0.2 g of TMG-(GlcNAc) 4 was injected into a 5th instar, day-3 O. furnacalis larva. The metamorphosis of the TMG-(GlcNAc) 4 -injected group was severely affected compared with the water-injected group (Fig. 4B). In the control group, 100% of the insects molted into normal pupa 5 days after injection compared with only 40% of the insects from the TMG-(GlcNAc) 4 -injected group. 23% of the insects in the TMG-(GlcNAc) 4 -injected group were arrested during the larva stage, whereas the other 37% molted into abnormal pupa. The abnormal pupa appeared to be prepupa trapped by undetached head capsules and thoracic legs (Fig. 4B). Eventually, most of the abnormal pupa and larvae died within 10 days after injection.

Comparison of OfChi-h with Its Bacterial Homolog Sm-ChiA-Insect
Chi-h is presumed to have been obtained from bacteria as it shares higher sequence identities with bacterial chitinases than insect chitinases (3,4). In this study we found SmChiA from S. marcescens had the highest sequence identity of 73% and the highest similarity of hydrolytic anomeric products profiles with OfChi-h and showed the highest structural similarity with OfChi-h (r.m.s.d. of 1.3 Å for 534 C ␣ atoms). Structure superimposition of OfChi-h and SmChiA (E315L) in complex with (GlcNAc) 8 (PDB code 1EHN) demonstrates that the aromatic residues for chitin binding at subsites Ϫ5, Ϫ3, Ϫ1, and ϩ1 are conserved, except the OfChi-h tryptophans at subsite ϩ2 (Trp 389 ) and the SmChiA-corresponding subsite, Ϫ6 (Trp 225 ), are substituted by phenylalanines (Phe 396 and Phe 232 ) in SmChiA (Fig. 6A). As previously shown, OfChi-h and SmChiA have similar substrate specificity (Table 2) and hydrolytic anomeric products composition (Table 3). Given their similar structural characteristics and enzymatic properties, insect Chi-hs and bacterial ChiAs may act similarly in their respective chitin degradation systems.
The mutation of Phe 232 to Ala in SmChiA has been reported to affect the hydrolytic activity but not the binding activity toward crystalline ␤-chitin. Phe 232 is thought to aid in guiding the chitin chain into the catalytic cleft (54). Similarly, the Phe 396 to Ala mutation in SmChiA was reported to decrease its hydrolytic activity toward crystalline ␤-chitin but increase its hydrolytic activity toward soluble chitosan (55). To explore the effect of the Phe to Trp substitutions in the chitin binding cleft of OfChi-h, Phe 232 and Phe 396 in SmChiA were mutated to tryptophan, and the substrate specificity of SmChiA-F232W/ F396W was tested using insect cuticle, ␣-chitin, CNW, and EGC as substrates. Compared with wild-type SmChiA, SmChiA-F232W/F396W showed higher hydrolytic activity for insoluble and crystalline substrates but lower hydrolytic activity for the soluble substrate (Table 2). Because Trp allows more aromatic interactions with chitin chains (56), we deduce that SmChiA-F232W/F396W may guide chitin chains into the substrate binding cleft more efficiently and may improve binding affinity for chitin. Given that the formation of the complex with the chitin chain is presumed to be the rate-limiting step for SmChiA (57), this may explain why F232W/F396W had a higher activity for insoluble chitin. This result also suggests that the substitution of Phe to Trp in OfChi-h increases its ability to degrade insect cuticles, which are highly insoluble and crystalline.
Structural Differences between OfChi-h and OfChtI-As do key chitinases during molting, we found that OfChi-h and OfChtI work synergistically according to their catalytic efficiency in vitro. Their differences in the architecture of substrate-binding sites were then discussed. Although the substrate binding clefts of both OfChi-h and OfChtI are long with both sides open, they have different structural characteristics. First, in OfChi-h, the distribution of aromatic residues aligned along the substrate binding cleft is highly asymmetric with regard to the enzymatic cleavage site (Fig. 6B). There are 13 solvent-exposed aromatic residues in the nonreducing end side but only two in the reducing end side. However, in OfChtI, the distribution of aromatic residues along the substrate binding cleft is symmetric (Fig. 6B); namely, five aromatic residues on both the non-reducing end side and the reducing end side of the cleavage site. Because oligosaccharide substrates binding to the enzyme rely largely onand/or hydrophobic interactions, these aromatic residues are likely crucial for substrate binding. And the asymmetric architecture is generally believed to be a feature of processive exo-chitinases (45). Second, a unique structural element in OfChi-h, but not in OfChtI (residues 188 -214), was observed on the wall of the substrate binding cleft. This structural element increases the depth of the substrate binding cleft and narrows the substrate binding cleft (the narrowest point between residue Ile 210 and Gln 466 is 6.6 Å) (Fig. 6B). This may further increase the binding affinity of OfChi-h for chitin chains and thus favor OfChi-h to hydrolyze crystalline substrate.
Taken together, we deduce that OfChi-h works synergistically with OfChtI, an endo-chitinase. Because both Chi-h and ChtI are highly conserved in lepidopteran species, this synergistic mechanism is likely generalizable.

Experimental Procedures
Gene Cloning and Construction of the Expression Plasmid-Total RNA was extracted from O. furnacalis during the prepupal state using RNAiso Reagent (TaKaRa, Japan) and was subjected to reverse transcription using the PrimeScript TM RT reagent Kit (TaKaRa). Based on the mRNA sequence of OfChi-h (GenBank TM accession number AB201281.1), two primers, 5Ј-CTGAAGCTTACGTAGAATTCGCGCCCC-CTGGCAAACCC-3Ј (forward) and 5Ј-GTGGTGGTGGTG-GTGGTGACTAGTCGCGCTGTTACCTAGACCCA-3Ј (reverse) were designed to amplify the gene fragment encoding mature OfChi-h and add a C-terminal His 6 tag. The resulting PCR products were digested with EcoRI/SpeI and then ligated into pPIC9 vector (Invitrogen) to generate the expression plasmid pPIC9-OfChi-h.
Protein Expression and Purification-The expression plasmid pPIC9-OfChi-h was linearized by PmeI (New England Biolabs) and transformed into Pichia pastoris GS115 cells by electroporation. Positive clones carrying His ϩ and Mut ϩ traits were selected on minimal methanol and minimal dextrose plates. The selected transformant was first cultured in minimal glycerol-complex medium at 30°C to an A 600 of 2.0. The cells were then collected and resuspended in 1 liter minimal methanol-complex medium and incubated at 30°C. Methanol (1%) was added at 24-h intervals. After 72 h of fermentation, the culture supernatant was harvested by centrifugation at 8000 ϫ g for 10 min.
OfChi-h was purified by ammonium sulfate precipitation and immobilized metal ion affinity chromatography (IMAC). Solid ammonium sulfate was added to the culture supernatant to 75% saturation. After incubation at 4°C for 24 h, the sample was centrifuged at 12,000 ϫ g for 30 min. Then the precipitate was dissolved in buffer A (20 mM sodium phosphate, 0.5 M sodium chloride, pH 7.4) and recentrifuged at 12,000 ϫ g for 15 min. Next, the resulting supernatant was loaded onto a HisTrap TM crude column (5 ml, GE Healthcare) pre-equilibrated with buffer A. Then the column was washed with buffer A containing 75 mM imidazole to remove nonspecific binding proteins. Finally, the recombinant OfChi-h was eluted with buffer A containing 250 mM imidazole. The protein was quantified using a BCA protein assay kit (TaKaRa) with bovine serum albumin as a standard protein, and its purity was analyzed by SDS-PAGE (supplemental Fig. S1). The molecular mass of the recombinant OfChi-h was determined to be 64.6 kDa, which was 4.4 kDa larger than the theoretical moleculare mass. Two N-glycans found in the crystal structure may account for this discrepancy.

JOURNAL OF BIOLOGICAL CHEMISTRY 2085
Enzymatic Assays-Three kinds of polymeric substrates, EGC (Wako Pure Chemicals, Osaka, Japan), CNW (prepared as described in Kuusk et al.;Ref. 57), and ␣-chitin (Sigma), were used as substrates for the chitinase activity assays. The 100-l reaction mixtures consist of 2 M enzyme and 3 mg/ml substrate in 20 mM sodium phosphate buffer, pH 6.0. After incubating at 30°C for an appropriate time, the amount of reducing sugars was determined by the potassium ferricyanide method (58).
The chitin from insect cuticle was prepared as follows: 50 of the 5th-instar day-3 larvae were dissected, and the integuments were collected. The integuments were milled into powder in liquid nitrogen and then washed twice with buffer B (20 mM sodium phosphate, 0.15 M sodium chloride, pH 7.4). To remove minerals and catechols, the powder was treated with 4 M hydrochloric acid at 75°C for 2 h and then rinsed thoroughly with buffer B. Next the powder was treated with 4 M sodium hydroxide for 20 h at 100°C to remove proteins and then rinsed thoroughly with buffer B before being placed in an oven at 60°C for 24 h to dry. At last, the insect chitin was suspended in buffer B to a concentration of 10 mg/ml. To evaluate the enzymatic activity and the synergism of OfChi-h and OfChtI, 3 mg/ml insect chitin was incubated with 2 M OfChi-h, 2 M OfChtI, or a mixture of 2 M OfChi-h and 2 M OfChtI. The reaction mixtures were incubated at 30°C, and 50-l samples were collected at different times to determine the production of reducing sugar.
Analysis of TMG-(GlcNAc) 4 Hydrolytic Products by ESI-TOF MS-Three copies of TMG-(GlcNAc) 4 at 10 M concentration were incubated with 0.1 nM OfChi-h, SmChiA, and SmChiB for 30 min. Then 20-l of hydrolysate was analyzed by ESI-TOF MS using an Agilent G6224A (Agilent) in positive-ion reflection mode.
In Vivo Bioevaluation of TMG-(GlcNAc) 4 by Injection-O. furnacalis larvae were reared using an artificial diet with 16 h of light and 8 h of darkness and a relative humidity of 70 -90% at 26 -28°C. Larvae at day 3 of the fifth instar were selected for the microinjection experiment. In the experimental group, 0.2 g of TMG-(GlcNAc) 4 (solved in water) was injected into the penultimate abdominal segment of larvae. In the control group, distilled water was injected instead. Each group contained 10 individual larvae with three independent replicates. After injection, all of the treated larvae were reared under identical conditions as described above. Mortality and developmental defects were recorded every day until eclosion.
Crystallization and Data Collection-Pure OfChi-h was spin-concentrated to 10 mg/ml in 20 mM bis-Tris (pH 6.5) containing 50 mM NaCl. Crystallization screening of recombinant OfChi-h was performed using the following commercially available screens: Index, Crystal Screen, and Crystal Screen 2 (Hampton Research). The hanging-drop vapor-diffusion crystallization experiments were set up at 4°C by mixing 1 l of OfChi-h and 1 l of reservoir solution. The protein crystallized after 1 month in 100 mM HEPES, pH 7.0, 30% (w/v) Jeffamine ED-2001.
Crystals of OfChi-h-ligand complexes were obtained by transferring native crystals to a reservoir solution consisting of 5 mM (GlcNAc) 6 , 10 mM (GlcN) 7 (Qingdao BZ Oligo Biotech Co., Ltd.), or 1 mM TMG-(GlcNAc) 4 . For (GlcNAc) 6 , the crystals were soaked for 5 min, 15 min, and 1 h at room temperature. For (GlcN) 7 or TMG-(GlcNAc) 4 , the crystals were soaked for 1 h at room temperature. Then the crystals were soaked for several minutes in a reservoir solution containing 25% (v/v) glycerol and subsequently flash-cooled in liquid nitrogen. Diffraction data were collected on the BL-18U1 at the Shanghai Synchrotron Radiation Facility in China, and the diffraction data were processed using the HKL-2000 package (60).
Structure Determination and Refinement-The structure of free OfChi-h was solved by molecular replacement with Phaser (61) using the structure of SmChiA (PDB code: 1EDQ) as a model. OfChi-h-(GlcN) 7 complexes were solved using the coordinates of free OfChi-h as a model. Structure refinement was performed using PHENIX (62). The molecular models were manually built and extended using Coot (63). The stereochemistry of the models was checked by PROCHECK (64). The data collection and structure refinement statistics are summarized in Table 1. The coordinates of OfChi-h and OfChi-h-(GlcN) 7 are deposited in the PDB with the codes 5GPR and 5GQB, respectively. All structural figures were prepared using PyMOL (DeLano Scientific LLC, San Carlos, CA).
Author Contributions-T. L. and Q. Y. designed the experiments. T. L., L. C., X. J., and Y. D. performed the experiments. T. L. and Y. Z analyzed the protein structures. T. L. and Q. Y. analyzed the data and wrote the paper.