Pyrethroid Carboxylesterase PytH from Sphingobium faniae JZ-2: Structure and Catalytic Mechanism

Pyrethroid pesticides are widely applied in agriculture and household; however, extensive use of these pesticides also causes serious environmental and health problems. The hydrolysis of pyrethroids by carboxylesterases is the major pathway of microbial degradation of pyrethroids, but the structure of carboxylesterases and its catalytic mechanism are still unknown. Carboxylesterase PytH from Sphingobium faniae JZ-2 could effectively hydrolyze a wide range of pyrethroid pesticides. The crystal structures of PytH are solved in this study. This showed that PytH belongs to the α/β-hydrolase fold proteins with typical catalytic Ser-His-Asp triad, though PytH has a low sequence identity (about 20%) with them. The special large hydrophobic binding pocket enabled PytH to bind bigger pyrethroid family substrates. Our structures shed light on the substrate selectivity and the future application of PytH and deepen our understanding of α/β-hydrolase members.

ABSTRACT

Carboxylesterase PytH, isolated from the pyrethroid-degrading bacterium Sphingobium faniae JZ-2, could rapidly hydrolyze the ester bond of a wide range of pyrethroid pesticides, including permethrin, fenpropathrin, cypermethrin, fenvalerate, deltamethrin, cyhalothrin, and bifenthrin. To elucidate the catalytic mechanism of PytH, we report here the crystal structures of PytH with bifenthrin (BIF) and phenylmethylsulfonyl fluoride (PMSF) and two PytH mutants. Though PytH shares low sequence identity with reported α/β-hydrolase fold proteins, the typical triad catalytic center with Ser-His-Asp triad (Ser78, His230, and Asp202) is present and vital for the hydrolase activity. However, no contact was found between Ser78 and His230 in the structures we solved, which may be due to the fact that the PytH structures we determined are in their inactive or low-activity forms. The structure of PytH is composed of a core domain and a lid domain; some hydrophobic amino acid residues surrounding the substrate from both domains form a deeper and wider hydrophobic pocket than its homologous structures. This indicates that the larger hydrophobic pocket makes PytH fit for its larger substrate binding; both lid and core domains are involved in substrate binding, and the lid domain-induced core domain movement may make the active center correctly positioned with substrates.

IMPORTANCE Pyrethroid pesticides are widely applied in agriculture and household; however, extensive use of these pesticides also causes serious environmental and health problems. The hydrolysis of pyrethroids by carboxylesterases is the major pathway of microbial degradation of pyrethroids, but the structure of carboxylesterases and its catalytic mechanism are still unknown. Carboxylesterase PytH from Sphingobium faniae JZ-2 could effectively hydrolyze a wide range of pyrethroid pesticides. The crystal structures of PytH are solved in this study. This showed that PytH belongs to the α/β-hydrolase fold proteins with typical catalytic Ser-His-Asp triad, though PytH has a low sequence identity (about 20%) with them. The special large hydrophobic binding pocket enabled PytH to bind bigger pyrethroid family substrates. Our structures shed light on the substrate selectivity and the future application of PytH and deepen our understanding of α/β-hydrolase members.

INTRODUCTION

Pyrethroids are ester-containing synthetic analogs of pyrethrins. They generally have a broad-spectrum of high insecticidal activity and relatively low mammalian toxicity, avian toxicity, and phytotoxicity (1). Pyrethroid pesticides have been widely used for more than 30 years in agricultural, household, and livestock applications (2). The most commonly used pyrethroid pesticides are permethrin, fenpropathrin, cypermethrin, cyhalothrin, fenvalerate, deltamethrin, and bifenthrin (BIF) (Fig. 1). Pyrethroid pesticides currently account for approximately 30% of the worldwide pesticide market, and their usage is continuing to grow (3). However, extensive application of pyrethroid pesticides results in serious residues in agricultural products and environment (4). Although pyrethroids are less toxic than organophosphate and organochlorine pesticides, prolonged exposure to pyrethroid pesticides might disrupt endocrine, damage the lymph nodes and spleen, and cause carcinogenesis (5, 6). Furthermore, pyrethroid pesticides are very toxic to some nontarget organisms, such as bees, fish, and aquatic invertebrates. It is therefore of significance to develop efficient strategies to remove the pyrethroid pesticide residues.

FIG 1.

Molecular structures of commonly used pyrethroids. The labile bonds are labeled with arrows.

Previous studies showed that pyrethroids could be degraded or detoxified by microorganisms and pyrethroid-resistant insects through oxidation by P450 monooxygenases, conjugation with glutathione S-transferases, and hydrolysis by phosphotriesterases or carboxylesterases (7, 8). Carboxylesterases are a family of enzymes that are important in the hydrolysis of a large number of endogenous and xenobiotic ester-containing compounds. Thus far, several pyrethroid-degrading carboxylesterases have been purified and characterized, i.e., pyrethroid-degrading carboxylesterase (PytH) from Sphingobium strains JZ-1 and JZ-2 (9, 10), permethrin carboxylesterase from Bacillus cereus SM3 (11), EstP from Klebsiella sp. strain ZD112, pyrethroid carboxylesterase from Aspergillus niger ZD11 (12), Aes from mouse liver microsomes (1), thermostable pyrethroid carboxylesterase Sys410 from the Turban Basin metagenomic library (13), and PtY from Ochrobactrum anthropi YZ-1 (14).

Carboxylesterase PytH is a monomer with a molecular mass of approximately 31 kDa and can hydrolyze a wide range of pyrethroid pesticides with hydrolysis rates descending as follows: permethrin (specific activity, 5.82 μmol/min/mg), fenpropathrin (5.02 μmol/min/mg), cypermethrin (4.86 μmol/min/mg), cyhalothrin (2.46 μmol/min/mg), fenvalerate (1.75 μmol/min/mg), deltamethrin (1.52 μmol/min/mg), and BIF (0.84 μmol/min/mg) (9). Although the molecular weight of PytH is smaller than other reported pyrethroid-hydrolyzing carboxylesterases, it possesses a relatively higher catalytic efficiency than those of previously reported pyrethroid carboxylesterases (9, 10). Furthermore, pyrethroids usually contain one to three chiral centers, and therefore they consist of two to eight isomers. Previous studies showed that some pyrethroid-hydrolyzing carboxylesterases exhibit significant isomer selectivity (1, 15); however, no isomer selectivity was found for PytH (9). These characteristics make PytH a good candidate for eliminating pyrethroid residues in the environment and in agricultural products.

Sequence alignment reveals that PytH shares no similarity with other reported pyrethroid carboxylesterases but shares 20 to 24% sequence identity to some α/β-hydrolase fold proteins in some species in all domains of life, and PytH does contain the typical catalytic Ser-His-Asp/Glu triad (Ser78, Asp202, and His230) and the highly conserved pentapeptide motif GXSXG of α/β-hydrolase fold proteins (9). However, known structures of α/β-hydrolases are not enough for understanding the enzymatic characteristics of this novel enzyme due to their low sequence similarity with PytH and substrate specificity. Furthermore, understanding the catalytic mechanisms of the key enzymes in pyrethroid degradation pathway is a critical prerequisite for successful bioremediation of pyrethroid residues. In this study, we report the crystal structures of BIF and phenylmethylsulfonyl fluoride (PMSF) binding PytH and two single-site mutants (S78A and H230A). These structures shed light on substrate selectivity, pyrethroid degradation mechanism, and structural relationship to other α/β-hydrolase members.

RESULTS

Overall structure of PytH.

Recombinant PytH from Sphingobium faniae JZ-2 was overexpressed in Escherichia coli and purified to homogeneity (see Fig. S1 in the supplemental material). The purified recombinant PytH was then crystallized in the space group P4₂2₁2. The 1.9-Å resolution crystal structure of BIF-PytH was determined by single wavelength anomalous dispersion (SAD) using selenomethionine-derived protein with a final R_work of 27.3% and an R_free of 32.2%. A summary of data collection and model refinement statistics is presented in Table 1. The relatively poor electron density map for molecules E and F led to the relatively high R factor for the overall structure.

TABLE 1.

Crystallographic data collection and model refinement statistics

Parameter	Se-Met	PytH_PMS	PytH_BIF	PytH_H230A	PytH_S78A
PDB code		5Y57	5Y5R	5Y51	5Y5V
Data collection	5/31/2013	7/12/2013	3/15/2013	10/15/2013	1/1/2015
Wavelength (Å)	0.9791	0.9792	0.9793	0.9793	0.9791
Temp (K)	100	100	100	100	100
Crystal-to-detector distance (mm)	300	300	250	300	400
Rotation range/image (°)	1.0	1.0	1.0	1.0	0.5
Total rotation range (°)	720	150	360	90	360
Space group	P42212	P42212	P42212	P42212	P42212
Cell dimensions
a, b, c (Å)	169.02, 169.02, 123.53	168.2, 168.2, 123.4	168.86, 168.86, 123.86	168.38, 168.38, 123.86	168.20, 168.20, 123.63
α, β, γ (°)	90	90	90	90	90
Resolution range (Å)	20.00–1.96 (2.07–1.96)	20.0–1.96 (2.07–1.96)	19.92–1.90 (2.00–1.90)	20.0–2.30 (2.42–2.30)	20.0–2.98 (3.14–2.98)
No. of observed reflections	7,152,139 (682,781)	1,467,707 (145,072)	4,159,477 (602,839)	580,253 (85,039)	358,416 (52,344)
No. of unique reflections	125,139 (16,305)	125,394 (16,979)	140,256 (20,268)	75,180 (11,027)	36,638 (5,259)
Multiplicity	57.2 (41.9)	11.7 (8.5)	29.7 (29.7)	7.7 (7.7)	9.8 (10.0)
Rpim (%)a	3.7 (28.6)	6.3 (69.9)	5.5 (41.8)	6.3 (37.4)	9.3 (33.2)
Completeness (%)	98.3 (88.9)	98.9 (92.9)	99.9 (100)	95.4 (97.0)	99.7 (100.0)
I/δ〈I〉	12.0 (2.6)	9.2 (1.2)	9.1 (2.1)	8.0 (1.9)	7.1 (2.2)
Structure refinement
Total no. of atoms		11,971	12,042	11,410	11,233
No. of reflections used		125,199	139,583	74,976	36,595
Rwork (%)		25.5	27.3	27.0	22.4
Rfree (%)		30.9	32.2	32.6	28.8
RMSD bond length (Å)		0.0093	0.011	0.0112	0.010
RMSD angles (°)		1.144	1.461	1.263	1.368
Ramachandran plot (%)
Favored		94.4	97.2	94.8	92.5
Allowed		5.6	2.8	5.2	7.3
Outlier		0	0	0.0	0.2

^aPrecision-indicating merging R factor, i.e., R_pim = ∑hkl[1/(N − 1)]^1/2 ∑i|Ii(hkl) − [I(hkl)]|/∑hkl ∑[I(hkl)], where Ii(hkl) and I(hkl) are the observed intensity of measurement i and the mean intensity of the reflection with indices hkl, respectively, and N represents increased redundancy of the measurement.

In the final model, there are six molecules in each asymmetric unit (Table 1). However, the results of native polyacrylamide gel electrophoresis mobility experiments and gel filtration chromatography indicate that the purified PytH forms monomer and dimer in solution (Fig. S2). In addition, only minimal contacts are found among the different monomers, indicating that the higher order of oligomerization of PytH in the crystal structure is due to crystal packing. This result is a little bit different from the previous characterization result of PytH, in which the authors found that PytH is a monomer in solution (9, 10). This difference may be due to the different concentrations of the samples. The sample concentration used for crystallization is much higher than that used for activity assay, which may cause the transition from the monomeric state to the dimeric state. The structure of the best-defined molecule B is used for the subsequent analysis. This model contains residues 2 to 254, while the C-terminal region from residues 255 to 280 is not resolved. This missing fragment is away from the active site and substrate; therefore, it may not be involved in the catalytic reaction. The sequence alignment also showed that the missing fragment is a specific fragment of PytH and is not present in other paralogues, which suggests that this fragment may be not essential for its catalytic function (16).

The overall structure of PytH unambiguously confirms that PytH belongs to the well-known α/β-hydrolase superfamily. The structure of PytH is comprised of a core domain and a lid domain (Fig. 2A). The core domain of PytH, including residues 2 to 104 and 176 to 254, consists of six parallel β-strands (β1 to β6) in the central β-pleated sheet flanked by seven α helices (αA to αG) and two additional 3₁₀ helices (η1 and η2) situated on both sides (Fig. 2A). A lid domain, residues 105 to 175, is inserted in the loop connecting β4 and αD and is comprised of two-stranded antiparallel β-strands (Lβ41 and Lβ42) and three α-helices (LαD1 to LαD3) and a short 3₁₀ helix (Lη1) (Fig. 2A). However, unlike other α/β-hydrolases, PytH has a novel function to hydrolyze pyrethroids, and strikingly, a deep open channel with two entrances on the surface of PytH could be easily observed (Fig. 2B and C).

FIG 2.

Overall structure of PytH. (A) Overall structure of PytH. Helices, β-strands, and coils in the lid domain are depicted in red, yellow, and green, respectively. Helices, β-strands, and coils in the core domain are cyan, magenta, and orange, respectively. (B and C) Electrostatic potential surface of the PytH reveals two entrances in PytH (the rotation around the x axis is 33°, the rotation around the y axis is −92°, and the rotation around the z axis is −46°). The lid domain is depicted in light blue, and the core domain is depicted in pale green. The carbon skeleton of the substrate is shown in yellow.

A crystal structure comparison of PMSF-PytH and BIF-PytH was also performed. The structural comparison revealed that they are almost identical, with a root mean square deviation (RMSD) of 0.216 Å, but some differences between these two structures were still observed, such as the C terminus (Fig. 3A and B).

FIG 3.

Active sites of PytH. (A) Active sites of PytH with BIF (A) and PMSF (B) in the overall structures. The lid domain and the core domain are depicted in salmon and cyan, respectively. The residues of the active sites (S78, D202, and H230) are displayed as sticks with green carbon atoms. (C and D) Active sites of BIF-PytH (C) and PMSF-PytH (D). The residues of the active sites (S78, D202, and H230) are shown as sticks with purple carbon atoms. BIF (C) and PMSF (D) are shown as sticks with carbon atoms colored yellow. D202 forms interactions with H230 through the hydrogen bond, BIF forms interactions with S78 through hydrogen bonds, and PMSF binds to S78 by a covalent bond directly. The corresponding electron density (2Fo_Fc) is light gray and contoured at 1.0 σ. (E) Ala12 and Leu79 are important for enzyme substrate structure stabilization. The residues of the active sites (S78, D202, and H230), Ala12 and Leu79, are shown as sticks with purple carbon atoms. (F) Active site comparison of wild-type PytH (purple) and S78A (light orange) and H230A (cyan). No interaction could be formed among S78A and H230A.

Active site of PytH.

The α/β-hydrolase typically uses a catalytic triad Ser-His-Asp, and the catalytic residues typically locate in highly conserved regions of the protein core (17–21). In our BIF-PytH and PMSF-PytH complex structures, PytH shows a similar catalytic triad consisting of Ser78, His230, and Asp202 at its active site (Fig. 3A to D), which is located at the C-terminal end of the parallel sheet in the core domain for hydrolyzing the ester bond of the substrates (Fig. 3A and B). A 3₁₀-helix Lη1, LαD3, and the long loop (residues 158 to 170) between them, LαD2, Lβ41, and Lβ42 in the cap domain (Fig. 2A) cover the exposed side of the active site. The key nucleophile Ser78 is canonically located just after the β3 strand projecting from a sharp turn termed the nucleophilic elbow, which is featured by a conserved GXSXG sequence motif. The second member of the triad, His230, lies in a loop connecting strand β6 and η2, and serves as the general base for activation of the nucleophilic Ser78. The acidic residue Asp202 is situated in a turn linking αF and β5. The acidic moiety of Asp202 normally provides the carboxylate group as a counterion for the catalytic His230. The imidazole ring of His230 is hydrogen bonded to the side chain of the catalytic Asp202, which stabilizes the conformation of His230 (Fig. 3A to D). However, in both BIF-PytH and PMSF-PytH structures, the distances between Ser78 and His230 are too long to form hydrogen bond (Fig. 3; see also Fig. 5B), which is different from the typical catalytic triads commonly found in various other hydrolases (18, 21).

FIG 5.

Structural comparison of PytH, SABP2, HNL, MES16, and PNAE. (A) Superimposed overall structures of PytH, SABP2, HNL, MES16, and PNAE. (B) Superimposed active sites of PytH, SABP2, HNL, MES16, and PNAE. (C and D) Superimposed lid domains of PytH, SABP2, HNL, and PNAE. PytH, SABP2, HNL, MES16, and PNAE are indicated in purple, cyan, yellow, gray, and orange, respectively.

To corroborate the importance of these three residues, point mutants of these residues were constructed, PytH mutants were exogenously expressed and purified (see Fig. S1 in the supplemental material), and the pyrethroid-hydrolyzing activities of different PytH mutants were examined. As expected, point mutation of any one of the catalytic triad residues Ser78, Asp202, or His230 abolished the enzymatic activity of PytH (Table 2). In addition to the necessity of the catalytic triad, as proven by site-directed mutagenesis, the enzyme activity of the two mutants G76S or G80C, which in the conserved GXSXG drops to a negligible level, indicates that not just Ser78 itself but also the right conformation of Ser78 is essential for its activity (Table 2). Thus, in the PytH structure, the identification of Ser78, His230, and Asp202 as the catalytic residues forming a triad is therefore unambiguous. To further elucidate how PytH catalyzes its substrate, the catalytically inactive mutants S78A and H230A were also crystallized (Fig. S2). As indicated by the structure of H230A and S78A, the hydrogen-bonding network among the catalytic triad residues is not formed due to the loss of His230 and Ser78, respectively (Fig. S3).

TABLE 2.

Specific activities of wild-type PytH and its mutants toward permethrin and BIF

Strain	Mean sp act (μmol/min/mg) ± SD
	Permethrin	BIF
WT	5.32 ± 0.63	0.75 ± 0.11
S78A	0	0
H230A	0	0
D202A	0	0
G76S	0.23 ± 0.06	0.05 ± 0.01
G80C	0.19 ± 0.05	0.06 ± 0.02

The substrate BIF is located in the channel formed by the lid domain and the core domain; the ester group of BIF directly forms a hydrogen-bonding contact with the catalytic Ser78 and His230 (Fig. 3A and C). In addition to the hydrogen bonds with the triad, the ester group of BIF is also hydrogen bonded to the backbone amides of Ala12 and Leu79, the neighbor of Ser78 (Fig. 3E); therefore, these two residues may be involved in oxyanion stabilization and forming an oxyanion hole during the actual course of the substrate hydrolysis as some other α/β-hydrolases (20, 22, 23).

Substrate binding pocket.

Since an opening deep channel on the surface could be observed in the PytH structure (Fig. 2B and C) and may therefore be involved in substrate binding, the surface of BIF-PytH was checked. As expected, the BIF bound to active sites could be clearly seen from both sides of the opening deep channel (Fig. 2B and C). Pisa (24) was used to analyze the interactions between the BIF and the peptide of PytH. Two hydrogen bonds between Ser78 (OG), His230 (NE2) and BIF, respectively, and one salt bridge between His230 (NE2) and BIF were identified, and 29 amino acid residues were found to surround the BIF and form a pocketlike structure (Fig. 4). Since most residues of the pocket are hydrophobic, this pocket could bind hydrophobic substrates, such as BIF or other pyrethroids, by forming hydrophobic interactions with these residues. Among 29 residues, 14 residues (Val116, Thr125, Ala128, Leu129, Ile132, Leu141, Ala143, Phe145, Leu151, Phe155, Met156, Phe170, Ile171, and Gln174) are located in the lid domain, while the remaining 15 residues (Gly11, Ala12, Leu13, Asn14, His77, Ser78, Leu79, Val104, Phe179, Val204, Leu205, His230, Ala231, Tyr233, and Tyr234) and the catalytic Ser78, Asp202, or His230 are components of the core domain. Therefore, both the lid domain and the core domain are involved in substrate binding; the catalytic site is deeply buried inside the protein to form a hydrophobic cavity constituted by nonpolar residues (Fig. 4) and is connected to the protein surface by a narrow channel (Fig. 2B and C).

FIG 4.

Substrate binding pocket of PytH. A total of 29 hydrophobic residues form a hydrophobic pocket to bind the substrates, such as BIF (carbon atoms colored green). The residues shown as sticks with cyan carbon atoms belong to the lid domain. Residues shown as sticks with carbon atoms colored light gray belong to the core domain. The active site residues (S78, D202, and H230) are colored purple.

Structural comparison with related proteins.

To establish the proper structural classification of PytH and identify its relatives, a comparison with available structures in the Protein Data Bank (PDB) was carried out using a DALI search (25). Four significant matches were found. The closest relative structures are salicylic acid-binding protein SABP2 from Nicotiana tabacum (17), hydroxynitrile lyase (HNL) from Baliospermum montanum (18), methylesterase family member 16 (MES16) from Arabidopsis thaliana (19), and polyneuridine aldehyde esterase (PNAE) from Rauvolfia serpentine (20). PytH shares 20 to 24% amino acid sequence identities with SABP2, HNL, MES16, and PNAE and shows a 2.0- to 2.4-Å root mean square deviation (RMSD) between the aligned Cα coordinates. Although the sequence identities are very low, the superimposition of PytH with SABP2, HNL, MES16, and PNAE showed that PytH superimposes well with them (Fig. 5A). All of these proteins possess an α/β core and an inserted lid domain, and a high degree of structural and topological conservation of the core domain was observed. The active sites of PytH and other α/β-hydrolases superimposed somewhat well, including the catalytic triad (Ser78, His230, and Asp202) (Fig. 5B). Notably, the nucleophile Ser78 of PMSF-PytH structure has two conformations compared to those of SABP2, HNL, MES16, and PNAE (Fig. 3B and D).

Although the overall structure of PytH superimposed well with SABP2, HNL, MES16, and PNAE, obvious differences could be seen among PytH and these enzymes in the active sites and lid domains (Fig. 5B and C). In the PytH structure, His230 of PMSF-PytH in the catalytic triad is much farther away from Ser78 (4.0 Å) (Fig. 5B). Though the 3₁₀-helix Lη1 and the long loop (residues 158 to 170) between LαD3s are positioned identically to the long α-helix LαD3s of SABP2, HNL, and PNAE (since residues 119 to 135 of MES16 are missing, MES16 has not been used in the lid domain superimposition), the LαD3 of PytH moves away from the hydrophobic binding pocket compared to the LαD2s of SABP2, HNL, and PNAE (Fig. 5C). There was an extra inserted helix LαD2 with a longer loop between LαD1 and LαD2 and a much shorter loop region in the inserted hairpin motif (β41-loop-β42) (Fig. 5C and D). In addition to the differences in the main chain, the marked difference between these structures is the substrate binding pocket (Fig. 6). In PytH, a large continuous binding pocket is formed in the protein, whereas in SABP2 the pocket region corresponding to the biphenyl moiety was blocked by residues Phe151, Leu160, and His15, resulting in a very narrow channel in this region of SABP2 (Fig. 6A). Similar to SABP2, this corresponding region is narrowed by two methionine residues, Met162 and Met166, and two leucine residues, Leu21 and Leu157, in PNAE (Fig. 6B). These different pockets coincide with the substrates of these enzymes. The substrates of SABP2 and PNAE are much shorter than that of PytH and just bind at the position corresponding to that of the acid moiety of BIF (Fig. 6). It is not necessary to occupy the region which is occupied by the alcohol moiety as BIF; therefore, the pockets of SABP2 and PNAE are shaped by blocking this region (Fig. 6).

FIG 6.

Structural comparison of the substrate binding pocket of PytH, SABP2, and PNAE. (A) Superimposed substrate binding pockets of PytH and SABP2. (B) Superimposed substrate binding pockets of PytH and PNAE. Corresponding amino acid residues composing the narrowed pocket in SABP and PNAE are indicated in PytH. Substrate BIF of PytH is depicted in green. PytH is depicted in purple in panels A and B, SABP2 is depicted in gray in panel A, and PNAE is depicted in gray in panel B.

It has been reported that the amino acid next to the histidine residue of the catalytic triad (M239 of SABP2 or M245 of PNAE) is important for structural significance because the hydrolysis activity of M239K of SABP2 and M245A of PNAE decreased to approximately 0.02 and 0.4% of the wild-type enzymes (20, 26). Interestingly, this residue naturally is an Ala231 in wild-type PytH, and PytH does have hydrolysis activity. Our former results also show that A231 is involved in hydrophobic pocket forming (Fig. 4). Further structural comparison showed that the positions occupied by the side chain of Met229 in SABP2 and M245 in PNAE were occupied by the side chain of Tyr233 (both are Ala in SABP2 and PNAE), which may compensate for the short side chain of Ala231 in PytH (Fig. S4). These data imply that PytH has a specific strategy to stabilize the structure or catalytic mechanism, though the catalytic triad of PytH structure is similar to SABP2 and PNAE.

DISCUSSION

PytH, a typical α/β-hydrolase, was demonstrated to have a typical catalytic triad Ser78-His230-Asp202, and the triad was demonstrated in this study to play a key role in the catalysis of the enzyme. This triad in PytH superimposes well with those of other homologs, such as SABP2 and PNAE. The well-superimposed conserved triad between PytH and other homologs and the functional assay of the point mutants enable us to propose that PytH has a similar catalytic mechanism to those of other homologs. Similar to SABP2, HNL, and PNAE (17, 18, 20), PytH uses a reaction channel to bind substrate. However, the reaction channel for binding the substrates of PytH has distinct differences compared to those of SABP2, HNL, and PNAE. PytH has a much bigger and longer reaction channel, with two obvious entrances, compared to a small and short reaction channel with one entrance in SABP2, HNL, or PNAE (Fig. 7A to C). The small entrance of PytH superimposes with the entrance of SABP2, HNL, and PNAE (Fig. 7B and D), but it could be clearly seen that the side chains of Leu117, Leu129, Pro134, Leu141, and Phe179 at the entrance of PytH constitute a larger entrance than those of SABP2, HNL, and PNAE (Fig. 7D). The big entrance of PytH is constituted by the side chain of Phe145, Leu151, Met164, Phe170, Ile171, and Gln172 (Fig. 7C). Figure 8A and B further clearly show that the binding sites of substrate salicylic acid of SABP2 (Fig. 8A) and product EVS (16-epivellosimine) of PNAE (Fig. 8B) are located closer to the entrances, whereas BIF occupies the whole reaction channel (Fig. 8C). In addition, the reaction channels of SABP2 and PNAE are narrower or shorter than that of PytH (Fig. 8A to C), although the reaction channel of SABP2 seems longer than that of PNAE, whereas a narrow neck divided the long channel into two short channels (Fig. 8A). Since the substrates of PytH are much larger and longer than the substrates of SABP2 and PNAE (Fig. 8D), PytH may evolve a wider and longer L-shaped hydrophobic tunnel with bigger entrances. However, since the active site is deeply buried inside PytH and separated from the solvent, even with two wider entrances in PytH, the width of the entrances is still not large enough to allow the substrate to directly enter and bind to the active site, so we speculated that the lid of the PytH should be expected to have an open/closed transition to allow substrate binding and product release. Actually, this kind of open/closed transition has been observed in SABP2 (17).

FIG 7.

Substrate binding pocket of PytH. (A to C) The substrate binding pocket of PytH is wider and deeper (A) with a small entrance (B) and a big entrance (C). The lid domain is depicted in light blue, and the core domain is depicted in pale green. (D) Superimposed small entrances of PytH (purple), SABP2 (gray), HNL (orange), and PNAE (cyan).

FIG 8.

Substrate binding pocket comparison of PytH, SABP2, and PNAE. (A to C) Substrate binding pockets of SABP2 (A), PNAE (B), and PytH (C). The lid domain is depicted in light blue, and the core domain is depicted in pale green. Substrates of SABP2 and PytH, the products of PNAE, are depicted in orange. (D) Substrates of PytH, PNAE, and SABP2. The labile bonds are labeled with arrows.

In addition, differences in the lid domain of PytH also contribute to the bigger hydrophobic pocket. It has been reported that the lid domain could modulate substrate recognition and could be utilized to adapt the core domain to a variety of diverse functions, such as interfacial activation, substrate recognition, and enzyme inactivation (22). Therefore, the lid domain of PytH (Fig. 5C and D) is expected to recognize and select the larger pyrethroid substrates and then force the conformational changes of the core domain to activate PytH. This big substrate binding pocket suggests that PytH could recognize and hydrolyze different pyrethroids (Fig. 1). Hydrolysis of different pyrethroids by PytH (Fig. 1) also indicates that the binding pocket of PytH would adjust when binding to different pyrethroids. However, although PytH could digest different pyrethroids, the hydrolyzing efficiency of PytH versus different pyrethroids varies (9). Here, we solved the complex structure of PytH-BIF, but we did not obtain the complex structures of PytH with other pyrethroids. Failure to obtain the complex structures of other pyrethroids may be due to the relatively high hydrolyzing efficiency of PytH versus these pyrethroids (9). To understand how PytH discriminates different substrates, we further analyzed the structure of pyrethroids, along with the hydrolyzing efficiency of PytH to different pyrethroids (Fig. 1). PytH hydrolyzes permethrin (5.82 μmol/min/mg) better than cypermethrin (4.86 μmol/min/mg), suggesting that PytH adapts its structure to the (3-phenoxyphenyl)methyl group better than to the cyano-3-phenoxybenzyl group, since permethrin and cypermethrin share the same acid moiety. A higher hydrolyzing efficiency for fenpropathrin (5.02 μmol/min/mg) than for cypermethrin (4.86 μmol/min/mg) suggested that PytH prefers to select 2,2,3,3-tetramethylcyclopropanecarboxyl group compared to 2,2-dimethyl-3-(2,2-dichlorovinyl) cyclopropanecarboxyl group because these two substrates share the same alcohol moiety. In addition, a higher activity of PytH to cyhalothrin (2.46 μmol/min/mg) than to bifenthrin (0.86 μmol/min/mg) indicated that PytH recognizes the cyano-3-phenoxybenzyl group better than the (2-methyl[1,1′-biphenyl]-3-yl) methyl group. These data, taken together, show that the size of the acid moiety obviously affects the activity of PytH, i.e., the large size of the acid moiety inhibits the activity of PytH. Hence, we speculate that the large size of the acid moiety pushes the substrate to the other side of the substrate binding pocket and so it is not located in the best position for hydrolyzing. At the same time, the size of the alcohol moiety would also affect the activity of PytH by influencing the position of the substrate of PytH. Since the size of both moieties of the substrate is important for the activity of PytH, we speculate that two medium-sized moieties of permethrin in the best position to be hydrolyzed resulted in the highest activity of PytH to permethrin, which was demonstrated in an activity assay (9).

In addition, we observed that the pocket is still larger than the substrate BIF in the PytH-BIF complex structure (Fig. 8C). This large binding pocket of PytH suggests that PytH could hydrolyze substrates other than pyrethroids.

The extensive use of pyrethroid pesticides has caused serious contamination of pyrethroid residues in soil and agricultural products. It is reported that the detection rate of pyrethroid residues in vegetable, tea, and fruit products in China reached 40 to 50%, and bifenthrin, cyhalothrin, deltamethrin, and fenvalerate were the most frequently detected pyrethroid pesticides (27–31). Pyrethroid residues have been detected in other countries, in addition to China, in soil, water systems, air or dust, river sediments, and agricultural products (31). Pyrethroid pesticides are hydrophobic and insoluble in water; thus, the pyrethroid residues tightly adhere to the hydrophobic surfaces of agricultural products and are very difficult to remove by physical or chemical methods. Thus, it is necessary to solve this problem. Our results indicated that PytH can hydrolyze various pyrethroid pesticides due to its deeper and wider hydrophobic pocket, which is an advantage for developing PytH as a bioremediation enzyme agent. Although PytH cannot completely mineralize pyrethroids into carbon dioxide and water, it could still transform pyrethroids into water-soluble chrysanthemic acid and 3-phenoxybenzaldehyde (or biphenylaldehyde) by cleaving the ester bond, resulting in easy removal from the surfaces of agricultural products. In addition, the rate-limiting step for the degradation of pyrethroid residues in soil and water is the hydrolysis of ester bonds (10, 16). PytH can also be applied for bioaugmentation to accelerate the degradation of pyrethroid residues in soil and water. However, there are also some limitations for the application of PytH. The hydrolysis efficiency of PytH versus several important pyrethroid insecticides such as bifenthrin, cyhalothrin, deltamethrin, and fenvalerate is relatively low, with only one-half to one-fifth of the hydrolysis efficiencies versus permethrin and cypermethrin (16). This disadvantage may reduce the application value of PytH. Therefore, it is necessary to improve the catalytic efficiency of PytH for these pyrethroid pesticides through directed evolution or protein rational design, based on the determined structures. By taking into consideration both the lid domain and the core domain, we may improve the hydrolyzing efficiency of PytH for bifenthrin, cyhalothrin, deltamethrin, and fenvalerate in the future.

In summary, the structure of the novel pyrethroid hydrolase PytH was determined, the catalytic Ser-His-Asp triad was confirmed by the structure and activity assay, and processes of substrate selection, cap domain movement, and the catalytic mechanism were also proposed. Our study sheds light on future applications in directed enzyme evolution and genetic engineering.

MATERIALS AND METHODS

Cloning, expression, and purification of PytH and PytH mutants.

The gene fragment encoding PytH (GenBank accession number GU119902.1) from Sphingobium faniae JZ-2 (10, 32) was amplified by PCR. H230A and S78A mutants of PytH were generated by two-step overlap PCR as described previously (33). The gene cloning, protein expression protocol for the mutated PytH gene is same as for wild-type pytH. The amplified DNA fragment was cloned into the expression vector pET24b (Novagen) at cloning sites of NdeI/XhoI without a translation stop codon. The construct plasmid was sequenced to confirm that there was no mutation present. The construct was then transformed into E. coli C43(DE3) cells for protein expression. When the culture density in Luria-Bertani medium containing 30 μg ml⁻¹ kanamycin reached an optical density at 600 nm of 0.9 to 1.0, induction with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) was performed, and cells were grown for a further 3 h at 30°C. The cells were harvested by centrifugation (5,000 rpm for 10 min). The activities of wild-type and mutant PytH were determined according to the method of Wang et al. (9).

Cells were resuspended in binding buffer (350 mM KCl, 50 mM K₂HPO₄/KH₂PO₄ [pH 7.6], 20 mM imidazole, 20 mM β-mercaptoethanol [β-ME], 5% glycerol, and 1 mM PMSF) and then sonicated for 30 min on ice. The cell lysate was clarified by centrifugation for 30 min at 15,000 rpm, and the recombinant protein with six histidines at the C terminus was purified by immobilized metal-affinity chromatography on a Ni-NTA Superflow apparatus (Qiagen). The target protein was then desalted with a P10 column (GE Life Sciences) preequilibrated in buffer containing 350 mM KCl, 10 mM EDTA, 1 mM dithiothreitol, and 20 mM imidazole (pH 7.0). A typical purification yielded about 20 mg of PytH with a purity of >95% from 1 liter of cell culture.

Since PytH shares very low sequence identity with solved crystal structures, selenomethionine-labeled PytH was produced to obtain crystals for determining the phase. The purification of selenomethione-containing protein was performed according to the same protocols used for the native protein with the exception that 5 mM β-ME was added to the buffers to prevent the oxidization of selenium. All the purification procedures were performed at 4°C.

Crystallization, data collection, and structure determination.

Crystallization trials were performed using the sitting-drop vapor diffusion method with the protein concentrations of 20 mg ml⁻¹ at 4°C. Commercial screening kits from Hampton Research (PEGRx1-2, Crystal Screen, Crystal Screen 2, MembFac, Crystal Screen Cryo, Crystal Screen Lite, and Index) and Microlytic (MCSG1-4) were used for the initial crystallization screening. The crystallization drops consisting of a 1:1 mixture of protein (20 mg ml⁻¹) and reservoir solution were equilibrated against 50-μl reservoir solutions.

The preliminary crystals were obtained using the following conditions: HR2-110-16 (Crystal Screen; 0.1 M HEPES sodium [pH 7.5], 1.5 M lithium sulfate monohydrate) and HR2-114-36 (MembFac; 0.1 mol/liter HEPES sodium [pH 7.5], 1 M sodium citrate tribasic dihydrate). With extensive optimization, the best crystals from condition HR2-114-36 could only diffract to 3.5 Å. After extensive optimization, the best crystal that diffracts to 1.9 Å was obtained from 0.1 mol/liter HEPES sodium (pH 7.7) and 1.4 mol/liter lithium sulfate monohydrate. For the cryogenic X-ray diffraction, the crystals were sequentially transferred to 10-μl drops of crystallization buffer containing 5, 10, and 15% (wt/vol) glycerol, respectively, and then flash frozen in liquid nitrogen. To obtain crystals of BIF-PytH, 1.0 mg/ml BIF was mixed with protein before crystallization. BIF complex crystals were obtained with the well buffer containing 0.1 M HEPES sodium (pH 7.7) and 1.5 M lithium sulfate monohydrate. The mutant proteins (H230A and S78A) were crystallized under the same conditions used for wild-type PytH.

Diffraction data sets of native, Se-Met SAD, BIF_PytH, and PytH_H230A mutant data were collected at the BL17U1 beamline at Shanghai Synchrotron Radiation Facility using an ADSC Quantum 315r CCD area detector. The diffraction data set of PytH_S78A was collected at BL19U1 beamlines at the National Center for Protein Sciences Shanghai using a Pilatus 6M detector. The data sets were processed using the XDS package (34) and scaled using SCALA (35). The structure was solved using the SAS protocol of Auto-Rickshaw, the EMBL-Hamburg automated crystal structure determination platform (36). Input diffraction data were prepared and converted for use in Auto-Rickshaw using programs of the CCP4 suite (37). Structure factor amplitude (FA) values were calculated using the program SHELXC (38). Based on an initial analysis of the data, the maximum resolution for substructure determination and initial phase calculation was set to 5.0 Å. A total of 41 heavy atoms out of the maximum number of 48 heavy atoms requested were found using the program SHELXD (39). The correct hand for the substructure was determined using the programs ABS (40) and SHELXE (41). The occupancy of all substructure atoms was refined, and initial phases were calculated using the program MLPHARE (37).

The 2-fold noncrystallographic symmetry (NCS) operator was found using the program RESOLVE (42). Density modification, phase extension, and NCS averaging were performed using the program DM (43). A partial model was produced using the program BUCCANEER (44). The partial model contained 614 residues out of a total of 1,728 residues. After several round of refinement (45) and manually building with Coot (46), a complete chain was then used for molecular replacement to solve the structures of apoenzyme and the mutants.

Structural comparisons were carried out using the DALI algorithm (25). Figures were produced using the program PyMOL (47).

Data availability.

The atomic coordinates and structure-factor amplitudes have been deposited in the PDB for the BIF-PytH complex (PDB 5Y5R), PMSF-PytH (PDB 5Y57), PytH-H230A (PDB 5Y51), and PytH-S78A (PDB 5Y5V). The PytH gene and the mutants of PytH are available from J. He (hejian@njau.edu.cn) or W. Wang (wwwang@njau.edu.cn) upon request.