Single-Component and Two-Component para-Nitrophenol Monooxygenases: Structural Basis for Their Catalytic Difference

ABSTRACT

para-Nitrophenol (PNP) is a hydrolytic product of organophosphate insecticides, such as parathion and methylparathion, in soil. Aerobic microbial degradation of PNP has been classically shown to proceed via the “hydroquinone (HQ) pathway” in Gram-negative degraders, whereas it proceeds via the “benzenetriol (BT) pathway” in Gram-positive ones. The “HQ pathway” is initiated by a single-component PNP 4-monooxygenase and the “BT pathway” by a two-component PNP 2-monooxygenase. Their regioselectivity intrigued us enough to investigate their catalytic difference through structural study. PnpA1 is the oxygenase component of the two-component PNP 2-monooxygenase from Gram-positive Rhodococcus imtechensis strain RKJ300. It also catalyzes the hydroxylation of 4-nitrocatechol (4NC) and 2-chloro-4-nitrophenol (2C4NP). However, the mechanisms are unknown. Here, PnpA1 was structurally determined to be a member of the group D flavin-dependent monooxygenases with an acyl coenzyme A (acyl-CoA) dehydrogenase fold. The crystal structure and site-directed mutagenesis underlined the direct involvement of Arg100 and His293 in catalysis. The bulky side chain of Val292 was proposed to push the substrate toward flavin adenine dinucleotide (FAD), hence positioning the substrate properly. An N450A variant was found with improved activity for 4NC and 2C4NP—probably because of the reduced steric hindrance. PnpA1 shows an obvious difference in substrate selectivity with its close homologues TcpA and TftD, which may be caused by the unique Thr296 and a different conformation in the loop from positions 449 to 454 (loop 449–454). Above all, our study allows structural comparison between the two types of PNP monooxygenases. An explanation that accounts for their regioselectivity was proposed: the different PNP binding manners determine their choice of ortho- or para-hydroxylation on PNP.

IMPORTANCE Single-component PNP monoxygenases hydroxylate PNP at the 4 position, while two-component ones do so at the 2 position. However, their catalytic and structural differences remain elusive. The structure of single-component PNP 4-monooxygenase has previously been determined. In this study, to illustrate their catalytic difference, we resolved the crystal structure of PnpA1, a typical two-component PNP 2-monooxygenase. The roles of several key amino acid residues in substrate binding and catalysis were revealed, and a variant with improved activities toward 4NC and 2C4NP was obtained. Moreover, through comparison of the two types of PNP monooxygenases, a hypothesis was proposed to account for their catalytic difference, which gives us a better understanding of these two similar reactions at the molecular level. In addition, these results will also be of further aid in rational design of enzymes in bioremediation and biosynthesis.

INTRODUCTION

para-Nitrophenol (PNP), which is widely used in the manufacture of dyes, drugs, pesticides, and other industrial products (1), has been listed as a priority pollutant by the U.S. Environmental Protection Agency (2). The decomposition of organophosphorus pesticides is one of the main sources of residual PNP in soil. Many bacterial strains have evolved to be able to degrade PNP, including Gram-negative Moraxella (3), Pseudomonas (4), and Burkholderia, as well as Gram-positive Nocardia (5), Rhodococcus (6, 7), and Arthrobacter (8). Two alternative degradation pathways for PNP have been elucidated: the “hydroquinone (HQ) pathway” is mostly found in Gram-negative bacteria and the “benzenetriol (BT) pathway” in Gram-positive bacteria. These two pathways are classified according to their different ring cleavage substrates as terminal aromatic intermediates (HQ and BT). They are respective products of PNP monooxygenations catalyzed by two different types of PNP hydroxylases/monooxygenases, which initiate the different degradation pathways. Both types belong to flavin-dependent monooxygenases (FDMs) that are involved in various biological processes, such as drug detoxification, biodegradation of natural and xenobiotic compounds, and biosynthesis of hormones, vitamins, and antibiotics (9). The HQ pathway is initiated by a single-component FDM (PNP 4-monooxygenase), while the BT pathway is initiated by a two-component FDM (PNP 2-monooxygenase).

The general catalytic difference between single-component and two-component flavin monooxygenases has been revealed in previous studies. As a paradigm for single-component ones, the catalytic mechanisms are summarized as the dynamic transformation between two conformations through many structural and kinetic studies (10, 11). Upon binding of a substrate, the protein conformation changes from an open form to a closed one. After binding NAD(P)H, the isoalloxazine ring of flavin adenine dinucleotide (FAD) swings to the protein surface (the out conformation), facilitating the transfer of electron from NAD(P)H to FAD. Subsequently, the enzyme and reduced FAD go back to the in/closed conformation. The reduced FAD reacts with oxygen, forming a C-4a-hydroperoxyflavin intermediate, and the substrate hydroxylation occurs (12). In contrast, the two-component monooxygenases also form the C-4a-hydroperoxyflavin intermediate, but they exhibit remarkable differences from the single-component systems. In a two-component system, reduced flavin is supplied by a reductase component. The oxygenase component first binds reduced flavin and reacts with an oxygen molecule, forming the C-4a-hydroperoxyflavin intermediate, which is stabilized until a substrate enters the active site. Accordingly, the formation and stabilization mechanism of this intermediate is important during the dynamic catalytic reaction, which is different from the single-component system (13–15). In contrast, in the single-component system a substrate must be bound to the enzyme before oxygen activation. The substrate may immediately react with the C-4a-hydroperoxyflavin intermediate formed upon binding of oxygen. These studies provide a rationale for the basic catalytic difference between single-component and two-component flavin-dependent enzymes, but are inadequate to explain the reason for the catalytic difference between the two types of aforementioned PNP monooxygenases.

In Gram-negative PNP utilizers, including Pseudomonas sp. strain WBC-3 (4, 16), Pseudomonas putida DLL-E4 (17), Moraxella sp. (3), and Burkholderia sp. strain SJ98 (18), single-component PNP 4-monooxygenases catalyze a one-step hydroxylation of PNP at the 4 position, giving a p-benzoquinone. This product was reduced by a p-benzoquinone reductase to produce HQ/quinol (Fig. 1a) (3, 4). A PNP 4-monooxygenase of this type from Pseudomonas putida DLL-E4 (PnpA_DLL-E4) has been structurally characterized as a class A flavin monooxygenase. Class A enzymes are almost all hydroxylases with aromatic compounds as substrates. Using AutoDock, PNP was modeled into the active site of the PnpA_DLL-E4 (19). The nitro group of PNP is directly hydrogen bonded to Val54 and Arg234, and the hydroxyl group is hydrogen bonded to Val223 and Cys236 (19).

FIG 1.

Comparison of the degradation pathways in Gram-negative and Gram-positive strains. (a) Degradation pathways of PNP and 4NC in Gram-negative strains. PNP and 4NC are oxidized to p-benzoquinone and hydroxyl-1,4-benzoquinone, respectively, and the latter two can be reduced and degraded by other enzymes. (b) Degradation pathways of PNP and 2C4NP in Gram-positive strains. PnpA1 oxidizes PNP to 4NC, and the latter is further oxidized to hydroxyl-1,4-benzoquinone, which can be nonenzymatically reduced to BT. 2C4NP can also be oxidized to produce chloro-1,4-benzoquinone, which is immediately converted to hydroxyl-1,4-benzoquinone by hydrolytic dechorination and reduced to BT.

In Gram-positive PNP utilizers, including Rhodococcus imtechensis RKJ300 (7), Rhodococcus opacus SAO101 (20), Rhodococcus sp. strain PN1 (6), Arthrobacter sp. strain JS443 (8), and Bacillus sphaericus JS905 (21), two-component PNP 2-monooxygenases catalyze a tandem two-step hydroxylation of PNP (Fig. 1b): first at the 2 position of PNP, producing 4-nitrocatechol (4NC), and subsequently at the 4 position of 4NC, producing hydroxyl-1,4-benzoquinone (8, 21), which is reduced by NADH nonenzymatically, giving a 1,2,4-benzenetriol/BT(4). A typical two-component PNP 2-monooxygenase (PnpA1A2) from Rhodococcus imtechensis RKJ300 was biochemically identified previously (22). It is also able to catalyze the hydroxylation of 2-chloro-4-nitrophenol (2C4NP) to produce hydroxyl-1,4-benzoquinone. However, the catalytic mechanism of PnpA1 remains to be studied.

The crystal structures of some of PnpA1’s close homologues have been analyzed—for example, 2,4,6- trichlorophenol (2,4,6-TCP) 4-monooxygenase (TcpA) from Cupriavidus necator JMP134 (23), 2,4,5-trichlorophenol (2,4,5-TCP) 4-monooxygenase (TftD) from Burkholderia cepacia AC1100 (24), 4-chlorophenol 4-monooxygenase (HadA) from Ralstonia pickettii DTPO602 (25), and 4-hydroxyphenylacetate 3-monooxygenase (HpaB) from Thermus thermophilus HB8 (12). Although their amino acid sequence identities with PnpA1 reach as high as 44, 44, 43, and 27%, respectively, the substrate specificities of these monooxygenases are obviously different. This prompted us to investigate and compare their structural features, particularly the active site.

In this study, structural biology methods were adopted to explain the catalytic difference between the two types of PNP monooxygenases and to understand the catalytic mechanism of the two-component PNP 2-monooxygenase. In particular, we determined the crystal structure of PnpA1, the oxygenase component of a two-component PNP 2-monooxygenase, from Rhodococcus imtechensis RKJ300. On the one hand, the roles of some pivotal residues in its active center were revealed. A loop covering one side of the substrate-binding pocket may account for the difference in substrate specificity between PnpA1 and its homologues, TcpA and TftD. On the other hand, structural comparison between single-component and two-component PNP monooxygenases, coupled with computational substrate docking, was conducted. It is proposed that their different FAD and substrate binding manners determine their regioselectivity and then result in different products.

RESULTS

Overall structure of PnpA1.

Recombinant PnpA1 from Rhodococcus imtechensis RKJ300 was crystallized in the space group I4₁22, with cell dimensions of a = 150.01 Å, b = 150.01 Å, and c = 321.30 Å. There were three PnpA1 molecules in an asymmetric unit. Through generation of symmetry mates, it can be observed that one tetramer is obtained by a 2-fold axis operation of the dimer of A and B molecules. Another tetramer is obtained by two 2-fold axis operations of the C molecule. Therefore, the functional unit of PnpA1 is a tetramer (Fig. 2a), and the tetramer state in solution was confirmed by gel filtration chromatography (see Fig. S1 in the supplemental material). A summary of the crystallographic data with a 2.5-Å resolution is given in Table 1. Through structural analysis, PnpA1 was identified as a member of group D flavin-dependent monooxygenases, with an acyl coenzyme A (acyl-CoA) dehydrogenase protein fold (26).

FIG 2.

The structure of PnpA1 from Rhodococcus imtechensis RKJ300. (a) The overall structure of PnpA1 is a tetramer, and the four subunits are shown in different colors. (b) The three segments in a monomer are shown in blue, red, and green, respectively.

TABLE 1.

Data collection and refinement statistics

Parameter	Value(s) for PnpA1a
Data collection statistics
Unit cell dimensions
a, b, c (Å)	150.0, 150.0, 321.3
α = β = γ (°)	90
Space group	I4122
Wavelength (Å)	0.97855
Resolution range (Å)	49.41–2.50 (2.54–2.50)
No. of reflections
Total	437,743
Unique	63,441
Completeness (%)	99.9 (99.9)
CC1/2	0.985
Redundancy	6.9 (6.5)
Avg I/σ〈I〉	13.1 (3.5)
Rmerge (%)b	0.113 (0.616)
Refinement statistics
No. of reflections in:
Working set	59,779
Test set	3,212
Rwork/Rfree (%)c	24.18/29.13
RMSD
Bond length (Å)	0.009
Bond angle (°)	0.959
Ramachandran plot (%)
Favored regions	98.04
Allowed regions	0.07
Outlier regions	0.00

^aValues in parentheses represent the highest-resolution shell for PnpA1.

^bR_merge = Σ_hΣ_i|I_i(h) − 〈I(h)〉|/Σ_hΣ_i〈I_i(h)〉, where I_i(h) is an individual intensity measurement and 〈I_i(h)〉 is the average intensity for all measurements of the reflection.

^cR_work = Σ‖F_obs| − |F_calc‖/Σ|F_obs|, where F_obs and F_calc are the observed and calculated structure factor amplitudes, respectively. R_free was calculated as R_work using a randomly selected subset of ∼5% of unique reflections not used for structure refinement.

The PnpA1 monomer consists of 15 α-helices, 13 β-strands, and four 3₁₀ helices and can be divided into three sequential segments based on their constituent secondary structural elements. Those three segments are tightly interconnected in the three-dimensional structure (Fig. 2b). Moreover, they form three faces of the hydrophobic substrate pocket. The first segment (residues 1 to 147) is composed of six helices (α1 to α6) and three short β-strands (β1 to β3). This area is located in the periphery of the tetramer and constitutes the inner side of the pocket. The second segment (residues 148 to 277) has eight β-strands (β4 to β11). These β-strands are arranged as a small β-barrel. The third segment, namely, the C-terminal segment, is composed of 8 helices (α7 to α15) and 2 short β-strands (β12 and β13). Six α-helices (α7, α8, α9, α11, α12, and α13) form a helix bundle, which is mainly maintained by hydrophobic interaction among them. This segment is located at the tetramerization interface and involved in interacting with neighboring subunits. Noticeably, α14 and α15 are the most extensive two of the eight helices (Fig. 2b). They clearly protrude from the protein core and extend into the neighboring subunit. No electron density was observed for residues 162 to 168 and residues 506 to 528 because of their high flexibility, which is common among PnpA1’s homologues.

FAD binding site of PnpA1.

The narrow gap constituted by the three segments clearly defines a hydrophobic pocket that accommodates reduced FAD and the substrate. However, because the oxygenase component of the two-component flavin monooxygenase has a much lower affinity for oxidized flavin than for the reduced one (13), the cocrystalization of PnpA1 and FAD failed. Here FAD was then docked into PnpA1 crystal structure (Fig. 3a). The pose that shows highest consistency with that in the HpaB·FAD complex was selected (12). HpaB shares 28% sequence identity with PnpA1 and conservation of nearly all secondary structural elements, with a root mean square deviation (RMSD) of 1.69 Å for Cα atoms. The si-side of FAD is exposed to solvent, and the re-side faces the substrate-binding site. FAD is embedded in a protein scaffold formed by several stretches of fragments: the loop from positions 449 to 454 (loop 449–454), loop 187–201, α13, α14, and loop 154–175, which are responsible for coordinating FAD. The former three interact with the isoalloxazine ring of FAD, while the latter two interact with the AMP group of FAD. As generally observed in flavoenzymes, FAD is kept in position by a network of hydrogen bonds and hydrophobic interactions. Phe446, Phe449, Met192, and Val156 have hydrophobic interactions with FAD. The side chains of Thr194, Gln159, Arg209, and Tyr455, as well as the main chains of Met192, Phe155, Val156, Phe449, and Gly451, are associated with polar interactions with FAD. It is noteworthy that the flexible side chain of Asn450 in PnpA1·FAD must be pushed away from its original location in PnpA1 in order to accommodate FAD (Fig. 3a). The side chain is a short distance from and orients to the substrate-binding site.

FIG 3.

The FAD and PNP binding sites of PnpA1. (a) AutoDock positioning of FAD in the active site of PnpA1. Asn450 performs a side-chain swing (from the pink stick to the green one) in order to accommodate FAD. Molecular surfacing of vacuum electrostatics features an electronegative cavity for the binding of FAD and substrates (Inset). (b) The substrate-binding pocket in the PnpA1·FAD·PNP model. Hydrogen bonds are shown as gray dashes.

Substrate binding in PnpA1.

Previous studies indicated that for two-component flavin-dependent monooxygenases, reduced flavin must be bound to the enzyme before the substrate’s entry (13–15). The PnpA1·FAD model shows that a substrate can only enter the binding pocket through a narrow hydrophobic tunnel, mainly consisting of Val156, Leu456, and Leu207, because of the prepositioned FAD. The perimeter of the substrate-binding site is mainly made with four fragments including α7, loop 449–454, loop 93–102, and loop 154–175. Among those fragments, loop 449–454 has a relatively high temperature factor and shows the most heterogeneous conformations among the four subunits, reflecting its flexibility. The residues lining the substrate-binding pocket are predominantly hydrophobic ones, including Phe446, Phe155, Phe289, Val292, and Leu207. In addition, the roles of some polar residues, Arg100, Asn450, His293, Thr452, and Thr296, are also discussed (see below). These stereochemical features in the interactions between hydroperoxyflavin and the substrate promote the monooxygenation that occurs in a solvent-protected environment. Through AutoDock, PNP was docked into PnpA1 with FAD prepositioned (Fig. 3b). The first pose with highest binding affinity was selected for analysis (see Table S1 in the supplemental material). PNP binds on the re-side of the flavin, and its aromatic ring lies on a plane roughly perpendicular to the plane of the FAD (Fig. 3b). The hydroxyl group of PNP forms hydrogen bond contact with Arg100, while the nitro group has hydrogen bond interactions with FAD, His293, and Thr452 (Fig. 3b).

Enzyme activity assays.

Wild-type PnpA1 showed an evidently higher specific activity for PNP (34.2 μmol/min·g) than for both 2C4NP and 4NC (12.8 and 17.9 μmol/min·g, respectively). Here the roles of residues in the substrate-binding site were investigated through site-directed mutagenesis (Fig. 4). Apart from binding substrates, Arg100 and His293 are also necessary for enzymatic reaction. Arg100, a conserved catalytic residue in the homologues of PnpA1, is proposed to act as a proton donor in the reaction between reduced flavin and oxygen that leads to the formation of the C-4a-hydroperoxyflavin intermediate (12). His293 (homologous to His289 in TftD) is proposed to be directly involved in catalysis (24). As predicted, both R100A and H293A completely lost activities against PNP, 2C4NP, and 4NC. Surprisingly, although Arg370 is not directly involved in PNP binding, the R370A mutant also lost all the activity for these substrates.

FIG 4.

The specific activities of PnpA1 and its variants for PNP, 4NC, and 2C4NP.

Among the close homologues of PnpA1, including TcpA, TftD, HadA, and HpaB, the consensus of hydrophilic Thr296 of PnpA1 is hydrophobic (Ala, Ile, Ile, and Val, respectively), exhibiting a significant difference between PnpA1 and its homologues. In our result, the T296A variant was insolubly expressed (see Fig. S2 in the supplemental material), indicating the essential role of Thr296 in maintaining the conformation of the whole protein—possibly through the interactions between its hydroxyl group and the surrounding residues in α7, such as Arg299, Glu295, and Val292.

A previous study showed that Ile292 in TftD and Ala293 in TcpA are possibly responsible for the differences in substrate specificities between these two monooxygenases (23). In PnpA1, the homologous residue is Val292, which is at the other end of the substrate pocket away from FAD (Fig. 3b). The V292A mutant lost most the activity for PNP, 2C4NP, and 4NC. Replacement of Val292 with a leucine, which has a side chain of similar size to Val, did not cause an obvious change in its activity (Fig. 4). It is proposed that a side chain with enough steric hindrance at this location is necessary, in order to push the substrate closer to FAD.

It is noteworthy, as mentioned above, that the side chain of Asn450 needs to swing toward the substrate-binding site in order to accommodate FAD. The mutation N450A had no influence on the hydroxylase activity for PNP but improved the activities for 4NC (1.4-fold) and 2C4NP (1.2-fold) (Fig. 4). 4NC and 2C4NP both have a substituent group (chlorine atom and hydroxyl group, respectively) at the 2 position, and these substituents are merely about 3.3 to 3.6 Å away from the side chain of Asn450. It is proposed that the replacement of Asn450 with an alanine reduced the steric hindrance for the binding of 2C4NP and 4NC.

Comparison of PnpA1 with its homologues.

Group D is a relatively small family among the A to H groups of flavin-dependent monooxygenases. Although a search using the Dali server (27) indicates that TcpA (PDB ID 4g5e) and TftD (PDB ID 3hwc) are structurally similar to PnpA1, with about 44% amino acid sequence identity, their substrate specificities vary to some extent. The most suitable substrate of PnpA1 is PNP, and it can also oxidize 4NC, 2C4NP, and 2,4-dinitrophenol, with much lower activities. The substrates of TcpA and TftD are trichlorophenols, which are larger than PNP. TcpA catalyzes the dechlorination of 2,4,6-TCP and 2,6-dichlorophenol, while TftD aims at 2,3,5,6-tetrachlorophenol, 2,4,5-TCP, 2,4,6-TCP, and 2,5-dichloro-p-hydroquinone (2, 5-DiCH). A structural comparison shows that two structural features in the active site may be responsible for their difference in substrate selectivity.

On the one hand, the Thr296 in PnpA1 is a unique residue among those homologous enzymes where an Ala293 (TcpA) or an Ile292 (TftD) is present (Fig. 5a). The hydroxyl group of Thr296 participates in hydrogen bonding with many other residues in the molecule, including Val292, Glu295, and Arg299. These interactions do not exist in TcpA and TftD. The mutation of Thr296 caused a change in the folding of the entire molecule in our study. These findings highly imply that Thr296 acts differently from its homologous residues. Considering its close distance from the nitro group of PNP, Thr296 may cause PnpA1’s preference for substrates with nitro at the 4 position, while Ala293 and Ile292 make TcpA and TftD choose substrates with chlorines.

FIG 5.

Two structural features of PnpA1 that may account for its substrate selectivity compared with those of the homolog proteins. (a) PnpA1 overlapped with TcpA (orange) and TftD (blue). Thr296, shown in sticks, is a distinct residue from the corresponding residue in TcpA (Ala293) or TftD (Ile292). (b) PnpA1 is present in b-factor putty, and loop 449–454 (green) shows relatively high flexibility.

On the other hand, the loop with residues 449 to 454 in PnpA1 shows relatively high flexibility, especially Asn450 through Thr452 (Fig. 5b). This loop forms one side of the substrate-binding pocket and causes great steric hindrance to substrate binding. The loop 449–454 may account for the accommodation of larger substrates by TcpA and TftD, while PnpA1’s pocket is more suitable for smaller ones.

DISCUSSION

Implications for the catalytic mechanism of PnpA1.

When using PNP, 4NC, and 2C4NP as substrates, the position 1 hydroxyl group decides that electrophilic substitution can happen at its ortho and para sites. PnpA1 oxidizes PNP to 4NC (Fig. 6a), and the latter is further oxidized to hydroxyl-1,4-benzoquinone (Fig. 6b), which can be nonenzymatically reduced to BT (Fig. 1b). 2C4NP can be oxidized to chloro-1,4-benzoquinone, which is immediately converted to hydroxyl-1,4-benzoquinone by hydrolytic dechorination and reduced to BT (Fig. 1b and Fig. 6c). Analysis of the crystal structure provides a framework for the catalytic events in C-2 and C-4 active sites. A preorganized binding site recruits free reduced FAD (supplied by PnpA2). The protein-bound reduced cosubstrate FAD reacts with an oxygen molecule to produce C-4a-hydroperoxy-FAD, which is housed in the C-4a cavity. Subsequently, the substrate binds to the enzyme. The hydrophilic groups of the substrates are contacted through Arg100, His293, and Thr452. As shown in Fig. 6, three types of substitution reactions are involved in the reactions of PNP, 4NC, and 2C4NP. (i) The first type is substitution of hydrogen with hydroxyl group at the C-2 position of PNP (Fig. 6a). (ii) The second is substitution of the nitro group with a hydroxyl one at the C-4 positions of 4NC and 2C4NP (Fig. 6b and c). In this process, His293 acts as a general base and abstracts a proton from the 4-hydroxyl group. (iii) The third type is substitution of chlorine with a hydroxyl group at the C-2 position of 2C4NP (Fig. 6c). Arg100 and Arg370 are proposed to play important roles in this hydrolytic dechlorination. Due to inductive effects caused by the positively charged guanidinium group of Arg100, the electrophilicity of the chlorine 2 position should be slightly elevated to facilitate the attack by a water molecule. Arg370 would lead to charge stabilization during the attack through the interaction with the para-hydroxy group (Fig. 6c). In addition, it is worth noting that 4NC and 2C4NP may adopt different orientations from PNP in a way.

FIG 6.

Proposed catalytic mechanism of PnpA1. (a) PnpA1 oxidizes PNP to 4NC, and (b) the latter is further oxidized to hydroxyl-1,4-benzoquinone. (c) PnpA1 oxidizes 2C4NP to chloro-1,4-benzoquinone, which is immediately converted to hydroxyl-1,4-benzoquinone by hydrolytic dechlorination.

The former two types are electrophilic attacks started by FAD C-4a-(hydro)peroxide, while the third type is a hydrolytic dechlorination. Finally, C-4a-hydroxy-FAD eliminates H₂O, and the resulting oxidized FAD dissociates from the enzyme, which is ready to undergo a new cycle.

Structural analysis of the catalytic regioselectivity of PNP monooxygenases.

For the overall structure, these two types of PNP monooxygenases share low amino acid sequence similarity (16%): PnpA1 adopts an acyl-CoA dehydrogenase fold and belongs to group D of flavin-dependent monooxygenases, while PnpA_DLL-E4 adopts a Rossmann (GR-2) fold and belongs to group A. Unlike PnpA1, single-component monooxygenase PnpA_DLL-E4 is not able to conduct the hydroxylation at the 2 position of PNP. The catalytic product from PnpA_DLL-E4 is a p-benzoquinone, whereas the catalytic products from PnpA1 are 4NC and BT in tandem. It is remarkable that enzymes that catalyze the hydroxylation of a same substrate by using a similar mechanism are so different in the geometry of enzyme-substrate interactions and active site architecture.

First, inspection of their three-dimensional structures shows that the geometries of the substrate-binding site are very different in the two enzymes. For PnpA1, PNP is located at the re-face of FAD and is perpendicular to the isoalloxazine ring of FAD, with its 2 position close to C-4a of FAD (Fig. 7a). In contrast, for PnpA_DLL-E4, PNP is also perpendicular to the isoalloxazine ring but binds above FAD, with its 4 position close to C-4a of FAD (Fig. 7b). Secondly, different residues coordinate the hydroxyl groups and nitro groups in the two enzymes. In PnpA_DLL-E4, the hydroxyl group of PNP has a hydrogen bond with Cys236, and the nitro group engages in hydrogen bonding with Arg234 and Val54. In contrast, in PnpA1 the hydroxyl group of PNP has hydrogen bond contact with Arg100, and the nitro group forms hydrogen bonds with His293, Thr452, and FAD, respectively. From the above analysis, it can be seen that the polar interactions between PNP and PnpA_DLL-E4 are weaker than those between PNP and PnpA1.

FIG 7.

Comparison of the FAD and PNP binding manners between single- and two-component PNP monooxygenases. (a) PnpA1 and (b) PnpA_DLL-E4 (19). Hydrogen bonds are shown as gray dashes.

Here, we proposed a hypothesis to account for the catalytic difference between PnpA_DLL-E4 and PnpA1. For PnpA_DLL-E4, the 4 position nitro group was firstly substituted because of its proximity to the reactive oxygen of C-4a-hydroperoxyflavin intermediate, producing p-benzoquinone. p-Benzoquinone was released from the binding pocket of PnpA_DLL-E4 due to the less polar characteristics of the carbonyl group. However, for PnpA1, the hydroxylation first occurs at the 2 position of PNP, which is close to the reactive oxygen of the C-4a-hydroperoxyflavin intermediate. The product 4NC can be held in the active site for a subsequent hydroxylation at the 4 position to generate 1,4-hydroxybenzoquinone, which will immediately be reduced to BT by NADH.

The structural study here would lay a foundation for the rational design of PnpA1 to obtain variants with improved enzymatic activity and engineered strains with high degradation efficiency. Moreover, the proposed explanation for the catalytic difference between single-component PNP 4-monooxygenase and two-component PNP 2-monooxygenase provides molecular insights into the regioselectivity of flavin-dependent monooxygenases. The catalytic differences between the two types of PNP monooxygenases also help us understand microorganisms’ different choices when degrading the same aromatic compound.

MATERIALS AND METHODS

Cloning, expression, and purification.

To overproduce the PnpA1 and PnpA2 proteins, pnpA1 and pnpA2 were amplified by PCR with the primers in Table 2 from genomic DNA of strain RKJ300. The PCR products were digested by NdeI and HindIII and then inserted into pET-28a to encode recombinant proteins with an N-terminal 6×His tag. The recombinant plasmids pET-pnpA1 and pET-pnpA2 (Table 2) were then transformed into Escherichia coli Rosetta (DE3)/pLysS, respectively, for protein expression and purification. The bacterial culture was then grown at 37°C in 2 liters of LB medium supplemented with 50 μg/ml kanamycin until it reached an optical density at 600 nm (OD₆₀₀) of 0.6. Protein overproduction was induced by 0.4 mM isopropyl-β-d-thiogalactopyranoside (IPTG), and the cells were subsequently incubated overnight at 16°C. Cells were harvested and resuspended in lysis buffer (50 mM Tris-HCl [pH 7.6], 300 mM NaCl, 20 mM imidazole, 1 mM dithiothreitol [DTT]), followed by disruption via sonication. The insoluble cellular material was removed by centrifugation. PnpA1 and PnpA2 were first purified using Ni-nitrilotriacetic acid (NTA) affinity chromatography and the ÄKTAxpress system (GE Healthcare Life Sciences, Boston, MA). After centrifugation, supernatant was applied to a nickel column. Then the column was washed with lysis buffer with 20, 40, 60, and 80 mM imidazole successively. The proteins were collected by applying elution buffer (50 mM Tris-HCl [pH 7.6], 300 mM NaCl, 250 mM imidazole, and 1 mM DTT) to the column. The PnpA1 protein for crystallization was further purified by a Superdex-200 gel filtration chromatography column (GE Healthcare Life Sciences, Boston, MA) in a mixture of 50 mM Tris-HCl (pH 7.6), 150 mM NaCl, and 1 mM DTT. Finally, it was concentrated to 9 mg/ml by ultrafiltration in solutions of 10 mM Tris-HCl (pH 7.6) buffer containing 25 mM NaCl and 1 mM DTT.

TABLE 2.

Bacterial strains, plasmids, and primers used in this study

Strain, plasmid, or primer	Relevant genotype, characteristic(s), or sequencea	Source or reference
Strains
Rhodococcus imtechensis RKJ300	PNP and 2C4NP utilizer, wild type	7
Escherichia coli
DH5α	supE44 lacU169 (ϕ80 lacZΔM15) recA1 endA1 hsdR17 thi-1 gyrA96 relA1	Novagen
Rosetta(DE3)/pLysS	F− ompT hsdS(rB− mB+) gal dcm lacY1(DE3)/pLysSRARE (Cmr)	Novagen
Plasmids
pET-28a	Expression vector, Kanr	Novagen
pET-pnpA1	NdeI-HindIII fragment containing pnpA1 inserted into pET-28a	This study
pET-pnpA2	NdeI-HindIII fragment containing pnpA2 inserted into pET-28a	This study
Primers
pnpA1-F	5′-AGGCACCATATGAGGACAGGACAGCAGTATC-3′	This study
pnpA1-R	5′-TTTAAGCTTTCAGACCTTCTCGGTCTCC-3′	This study
pnpA2-F	5′-AGGCACCATATGTTGGAGGATCCGATGAAAC-3′	This study
pnpA2-R	5′-TTTAAGCTTTCAGTGGGGGGAAACTGC-3′	This study

^aCm, chloramphenicol; Kan, kanamycin.

Crystallization, data collection, and structural determination.

Crystals of PnpA1 were grown using the sitting-drop vapor diffusion method at 20°C, by mixing equal volumes of protein solution and reservoir solution. The reservoir solution contained 0.1 M Tris (pH 7.25), 1.5 M (NH₄)₂SO₄, and 4% (vol/vol) glycerol. Diffraction-quality crystals of PnpA1 appeared within 5 days.

The crystals above were flash cooled by liquid nitrogen and stored in liquid nitrogen for data collection. The reservoir solution with 20% glycerol was used as a cryoprotectant, and data were collected at 100K on beamline BL19U of the Shanghai Synchrotron Radiation Facility. All X-ray diffraction intensity data were integrated, scaled, and merged with the program HKL3000 (28). For structure determination of PnpA1, the phasing problem was solved by the MR method using the program Phaser in the CCP4 program suite (29). TcpA was employed as a search model (23). The program Coot (30) was used for manual adjustment of the model, and the programs Refmac5 (31) and Phenix (32) were used for refinement of the model.

Molecular docking of FAD and substrates.

To locate the binding sites of FAD and the substrate in PnpA1 crystal structure, in silico substrate docking was performed. The molecular docking process was performed with AutoDock (33). First, FAD was docked into PnpA1, generating a PnpA1·FAD model, which was used as a search template for the substrate-binding site. In the result of docking PNP into PnpA1·FAD model (Table S1), the pose with highest binding affinity (−6.5 kcal/mol) was used for subsequent analysis.

Site-directed mutagenesis.

Site-directed mutagenesis of PnpA1 was conducted using the Fast site-directed mutagenesis kit (Tiangen Biotech Co., Ltd., Beijing, China). Mutagenic oligonucleotides were designed following guidelines provided by the manufacturer.

Enzyme activity assay.

Substrate depletion was quantified to determine the enzyme activity, as an earlier study proved that in this enzymatic reaction substrate consumption and product formation are nicely coupled and no other side-products were detected (22). The enzyme activity assay was performed according to the method described previously (22). Briefly, activity was determined based on the enzyme’s capability to consume the substrates (PNP, 4NC and 2C4NP), and the concentration of the residual substrate was determined. The reactions were performed at room temperature in 0.5 ml 50 mM Tris-HCl buffer (pH 7.6) with 73.6 μg/ml PnpA1, 4.4 μg/ml PnpA2, 0.02 mM FAD, 0.2 mM NADPH, and 0.1 mM substrate. The reaction was stopped after 1 min by adding HCl, and the supernatant was obtained by centrifugation (20,000 × g for 10 min). The concentration of residual substrate was monitored by high-performance liquid chromatography (HPLC) (Waters Alliance 2695, Milford, MA) with a chromatograph equipped with a UV detector (set at 315 nm for PNP, 347 nm for 4NC, and 315 nm for 2C4NP). A Poroshell 120 EC-C₁₈ column (150 mm long, 4.6-mm inner diameter; Agilent Technologies, Tokyo, Japan) was used, and the temperature was set as 30°C. The mobile phase was a mixture of methanol and 0.1% acetic acid in water (70:30), and the pump was set to run in isocratic mode with a flow rate of 1.0 ml/min. The retention times of PNP, 4NC, and 2C4NP were 3.6, 3.1, and 5.0 min respectively. In each group of experiments, the reaction solution without NADPH was used as a negative control that was treated exactly as the experimental group to calculate the spike recovery. One unit of enzyme activity was defined as the amount required for the disappearance of 1 μmol of the substrate per minute at room temperature.

Data availability.

The protein structure of PnpA1 and associated diffraction data have been deposited in the Protein Data Bank under accession no. 7DBW.