PcpA (2,6-dichloro-p-hydroquinone 1,2-dioxygenase) from Sphingobium chlorophenolicum, a non-haem Fe(II) dioxygenase capable of cleaving the aromatic ring of p-hydroquinone and its substituted variants, is a member of the recently discovered p-hydroquinone 1,2-dioxygenases. Here we report the 2.6 Å structure of PcpA, which consists of four βαβββ motifs, a hallmark of the vicinal oxygen chelate superfamily. The secondary co-ordination sphere of the Fe(II) centre forms an extensive hydrogen-bonding network with three solvent exposed residues, linking the catalytic Fe(II) to solvent. A tight hydrophobic pocket provides p-hydroquinones access to the Fe(II) centre. The p-hydroxyl group is essential for the substrate-binding, thus phenols and catechols, lacking a p-hydroxyl group, do not bind to PcpA. Site-directed mutagenesis and kinetic analysis confirm the critical catalytic role played by the highly conserved His10, Thr13, His226 and Arg259. Based on these results, we propose a general reaction mechanism for p-hydroquinone 1,2-dioxygenases.
Pentachlorophenol (PCP) is introduced into the environment mainly through its massive use as a wood preservative, a practice that started in the late 1930s (Crosby, 1981). PCP is very toxic due to its ability to disrupt the membrane-dependent functions of most organisms. Acute exposure to PCP can lead to hyperthermia, convulsions and rapid death, with long-term health effects including mutagenicity and carcinogenicity in humans (Crosby, 1981). Further health concerns exist with the highly carcinogenic chlorinated dibenzo-p-dioxins and dibenzofurans (Firestone, 1978), which are produced from PCP by means of biotransformation in the environment or as impurities from industrial production processes (Hoekstra et al., 1999). Uncontrolled production, storage and wide range of application have made PCP a leading environmental pollutant detrimental to public health.
Several bacteria have been reported for their capacity to mineralize PCP, such as Mycobacterium chlorophenolicum (Häggblom et al., 1994, Briglia et al., 1994), Sphingobium chlorophenolicum (Saber and Crawford, 1985) and Novosphingobium sp. (Tiirola et al., 2002), and some have been applied to the field for bioremediation (Miethling and Karlson, 1996, Crawford and Mohn, 1985).
The PCP degradation pathway has been intensively investigated in S. chlorophenolicum (Cai and Xun, 2002; Sinnokrot and Sherril, 2004). In S. chlorophenolicum, PCP is converted to 2,6-dichloro-p-hydroquinone by the concerted actions of a monooxygenase (PcpB) (Xun and Orser, 1991), a quinone reductase (PcpD) (Xun and Orser, 1991; Dai et al., 2003), and a reductive dechlorinase (PcpC) (Orser et al., 1993). Following these conversions, the aromatic ring of 2,6-dichloro-p-hydroquinone is opened by 2,6-dichloro-p-hydroquinone 1,2-dioxygenase (PcpA, AAC64295) to produce 2-chloromaleylacetate (Ohtsubo et al., 1999; Xun et al., 1999), which is ultimately channelled to the tricarboxylic acid cycle for complete mineralization.
Typically, aromatic ring-cleavage dioxygenases are non-haem Fe-containing intradiol or extradiol dioxygenases, which use catechols as substrates (Harayama et al., 1992; Que and Ho, 1996). PcpA, a ∼ 37 kD dioxygenase from S. chlorophenolicum (AAC64295), contains a non-haem Fe(II)-catalytic centre like the extradiol dioxygenase, but is unique in that it does not catalyse dioxygenation of 1,2-dihydroxylated rings or have any significant sequence similarity with catechol dioxygenases. In addition, PcpA uniquely catalyses the ring-opening of 2,6-dichloro- or other 2,6-substituted p-hydroquinones (Machonkin and Doerner, 2011). Two homologues of PcpA have been reported: 2-chlorohydroquinone 1,2-dioxygenase (LinE), which is involved in γ-hexachlorocyclohexane biodegradation by Sphingomonas paucimobilis (Miyauchi et al., 1999), and hydroquinone 1,2-dioxygenase (MnpC), which is involved in 3-nitrophenol biodegradation by Cupriavidus necator JMP134 (Yin and Zhou, 2010).
The known structures of intra- or extradiol dioxygenases have not been helpful for understanding the enzymatic activity of this new class of enzymes due to their low sequence similarity with PcpA. A threaded model of PcpA has been constructed using a protein of unknown function (Machonkin et al., 2010), but details of the substrate-binding pocket are missing due to low sequence similarity. However, the model suggests that PcpA and type I catechol extradiol dioxygenases are significantly different on a structural level despite being evolutionarily related.
Understanding the key enzymes involved in PCP degradation is a critical prerequisite for successful bioremediation. In this report, we present the structure of PcpA to shed light on its biological function, reaction mechanism and structural relationship to other members in this new functional class.
Purified recombinant PcpA was crystallized in the P41212 space group and solved at 2.6 Å resolution. The structure of the PcpA molecule consisted of four similarly sized βαβββ motifs, a structural motif strongly conserved among the vicinal oxygen chelate (VOC) superfamily (He and Moran, 2011), connected by long loops between them (Fig. 1A). These repeating motifs approximately spanned residues 11–67, 84–135, 160–209 and 227–279. Although there was no significant sequence similarity among the four motifs, the Cα carbons of each motif were superimposable with rmsd values of 0.2–0.3 Å without including the intervening loop (Fig. 1B).
The β-strands of the first and fourth βαβββ motifs interacted to form an incompletely closed barrel with an 8-stranded β-sheet. Likewise, the β-strands of the second and third motifs also formed an 8-stranded β-sheet. The two twisted β-sheets constituted the core of PcpA with four flanking helices covering two open sides of the twisted β-barrel core (Fig. 1A). Most residues from the two β-sheets facing each other were distinctively hydrophobic in nature. The resulting hydrophobic core of the β-barrel was in contact with the hydrophobic side of the amphiphilic α-helices, whose hydrophilic faces together with hydrophilic loops were exposed to solvent. Consequently, the surface residues of PcpA were very hydrophilic and explained its high solubility. In addition to four similarly shaped βαβββ motifs, the C-terminal tail, comprised of one α-helix and one β-strand, participated in forming the substrate entrance site of the enzyme.
The temperature factors for most of the loop region were slightly elevated compared with those constituting secondary structural elements. In particular, the loop areas with a deletion or addition of residues among proteins of high similarity (Fig. 2) had the highest temperature factors, reflecting their intrinsic flexibility.
PcpA uses Fe(II) for catalysis (Xun et al., 1999; Sun et al., 2011) which is partially oxidized to Fe(III) during purification (Xun et al., 1999; Sun et al., 2011). Purified PcpA lost activity in a time-dependent manner (Supplementary Fig. S1) by autoxidation of Fe(II) to Fe(III) (Xun et al., 1999; Sun et al., 2011), which could be completely recovered in 48 h by reduction with 100 μM ascorbic acid. PcpA recovered from crystals that were dissolved in solution were found to be catalytically inactive, but regained full activity after reduction by ascorbic acid (Supplementary Fig. S1). The ferric ion was located inside a cleft formed between the first and fourth βαβββ motifs, and was co-ordinated by three residues, His11, His227 and Glu276. These residues occupied one face of the co-ordination sphere similar to those of Fe(II)-containing extradiol dioxygenase (Fig. 1C). The location of participating residues of the 2-His-1-carboxylate facial triad motif (Koehntop et al., 2005) observed in PcpA was different from those in other extradiol dioxygenases, considering both His11 and His 227 were the beginning residues in the first and fourth βαβββ motifs, respectively, and Glu276 was located near the end of the fourth motif (Fig. 2). The opposite face of Fe co-ordination sphere was occupied by a SO42− ion (Fig. 1C), which was present in the crystallization buffer at 1.36 M. The SO42− ion was co-ordinated to Fe in a bidentate mode and the two Fe-co-ordinating oxygen atoms of SO42− were also within hydrogen bonding distance to the hydroxyl groups of Thr13 and Tyr266 (Fig. 1C). In addition, the hydroxyl group of Thr13 established a hydrogen bond network with the hydroxyl group of Thr64 and the Nε2 atom of His49, which was in turn connected to the bulk solvent (Fig. 3).
Both Fe-co-ordinating imidazole rings of His11 and His227 were within a hydrogen bond distance from the neighbouring imidazole rings of His10 and His226 respectively. That is, the Nδ1 atoms of their imidazole rings were within hydrogen bond distance from each other. The Nε2 atoms of the imidazole rings of both His10 and His 226 were exposed to the solvent region (Fig. 3). The carboxyl side-chain of Glu276 was involved in Fe-co-ordination in a monodentate mode. The other oxygen atom of Glu276 was within a hydrogen bond distance from the hydroxyl group of both Ser264 and Thr278.
Oligomeric structure of PcpA
A crystallographic symmetry operation assembled one PcpA molecule in the asymmetric unit into a dimeric unit in the crystal lattice. This oligomeric status of PcpA as a dimer was verified in solution by a multi-angle laser light scattering (MALLS) experiment (Fig. 4A). The dimer interface had four symmetrically oriented inter-subunit salt bridges including Asp255–Lys242 and Glu169–Arg214 (Fig. 3C) suggesting that dimer formation is pH-dependent. The observed dimer interface contributed to one face of the substrate entrance, increasing the binding pocket depth (Fig. 3B).
Isothermal titration calorimetry (ITC) data
We used ITC to confirm the differential binding affinities among substituted benzene molecules. Auto-oxidized PcpA-Fe(III) was used for the titration. Although PcpA-Fe(III) did not catalyse enzymatic turnovers, it bound several tested hydroquinones. A significant amount of heat was released when hydroquinone or 2,6-dimethyl-p-hydroquinone was titrated into a PcpA containing solution (Fig. 5), indicating that the binding interactions had significant enthalpy contributions, ΔH = −39.6 and −67.2 kcal mol−1 respectively. Analysis of the corresponding ITC data also revealed unfavourable entropic contributions for the two compounds, ΔS = −107 and −199 cal mol−1 degree−1, respectively, indicating that the binding pocket was rearranged upon binding and/or very few solvent molecules were freed from the pocket. These calorimetric data were consistent with structural data that indicated there were very few solvent molecules in the substrate-binding pocket. The calculated Kd values for hydroquinone and 2,6-dimethyl-p-hydroquinone (Fig. 5, Table 1) were 5.8 and 2.1 μM respectively. However PcpA-Fe(III) did not show any apparent affinity for benzoic acid, 2,5-dihydroxybenzoquinone, 1,2-dihydroxy benzene (catechol), 4-nitrocatechol, phenol, 2,6-dichlorophenol, 2,4,6-dichlorophenol and 2,4,6-dibromophenol.
Table 1. ITC analysis of 2,6-dimethylhydroquinone (DMHQ) and hydroquinone (HQ) binding to PcpA
ΔH (kcal mol−1)
ΔS (cal mol−1 degree−1)
a Both wild-type and R259A mutant PcpA are in the Fe(III) state.
Characterization of steady-state kinetics of wild-type and mutant PcpA
To confirm the catalytic roles of the residues suggested by the crystal structure, the corresponding mutants were generated and characterized using 2,6-dichloro-p-hydroquinone as the substrate (Fig. 6A, Table 2). The similar shapes and intensities of the far-UV CD spectra indicated that the wild-type and five mutants, H10A, T13A, H49A, H226A and R259A, contained a similar amount of secondary structure (Fig. 4B). The calculated Km, kcat and kcat/Km values for wild-type PcpA were 6.7 ± 0.8 μM, 1.8 ± 0.1 s−1 and 270 s−1 mM−1 respectively, which were close to the previously reported values (Xun et al., 1999; Machonkin et al., 2010). The H10A and T13A mutants showed a substantially reduced activity compared with that of wild-type with kcat/Km values of 3.2 and 0.9 s−1 mM−1 respectively (Table 2). In addition, the H49A, H226A and R259A did not show any activity (Table 2). Contrary to previous reports (Machonkin et al., 2010), hydroquinone was also confirmed to be a substrate (Fig. 6B, Table 2) with Km and kcat values for wild-type PcpA of 242 ± 40.4 μM and 0.28 ± 0.01 s−1 respectively.
Table 2. Steady-state kinetic parameters for PcpA measured with 2,6-dichloro-p-hydroquinone (DCHQ) and hydroquinone (HQ)
kcat/Km (s−1 mM−1)
NA, no activity.
6.7 ± 0.8
1.8 ± 0.1
38.0 ± 9.0
0.12 ± 0.01
27.0 ± 5.0
0.023 ± 0.001
242 ± 40.4
0.28 ± 0.01
To establish the proper structural classification of PcpA and identify its homologues, a comparison with available structures in the protein data bank (PDB) was carried out using a Dali search (Holm and Sander, 1993) and only two significant matches were found. The most similar structure was a putative dioxygenase from Bacillus subtilis (PDB 3OAJ) with a high Z score of 39.5, followed by a metalloprotein from Bacillus cereus (PDB 1ZSW) with a Z score of 36.7. The Z score for all subsequent proteins was significantly lower than the first two hits. A blast (Altschul et al., 1997) search of the PDB revealed again that only 3OAJ and 1ZSW had significant sequence similarity to PcpA. The putative dioxygenase from B. subtilis (3OAJ) showed the highest score (144 bits), followed by the metalloprotein from B. cereus (1ZSW; 94 bits). The level of sequence identity for other identified proteins was very low and most of their blast alignments were functionally inconsistent and rather random.
A careful inspection of the superimposed 3D structures of PcpA, 3OAJ and 1ZSW displayed that most of the regions with high sequence similarity were located around the active-site pocket. Consequently, the first and fourth βαβββ motifs generally showed higher sequence similarity than the second and third motifs (Fig. 2). Specifically, the Fe(II)-co-ordinating residues and critical residues in the second co-ordination sphere were completely conserved. The intricate hydrogen bond network observed in PcpA is conserved between these two putative dioxygenases. Although those two enzymes (3OAJ and 1ZSW) have no known functions, they likely share similar activity and substrate specificity with PcpA, thus constituting a unique group of dioxygenases.
As shown in Fig. 2, PcpA also shares a high level of sequence identity with LinE (50%) from Sphingobium japonicum (Miyauchi et al., 1999), and MnpC (41%) from Cupriavidus necator JMP134 (Yin and Zhou, 2010). As predicted, our data showed that PcpA can catalyse hydroquinone ring cleavage like LinE and MnpC. Although the structures of LinE and MnpC have yet to be established, they likely share a similar substrate-binding pocket and ring-cleavage mechanism as they all can catalyse the ring-opening of hydroquinone.
Confirming the results of Dali and blast searches, the tertiary structure and amino acid sequence of PcpA were quite different from those of the type I extradiol dioxygenases. Although the Fe(II) in PcpA was co-ordinated by a 2-His-1-carboxylate facial triad like the type I extradiol dioxygenases (Lipscomb, 2008), the residues were originated from different βαβββ motifs. The VOC superfamily consists of proteins with βαβββ repeats, such as extradiol dioxygenases, glyoxylase I, fosfomycin resistance proteins and bleomycin-binding proteins (He and Moran, 2011). We propose that PcpA, LinE, MnpC and the two putative dioxygenases (1ZSW and 3OAJ) constitute a new class of VOC dioxygenases, the hydroquinone 1,2-dioxygenases.
There are two other reported hydroquinone 1,2-dioxygenases that can catalyse ring-opening reactions for unsubstituted hydroquinone: HapCD from the 4-hydroxyacetophenone degradation pathway of Pseudomonas fluorescens ACB (Moonen et al., 2008), and PnpC1C2 from the p-nitrophenol degradation pathway of Pseudomonas putida DLL-E4 (Shen et al., 2010). Although their amino acid sequences are more than ∼ 60% identical from each other, neither shares any significant sequence similarity with PcpA.
Active site of PcpA
The structure of PcpA suggested a potential mechanism by which PcpA performs its catalysis uniquely from other extradiol dioxygenases that catalyse the ring-opening of catechol (1,2-dihydroxybenzene). The substrate-binding pocket of PcpA, located on the face of the triad (His11, His227 and Glu276), was rather apolar. In particular, the somewhat symmetrically arranged hydrophobic residues, Phe66 and Phe262 were located across the Fe(II) from each other (Fig. 3A), with Phe66 especially important due to its side-chain orientation perpendicular to the ring of the modelled hydroquinone. This edge-to-face interaction between the two aromatic rings might significantly stabilize the bound substrate (Sinnokrot and Sherril, 2004; Emsley et al., 2010). Here, 2,6-dichloro-p-hydroquinone or 2,6-dimethyl-p-hydroquinone could fit with its 2,6-substitutents inserted into hydrophobic cavities (Fig. 3), which could induce a minor rearrangement of the local conformation, as suggested by the substantial decrease of both ΔS and ΔH of the di-substituted p-hydroquinones relative to those values of hydroquinone.
Hydrogen bond network
Nearly all the residues in the second co-ordination sphere were engaged in a hydrogen bond network connected to the Fe(II)-co-ordinating 2-His-1-carboxylate facial triad. For example, the carboxylate side-chain of Glu276 was hydrogen bonded to the hydroxyl group of Ser264. In addition, the imidazole rings of His10 and His226 (Fig. 3A) were hydrogen bonded to His11 and His227, respectively, establishing a His-His-Fe(II)-His-His-type network piercing through the enzyme which connected the facial triad to the solvent. Both His10 and His226 are conserved in all the aforementioned hydroquinone 1,2-dioxygenases (Fig. 2). Those arrangements could be the reason for the reported maximal activity of PcpA at pH 7.0 and its sensitivity to acidic pH (Sun et al., 2011), as the imidazole rings of the solvent-exposed His10 and His226 would become protonated at a low pH and lose their orientation effect for the imidazole rings of His11 and His227. In addition, a low pH could significantly weaken the dimer interface, which was established mainly through salt-bridges (Fig. 3C).
Both the hydroxyl and phenolic oxygen of Thr13 and Tyr266 were within hydrogen bonding distance from the two co-ordinating oxygen atoms of the SO42− ion (Fig. 1C). However, a small side-chain dihedral motion of Tyr266 could bring its phenolic oxygen within proper co-ordination distance from Fe(II), making it possible that Tyr266 could be an inner-sphere ligand. The solution of PcpA in its Fe(III) form displayed strong purple colour, but Fe(II) did not, which supports the possibility of a charge-transfer transition between a tyrosine residue and a ferric iron as in other tyrosinate-ferric proteins (Bradley et al., 1986; Klabunde et al., 1996). Alternatively, Tyr266 could co-ordinate Fe ion indirectly through a water or hydroxide ion. Thus, it is also plausible that three H2O molecules bind to Fe(II), as is generally accepted for extradiol dioxygenases (Kovaleva and Lipscomb, 2007), assisted by hydrogen bonds from the side-chains of Thr13 and Tyr266.
Both Thr13 and Tyr266 are highly conserved, and mutation of Thr13 (Table 2, Fig. 6A) and Tyr266 (Machonkin et al., 2010) resulted in a substantial decrease of enzymatic activity (Table 2, Fig. 6A). In addition, the hydroxyl oxygen of Thr13 was hydrogen bonded to the solvent-exposed imidazole of His49 through the side-chain of Thr64 and thus the pKa of its hydroxyl group could be decreased. The H49A mutant of PcpA was inactive (Table 2, Fig. 6A). In addition, there was a conserved Arg259 (Fig. 2), whose guanidinium side-chain was within a hydrogen bond distance from the sulphate oxygen that was not directly co-ordinating the ferric ion. Consequently, the side-chain of Arg259 in PcpA was inserted into the pocket probably due to SO4−2, which is not seen in the crystal structures of two closely related proteins with conserved arginine, 3OAJ and 1ZSW (Fig. 2). ITC data illustrated that the R259A mutant has substantially decreased affinity to both hydroquinone and 2,6-dimethylhydroquinone (Table 1, Supplementary Fig. S2), and had no measurable enzymatic activity (Table 2, Fig. 6A).
Contrary to the five-co-ordinate SO42−-complex Fe(III) structure, it is likely that the resting state of the Fe(II) centre of PcpA is a six-co-ordinate complex involving Tyr266, two waters (or one water and one OH−), and the 2-His-1-carboxylate facial triad (Fig. 7A). Unlike the bidentate catechol binding mode to the ferrous ion of extradiol dioxygenases (Senda et al., 1996), the hydroquinolic substrates of PcpA can only bind with one hydroxyl group with the 2,6-substituents accommodated by the hydrophobic pocket (Figs 3 and 5). Upon being docked to the hydrophobic binding pocket through their relatively polar tip (Fig. 8), the hydroquinolic ligand is co-ordinated to iron with its 1-hydroxyl group deprotonated, which is facilitated by both the departing hydroxyl group and the resulting lower dielectric constant of the active site (Fig. 7B).
The substrate-binding reaction causes changes in the electronic properties of Fe(II) together with hydrogen bond network, as noticed in the case of extradiol dioxygenases (Arciero et al., 1985; Arciero and Lipscomb, 1986; Shu et al., 1995; Sato et al., 2002; Kovaleva and Lipscomb, 2007). Similar to the proposed catalytic mechanism of extradiol dioxygenases (Kovaleva and Lipscomb, 2007), this transient change in electronic properties of Fe(II) triggers binding of O2 in proper orientation for electron transfer from hydroquinone to O2 through Fe(II), forming semiquinone and superoxide (Fig. 7C). It is likely that O2 binding is mediated by Thr13 that is located at the opposite site of Glu276. The pKa of its hydroxyl group should become favourable due to the observed network of the hydrogen bonds with Thr64 and His49 in contact with the solvent.
Unlike the observed pKa range of hydroquinone, ∼ 9–11, the pKa for p-hydroxyl group of semiquinone is typically in the range of 4–5 (Song and Buettner, 2010) and thus the semiquinone radical exists primarily in anionic form (Fig. 7C). Thus, a base is not required to deprotonate the para-hydroxyl group of the semiquinone. The adjacent Arg259 could play a role in orienting the p-oxygen in substrate/intermediate of semiquinone and benzoquinone nature.
The two radicals react to generate an alkylperoxo intermediate in a spin-allowed reaction, followed by O–O bond fission and C–C bond cleavage in the hydroquinone ring by a Criegee rearrangement (Fig. 7D) to yield a seven-membered lactone (Lipscomb, 2008) (Fig. 7E). The hydroxyl group of Thr13 could serve an acid/base role during the hydrolysis of the lactone by the second oxygen atom of O2 (Fig. 7F and G).
Substrate specificity of PcpA
The 1,4-hydroxyl groups of hydroquinone seem essential for affinity, and thus catechol, 4-nitrocatechol, phenol, 2,6-dichlorophenol, 2,4,6-trichlorophenol and 2,4,6-tribromophenol did not display any apparent affinity to PcpA (Fig. 5). The oxidation state of iron did not make any difference in the affinity of those non-binders (Supplementary Fig. S2).
The facts that ΔH of 2,6-dimethyl-p-hydroquinone binding (−67.2 kcal mol−1) was more negative than ΔH of hydroquinone binding (−39.6 kcal mol−1) (Fig. 5, Table 1) and the Km for 2,6-dimethyl-p-hydroquinone (6.65 μM) was only 2.7% of the Km for hydroquinone (242 μM) (Table 2) are indicative of increased hydrophobic interactions with the binding pocket due to its methyl substituents. In addition, binding of 2,6-dimethyl-p-hydroquinone had an extra entropic cost of −92 cal mol−1 degree−1 relative to that of the hydroquinone suggesting rearrangement of the active-site pocket by the two methyl groups.
Considering the arrangement of functional side-chains and hydrophobic nature of the PcpA active site, the varied levels of hydrophobicity between dichloro-p-hydroquinone and dimethyl-p-hydroquinone could be responsible for the apparent difference in their catalytic rates (Machonkin and Doerner, 2011). As shown in Fig. 8, the ring surface of 2,6-dichloro-p-hydroquinone is less polar, exhibiting horizontally diffused electronegativity at one side of the molecule compared with that of 2,6-dimethyl-p-hydroquinone. In addition, the resulting pKa's caused by the substituted groups on the hydroquinone could affect both binding affinity and reaction rate. For example, substituted electron-withdrawing groups on the hydroquinone ring lead to lower pKa's. The inhibitory nature of asymmetrically substituted compounds could be due to the asymmetric mode of binding into the symmetrically shaped pocket. Such binding can produce improper local rearrangement resulting in unproductive positioning for the hydroxyl group of Thr13 or Tyr266 or allow a water molecule to stay bound to Fe(II). Any of these scenarios could prevent proper O2 binding or proposed acid/base catalysis by Thr13. Alternatively, the alkylperoxo intermediate cannot proceed to Criegee rearrangement due to improper orientation.
For the first time, our results for PcpA propose a comprehensive picture of its unique catalytic mechanism that is very likely conserved among the widely distributed p-hydroquinone 1,2-dioxygenases. The p-hydroquinone and its 2,6-substituents enter a deep hydrophobic pocket formed near the dimer interface of PcpA, in which the 1-hydroxyl group binds to the Fe(II) centre. The hydrophobic substrate-binding pocket does not accommodate either hydroxy-p-hydroquiones or catechols. Knowledge of the structure reported here will be beneficial not only for the basic biochemistry of p-hydroquinone metabolism, but also for bioremediation of polychlorinated phenolic pollutants.
Chemicals were obtained from Sigma Aldrich or Fisher Scientific. Crystallization screens were obtained from Hampton Research.
Expression clones of S. chlorophenolicum PcpA in pET30 were grown in Escherichia coli BL21(DE3). Cultures were grown at 37°C to OD600 = 0.6 and induced for 4 h at 22°C with 0.5 mM isopropyl-β-d-thiogalactopyranoside. Cells were harvested by centrifugation at 4200 g and resuspended in 20 mM Tris pH 8.5. Cells were lysed by sonication and cleared by centrifugation at 27 000 g. The cleared lysate was applied to a Toyopearl DEAE-650 M column (TOSOH Biosciences) and eluted with a gradient of NaCl in the same buffer. The fraction containing PcpA (200 mM NaCl) was concentrated and exchanged into 20 mM Tris pH 7.5, applied to a Mono-Q HR 16/10 column (GE Healthcare) and eluted with a NaCl gradient in the same buffer. PcpA containing fractions eluting at about 300 mM NaCl were pooled, buffer exchanged into 20 mM Tris pH 7.5 and concentrated. For enzyme activity assays, cells expressing C-terminal His-tagged wild-type and mutant PcpA were induced and cleared cell lysate prepared as described above but in 50 mM Tris pH 8, 300 mM NaCl and 15 mM imidazole. The lysates were applied to a Ni-NTA column, the column was washed extensively with the lysis buffer, and the purified protein was eluted with the elution buffer containing 250 mM imidazole. The proteins were concentrated and buffer exchanged as described above. Purity was monitored for all protein preparations by SDS-PAGE and protein concentrations were determined by BCA assay (Thermo Scientific) using BSA as a standard.
Protein crystallization and structure determination
Crystals of PcpA were grown using the hanging drop vapour diffusion method. Purified protein at 15 mg ml−1 in 20 mM Tris pH 7.5 was mixed with an equal volume of reservoir solution and equilibrated against the same solution at 4°C or 21°C. The reservoir solution was 85 mM HEPES pH 7.5, 85 mM NaCl, 1.36 M (NH4)2SO4 and 15% glycerol. Crystals of PcpA typically appeared within 5 days. The presence of 15% v/v glycerol in the crystallization condition alleviated the need for subsequent cryoprotection, and thus crystals were directly looped and flash frozen in liquid nitrogen. The space group of PcpA was P41212 with one molecule in the asymmetric unit. Diffraction data up to 2.5 Å resolution were collected at the Berkeley Advanced Light Source (ALS, beam line 8.2.1) and were processed with the HKL2000 package (Otwinowski and Minor, 1997). The statistics for the diffraction data are listed in Table 3. Initial phasing of apo-form PcpA diffraction data was conducted by molecular replacement with the PDB co-ordinates, 3OAJ, using phenix phaser (Adams et al., 2002). Iterative model building and refinement took place using the programs coot (Emsley et al., 2010) and phenix. The co-ordinate and diffraction data have been deposited in the Protein Data Bank (RCSB ID code rcsb075953 and PDB ID code 4HUZ).
Table 3. Crystallographic data for PcpA
a Numbers in parentheses refer to the highest resolution shell.
b Rsym = ΣIh − <Ih>|/ΣIh, where <Ih> is the average intensity over symmetry equivalent reflections.
c Rcryst = Σ|Fobs − Fcalc|/ΣFobs, where summation is over the data used for refinement.
d Rfree was calculated as for Rcryst using 5% of the data that was excluded from refinement.
Multi-angle light scattering (MALS) and isothermal titration calorimetry (ITC)
Static light scattering was performed as previously described (Webb et al., 2010); however, the isocratic elution buffer for PcpA was 20 mM Tris, pH 7.5.
Isothermal titration calorimetric reactions were carried out in a VP-ITC instrument (MicroCal). The protein was prepared for ITC by dialysing into titration buffer (20 mM Tris, pH 7.5) for 24 h at 4°C. The concentration of protein in the calorimetric reaction cell was diluted to 100 μM. All titrations were performed at 25°C with a stirring speed of 300 r.p.m and 29 injections (10 μl each). Ligands were diluted into the same titration buffer and injected into the cell containing protein solution, and the heats of binding were recorded. Ligands were also titrated against buffer to account for the heats of dilution. Ligand concentrations were adjusted to obtain significant heats of binding, and the time intervals between injections were also adjusted to ensure proper baseline equilibration. All samples were degassed prior to titration.
Molecular docking was performed using AutoDock 4.2 (Morris et al., 2009) and Python Molecule Viewer (PMV) suite (Sanner, 1999). The co-ordinate of 2,6-dichlorohydroquinone was converted from pdb to pdbqt format and partial charges were added. The charge on Fe(II) was determined via gas-phase Mulliken population analysis at the CAM-B3LYP level of theory (cc-pVDZ on H and C; aug-cc-pVDZ on N, O and Cl; and a double-zeta basis with an effective core potential and relativistic pseudopotential on Fe) in Gaussian 09 using a small model of the Fe(II) site comprised of His11 and His226 approximated as 5-ethylimidazole; Glu276 as propanoate; and the ligand included as the 2,6-dichloro-p-hydroquinone dianion, with a total charge on the system of −1 and quintet spin state. Docking was performed in a rigid receptor-flexible ligand model, where full flexibility was given to the small molecule while the whole protein was kept rigid. The crystallographic sulphate ion was removed from the active site and Arg259 was mutated to alanine to decrease steric hindrance. Individual atomic affinity grids were calculated for each substrate atom type versus every atom type present in the protein. Grids were developed with uniform spacing of 0.375 Å between each grid point using the autogrid module. One hundred docked conformations were generated using Lamarckian genetic algorithm search method employed in AutoDock. The default torsion angle rotation step size was decreased to 25° in order to reduce missing the quantity of intermediate conformations and to increase the exhaustiveness of sampling with affordable computational cost. All other conformation search parameters were kept default.
Substrate surface calculations
Hydroquinone, 2,6-dichloro-p-hydroquinone, 2,6-dimethyl-p-hydroquinone, and their respective singlet monoanions (deprotonated at the 1-hydroxyl) were generated using GaussView 3.09 and optimized in Gaussian 09 (G09) at the B3LYP/cc-pVDZ level of theory with augmented functions (aug-cc-pVDZ) on all heavy atoms. All optimizations used tight convergence criteria, point group symmetry constraints (C2v for hydroquinone and Cs for all others), and were confirmed as minima via frequency analysis at the B3LYP/(aug)-cc-pVDZ level. Electrostatic potential and total electronic density grids were generated at 12 points/Bohr from the molecular self-consistent field densities using the G09 cubegen utility. Electrostatic potential surfaces were generated in GaussView 3.09 by mapping the electrostatic potential grids onto their corresponding total electron densities at isodensities of 0.01 electrons/Bohr3.
Enzyme assays of mutant PcpA
Mutant enzymes were cloned using the QuikChange Lightning Site-directed Mutagenesis Kit (Agilent Technologies), and were sequenced. Confirmed mutants were purified with the same protocol as C-terminal His-tagged wild-type PcpA. Enzymatic assays were performed in 40 mM sodium phosphate buffer (pH 7) according to an established method (Xun et al., 1999; Machonkin et al., 2010). Briefly, 2,6-dichloro-p-benzoquinone was reduced with 1 mM sodium borohydride to 2,6-dichloro-p-hydroquinone. In order to initially determine enzyme activity, assays were performed with wild-type and mutant PcpA enzyme concentrations at 2.74 μM and with 200 μM 2,6-dichloro-p-hydroquinone. For wild-type and two mutant enzymes (H10A and T13A) that showed activity, the steady-state kinetics experiments were performed by holding the enzyme concentration at 137 nM for wild-type, 2.10 μM for H10A and 6.39 μM for T13A. The concentrations of the substrate were varied from 0 to 400 μM. Ten minutes prior to reaction, enzymes were treated with 1 mM ferrous sulphate and 750 mM imidazole. Reactions were monitored at 253 nm for 30 seconds and were repeated in triplicate. The enzymatic assay for hydroquinone was performed following the previous report (Yin and Zhou, 2010) except for substrate range of 0–4000 μM.
CD spectra for each wild-type and mutant PcpA protein was measured using an AVIV 202SF spectropolarimeter (AVIV Biomedical) at 25°C, at a concentration of 5 mM in phosphate buffer saline (PBS), was recorded from 200 to 300 nm.
Original research was supported by NSF (MCB 1021148, DBI 0959778) and M.J. Murdock Charitable Trust.