Crystal structure of the γ-hydroxymuconic semialdehyde dehydrogenase from Pseudomonas sp. strainWBC-3, a key enzyme involved in para-Nitrophenol degradation

Background para-Nitrophenol (PNP) is a highly toxic compound with threats to mammalian health. The pnpE-encoded γ-hydroxymuconic semialdehyde dehydrogenase catalyzes the reduction of γ-hydroxymuconic semialdehyde to maleylacetate in Pseudomonas sp. strain WBC-3, playing a key role in the catabolism of PNP to Krebs cycle intermediates. However, the catalyzing mechanism by PnpE has not been well understood. Results Here we report the crystal structures of the apo and NAD bound PnpE. In the PnpE-NAD complex structure, NAD is situated in a cleft of PnpE. The cofactor binding site is composed of two pockets. The adenosine and the first ribose group of NAD bind in one pocket and the nicotinamide ring in the other. Conclusions Six amino acids have interactions with the cofactor. They are C281, E247, Q210, W148, I146 and K172. Highly conserved residues C281 and E247 were identified to be critical for its catalytic activity. In addition, flexible docking studies of the enzyme-substrate system were performed to predict the interactions between PnpE and its substrate γ-hydroxymuconic semialdehyde. Amino acids that interact extensively with the substrate and stabilize the substrate in an orientation suitable for enzyme catalysis were identified. The importance of these residues for catalytic activity was confirmed by the relevant site-directed mutagenesis and their biochemical characterization.

Background para-Nitrophenol (PNP), which is widely used in the manufacture of medicines, pesticides, dyes, explosives, leather coloring, wood preservatives and rubber chemicals, is highly toxic to animal's health. Through the respiratory system and skin it affects the blood, liver and central nervous system in the body. It can cause dizziness, rash, itching spam, anemia and various neurological symptoms [1]. It contains a nitroso group in the structure that strongly attracts the electron of phenyl, thus PNP is difficult for degradation in the nature [2]. Therefore, a number of studies have been triggered focusing on the microbial degradation of PNP [3][4][5][6] and two pathways have been clearly illustrated so far. One is the hydroquinone pathway that is usually found in Gramnegative bacteria. For example, in Moraxella sp, PNP monooxygenase converts PNP to hydroquinone via the potential intermediate p-benzoquinone [7,8]. The other is the hydroxyquinol (1, 2, 4-trihydroxybenzene) pathway that is preferentially found in Gram-positive bacteria. For example, in Rhodococcus opacus SAO101, PNP is converted to hydroxyquinol via 4-nitrocatechol [9,10].
In contrast to the reasonably thorough biochemical and genetic characterization of the PNP degradation, the structural mechanism of each reaction step in the pathway remains unclear. In order to understand the structural basis required for the activity of the enzyme, we embarked on structure determination of these enzymes that catalyze the PNP degradation process.
Recently, the primary X-ray analysis of PnpA in P.putida DLL-E4 was reported. In which PnpA was crystallized and the diffraction data was obtained [12]. However, there was no further structural report about the enzyme or other molecules involved in the hydroquinone pathway. PnpE belongs to NAD dependent aldehyde dehydrogenase (ALDH) super-family which can catalyze the aldehydes oxidation to corresponding carboxylic acids [13][14][15]. To date, some crystal structures of NAD (P) + -dependent ALDH family have been reported [16][17][18][19][20]. These enzymes exhibit wide differences in substrate and cofactor specificity. Sequence alignments of them with PnpE gave amino acid sequence identities ranging from 23% to 45%. Among them mitochondrial aldehyde dehydrogenase from bovine gives the highest sequence identity with PnpE.
This super family involves highly conserved residues Cys which was essential nucleophile and Glu that was the general base necessary to activate Cys for the dehydrogenase reaction. The substrate of PnpE is γhydroxymuconic semialdehyde which is unstable in air. So far the catalytic mechanism of this substrate has not been reported. In the current study, we determined the crystal structure of Pseudomonas sp. strain WBC-3 apo-PnpE at 2.7 Å resolution and its complex with NAD at 3.1 Å resolution. Through enzyme-substrate dockingguided point mutations, we identified the active site of PnpE and studied its catalytic mechanism.

Methods
Bacterial strains, chemicals, media and culture conditions Pseudomonas sp. Strain WBC-3 genomic DNA obtained from Wu Han Institute of Virology, Chinese Academy of Science. Escherichia coli strain BL21(DE3) used as expression host, which was cultured at 37°C in lysogeny broth (LB) and transformed as described by Sambrooket al. [21].
All chemical products used in the experiment were purchased from Sigma Chemical Co. (St Louis, MO, USA).

Protein expression and purification
The E.coli cells were cultured in the LB medium containing 100 μg/mL ampicillin until OD 600 reached 0.8 and were then induced with 1 mM IPTG overnight at 15°C. The cells were harvested by centrifugation. Cells lysis was achieved by sonication method. After centrifugation at 28,000 × g for 45 min, the supernatant was applied to Ni-NTA column. The His-tagged PnpE was eluted with elution buffer (10 mM Tris-HCl pH 8.0, 100 mM NaCl, and 250mM imidazole). The purification process was then followed by anion exchange on a Source-Q column and finally applied to size-exclusion chromatography with Superdex-200 column.

Enzyme activity assays
The catalytic activity of PnpE was measured by cascade reactions. The biosynthesis of γ-hydroxymuconic was performed in a total volume of 100 μL system containing 0.1 mM hydroquinone, 5.6μg hydroquinone dioxygenase PnpCD (PnpCD was expressed and purified in the same way as described in section 2.3) and 0.1 mM FeSO4 and 100mM Tris-HCl (pH 8.0) at 25°C. The absorbance of the reaction mixture was monitored at 290 nm until no further decrease was observed (about 10 min). At this point, the product solution can be used as the substrate for the next step to test PnpE activity. The assay of PnpE was performed immediately after the production of γhydroxymuconic in 100μL reaction system containing 5.2 μg purified PnpE, 10 μL γ-hydroxymuconic (produced by biosynthesis method), 0.1 mM NAD and 100 mM Tris-HCl (pH 8.0). The reaction was incubated at 25°C for 15 min. PnpE activity was determined by measuring the absorbance increase at 340 nm due to NADH formation.

Crystallization and data collection
Protein concentration was adjusted to 10 mg/mL before setting up crystallization screens at 20°C. Initially the native crystals were grown from 20% (w/v) PEG3350, 0.2 M KNO 3 at 20°C using the sitting-drop, vapor-diffusion technique. The crystallization condition for the NAD bound PnpE was 0.1 M Bicine pH 8.5, 20% PEG10000 and 1 mM NAD. Both data of the two kinds of crystals were collected on BL17U beam line at the Shanghai Synchrotron Radiation Facility (Shanghai, China) using a MAR 225 CCD detector. Crystals were equilibrated in a cryoprotection buffer containing 15% glycerol (v/v) plus reservoir buffer and then flash frozen in a 100K nitrogen stream. The diffraction images were processed with HKL2000 [23].

Structure determination and refinement
The crystal structure was solved by molecular replacement methods using the program Phaser of the CCP4 program suite [24] with bovine mitochondrial aldehyde dehydrogenase (PDB code: 1A4Z) [20] as the search model. The initial phase obtained from molecular replacement was further refined using PHENIX [25]. The model was rebuilt using COOT [26]. Data collection and structure refinement statistics are summarized in
Among all the structure-known homologues of PnpE, the bovine aldehyde dehydrogenase [20] shares the highest sequence identity (45%) with PnpE. The human aldehyde dehydrogenase [16] follows close behind with a sequence identity of 44%. The result of sequence alignment indicates these proteins may share similar structures. Structure superimposition shows bovine aldehyde dehydrogenase has a rmsd of 0.925 Å for C α with PnpE, meanwhile the C α rmsd between human aldehyde dehydrogenase and PnpE is 0.943 Å.

Structure of PnpE-NAD complex
There are also eight PnpE-NAD complexes in asymmetric unit. NAD molecule was found in all PnpE monomers (Figure 2A). The cofactor binding domain of PnpE is gathered together like a clamp that can hold the NAD ( Figure 2B).
In the PnpE-NAD complex structure, NAD is situated in a cleft of PnpE. The cofactor binding site is composed of two pockets. The adenosine and the first ribose group of NAD bind in one pocket and the nicotinamide ring in the other. There are several hydrogen bond interactions between PnpE and NAD. The adenine N1A amino group of NAD forms a hydrogen bond with OE1 of Q210 of PnpE. The ribose hydroxyl groups O2B and O3B receive hydrogen bonds from the side chain of K172 and V146, respectively. The diphosphate O2N donates a hydrogen bond to W148. The C4N accepts a hydrogen bond from C281. N7N accepts a hydrogen bond from E247 ( Figure 2C).
The PnpE structure of the PnpE-NAD complex is quite similar to apo-PnpE with a C α rmsd of 0.52 Å. The most  significant structural discrimination is observed at the loop between L246 and G252, in which the spatial position of the E247 moved 4.87 Å. This segment belongs to NAD binding region. In the apo-PnpE structure, this loop blocks the NAD binding site. In the complex structure, however, it moves to open the binding site to create enough space for NAD binding ( Figure 2C).

Structural bases for the specificity of cofactor binding
So far, both NAD and NADP have been known to be cofactors of the ALDH family. PnpE specifically uses NAD as cofactor. In order to elucidate the structural basis for the coenzyme specificity of PnpE, we compared structures of PnpE-NAD and NADP bound Betaine Aldehyde Dehydrogenase (BADH) (PDB code: 2wme) from Pseudomonas Aeruginosa [29] (Figure 3).
Among NADP dependent proteins with known structures, BADH shares the highest sequence identity of 41% with PnpE. Since the only difference between NAD and NADP is that the later has an extra phosphate group, PnpE and BADH should be able to distinguish them by this point. Both PnpE and BADH contain a two-pocket NAD (P) binding site. The first pocket is close to the molecule surface and accommodates the adenosine moiety of NAD, whereas the second, which is deeply located in the active site, can accommodate the nicotinamide ribose moiety. From the structures we can see clearly that the second pocket of BADH is wider than that of PnpE. This means BADH has extra space to accommodate the phosphate group of NADP whereas PnpE does not. On the other hand, in BADH the cofactor bind pocket surface around phosphate group is highly positive charged to match the negative charge of the phosphate group. This is not the case in PnpE. Therefore, PnpE can only use NAD rather than NADP as its cofactor.

Molecular docking results and mutagenesis analysis
The substrate γ-hydroxymuconic semialdehyde for PnpE contains an enol structure, which tends to convert into keto under aerobic condition. So it is impracticable to get holo-PnpE structure. To figure out the catalytic mechanism of PnpE without holo-PnpE structure we performed flexible docking for NAD bound PnpE and γ-hydroxymuconic semialdehyde using AUTODOCK (experimental procedure detailed in section 2.7). The docking result is illustrated in Figure 4. The substrate located in a narrow cleft close to NAD. There are nine residues (C281, E247, F150, I282, F154, H275, F447, W157, N149) probably involved in the substrate binding. C281 forms hydrogen Figure 6 Multiple sequence alignment of PnpE with Human mitochondrial Aldehyde Dehydrogenase (PDB entry 1cw3) [33], rat 10′ formyltetrahydrofolate dehydrogenase (PDB entry 2o2q) [34], sheep Aldehyde Dehydrogenase (PDB entry 1bxs) [17], Pseudomonas putida BADH (PDB entry 2wme) [29] and Thermus thermophilus Hpcc (PDB entry 2d4e). The highly conserved catalytic residues are labeled with red asterisks. bond with the aldehyde group of the substrate directly and may play a key role in the enzymatic reaction. F150, F154, F447, W157, N149, I282 and H275 probably composed a hydrophobic pocket to accommodate the substrate. However, W157 and F447 are a little farther from the substrate compared to other five residues.
To confirm the molecular docking result, nine single amino acid mutants (C281A, E247K, N149D, F150A, F154A, W157A, H275E, F447A, I282D) were generated using a two-step PCR strategy. Their enzyme activities were then measured and repeated four times ( Figure 5). The results showed the mutants of C281A, E247K, N149D, F150A, H275E and I282D completely lost their activity. The activity of F154A decreased about 60%. The activity of W157Aand F447A decreased about 20%. C281 and E247, which are highly conserved across the ALDH family members, have been known to play essential roles in the catalytic process. The loss of PnpE activity by N149D, F150A, H275E, I282D, and the significant decrease of the activity by F154A also argue that these amino acids play important roles in the substrate binding. W157 and F447 do not contact substrate directly, thus W157A and F447A only affected enzymatic activity modestly, compared to the significant effects of other seven mutations. These data are consistent with the result of AUTODOCK calculations.

Conclusions
Sequence-based alignment of PnpE and previously determined structures of ALDHs shows that two residues (E247 and C281) are highly conserved across different species ( Figure 6). These two amino acids act as the catalytic residues that transport electron, according to the mechanism elucidated in other homologous proteins [30][31][32]. In this regard, PnpE may have a similar catalytic mechanism (Figure 7). The catalytic C281 plays a role as a nucleophilic reagent. It attacks the hydroxymuconic semialdehyde substrate and forms a covalent intermediate. Next, the hydride from this intermediate transfers to NAD to form NADH and the thioacyl enzyme intermediate. The other catalytic residue E247 directly attacks the acyl-sulfur bond and releases the acid product prior to NADH dissociation.

Accession codes
The atomic coordinates and structure factors have been deposited under the accession codes: 4GO3 and 4GO4 in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/). Figure 7 The oxidation process of γ-hydroxymuconic semialdehyde to maleyacetate with NAD as cofactor by PnpE.