The Escherichia coli Periplasmic Aldehyde Oxidoreductase Is an

Articles pubs.acs.org/acschemicalbiology

The Escherichia coli Periplasmic Aldehyde Oxidoreductase Is an Exceptional Member of the Xanthine Oxidase Family of Molybdoenzymes Márcia A. S. Correia,†,⊥ Ana Rita Otrelo-Cardoso,†,⊥ Viola Schwuchow,‡ Kajsa G. V. Sigfridsson Clauss,§,∥ Michael Haumann,§ Maria Joaõ Romaõ ,† Silke Leimkühler,*,‡ and Teresa Santos-Silva*,† †

UCIBIO/Requimte, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal ‡ Institut für Biologie und Biochemie, Universität Potsdam, Am Neuen Palais 10, 14469 Potsdam, Deutschland § Freie Universität Berlin, Fachbereich Physik, 14195 Berlin, Germany S Supporting Information *

ABSTRACT: The xanthine oxidase (XO) family comprises molybdenum-dependent enzymes that usually form homodimers (or dimers of heterodimers/trimers) organized in three domains that harbor two [2Fe-2S] clusters, one FAD, and a Mo cofactor. In this work, we crystallized an unusual member of the family, the periplasmic aldehyde oxidoreductase PaoABC from Escherichia coli. This is the first example of an E. coli protein containing a molybdopterin-cytosine-dinucleotide cofactor and is the only heterotrimer of the XO family so far structurally characterized. The crystal structure revealed the presence of an unexpected [4Fe-4S] cluster, anchored to an additional 40 residues subdomain. According to phylogenetic analysis, proteins containing this cluster are widely spread in many bacteria phyla, putatively through repeated gene transfer events. The active site of PaoABC is highly exposed to the surface with no aromatic residues and an arginine (PaoC-R440) making a direct interaction with PaoC-E692, which acts as a base catalyst. In order to understand the importance of R440, kinetic assays were carried out, and the crystal structure of the PaoCR440H variant was also determined.

M

believed that E. coli was unable to synthesize other forms of dinucleotide cofactor besides MGD. However, three E. coli enzymes containing MCD were reported recently: XdhABC, XdhD, and PaoABC.9,10 Moco-dependent enzymes have been grouped into three broad families11 based on structural data, sequence alignment, and spectroscopic and biochemical data: (a) the xanthine oxidase (XO) family, which usually contain a cyanolyzable equatorial sulfur ligand coordinated to the Mo atom; (b) the sulfite oxidase (SO) family, with two oxo groups and a cysteine ligand at the Mo center; and (c) the dimethyl sulfoxide reductase (DMSOR) family, where one Mo atom is coordinated by two dithiolene groups and, in most cases, also to an amino acid side chain (Ser, Cys, Se-Cys, Asp) and an additional sulfido or oxo ligand.12−14 With few exceptions, enzymes of the XO family catalyze the oxidative hydroxylation of a diverse range of aldehydes and aromatic heterocycles in reactions that involve the cleavage of a

olybdenum is an essential element for microbial, animal, and plant life. With the exception of nitrogenase,1 all molybdoenzymes carry the molybdenum cofactor (Moco), where Mo is coordinated to the unique dithiolene moiety of a conserved tricyclic pyranopterin cofactor called molybdopterin (MPT).2,3 In eukaryotes, the pyranopterin is found in the monophosphate form (MPT), while in prokaryotes it is often conjugated to nucleosides, usually cytosine (MCD, molybdopterin cytosine dinucleotide) or guanosine (MGD, molybdopterin guanosine dinucleotide), and occasionally adenosine or inosine.2,4−6 The diversity of the pterin cofactors within the Mo-containing enzymes seems to be related to the specific family of molybdoenzymes to which it belongs. The biosynthesis of Moco and its assembly in the corresponding enzymes involves a rather complex pathway.7 In eukaryotes, the cofactor synthesis ends with the addition of molybdenum to the MPTAMP precursor, forming Mo-MPT. In prokaryotes, a further step is required for the addition of the nucleosides with the recruitment of specific proteins. In Escherichia coli, MGD is formed by the covalent addition of guanosine 5′-phosphate, GMP, to Moco by the MobA protein, while MCD is formed with the participation of MocA.8 For many years, it was © XXXX American Chemical Society

Received: July 4, 2016 Accepted: August 25, 2016

A

DOI: 10.1021/acschembio.6b00572 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Articles

ACS Chemical Biology

Figure 1. Enzymes of the xanthine oxidase family and reactions catalyzed: (a) scheme of the PaoABC heterotrimer from E. coli; (b) scheme of the Thauera aromatica 4-HBCR; (c) scheme of the Bos tauros XO.

C−H and the formation of a C−O bond (Figure 1).3 Enzymes that belong to this family are most commonly organized either as homodimers (α2) or as dimers of heterotrimers (αβγ)2. In the multisubunit modular enzymes, the redox centers are organized within independent subunits. In these cases, the α subunit usually harbors the two [2Fe-2S] centers, the β subunit harbors the flavin, and the γ subunit harbors the catalytic center (Figure 1). A few other examples have been reported for different quaternary structures of enzymes of this family such as α3 (Pseudomonas putida15), (αβ)2 (Comamonas acidovorans16), and (αβ)4 (Pseudomonas putida 8617). PaoABC is a periplasmic aldehyde oxidoreductase that oxidizes aldehydes to the corresponding carboxylic acids with a preference for aromatic aldehydes, but excluding purines (see Figure 2).18 Aromatic aldehydes are common in nature, present in plants and fruits (e.g., cinnamaldehyde), but also arise from fuel sources and air pollutants. At high doses, they can act as antimicrobial agents and are often used as food preservatives. PaoABC has high specificity toward cinnamaldehyde with a kcat of 84 ± 5 s−1.18 In 2009, Neumann et al. showed that a complete growth inhibition of E. coli devoid of genes derived from the paoABCD operon was achieved through the addition of cinnamaldehyde, probably damaging the bacterial cell surface.18 The results obtained suggest that PaoABC is part

of the immune system of E. coli with an important protective role against aromatic aldehydes. Size exclusion chromatography and small-angle X-ray scattering (SAXS) studies revealed that this 135 kDa enzyme is an αβγ heterotrimer with a large (78.1 kDa) Mococontaining PaoC subunit, a medium (33.9 kDa) FADcontaining PaoB subunit, and a small (21.0 kDa) 2×[2Fe2S]-containing PaoA subunit.10,18 The PaoA contains a twin arginine protein transport (Tat) leader peptide for translocation to the periplasm. The amino acid sequences of the three subunits of PaoABC show significant similarities to enzymes of the xanthine oxidase (XO) family (30−40% identity). The structurally characterized members of this family with higher sequence homology with PaoABC are bovine milk XO,19 XDH from Rhodobacter capsulatus,20 CO dehydrogenase from Oligotropha carboxidovorans21 and from Hydrogenophaga pseudof lava,22 quinoline oxidoreductase from Pseudomonas putida (QoR),23 and 4hydroxybenzoyl-CoA reductase (Ta4-HBCR) from Thauera aromatica.24 Ta4-HBCR is a central enzyme in the anaerobic degradation of phenolic compounds. It displays unique characteristics within the XO family such as the presence of an additional [4Fe-4S] cluster thought to participate in the inversion of the electron flow, allowing the enzyme to reduce 4hydroxybenzoyl-CoA to benzoyl-CoA.24 B



■

Articles

RESULTS AND DISCUSSION Overall Structure. The structure of PaoABC was determined at a high resolution (1.7 Å) using synchrotron radiation data. PaoABC is an αβγ heterotrimeric enzyme embodying a set of redox centers involved in electron transfer and a Mo active site, with overall dimensions of 92 × 78 × 70 Å3 and an accessible surface area of 38 035 Å2. The overall structure of the protein (Figure 3) is similar to other

Figure 3. Crystal structure of E. coli PaoABC. The cofactors are displayed as spheres and are color-coded as atom types.

structurally characterized members of the XO family with a RMSD of 2.1 Å, 2.2 Å, and 2.3 Å for the superposition with Ta4-HBCR, bovine XO, and human aldehyde oxidase (HsAOX1),27 respectively. Sequence identities are detailed in Supporting Information Figure 2. The iron−sulfur subunit (PaoA) of the protein comprises 179 residues (depicted in orange in Figure 3), with dimensions of 44 × 37 × 32 Å3 and can be divided into two subdomains, each carrying one [2Fe-2S] cluster. According to EPR experiments, the clusters were termed type I and type II as determined for the other structurally characterized homologues.18 The N-terminal domain (residues 53−134) is similar to plant-type ferredoxins and harbors the type II [2Fe-2S] cluster, which is localized next to the FAD in the PaoB subunit about 9 Å away. The type I [2Fe-2S] in subunit PaoA (residues 135−226) is deeply buried about 19 Å from the protein surface (Figure 3). The PaoC subunit (in blue in Figure 3) comprises 729 residues, contains the Moco active site, and has overall dimensions of 75 × 67 × 48 Å3. The PaoB subunit comprises 316 residues and has dimensions of 51 × 48 × 32 Å3 (in green in Figure 3). It exhibits a typical FAD binding motif conserved within the XO family and in the vicinity of the isoalloxazine moiety, toward the solvent, the structure is very similar to that in BtXDH, HsAOX1, and Ta4-HBCR. There is no deviation of the FAD variable loop (loop 430−440) in XDH that adopts a different conformation in the XDH to XO interconversion.19,27 During structure determination, the presence of an unexpected [4Fe-4S] cluster was identified in the protein. Comparison with homologous proteins showed that this cluster corresponds to the [4Fe-4S] center present in the crystal structure of Ta4-HBCR.24 The [4Fe-4S] cluster of PaoABC is

Figure 2. XAS spectra of PaoABC. (a) Normalized molybdenum Kedge spectra of PaoABC(O) and PaoABC(S) in comparison to the spectrum of XDH protein. Edge energies determined at 50% level of the normalized XANES: XDH, 20014.3 eV; PaoABC(S), 20014.4 eV; PaoABC(O), 20015.3 eV. The inset shows the pre-edge peak features in magnification. (b) Fourier-transforms (FTs) of experimental EXAFS spectra (in the inset) of PaoABC(O), PaoABC(S), and XDH. Spectra were vertically displaced for comparison. Inset: experimental data (thin lines) and fit curves (strong lines) corresponding to parameters of fits 2, 5, and 7 in Supporting Table 1.

Biochemical analyses have shown that PaoABC binds the MCD form of Moco. This is the first enzyme identified from E. coli that binds this form of the cofactor.18 So far, almost all of the characterized E. coli molybdoenzymes belong to the DMSOR family and have been shown to bind bis-MGD. The only enzyme that does not belong to this class is the E. coli MsrP protein (formerly called YedY), which belongs to the SO family and binds the Mo-MPT form of Moco25 and was recently shown to be involved in repairing proteins containing methionine sulfoxide in the bacterial cell envelope.26 In the present work, the PaoABC crystal structure solved at 1.7 Å resolution is reported and compared to the 3D structures of other members of the XO family. EXAFS studies suggest that the Mo site in PaoABC is similar to the site in XDH, showing the presence of a MoS bond in the as-isolated state. Kinetic studies and site directed mutagenesis have been carried out for a better understanding of the mode of action of this enzyme. C


Articles


Figure 4. continued

D


Articles


Figure 4. (a) Sequence alignment of the FAD domain of 16 bacterial members of the molybdenum hydroxylase family. Sequences were aligned using the Clustal omega. The blue arrows represent the four cysteines that coordinate the [4Fe-4S] cluster; the red arrows, the Arg 118 and Phe 234 probably involved in the electronic path from FAD to [4Fe-4S] cluster; and the star symbol (★), the proteins with X-ray structure deposited in the PDB database. Abbreviations: EcPaoABC, E. coli periplasmic aldehyde oxidoreductase B subunit; SbPaoABC, Shigella boydii aldehyde oxidoreductase family protein FAD subunit (UniParcKB: UPI00066BBD47); KpPaoABC, Klebsiella pneumoniae aldehyde oxidoreductase family protein FAD subunit (GeneBank: KME74100); SpPaoABC, Sodalis praecaptivus aldehyde oxidoreductase family protein FAD subunit (UniProtKB: AHF78137.1); BrPaoABC, Bradyrhizobium retamae aldehyde oxidoreductase FAD subunit (UniProtKB: A0A0R3MPP7); Af PaoABC, Asanoa ferruginea aldehyde oxidoreductase family protein FAD subunit (GeneBank: KOX54053); TaHBCR, Thauera aromatica 4-hydroxybenzoyl-CoA reductase B subunit (PDB: 1rm6); MmHBCRB, Magnetospirillum magneticum 4-hydroxybenzoyl-CoA reductase B subunit (UniProtKB: Q2W5Q0); RpHBCRB, Rhodopseudomonas palustris 4-hydroxybenzoyl-CoA reductase B subunit (UniProtKB: Q21AK7); PpQor, Pseudomonas putida quinoline 2oxidoreductase C subunit (PDB: 1t3q); HpCODH, Hydrogenophaga pseudof lava carbon monoxide dehydrogenase (CODH) C subunit (PDB: 1ffv); BtXO, Bos Tauros xanthine oxidase B subunit (PDB: 1fiq); HsAOX1, Homo sapiens aldehyde oxidase (PDB: 4uhw); HsXDH, Homo sapiens xanthine oxidase (oxidoreductase; XOR) FAD domain (PDB: 2e1q); EcXDH, E. coli xanthine oxidase FAD subunit (UniProtKB: Q46800). (b) Stereo representation of the insertion segment of the [4Fe-4S] center domain, for PaoABC (green) and 4-HBCR (gray). The insertion segment is composed of 43 amino acids from Cys(B)119 to His(B)161 and wraps the [4Fe-4S] cluster. The distances between the irons and the cysteine sulfur atoms vary between 2.2 and 2.4 Å. The anomalous map in red is at the 5 σ level.

subunits mediate the dimer formation. Ionic interactions are essential for dimerization (e.g., R802(α)/E765(α′) and E768(α)/R801(α′) in HsAOX1, R793(α)/E756(α′) and E759(α)/K792(α′) in BtXO). However, the corresponding residues are not conserved in PaoABC. It is commonly accepted that for HsAOX127 and Ta4HBCR,28 the individual subunits function independently. It was reported previously for R. capsulatus XDH that dimerization is required for Moco insertion.29,30 In general, proteins form dimers ensuring a higher stability for the protein. In the case of bacterial molybdoenzymes, there seems to be a difference in the oligomerization state depending on the cellular localization of the protein. While cytoplasmic molybdoenzymes like the bacterial XDH require formation of dimers “via” the catalytic subunit for Moco insertion, it does not seem to be necessary for all periplasmic enzymes. Here, the periplasmic enzymes often form multimers of two different subunits, with no dimerization via the Moco subunit. Examples include the formate dehydrogenase (αβ) from Desulfovibrio gigas, the formate dehydrogenase 2 (αβγ) from Desulfovibrio vulgaris, the sulfite dehydrogenase (αβ) from Starkeya novella, the TMAO reductase (α) from E. coli, and the DMSO reductase (αβγ)

embedded in a 43-residue-long polypeptide segment (PaoBC119 to PaoB-H161) positioned close to the si-face of FAD. Multiple-sequence alignments gave the highest homology scores for those Moco enzymes that also contain this insertion segment and the cysteine residues responsible for the binding of the [4Fe-4S] cluster (eg PaoABC from Shigella boydii and Klebsiella pneumonaie with 99% sequence identity; Figure 4). The majority of these enzymes are classified as XDH, and only a few as periplasmic aldehyde oxidases or 4-HBCRs. These findings suggest that the presence of a [4Fe-4S] cluster in members of the XO family of enzymes is more common than thought before and may be important for their respective physiological activity. Currently, the mode of function for most of the enzymes mentioned above is not fully clear, except for the 4-HBCRs and CODH. The three subunits of PaoABC form a stable heterotrimer, unlike the other members of the XO family so far structurally characterized, which are organized as dimers of heterotrimers (αβγ)2 (e.g., Ta4-HBCR, CODH) or as homodimers α2 (e.g., BtXOR, HsAOX1). In the case of XO and AOX, the dimer interface is predominantly between the two catalytic subunits, whereas in Ta4-HBCR, both the catalytic and the 2Fe-2S E


Articles


Figure 5. (a) Sequence alignment of the Moco domain of 14 bacterial members of the molybdenum hydroxylase family. Sequences were aligned using the Clustal omega. The blue arrows represent the most important conserved amino acids that are involved in the coordination of the MCD; the red arrows, the nonconserved amino acids; and the star symbol (★), the proteins with X-ray structure deposited on PDB database. Abbreviations: EcPaoABC, E. coli periplasmic aldehyde oxidoreductase C subunit; SbPaoABC, Shigella boydii molybdopterin-binding domain of aldehyde oxidoreductase family protein (UniParcKB: UPI00066A1DF4); KpPaoABC, Klebsiella pneumoniae molybdopterin-binding domain of aldehyde oxidoreductase family protein (UniProtKB: KME74099); SpPaoABC, Sodalis praecaptivus molybdopterin-binding domain of aldehyde oxidoreductase family protein (GeneBank: AHF78138); BrPaoABC, Bradyrhizobium retamae molybdopterin-binding domain of aldehyde oxidoreductase family protein (GeneBank: KRR22102); Af PaoABC, Asanoa ferruginea molybdopterin-binding domain of aldehyde oxidoreductase family protein (GeneBank: KOX54052); TaHBCR, Thauera aromatica 4-hydroxybenzoyl-CoA reductase A subunit (PDB: 1rm6); PpQor subB, Pseudomonas putida quinoline 2-oxidoreductase B subunit (PDB: 1t3q); HpCODH subB, Hydrogenophaga pseudof lava carbon monoxide dehydrogenase B subunit (PDB: 1ffv); DgAOR subA, Desulfovibrio gigas aldehyde oxidoreductase Moco domain (PDB: 1vlb); BtXO, Bos Tauros xanthine oxidase C subunit (PDB: 1fiq); HsAOX1B, Homo sapiens aldehyde oxidase (PDB: 4uhw); HsXDH, Homo sapiens xanthine oxidase (oxidoreductase; XOR) FAD domain (PDB: 2e1q); EcXDH, E. coli xanthine oxidase FAD subunit (UniProtKB: Q46800). (b) Scheme of the superposition of the Moco domain from PaoABC (blue), HsAOX1 (pink), Ta4-HBCR (green), BtXO (orange). PaoABC is present in a more open state with the active site easily accessible to the solvent while Ta4HBCR and BtXO are present in a more closed state.

highly conserved, especially in the five-stranded β sheets, while some divergence is observed in the C-terminal subdomain, especially in the two loops that are connecting strands β20 to β21 and α16 to β28, designated as loop 1 and loop 2, respectively (Figure 5). Sequence alignment of this region shows that these loops are shorter in PaoABC and align with the closely related molybdopterin enzymes that contain the [4Fe-4S] cluster. Loops 1 and 2 correspond to a cap at the surface of the protein that controls the solvent exposure of the active site. While in most Mo enzymes the cap protects the catalytic site from the solvent, with the metal deeply buried at the bottom of a 10−14-Å-long funnel, the same is not true for PaoABC. In this case, the cap is much shorter and the Mo active site is very exposed to the solvent. The substrate channel

from E. coli. Explanations for the lack of dimerization might be that either structural stabilization by dimerization of these proteins is not required in the periplasmic environment or that the translocation via the Tat-translocase has an impact on the oligomerization state. In the periplasmic enzymes, Moco insertion occurs before the translocation to the periplasm and includes a “proofreading” mechanism which ensures that only the fully matured protein is transported in its folded state after the cofactor insertion. The catalytic subunit of PaoABC has a heart-like shape similar to the other family members that can be further divided into two subdomains running almost perpendicular to each other. Analysis of the superposition of members of the XO family suggest that the N-terminal subdomain of the subunit is F


Articles


sequence coverage, 54 hits were found with over 80% identity, after excluding all Escherichia species from the search. For most of the enzymes, a high degree of sequence identity is also found in the Mo containing subunit (eg Shigella boydii, Klebsiella pneumonaie, Bradyrhizobium retamae, and Asanoa ferruginea with 99%, 99%, 71%, and 66% identity, respectively). To further characterize the phylogenetic relationships between these enzymes, a dendrogram was created with Phylogeny.f r31,32 using PaoB-related sequences from 67 bacterial strains, belonging to the five phyla. The sequence of the FAD containing subunit of XDH from E. coli and from Ta4-HBCR was also included as negative controls, since the first does not have a [4Fe-4S] cluster and the latter, although containing such a cluster, catalyzes a reduction reaction unrelated to aldehyde oxidation. The resulting unrooted tree is shown in Supporting Figure 3. Bootstrap analysis was also performed, and the obtained values are shown in the branch points of the tree. This analysis suggests that the function of the periplasmic enzymes containing the [4Fe-4S] cluster is markedly different from XDHs and HBCRs, which appear in a distant and separate branch of the dendrogram. In addition, this dendrogram suggests horizontal gene transfer events between unrelated species from five different bacteria phyla. Most of the enzymes used are classified as periplasmic with the corresponding Moco subunit displaying 35% to 72% sequence identity to PaoC. These findings point to the existence of an extensive number of XO-type enzymes harboring an additional [4Fe-4S] cluster. Thus, we propose that these enzymes will have a similar function and subcellular localization as PaoABC and will be involved in periplasmic oxidation of aldehydes and acting as detoxifying enzymes. A similar analysis was done for the housekeeping gene 16S RNA from the same bacterial strains as a control (data not shown). As expected, the dendrogram shows the distribution of the strains according to the corresponding phyla with no evidence for the same type of gene transfer events. The Active Site of PaoABC. PaoABC has been characterized as the only example of an E. coli Moco enzyme where cytosine is found in the dinucleotide form of the cofactor. In the present crystal structure, MCD is very welldefined in the electron density maps, where a net of hydrogen bonds contributes to the stabilization of the entire cofactor. Like in most Moco enzymes deposited in the PDB, the tricyclic pyranopterin ring system forces a conserved arginine residue in the catalytic domain, PaoC-R350, to adopt a nonfavorable conformation that is easily spotted in the Ramachandran plot, bringing its guanidinium moiety coplanar with the pterin aromatic system. In PaoABC, the Mo atom is coordinated to the dithiolene moiety, to an apical oxo ligand, an equatorial sulfido ligand, and a labile hydroxo group in a distorted square pyramidal geometry, similar to most of the enzymes of the XO family. Unlike other ligands of the metal, a high B factor was observed for the sulfur ligand of the equatorial position. This discrepancy led us to suspect a low occupancy of the sulfido ligand or its replacement by an oxo group, which would correspond to the desulfo form of the protein. Considering the cyanolyzable content of the protein (58%) previously reported,18 the sulfido ligand of PaoABC has been modeled with 60% occupancy, and the corresponding B factor is in the same range as the other ligands (12 Å2). To verify the presence of the sulfido ligand at the active site, XAS was employed to compare the molybdenum site structures

commonly observed in the structurally characterized enzymes is absent in PaoABC. The active site is located in a shallow groove, very close to the surface of the protein, unlike in Ta4HBCR where the long and narrow channel accounts for its high substrate specificity. The easy access and the absence of aromatic residues lining the active site makes this protein unique in the XO family of enzymes. The [4Fe-4S] Cluster of PaoABC. In PaoB, the [4Fe-4S] cluster is buried approximately 12 Å beneath the protein surface and is coordinated by cysteines 119, 129, 138, and 157 that belong to a 43 residue insert in the polypeptide chain. This insertion corresponds to a coiled region with no secondary structure elements except a short α-helix that is not structurally related to Ta4-HBCR (the RMSD value for the superposition of 38 Cα atoms is 6.1 Å). In PaoABC, the [4Fe-4S] cluster is ca. 17 Å distant from the isoalloxazine ring of FAD, which is a large distance for an electron transfer process. Nevertheless, as found in Ta4-HBCR, key residues may mediate the putative electron path, connecting the [4Fe-4S] cluster and the remaining redox centers toward the Mo active site. As depicted in Figure 4, PaoB-R118 and PaoB-F239 are in close distance to the [4Fe4S] cluster and the FAD, respectively, and may facilitate a possible electron transfer route between the two centers. For the case of Ta4-HBCR, the equivalent residues (R121 and F233, Ta4-HCBR numbering) were considered to be important for the reductase activity of the enzyme, contributing to the electron flux from the [4Fe-4S] center to the substrate.24 PaoABC has been classified as a detoxifying enzyme in the metabolism of aldehydes to less toxic carboxylic acids, catalyzing the hydroxylation reaction of aromatic aldehydes. The reverse reaction of the enzyme, the reduction of carboxylic acids into aldehydes was not observed for this enzyme (data not shown), and so far, the involvement of the [4Fe-4S] cluster during catalysis is not clear. One possible involvement of the [4Fe-4S] cluster might be to prevent the formation of a flavin semiquinone at the FAD site, since the formation of the semiquinone has never been detected during the enzyme reduction.18 A stable protein could not be obtained when variants with one of the four coordinating cysteines (e.g., mutations at PaoBC138 or PaoB-C157) were made, indicating that the [4Fe-4S] cluster is required to maintain the integral fold of the PaoB subunit. The amino acid sequence of PaoB was used as a query to search for other enzymes containing an FAD binding subunit where the four cysteines binding motif (CX9CX7CX18C) would also be present. It appears that the existence of subunits harboring FAD and the cysteine binding motif is common in bacterial strains even though some deviation in the number of residues separating the Cys residues is observed, as expected. On the basis of a BLAST search, these subunits belong to Mo containing enzymes predominantly found in Gram negative proteobacteria in all its subdivisions and also in other bacteria phyla such as Green Non-Sulfur bacteria (e.g., Ktedonobacter racemifer, 51% identity), nitrogen-fixing bacteria (Bradyrhizobium retamae, 80% identity), cyanobacteria (e.g., Tolypothrix campylonemoides, 53% identity), verrucomicrobia (e.g., Chthoniobacter f lavus, 52% identity), or in Gram positive species from firmicutes (e.g., Paenibacillus sophorae, 53% identity), or actinobacteria (e.g., Asanoa ferruginea and Paenibacillus sophorae, 73 and 53% identity, respectively). Considering up to 98% of G


Articles


Figure 6. (a) The Mo active site of EcPaoABC for the wildtype, (b) EcPaoABC R440H variant, (c) HsAOX1, (d) TaHBCR, (e) BtXO. The electron density map at a contour level of 1.0 σ is shown in blue. Moco and selected amino acids are displayed as balls-and-sticks and are atom color-coded.

num−oxygen interaction into two shells for PaoABC(S) yielded a short MoO bond and a longer Mo−O distance of 1.76 Å, presumably attributable to a deprotonated hydroxo group (Mo−O−). Inclusion of a long-distance Mo−C/N/O ligand improved the fit quality, which may stem from amino acid residues, solvent molecules, or oxygenated S(pterin) species in the second coordination sphere of molybdenum. For further details, see Supporting Table 1. For PaoABC(O), inclusion of a short MoS bond in the EXAFS fit yielded a low coordination number, meaning that a double-bonded sulfur was not likely to be present. Refinement of the fit approach led to a model comprising two MoO bonds, one Mo−OH bond (∼2.00 Å), and two more heterogeneous Mo−S bonds of the pterin ring for PaoABC(O). Accordingly, the sulfido ligand in PaoABC(S) was likely to be replaced by an oxygen (oxo or hydroxo) ligand in PaoABC(O). The structural attributions to the Mo sites in the PaoABC proteins (sulfurated and nonsulfurated) from XDH and XAS data are summarized in Supporting Figure 1. Analysis of active site contacts shows the Mo equatorial hydroxyl ligand within hydrogen-bonding distance (2.9 Å) to the backbone nitrogen of PaoC-G508 (for clarity, this residue is not shown in Figure 6) while the apical oxo-group (OM1) is hydrogen-bonded to the Nϵ2 of the highly conserved PaoCQ211 (2.9 Å). Moreover, PaoC-E692, implicated in the catalytic reaction mechanism of all XO-related enzymes,

in PaoABC in its as isolated state (sample referred to as PaoABC(S)) and in a sample where the equatorial sulfido ligand was removed by cyanide treatment (sample referred to as PaoABC(O); Figure 2). Both samples were compared to the well characterized XDH from R. capsulatus at pH 7.5. The XANES spectrum of PaoABC(S) was almost identical to the spectrum of XDH, suggesting a very similar first-sphere coordination of the Mo ion in both proteins (Figure 2A). The oxidation state of molybdenum in PaoABC(S) was close to the Mo(VI) level, in comparison to the Mo(VI) in XDH or other molybdenum enzymes,33−35 and further Mo compounds.36 A steeper edge slope and higher edge energy suggested significant coordination changes, i.e., formation of more molybdenum−oxygen bonds at the expense of molybdenum−sulfur bonds in PaoABC(O). The magnitudes of the pre-edge features in all protein samples suggested 5coordinated Mo ions. Analysis of EXAFS spectra (Figure 2B) of the protein samples (Supporting Table 1) revealed the structural differences of the Mo sites in PaoABC(O) and PaoABC(S). The EXAFS spectrum of PaoABC(S) was similar to the spectrum of XDH. The best-fit result for both spectra suggested two molybdenum−oxygen bonds, close to one Mo S bond (∼2.17 Å), and two Mo−S interactions from the pterin moiety (Supporting Table 1). This result was in agreement with almost quantitative sulfuration (MoS) of the molybdenum ion in PaoABC(S). Splitting of the molybdeH


Articles

ACS Chemical Biology occupies the same position as in the other structurally characterized enzymes (Figures 5 and 6). Site directed mutagenesis studies have already been performed for this particular residue, and the exchange of E692 for glutamine resulted in an inactive enzyme.18 Despite the similarities with other enzymes from the XO family, the active site of PaoABC is remarkably different. The presence of a proline (PaoC-P352) occupying a similar position where a phenylalanine (F923/F914 in HsAOX1/BtXO) or a histidine (H360 in Ta4-HBCR) are found is unexpected (Figures 5, 6 and Supporting Table 2). Those planar residues are often involved in stacking interactions with the substrate molecules. Also notable is PaoC-R440 positioned close to the catalytic PaoC-E692. The side chain of PaoC-R440 is replacing the aromatic residues typical of related enzymes (F1014/F1005 in HsAOX1/BtXOs; Figure 5 and 6). In fact, there are no aromatic side chains that could be responsible for substrate stabilization during catalysis in the active site of PaoABC. The only aromatic residues in the vicinity of the cofactor are at ca. 7−9 Å from the Mo (PaoC-H187, F247, W215, and W510), and no stacking interactions are expected. PaoC-R440 is at 2.77 Å from PaoC-E354, and these two residues are highly conserved in other putative Moco enzymes with the [4Fe-4S] binding motif at the FAD subunit (Figures 4, 5 and Supporting Table 2). In order to understand the importance of PaoC-R440 in the reaction mechanism, site directed mutagenesis was carried out, and two variants of PaoABC, PaoC-R440H and PaoC-R440 K, were produced. The variants were successfully expressed and purified and used for kinetic studies. The pH dependencies of the reaction of PaoABC variants (wild-type, PaoC-R440 K, PaoC-R440H) for the substrate benzaldehyde were determined in the presence of the electron acceptor ferricyanide. For all protein variants, the activity with benzaldehyde was detectable between pH 4.0 and pH 5.5, which is in accordance with previous reports by Neumann18 and Badalyan.37 The enzymes were not active with ferricyanide as the electron acceptor in the pH range above pH 6.0. The kinetic constants determined for the purified enzymes show that the KM values increase with increasing pH, while the kcat values are higher at low pH values. Consequently, the pH optimum for the catalytic efficiency is at pH 4.0 for the wildtype and the PaoC-R440H variant and at pH 4.5 for the PaoCR440 K variant. The KM values of the PaoC-R440 K variant were twice as high in comparison to the wild-type, while the kcat for this variant was about 3-times reduced. The PaoC-R440H variant showed a slightly increased KM and a slightly lower kcat in comparison to the wild-type (Figure 7, Table 1). The crystal structure of the PaoC-R440H variant was also determined to 2.3 Å resolution, and the refined model is almost identical to the wild-type structure (RMSD of 0.17 Å for the superposition of 1212 Cα atoms). Clear electron density was found for the imidazole side chain (H440), which is hydrogen bonded to the PaoC-E692 at 2.88 Å (Figure 6). In the crystal structure of this variant, the side chain of PaoC-E354 is in a different position compared to the wild-type, which may suggest the involvement of this conserved glutamate in the reaction mechanism. Furthermore, additional electron density was found close to the active site, filling part of the substrate groove, and it could be interpreted and refined as an acetate ion, probably from the buffer solution. Other possible anions like bicarbonate cannot

Figure 7. (a) pH dependencies of overall reaction for EcPaoABC wildtype, (b) R440H variant, and (c) R440 K variant. Steady-state kinetic analyses were carried out with ferricyanide as electron acceptor, benzaldehyde as substrate, and 10 nM EcPaoABC. The consumption of ferricyanide was followed by recording the decrease of absorbance at 420 nm under anaerobic conditions. KM (open triangle) and kcat (filled circles) were calculated according to the Michaelis−Menten equation, and pH optima were fitted by the Gauss equation.

be excluded. This negatively charged ion is 2.94 Å from the positively charged PaoC-R440 and interacting with the labile hydroxyl ligand of the active site through hydrogen bonds mediated by a water molecule (Figure 6). The position and orientation of the ion suggests the position of the substrate in the active site during catalysis, and in fact, this anion is absent in the crystal structure of the PaoC-R440H mutant. In order to understand the interaction between PaoABC and its physiological substrates, several attempts have been carried out to obtain the crystal structure of a protein−ligand complex using substrate analogs or putative inhibitors. However, structures of complexes have not been obtained yet, and all attempts resulted in the ligand-free form of PaoABC. The presence of a negatively charged ion putatively blocking the access toward the active site and/or its high degree of solvent I


Articles


Table 1. pH Dependence of the Overall Reaction of the E. coli PaoABC Variants with Ferricyanide As Electron Acceptora EcPaoABC wildtype pH 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0

KM (μM) 659.3 306.7 158.9 129.2 n.d. n.d. n.d. n.d. n.d.

± ± ± ±

80.2 44.21 35.4 15.5

kcat (s−1) 1097.6 486.0 217.2 150.7 n.d. n.d. n.d. n.d. n.d.

± ± ± ±

41.2 19.5 11.6 7.6

EcPaoABC-R440H kcat/KM (μM−1 s−1) 1.66 1.58 1.37 1.17

KM (μM) 513.9 395.9 288.0 59.8 n.d. n.d. n.d. n.d. n.d.

± ± ± ±

86.5 89.0 60.1 6.7

EcPaoABC-R440 K

kcat (s−1)

kcat/KM (μM−1 s−1)

± ± ± ±

0.79 0.74 0.52 0.81

404.5 293.0 150.0 48.7 n.d. n.d. n.d. n.d. n.d.

19.2 17.4 7.7 1.0

KM (μM) 1517.4 520.1 361.6 204.2 n.d. n.d. n.d. n.d. n.d.

± ± ± ±

252.2 59.2 96.3 45.1

kcat (s−1)

kcat/KM (μM−1 s−1)

± ± ± ±

0.20 0.62 0.22 0.20

306.8 324.5 80.48 41.13 n.d. n.d. n.d. n.d. n.d.

17.6 9.0 4.8 1.7

a The kinetic constants were determined under steady state conditions and anaerobic conditions with different concentrations of the substrate benzaldehyde. The reduction of ferricyanide was measured at 420 nm. n.d. = not detectable.

■

exposure might be responsible for the difficulties in obtaining the structure of a ligand-bound form of the enzyme. Conclusions. Escherichia coli PaoABC is an enzyme with potential application in industrial biocatalysis,38 bioanalytical applications, and development of bioelectronic devices.37,39 In this work, we structurally characterized the molybdenumcontaining iron−sulfur flavoprotein PaoABC from E. coli. During refinement of the crystal structure, we could identify the presence of an unanticipated [4Fe-4S] cluster in the FAD subunit (PaoB), whose presence was unambiguous after calculation and interpretation of the anomalous maps. This cluster has also been found in Ta4-HBCR, whose structure is known for the Thauera aromatica enzyme.24 However, the two enzymes are unrelated in terms of cellular localization and reaction catalyzed. Nevertheless, a high sequence similarity with other Mococontaining enzymes with unknown functions has been shown. These enzymes possess the four cysteine residues required for the binding of the [4Fe-4S] center in an approximately 40 residue insert of the FAD subunit. Moreover, several of those enzymes have a signaling peptide for transport to the periplasm, show shortened loops 1 and 2 at the substrate channel entrance, and contain an arginine and a glutamate close to the active site (corresponding to R440 and E354 in PaoC). The similarities suggest a common function for this group of enzymes, that is, the remarkably efficient oxidation of aromatic aldehydes, as observed for PaoABC, although its real physiological function is yet to be determined. The numerous enzymes identified as putative XO-related and likely to bind a [4Fe-4S] cluster largely exceeded what was known in the field. Up to now, 4-HBCRs were the only examples reported in the literature. Phylogenetic analysis suggests that horizontal gene transfer events spread the PaoABC gene throughout different bacterial phyla over time. These events are usually associated with adaptation, conferring some evolutionary advantage to the organism. In the case of HBCRs, the importance of the cluster in the electron transfer chain for the reduction reaction is very clear, but not for the oxidation reaction, and further studies are necessary. This work shows that PaoABC is widespread in bacteria, commonly, but not exclusively found in proteobacteria. Its putative role in detoxification activities makes this enzyme an interesting target for fermentation of products from lignocellulosic biomass.38

EXPERIMENTAL METHODS

Site-Directed Mutagenesis, Protein Purification, and Quantification. Site-directed mutagenesis was used to produce two variants of PaoABC at the position arginine 440 in subunit C by exchanging it to a histidine (PaoC-R440H) or a lysine (PaoC-R440 K). The base pair exchanges were introduced in the paoC gene on plasmid pMN10018 by PCR mutagenesis using a Herculase II-polymerase kit (Fermentas Life Science) resulting in the plasmids pVS21/paoCR440H and pVS22/paoC-R440 K. The complementary primers containing the base pair exchanges were the following: PaoCR440H-fw primer cgg tgt tgc ggc ggg ctt tca caa taa tct gc, PaoCR440H-rv primer gca gat tat ttt taa agc ccg ccg caa cac cg, PaoC-R440 K-fw primer cgg tgt tgc ggc ggg ctt taa aa taa tct gc, and PaoC-440 K-rv primer gca gat tat tgt gaa agc ccg ccg caa cac cg. The E. coli TP1000 (ΔmobAB) strain was used for homologous expression of the PaoABC wildtype and the generated variants, as reported previously.18 The purification of the enzyme variants was performed after the protocol established for PaoABC wildtype protein.18 The wildtype enzyme and variants were concentrated to 20 mg mL−1 in 50 mM Tris-HCl (pH 7.5), 250 mM NaCl, and 1 mM EDTA using a Vivaspin 20 ultrafiltration device (Sartorius Stedim Biotech S.A). The final concentration was determined from the absorbance at 445 nm, using an extinction coefficient of 23 686 M−1cm−1. Crystallization. Crystallization experiments were performed using the hanging drop vapor-diffusion method, and drops were prepared at 277 K by adding 1 μL of pure protein to 1 μL of precipitating solution containing 20% (w/v) polyethylene glycol 3350 and 0.2 M ammonium iodide from the PEG/ION 1 screen (Hampton).10 Yellow-brownish crystals with maximum dimensions of 0.08 mm × 0.2 mm × 0.2 mm appeared within three to four days of both the wildtype and the PaoCR440H variant. Data Collection, Structure Determination, and Refinement. The harvested crystals were flash-cooled in liquid nitrogen using a cryoprotectant with 30% (v/v) glycerol, 0.2 M ammonium iodide, and 22% (w/v) polyethylene glycol 3350 and maintained at 100 K under a stream of nitrogen gas during data collection. Complete data sets were collected at the PXIII (X06DA) beamline of the Swiss Light Source (SLS, Villigen PSI, Switzerland) and Proxima I beamline at Soleil (Saint Aubin, France) for the wildtype and the BM14 beamline of the European Synchrotron Radiation Facility (ESRF, Grenoble, France) for the PaoC-R440H variant. The crystals diffracted up to 1.7 Å at the Soleil synchrotron for the wildtype and up to 2.3 Å at the ESRF for the PaoC-R440H variant. The data were processed with iMOSFLM v.1.0.740 and SCALA41 from the CCP4 program package v. 6.3.0.42 The data-collection and processing statistics are presented in Table 2. Structure determination of wildtype PaoABC was performed with PHASER43 using several molecular models according to sequence alignment homologies.10 A molecular replacement solution was obtained for the three subunits, and clear electron density was observed for the entire protein as well as for the expected cofactors MCD, two distinct [2Fe-2S] clusters and one FAD, that had been J


Articles


isolated sample is designated as PaoABC(S), while the second sample was treated with 1 M KCN for 15 min at RT to remove the cyanolysable sulfur. This inactive protein is designated as PaoABC(O). Both samples were concentrated by ultrafiltration (Amicon Ultra; Merck Millipore Corporation) to a molybdenum concentration of 1 mM. The molybdenum concentration was quantified by inductively coupled plasma−optical emission spectroscopy (ICP-OES; Optima 2100 DV, PerkinElmer Life and Analytical Sciences) against the multielement Standard XVI (Merck), as described previously.18 XAS spectra were measured at the Samba beamline at Soleil (Paris, France) and at the SuperXAS beamline at Swiss Light Source (SLS at Paul Scherrer Institute, Villigen, Switzerland) with the synchrotrons operated in top-up mode (400 mA). At Samba, a Si[111] doublecrystal monochromator and 7-element Ge-detector and at SuperXAS a Si[111] channel-cut monochromator and 13-element Ge-detector were used. Samples were held in liquid-helium cryostats at 20 K. A total of 4−7 monochromator scans, each on a fresh sample spot, were averaged after energy calibration using a Mo foil as a standard. XAS spectra were normalized, and EXAFS oscillations were extracted and simulated using the in-house software SimX as previously described.33 Fourier-transforms (FTs) of EXAFS spectra were calculated for k values of 2.2−14.1 Å−1 using cos2 windows extending over 10% at both k-range ends. Kinetic Analysis. To determine the enzymatic activity of PaoABC, the time-dependent reduction of ferricyanide was recorded for 1 min at 420 nm using a Shimadzu UV-2180 photometer (Shimadzu). The reaction was carried out under anaerobic conditions in a Coy-chamber. The final reaction mix of 1 mL containing 1 mM ferricyanide and benzaldehyde as a substrate in McIlvaine buffer (consisting of varying amounts of Na2HPO4 and C6H8O7) at pH 4.0−8.0 was incubated at RT, and the enzyme activity was measured after the addition of 10 nM recombinant protein. For steady state kinetic analyses, eight concentrations of benzaldehyde (0.1 mM to 10 mM) were used. Data from three different measurements were averaged and evaluated according to the Michaelis−Menten equation using the OriginPro 8.1 g SR1 software (OriginLab Corporation).

Table 2. Crystallographic and Refinement Data of PaoABC Wildtype and PaoC-R440H Variant from E. colia EcPaoABC wildtype detector wavelength (Å) a (Å) b (Å) c (Å) β (deg) space group molecules per ASU Matthews coefficient (Å3 Dalton−1) mosaicity (deg) resolution range (Å) ⟨I/σI⟩ Wilson B factor Rmerge (%)* Rpim (%)+ multiplicity no. of observed reflections no. of unique reflections completeness (%) solution method Rwork Rfree RMSD bonds (Å) RMSD bonds (deg) average B (Å2) residues per ASU ligands per ASU no. of water molecules total no of atoms Ramachandran favored (%) Ramachandran outliers (%)

EcPaoABC-R440H

PILATUS 6M 0.976 109.68 78.34 151.91 99.69 C2 1 2.39

PILATUS 2M 0.979 109.84 78.26 151.73 99.93 C2 1 2.39

0.14 48.32−1.70 (1.73−1.70) 9.69 (1.90) 16.05 10.9 (86.1) 8.6 (69.5) 4.4 (4.4) 605876 (30312)

0.35 48.27−2.37(2.30−2.30) 8.7 (1.81) 22.76 21.1 (82.6) 12.8 (65.1) 7.2 (4.9) 397602 (19057)

137343 (6876)

55360 (3851)

98.53 (99.04) MR 0.1372 (0.2509) 0.1660 (0.2676) 0.015 1.653 21.20 1222 [4Fe-4S], 2[2Fe-2S], FAD, MCNb, MOSc 987 10499 98

98.0 (84.2) MR 0.1683 (0.2588) 0.2165 (0.2880) 0.010 1.399 25.30 1223 [4Fe-4S], 2[2Fe-2S], FAD, MCNb, MOSc 765 10041 98

0.16

0.16

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acschembio.6b00572. Supporting Figure 1−3 and Tables 1 and 2 (PDF)

a

Statistics for the highest-resolution shell are shown in parentheses. MCN = molybdopterin cytosine dinucleotide. c MOS = dioxothiomolybdenum(VI) ion. b

Accession Codes

Coordinates and observed structure factor amplitudes have been deposited in the Protein Data Bank under the accession codes 5G5G for wildtype PaoABC and 5G5H for PaoC-R440H variant.

removed from the search models. A density modification protocol was applied using DM44 giving initial phases with a ca. 0.7 mean figure of merit. Restrained refinement was performed with REFMAC 5.2,45 and inspection of the electron density maps was carried out using COOT.46 During refinement, extra electron density could be observed not far from the FAD site. Inspection of the Fo−Fc and the anomalous maps (3.0σ and 5.0σ, respectively) revealed strong electron density peaks. We interpreted the peaks as corresponding to four iron atoms of a [4Fe-4S] center, although its presence had not been anticipated for PaoABC. In the final stages of refinement, the R-work and R-free converged to 14.2% and 17.5% for the wildtype enzyme and 16.1% and 21.5% for the PaoC-R440H variant, respectively. Geometrical validation and model improvement were carried out using PDBREDO47,48 and several validation programs such as PROCHECK49 and MOLPROBITY.50 Analysis of both models (wildtype and variant) showed that 98.0% of the protein residues are in the most favored or additionally allowed regions of the Ramachandran plot, while only 0.16% are in disallowed regions. Refinement statistics are summarized in Table 2. X-Ray Absorption Spectroscopy. For XAS studies, two samples of PaoABC were prepared in 50 mM Tris-HCl at pH 7.5. The as-

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Present Address ∥

MAX IV Laboratory, Lund University, 22100 Lund, Sweden.

Author Contributions ⊥

These authors contributed equally to this work

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was financially supported by Fundaçaõ para a Ciência e Tecnologia, through project PTDC/BIA-PRO/ 118377/2010, PTDC/BBB-BEP/1185/2014, PEst-C/EQB/ LA0006/2013, UID/Multi/04378/2013, and grants SFRH/ K


Articles


linked changes in molybdopterin coordination. Science 272, 1615− 1621. (15) Woolfolk, C. A. (1985) Purification and properties of a novel ferricyanide-linked xanthine dehydrogenase from Pseudomonas putida 40. J. Bacteriol. 163, 600−609. (16) Xiang, Q., and Edmondson, D. E. (1996) Purification and characterization of a prokaryotic xanthine dehydrogenase from Comamonas acidovorans. Biochemistry 35, 5441−5450. (17) Hettrich, D., and Lingens, F. (1991) Microbial metabolism of quinoline and related compounds. VIII. Xanthine dehydrogenase from a quinoline utilizing Pseudomonas putida strain. Biol. Chem. HoppeSeyler 372, 203−211. (18) Neumann, M., Mittelstädt, G., Iobbi-Nivol, C., Saggu, M., Lendzian, F., Hildebrandt, P., and Leimkühler, S. (2009) A periplasmic aldehyde oxidoreductase represents the first molybdopterin cytosine dinucleotide cofactor containing molybdo-flavoenzyme from Escherichia coli. FEBS J. 276, 2762−2774. (19) Enroth, C., Eger, B. T., Okamoto, K., Nishino, T., and Pai, E. F. (2000) Crystal structures of bovine milk xanthine dehydrogenase and xanthine oxidase: structure-based mechanism of conversion. Proc. Natl. Acad. Sci. U. S. A. 97, 10723−10728. (20) Truglio, J. J., Theis, K., Leimkühler, S., Rappa, R., Rajagopalan, K. V., and Kisker, C. (2002) Crystal structures of the active and alloxanthine-inhibited forms of xanthine dehydrogenase from Rhodobacter capsulatus. Structure 10, 115−125. (21) Dobbek, H., Gremer, L., Kiefersauer, R., Huber, R., and Meyer, O. (2002) Catalysis at a dinuclear [CuSMo(O)OH] cluster in a CO dehydrogenase resolved at 1.1-A resolution. Proc. Natl. Acad. Sci. U. S. A. 99, 15971−15976. (22) Hänzelmann, P., Dobbek, H., Gremer, L., Huber, R., and Meyer, O. (2000) The effect of intracellular molybdenum in Hydrogenophaga pseudoflava on the crystallographic structure of the seleno-molybdoiron-sulfur flavoenzyme carbon monoxide dehydrogenase. J. Mol. Biol. 301, 1221−1235. (23) Bonin, I., Martins, B. M., Purvanov, V., Fetzner, S., Huber, R., and Dobbek, H. (2004) Active site geometry and substrate recognition of the molybdenum hydroxylase quinoline 2-oxidoreductase. Structure 12, 1425−1435. (24) Unciuleac, M., Warkentin, E., Page, C. C., Boll, M., and Ermler, U. (2004) Structure of a xanthine oxidase-related 4-hydroxybenzoylCoA reductase with an additional [4Fe-4S] cluster and an inverted electron flow. Structure 12, 2249−2256. (25) Loschi, L., Brokx, S. J., Hills, T. L., Zhang, G., Bertero, M. G., Lovering, A. L., Weiner, J. H., and Strynadka, N. C. J. (2004) Structural and biochemical identification of a novel bacterial oxidoreductase. J. Biol. Chem. 279, 50391−50400. (26) Gennaris, A., Ezraty, B., Henry, C., Agrebi, R., Vergnes, A., Oheix, E., Bos, J., Leverrier, P., Espinosa, L., Szewczyk, J., Vertommen, D., Iranzo, O., Collet, J. F., and Barras, F. (2015) Repairing oxidized proteins in the bacterial envelope using respiratory chain electrons. Nature 528, 409−412. (27) Coelho, C., Foti, A., Hartmann, T., Santos-Silva, T., Leimkühler, S., and Romão, M. J. (2015) Structural insights into xenobiotic and inhibitor binding to human aldehyde oxidase. Nat. Chem. Biol. 11, 779−83. (28) Boll, M. (2005) Key enzymes in the anaerobic aromatic metabolism catalysing Birch-like reductions. Biochim. Biophys. Acta, Bioenerg. 1707, 34−50. (29) Leimkühler, S., and Klipp, W. (1999) Role of XDHC in Molybdenum Cofactor Insertion into Xanthine Dehydrogenase of Rhodobacter capsulatus. J. Bacteriol. 181, 2745−2751. (30) Schumann, S., Saggu, M., Möller, N., Anker, S. D., Lendzian, F., Hildebrandt, P., and Leimkühler, S. (2008) The mechanism of assembly and cofactor insertion into Rhodobacter capsulatus xanthine dehydrogenase. J. Biol. Chem. 283, 16602−16611. (31) Dereeper, A., Audic, S., Claverie, J. M., and Blanc, G. (2010) BLAST-EXPLORER helps you building datasets for phylogenetic analysis. BMC Evol. Biol. 10, 1−6.

BPD/64917/2009 (M.C.) and SFRH/BD/85806/2012 (A.O.C.) and cofinanced by the ERDF under the PT2020 Partnership Agreement (POCI-01-0145-FEDER-007728). Further financial support was provided by the Deutsche Forschungsgemeinschaft (DFG) grant LE1171/8-1 (S.L.) and the DAAD-PPP programme (M.J.R., T.S.-S., S.L.). M. Haumann thanks the Deutsche Forschungsgemeinschaft (DFG) for a Heisenberg Fellowship and for funding (grants Ha3265/3-1 and Ha3265/6-1). K.C. thanks “Stiftelsen Bengt Lundqvist minne” and the Wenner-Gren Foundation for fellowships. The authors thank the BM14 staff of the ESRF (France), PXIII staff from SLS (Switzerland), and Proxima I staff beamline at Soleil (France) for assistance during diffraction data collection. We thank M. Nachtegaal at SuperXAS of SLS and V. Briois and E. Fonda at Samba of Soleil for excellent technical support in XAS. We acknowledge S. Najmudin for the critical reading and reviewing of the manuscript.

■

REFERENCES

(1) Lawson, D. M., and Smith, B. E. (2002) Molybdenum nitrogenases: a crystallographic and mechanistic view, Metal Ions in Biological Systems, CRC Press, Boca Raton, FL, Vol 39, pp 75−119, DOI: 10.1201/9780203909331.ch3. (2) Rajagopalan, K. V., and Johnson, J. L. (1992) The pterin molybdenum cofactors. J. Biol. Chem. 267, 10199−10202. (3) Romão, M. J. (2009) Molybdenum and tungsten enzymes: a crystallographic and mechanistic overview. Dalton Trans. 21, 4053− 4068. (4) Bö r ner, G., Karrasch, M., and Thauer, R. K. (1991) Molybdopterin adenine dinucleotide and molybdopterin hypoxanthine dinucleotide in formylmethanofuran dehydrogenase from Methanobacterium thermoautotrophicum (Marburg). FEBS Lett. 290, 31−34. (5) Johnson, J., Indermaur, L., and Rajagopalan, K. (1991) Molybdenum cofactor biosynthesis in Escherichia coli. Requirement of the chlB gene product for the formation of molybdopterin guanine dinucleotide. J. Biol. Chem. 266, 12140−12145. (6) Johnson, J. L., Bastian, N. R., and Rajagopalan, K. V. (1990) Molybdopterin guanine dinucleotide: a modified form of molybdopterin identified in the molybdenum cofactor of dimethyl sulfoxide reductase from Rhodobacter sphaeroides forma specialis denitrificans. Proc. Natl. Acad. Sci. U. S. A. 87, 3190−3194. (7) Leimkühler, S., Wuebbens, M. M., and Rajagopalan, K. V. (2011) The History of the Discovery of the Molybdenum Cofactor and Novel Aspects of its Biosynthesis in Bacteria. Coord. Chem. Rev. 255, 1129− 1144. (8) Neumann, M., and Leimkühler, S. (2011) The Role of SystemSpecific Molecular Chaperones in the Maturation of Molybdoenzymes in Bacteria. Biochem. Res. Int. 2011, 1−13. (9) Neumann, M., Seduk, F., Iobbi-Nivol, C., and Leimkühler, S. (2011) Molybdopterin dinucleotide biosynthesis in Escherichia coli: identification of amino acid residues of molybdopterin dinucleotide transferases that determine specificity for binding of guanine or cytosine nucleotides. J. Biol. Chem. 286, 1400−1408. (10) Otrelo-Cardoso, A. R., da Silva Correia, M. A., Schwuchow, V., Svergun, D. I., Romão, M. J., Leimkühler, S., and Santos-Silva, T. (2014) Structural data on the periplasmic aldehyde oxidoreductase PaoABC from Escherichia coli: SAXS and preliminary X-ray crystallography analysis. Int. J. Mol. Sci. 15, 2223−2236. (11) Hille, R. (1996) The Mononuclear Molybdenum Enzymes. Chem. Rev. 96, 2757−2816. (12) Hille, R., Hall, J., and Basu, P. (2014) The mononuclear molybdenum enzymes. Chem. Rev. 114, 3963−4038. (13) Kisker, C., Schindelin, H., and Rees, D. C. (1997) Molybdenumcofactor-containing enzymes: structure and mechanism. Annu. Rev. Biochem. 66, 233−267. (14) Schindelin, H., Kisker, C., Hilton, J., Rajagopalan, K. V., and Rees, D. C. (1996) Crystal structure of DMSO reductase: RedoxL


Articles

ACS Chemical Biology (32) Dereeper, A., Guignon, V., Blanc, G., Audic, S., Buffet, S., Chevenet, F., Dufayard, J. F., Guindon, S., Lefort, V., Lescot, M., Claverie, J. M., and Gascuel, O. (2008) Phylogeny.fr: robust phylogenetic analysis for the non-specialist. Nucleic Acids Res. 36, W465−W469. (33) Havelius, K. G. V, Reschke, S., Horn, S., Döring, A., Niks, D., Hille, R., Schulzke, C., Leimkühler, S., and Haumann, M. (2011) Structure of the molybdenum site in YedY, a sulfite oxidase homologue from Escherichia coli. Inorg. Chem. 50, 741−748. (34) Reschke, S., Sigfridsson, K. G., Kaufmann, P., Leidel, N., Horn, S., Gast, K., Schulzke, C., Haumann, M., and Leimkühler, S. (2013) Identification of a bis-molybdopterin intermediate in molybdenum cofactor biosynthesis in Escherichia coli. J. Biol. Chem. 288, 29736− 29745. (35) Schrapers, P., Hartmann, T., Kositzki, R., Dau, H., Reschke, S., Schulzke, C., Leimkühler, S., and Haumann, M. (2015) Sulfido and cysteine ligation changes at the molybdenum cofactor during substrate conversion by formate dehydrogenase (FDH) from Rhodobacter capsulatus. Inorg. Chem. 54, 3260−3271. (36) Samuel, P. P., Horn, S., Döring, A., Havelius, K. G. V., Reschke, S., Leimkühler, S., Haumann, M., and Schulzke, C. (2011) A crystallographic and Mo K-Edge XAS study of molybdenum oxo bis-, mono-, and non-dithiolene complexes − First-sphere coordination geometry and noninnocence of ligands Eur. Eur. J. Inorg. Chem. 2011, 4387−4399. (37) Badalyan, A., Neumann-Schaal, M., Leimkühler, S., and Wollenberger, U. (2013) A biosensor for aromatic aldehydes comprising the mediator dependent PaoABC-Aldehyde oxidoreductase. Electroanalysis 25, 101−108. (38) McKenna, S. M., Leimkühler, S., Herter, S., Turner, N. J., and Carnell, A. J. (2015) Enzyme cascade reactions: synthesis of furandicarboxylic acid (FDCA) and carboxylic acids using oxidases in tandem. Green Chem. 17, 3271−3275. (39) Badalyan, A., Dierich, M., Stiba, K., Schwuchow, V., Leimkühler, S., and Wollenberger, U. (2014) Electrical wiring of the aldehyde oxidoreductase PaoABC with a polymer containing osmium redox centers: biosensors for benzaldehyde and GABA. Biosensors 4, 403− 421. (40) Powell, H., Leslie, A., and Battye, G. (2007) Mosflm 7.0.1 and its new interface: iMosflm 0.5.3, CCP4 Newsl Protein Crystallogr., BBSRC, Swindon, England, vol 46. (41) Kabsch, W. (1988) Automatic indexing of rotation diffraction patterns. J. Appl. Crystallogr. 21, 67−72. (42) Collaborative Computational Project, N. 4. The CCP4 suite: programs for protein crystallography. Acta Crystallogr., Sect. D: Biol. Crystallogr. 1994, 50, 760−3, DOI: 10.1107/S0907444994003112 (43) Read, R. J. (2001) Pushing the boundaries of molecular replacement with maximum likelihood. Acta Crystallogr., Sect. D: Biol. Crystallogr. 57, 1373−1382. (44) Cowtan, K. D., and Zhang, K. Y. (1999) Density modification for macromolecular phase improvement. Prog. Biophys. Mol. Biol. 72, 245−270. (45) Murshudov, G. N., Vagin, A. A., and Dodson, E. J. (1997) Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr., Sect. D: Biol. Crystallogr. 53, 240−255. (46) Emsley, P., and Cowtan, K. (2004) Coot: model-building tools for molecular graphics. Acta Crystallogr., Sect. D: Biol. Crystallogr. 60, 2126−2132. (47) Joosten, R. P., Joosten, K., Murshudov, G. N., and Perrakis, A. (2012) PDB_REDO: constructive validation, more than just looking for errors. Acta Crystallogr., Sect. D: Biol. Crystallogr. 68, 484−496. (48) Joosten, R. P., Salzemann, J., Bloch, V., Stockinger, H., Berglund, A.-C., Blanchet, C., Bongcam-Rudloff, E., Combet, C., Da Costa, A. L., Deleage, G., Diarena, M., Fabbretti, R., Fettahi, G., Flegel, V., Gisel, A., Kasam, V., Kervinen, T., Korpelainen, E., Mattila, K., Pagni, M., Reichstadt, M., Breton, V., Tickle, I. J., and Vriend, G. (2009) PDB_REDO: automated re-refinement of X-ray structure models in the PDB. J. Appl. Crystallogr. 42, 376−384.

(49) Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton, J. M. (1993) PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr. 26, 283−291. (50) Chen, V. B., Arendall, W. B., Headd, J. J., Keedy, D. A., Immormino, R. M., Kapral, G. J., Murray, L. W., Richardson, J. S., and Richardson, D. C. (2010) MolProbity: all-atom structure validation for macromolecular crystallography. Acta Crystallogr., Sect. D: Biol. Crystallogr. 66, 12−21.

M


The Escherichia coli Periplasmic Aldehyde Oxidoreductase Is an

Recommend Documents