Structural Basis of Stereospecific Vanadium-Dependent Haloperoxidase Family Enzymes in Napyradiomycin Biosynthesis

Abstract

Vanadium-dependent haloperoxidases (VHPOs) from Streptomyces bacteria differ from their counterparts in fungi, macroalgae, and other bacteria by catalyzing organohalogenating reactions with strict regiochemical and stereochemical control. While this group of enzymes collectively uses hydrogen peroxide to oxidize halides for incorporation into electron-rich organic molecules, the mechanism for the controlled transfer of highly reactive chloronium ions in the biosynthesis of napyradiomycin and merochlorin antibiotics sets the Streptomyces vanadium-dependent chloroperoxidases apart. Here we report high-resolution crystal structures of two homologous VHPO family members associated with napyradiomycin biosynthesis, NapH1 and NapH3, that catalyze distinctive chemical reactions in the construction of meroterpenoid natural products. The structures, combined with site-directed mutagenesis and intact protein mass spectrometry studies, afforded a mechanistic model for the asymmetric alkene and arene chlorination reactions catalyzed by NapH1 and the isomerase activity catalyzed by NapH3. A key lysine residue in NapH1 situated between the coordinated vanadate and the putative substrate binding pocket was shown to be essential for catalysis. This observation suggested the involvement of the ε-NH₂, possibly through formation of a transient chloramine, as the chlorinating species much as proposed in structurally distinct flavin-dependent halogenases. Unexpectedly, NapH3 is modified post-translationally by phosphorylation of an active site His (τ-pHis) consistent with its repurposed halogenation-independent, α-hydroxyketone isomerase activity. These structural studies deepen our understanding of the mechanistic underpinnings of VHPO enzymes and their evolution as enantioselective biocatalysts.

Graphical Abstract

Halogenated compounds are common in nature¹ and include many prominent examples such as the thyroid hormone triiodothyronine, the clinical antibiotic vancomycin, and ozone-depleting bromoform. In each case, halogen groups are critical for the molecule’s biological and chemical properties. Several classes of enzymes that install halogens on organic molecules have been characterized,^2,3 including halogenating peroxidases (haloperoxidases) that oxidize chloride, bromide, and/or iodide at the expense of hydrogen peroxide to generate highly reactive halohydrins or halonium ion equivalents that react with electron-rich organic substrates. Two types of haloperoxidases are known and contain either heme⁴ or vanadate^5–8 prosthetic groups. Both enzyme classes produce a diffusible hypohalite that results in generally nonspecific halogenating enzymes during halogen delivery.⁹

While this observation is generally true for characterized vanadium-dependent haloperoxidases (VHPOs) from brown and red macroalgae, fungi, and bacteria,¹⁰ a small group of vanadium-dependent chloroperoxidases (VCPOs) from Streptomyces bacteria catalyze selective halogen delivery in the course of meroterpenoid antibiotic biosynthesis, including the napyradiomycins and merochlorins.¹¹ The biosynthetic conversion of 1,3,6,8-tetrahydroxynaphthalene (THN, 1) to napyradiomycin B1 (9) involves six enzyme-catalyzed reactions (Figure 1), four of which are catalyzed by three VHPO family enzymes: NapH1, NapH3, and NapH4.^12,13 NapH1 plays a dual biosynthetic role catalyzing the asymmetric arene chlorofunctionalization second step reaction (2 → 3) and the asymmetric alkene chlorofunctionalization fifth step reaction (5 → 7 or 6 → 8). NapH4 similarly catalyzes a high-yield, chloronium-induced diastereoselective cyclization of the geranyl side chain in the sixth and final reaction to afford the napyradiomycin B1 (9) antibiotic. Curiously, the homologous NapH3 catalyzes a vanadate-, hydrogen peroxide-, and halide-independent α-hydroxyketone rearrangement in the fourth reaction step from 4 to naphthomevalin (5).¹⁴ This trio of VHPO family enzymes from a single biosynthetic pathway that performs a diverse suite of high-yield, asymmetric reactions provides an opportunity to dissect the mechanistic and structural basis of halogenation selectivity and catalytic diversity in these closely related enzymes.¹⁵

Figure 1.

Biosynthesis of the napyradiomycin meroterpenoids involves several asymmetric VHPO-catalyzed reactions. The crystal structures of the homologous VHPO enzymes NapH1 (blue) and NapH3 (red) were determined. NapH1 is a VCPO that catalyzes arene and alkene chlorofunctionalization reactions, while NapH3 is a VHPO family member that catalyzes a vanadium-independent, α-hydroxyketone rearrangement. Abbreviations: GPP, geranyl diphosphate; DMAPP, dimethylallyl diphosphate.

Herein we report high-resolution X-ray crystal structures of NapH1 and NapH3 and propose, with accompanying mutagenesis and protein mass spectral data, a unifying mechanism for halogenation selectivity in the Streptomyces VCPO enzymes, setting the stage for the future development of VHPOs as biocatalysts in organic synthesis.

MATERIALS AND METHODS

Plasmid Constructs and Site-Directed Mutagenesis.

Expression constructs of NapH1 and NapH3 are identical to those of previously described studies.^12,14 Point mutations of NapH1 and NapH3 were prepared using the Q5 site-directed mutagenesis kit (NEB). The primers that were used are listed in Table S1. All generated mutants were confirmed by Sanger sequencing (Genewiz, Inc.).

Protein Expression and Purification.

Each protein or its variant with an N-terminal polyhistidine tag and thrombin cleavage site was expressed and purified as previously described.¹⁶ For crystallization, the N-terminal polyhistidine tags were removed by thrombin digestion. The tag-free proteins were loaded on a HiLoad 16/600 Superdex 200 size exclusion chromatography column (GE Healthcare) and eluted over 1 column volume of protein storage buffer [20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) (pH 7.5), 150 mM NaCl, and 2 mM dithiothreitol (DTT)]. For holo-NapH1 crystallization experiments, NapH1 was dialyzed against storage buffer supplemented with 100 μM sodium orthovanadate.

NapH1 Enzyme Assay.

Discontinuous, high-performance liquid chromatography (HPLC)-based assays of the wild-type (WT) and mutant NapH1 enzymes were performed using methods modified from what was previously described.¹² The LC-MS assay contains 1 mM H₂O₂, 0.2 mM synthetic, racemic naphthomevalin (5), and 1 μM enzyme in a reaction buffer containing 200 mM KCl or KBr, 50 mM MES (pH 6.0), and 10 μM vanadate. The reaction was quenched after 0.5 h with 2 volumes of methanol and analyzed using a Bruker amaZon Ion Trap and Agilent 1200 LC-MS instrument with a Phenomenex Luna RP 5 μ C18(2) 150 mm × 4.6 mm column. The following method was used at a flow rate of 0.75 mL/min: 50% to 100% B (15 min), 100% B (5 min), 100% to 50% B (3 min), and 50% B (5 min), where A is 0.1% (v/v) aqueous formic acid and B is 0.1% (v/v) formic acid in acetonitrile.

NapH1 Monochlorodimedone (MCD) Assay.

The MCD assay was modified from the procedure described previously.¹⁶ Briefly, the reaction mixture contained 50 μM MCD and 1 μM enzyme in a reaction buffer containing 50 mM MES (pH 6.0), 10 μM vanadate, and 200 mM KCl or KBr. The reaction was initiated by supplying H₂O₂ to a final concentration of 1 mM. The decrease in absorbance at 290 nm was monitored with a Cary spectrophotometer (Varian/Agilent) while keeping the temperature at 30 °C. Data were analyzed with the Cary WinUV Kinetics program (Varian/Agilent).

Initial Velocity Measurements of the NapH3 (wild type vs H445A mutant)-Catalyzed Reaction.

The reaction of a 500 μL mixture containing 150 mM KCl and 0.25 mM synthetically prepared racemic 4¹⁴ in 50 mM HEPES-NaOH (pH 8.0) was initiated by addition of either 15 μM NapH3 or 15 μM NapH3 H445A. Then, 75 μL aliquots of the reaction mixture were removed at various time points, and reactions were quenched with an equal volume of ice-cold methanol. The reaction mixtures were then centrifuged at 15000g for 10 min to pellet the precipitated enzyme. The samples were analyzed using an Agilent Technologies 1200 series HPLC instrument equipped with an Agilent Eclipse XDB-C18 5 μm, 4.6 mm × 150 mm column. The following method was used at a flow rate of 0.75 mL/min: 5% B (1 min), 5% to 95% B (5 min), 95% to 100% B (3 min), 100% B (5 min), 100% to 5% B (2 min), and 5% B (1 min), where A is 0.1% (v/v) aqueous formic acid and B is 0.1% formic acid in acetonitrile. Initial velocities were obtained from the linear portion of the reaction by integration of the naphthomevalin (5) product peak at 280 nm and comparison to a standard set of synthetic 5 concentrations.

Intact Mass Spectroscopy of NapH1 and NapH3.

An Agilent 6230 time-of-flight mass spectrometer (TOFMS) with a Jet Stream electrospray ionization source (ESI) was used for LC-ESI-TOFMS analysis. Chromatographic separations were performed at room temperature on a Phenomenex Aeris Widepore C4 column (2.1 mm inside diameter × 50 mm length, 3.6 μm particle size). Mobile phase A was HPLC grade water with 0.1% TFA, and HPLC grade acetonitrile with 0.1% TFA was used as mobile phase B. The mobile phase was delivered at a rate of 0.3 mL/min with the following gradient conditions: increased from 5% to 90% mobile phase B in 12 min, held at 90% mobile phase B for 2 min, returned to 5% mobile phase B in 1 min, and equilibrated with 5% mobile phase B for 7 min. Agilent MassHunter software was used to acquire and analyze data.

Trypsin Digestion and LC-MS Analysis of NapH3 and NapH3 H445A.

To identify the site of phosphorylation, NapH3 and NapH3 H445A proteins were trypsin digested using a Trypsin Single Proteomics grade kit (Sigma-Aldrich). The resulting samples were analyzed by high-resolution LC-MS on an Agilent 1260 Infinity LC system coupled to an Agilent 6530 Accurate-Mass Q-TOF instrument. The following method was used on a Phenomenex Aeris WIDEPORE XB-C18 200 Å LC column (3.6 μm, 250 mm × 4.6 mm) at a flow rate of 0.75 mL/min: 5% to 30% B (5 min), 30% to 60% B (15 min), 60% to 100% B (5 min), 100% B (6 min), and 100% B to 5% B (12 min), where A is 0.1% (v/v) aqueous formic acid and B is 0.1% (v/v) formic acid in acetonitrile. Peptide fragments were detected using ESI-MS with a collision energy of 30 kV. Data were collected and analyzed using MassHunter Workstation Software version B.05.01.

Crystallization and Heavy-Atom Derivatization of Apo-NapH1.

Apo-NapH1 was crystallized by the hanging-drop vapor diffusion method at 4 °C. A 1 μL aliquot of 8.5 mg/mL NapH1 and 1 μL of the reservoir solution [0.1 M HEPES (pH 7.5), 0.3 M ammonium acetate, 7% (w/v) PEG 8000, and 2 mM DTT] were mixed in a sealed well with 150 μL of the reservoir solution. Transparent plate crystals formed within a week. Gold and tungstate derivatives were prepared by soaking crystals in reservoir solutions containing 15 mM KAu(CN)₂ and 5 mM Na₂WO₄, respectively, for 2 days. Both underivatized and derivatized crystals were transferred into a cryoprotectant [0.1 M HEPES (pH 7.5), 0.3 M ammonium acetate, 7% (w/v) PEG 8000, 2 mM DTT, and 25% (v/v) ethylene glycol] and flash-cooled in liquid nitrogen.

Crystallization of Vanadate-Enriched Holo-NapH1.

Holo-NapH1 was crystallized by the hanging-drop vapor diffusion method at 4 °C. A 1 μL aliquot of 9.0 mg/mL holo-NapH1 supplemented with 2.5 mM SF2415B1 and 1 μL of the reservoir solution [0.1 M HEPES (pH 7.5), 0.3 M ammonium acetate, 8% (w/v) PEG 8000, and 2 mM DTT] were mixed in a sealed well with 150 μL of the reservoir solution. Transparent plate crystals formed within a week. The crystals were transferred into a cryoprotectant [0.1 M HEPES (pH 7.5), 0.3 M ammonium acetate, 8% (w/v) PEG 8000, 2 mM DTT, 25% (v/v) ethylene glycol, and 1 mM SF2415B1] and flash-cooled in liquid nitrogen.

Crystallization of NapH3.

NapH3 was crystallized by the hanging-drop vapor diffusion method at 6 °C. A 1 μL aliquot of 10.0 mg/mL NapH3 and 1 μL of the reservoir solution [0.1 M bis(2-hydroxyethyl)amino-tris(hydroxymethyl)methane (Bis-Tris) (pH 6.5), 0.2 M magnesium chloride, and 17% (w/v) PEG 8000] were mixed in a sealed well with 150 μL of the reservoir solution. Transparent clusters of plate crystals formed in 5–7 days. Crystals were transferred into a cryoprotectant [0.1 M Bis-Tris (pH 6.5), 0.2 M magnesium chloride, 17% (w/v) PEG 8000, and 20% (v/v) 2-methyl-2,4-pentanediol (MPD)] and flash-cooled in liquid nitrogen.

Data Collection and Processing.

All data sets were collected at the Advanced Light Source (ALS, Berkeley, CA) on beamlines 8.2.1 and 8.2.2. Data sets of heavy-atom-derivatized NapH1 and other data sets were collected using the static collection method. Diffraction data were indexed, integrated, and scaled using HKL2000¹⁷ (heavy-atom-derivatized NapH1), iMosflm/Scala¹⁸ (apo- and holo-NapH1), or XDS (NapH3). R_merge was used as an indicator of where to trim the high-resolution data of NapH1. A CC_1/2 of ~0.7 was used as the indicator of where to trim the high-resolution data of NapH3. Data statistics are listed in Table S2.

Structure Determination and Refinements.

The structure of apo-NapH1 was determined using multiple isomorphous replacement (MIR). Both derivatized and underivatized NapH1 crystals belong to the P2₁2₁2₁ space group with similar cell dimensions (Table S2). The program SOLVE¹⁹ identified four gold sites or two tungsten sites in the asymmetric unit (ASU) of KAu(CN)₂− or Na₂WO₄-soaked NapH1, respectively. The initial figures of merit for these two data sets are 0.27 and 0.32. The programs SOLVE¹⁹ and RESOLVE²⁰ were used for the phase calculation and density modification. ARP/wARP²¹ was then used for the automatic model building based on the experimental electron density map. The model from MIR contains one NapH1 homodimer per ASU.

The model built from the MIR method was used as the initial model for determining the apo-NapH1 structure. Iterative rounds of positional refinement and individual ADP-factor refinement were performed in Refmac,²² and model building and model adjustment were performed in Coot.²³ In the later stages of refinement for the apo-napH1 model, translation/libration/screw (TLS) parameters were additionally refined. Water molecules were added manually. Two-fold noncrystallographic symmetry (NCS) restraints were used throughout refinement. The final structure of apo-NapH1 contains one NapH1 homodimer per ASU.

The structure of holo-NapH1 was determined using molecular replacement (MR) with MOLREP²⁴ and the apo-NapH1 structure as the MR search model. The model was then refined using the same procedure that was used for the apo-NapH1 structure with PHENIX²⁵ being the program for refinement. The final structure of holo-NapH1 contains one NapH1 homodimer per ASU with a vanadate molecule bound to His494 in each monomer. Despite the concentration of SF2415B1 being >1 mM throughout the crystallization process, no electron density was found corresponding to the additive.

The structure of NapH3 was determined by MR. The structure of dimeric holo-NapH1 [Protein Data Bank (PDB) entry 3W36], which is 49.8% identical to NapH3, was truncated with Phenix Sculptor²⁵ and used as the MR search model. A solution for a NapH3 dimer was found with an LLG of 995 and a TFZ of 34.8. Rigid-body refinement and simulated annealing were performed, and atomic coordinates and ADP factors were then iteratively refined in Phenix Refine²⁵ with model building and manual adjustment in Coot. Water molecules were added manually using F_o − F_c electron density contoured to 5.0σ as the criteria. Two-fold NCS restraints were used throughout refinement. The final structure contains one NapH3 homodimer per ASU with His445 in each monomer refined as 1-phosphohistidine.

Restraints for 1-phosphohistidine were from grade Web Server (Global Phasing). Restraints for vanadate were from the CCP4 ligand library.²⁶ The composite omit electron density map was calculated by Phenix Composite_omit_map,²⁵ which was used to verify all models. The refinement statistics are listed in Table S2, and residues built into each chain are listed in Table S3. All structure figures were rendered in PyMOL.

Phylogenetic Analysis.

Functionally and structurally characterized sequences were selected to build the framework of a focused phylogenetic tree. To further fill in sections, the top homologues of characterized sequences were selected by using the Basic Local Alignment Search Tool (BLAST) within UniProt, using the UniProtKB reference proteomes plus Swiss-Prot database and default BLAST parameters. This resulted in a representative list comprising a total of 36 sequences. Amino acid sequence alignments were completed using kalign (version 2.04)²⁷ using default parameters, and all position sites obtained by the alignments are used in final phylogenetic tree construction. A maximum likelihood tree was constructed with IQ-TREE (version 1.6.12)²⁸ with the following parameters using the best-fit model suggested by the IQ-TREE ModelFinder feature: iqtree -s file_aln.fasta -m WAG+I +G4 -bb 1000.

RESULTS

Structure of NapH1 and Lysine-Mediated Stereospecific Halocyclization.

We first determined the structure of apo-NapH1 and used gold and tungstate derivatives to determine the phases using multiple isomorphous replacement (Tables S2 and S3). Holo-NapH1 was prepared by growing crystals in the presence of vanadate and the napyradiomycin analogue SF2415B1 (6),¹² and the crystal structure was determined by molecular replacement using the apo-NapH1 structure (PDB entry 3W35). The apo and holo forms of NapH1 are nearly superimposable [root-mean-square deviation (RMSD) of 0.63 Å], and no electron density was observed for SF2415B1 (or any other napyradiomycin derivative attempted) in the holo-NapH1 structure (PDB entry 3W36).

The asymmetric unit of NapH1 comprises a physiological NapH1 homodimer with the vanadate cofactor bound in each monomer (Figure 2A and Figures S1A and S2). Each NapH1 monomer comprises two helical bundles with the N- and C-terminal bundles interconnected by an ~100-residue predominately loop region that forms the dimeric interface (Figure 2A and Figure S1A). The N-terminal bundle, which contains helices 1–5, is packed against the C-terminal bundle. The C-terminal bundle, which includes helices 7–14, contains all of the vanadate binding residues. The first 54 residues at the N-terminus of NapH1 are absent in our crystal structure. Hence, we constructed a NapH1 variant named “NapH1 d54” with the identical truncation. The variant was fully active (Figures S3 and S4 and Table S4), indicating our structure is representative of an active NapH1 construct.

Figure 2.

Structure of NapH1. (A) NapH1 is a homodimer. Each monomer contains three parts, the N-terminal helix bundle (helices 1–5, residues 55–194, purple), the interconnecting region (residues 195−290, teal), and the C-terminal helix bundle (helices 7–14, residues 291–523, green). The monomer on the left is colored gray. Each monomer binds a vanadate cofactor, which is shown as sticks. (B) In NapH1, the vanadate cofactor coordinates to H494 (yellow sticks), stabilized by the dipole interaction from helix 12 (green ribbon), forming charge–charge interactions with K372, R379, and R448 (yellow sticks). S427 (green sticks), which corresponds to a conserved histidine residue in other structurally characterized VHPOs (Figure S5), forms hydrogen bonds with the vanadate cofactor and K324 (green sticks). The C_α atoms of amino acids are shown as spheres.

The vanadate bound in NapH1 is coordinated to H494 (vHis) and positioned at the C-terminus of helix 12 (Figure 2B), which binds vanadate via helical dipole interactions. The vanadate cofactor is further stabilized by charge–charge interactions with K372, R379, and R488 (Figure 2B). These vanadate binding residues are like those observed in other structurally characterized VHPOs (Figure S5).^29–35 Notably, the vanadate binding site of NapH1 is enclosed by loops between helices 1 and 2 (Figures S6A and S7A). To validate if vanadate stabilization is critical for NapH1 activity, we constructed mutants that replaced either a positively charged residue with a charge-neutral residue (K372F and R379V) or a charge-neutral residue with a negatively charged residue (W369E). These variants lost their WT chlorinating activity in converting SF2415B1 (6) to SF2415B3 (8)¹² (Figure 1) as well as their ability to oxidize bromide to hypobromite in the standard VHPO-based monochlorodimedone (MCD) assay.¹⁶

Interestingly, K324 of NapH1 extends toward the vanadate cofactor and S427 on helix 12 (Figure 2B). In contrast, the lysine residue is absent in all other structurally characterized VHPOs, and a conserved histidine takes the place of S427 in these latter VHPOs (Figure 2B and Figure S5). To evaluate the roles of the K324 and S427 residues, we designed the K324M, K324R, S427T, and S427A mutants and measured their hypobromite formation activities using the MCD assay and their chloro- or bromocyclization activity with SF2414B1 (6) using HPLC. Both K324 variants are active for hypobromite formation (Table S4), but they are incapable of catalyzing the chloro- or bromocyclization of naphthomevalin (5)¹⁴ (Figures S3 and S4), which suggests K324 is essential for stereospecific chlorofunctionalization. In addition, the S427A variant shows diminished hypobromite formation activity, indicating the hydrogen bond between S427 and vanadate is important for halide oxidation.

The proposed chloramine-forming residues K324 and S427 are conserved in all other biochemically validated VHPO enzymes from Streptomyces biosynthetic pathways that catalyze enantioselective halogenating reactions (Figure S14). Examples include the remaining napyradiomycin VHPO NapH4 that catalyzes the conversion of napyradiomycin A1 to B1 (7 → 9) (Figure 1) as well as VCPOs associated with marinone (MarH1 and MarH3)³⁶ and merochlorin (Mcl24 and Mcl40)^37,38 biosynthesis. The third co-clustered marinone VHPO gene product MarH2 lacks the conserved lysine residue as well as the catalytically necessary vHis, which clarifies our earlier observation of no VHPO activity under standard MCD assay conditions.³⁶

Structures of NapH3.

To complete our structural analysis of VHPO homologues involved in napyradiomycin A1 (7) biosynthesis, we determined the structure of NapH3 (PDB entry 8CXL), a VHPO homologue catalyzing an α-hydroxyketone rearrangement reaction and not halogenation^13,14 (Figure 1 and Tables S2 and S3). The asymmetric unit of NapH3 contains a physiological NapH3 homodimer (Figure 3A). The amino acid sequence of NapH3 is 49.8% identical to that of NapH1 with a similar helical bundle composition (Figures S1B and S8). The overall RMSD between NapH3 and NapH1 is 1.0 Å over 791 pairs of C_α atoms in a dimer. The active site of NapH3 is concealed from the solvent by the loop between helices 1 and 2, similar to NapH1 (Figures S6B and S7B).

Figure 3.

Structure of NapH3. (A) NapH3 is a homodimer. Each monomer contains three parts, the N-terminal helix bundle (helices 1–5, residues 5–148, purple), the interconnecting region (residues 149–246, teal), and the C-terminal helix bundle (helices 7–14, residues 247–473, green). The monomer on the left is colored gray. Each monomer contains a phosphorylated H445 (τ-pHis), which is shown as sticks. (B) τ-pHis is stabilized by the dipole interaction from helix 12 (green ribbon) and forms charge–charge interactions with K322, R329, and R439 (yellow sticks), which are akin to other VHPOs. F378 and T278 (green sticks) take the place of S427 and K324 in NapH1, but the roles of F378 and T278 for NapH3 reactivity are unclear. The C_α atoms of amino acids are shown as spheres.

To our surprise, H445 in the active site, which corresponds to the vHis in NapH1, is post-translationally modified to a τ-phosphohistidine (τ-pHis) residue (Figure 3B and Figure S9). We verified the chemical composition of the modified histidine residue by intact protein mass spectrometry in the form of the +80 Da protein (Figure S10). We did not observe the same +80 Da modification in the NapH3 H445A mutant or in NapH1, thereby supporting the unexpected presence of the τ-phospho group of NapH3 at H445. Post-translational phosphorylation of H445 was independently confirmed by trypsin digestion of NapH3 followed by high-resolution LC-MS analysis of the resulting peptide sequences (Figure S11). This post-translational modification is positioned at the C-terminus of helix 12, which is similar to the location of the vanadate cofactor in NapH1. The τ-phospho group is stabilized by charge–charge interactions with K322, R329, and R439 (Figure 3B). K324 and S427 in NapH1, which may facilitate the formation of the chloramine intermediate, are replaced by T278 and F378 in NapH3. The steric hindrance from F378 may prevent NapH3 from binding vanadate (Figure 4A,B); however, it is unclear whether and how T278 and F378 contribute to the NapH3 reactivity or the phosphorylation of τ-pHis. Overall, τ-pHis prevents vanadate from binding to NapH3 and explains the loss of haloperoxidase activity in NapH3.

Figure 4.

Comparison of active site cavities of NapH1, NapH3, and a flavin-dependent chlorinase, RebH. (A) K324 is 5.1 Å from the vanadate cofactor in NapH1. The structure is shown in the same color scheme as in Figure 2. The active site cavity is shown as yellow mesh. (B) The space near pHis is blocked by F378 in NapH3. The structure is shown in the same color scheme as in Figure 3. The active site cavity is shown as yellow mesh. (C) Structure of RebH with FAD bound (PDB entry 2OAL, yellow) overlaid with the structure of RebH with the tryptophan substrate bound (PDB entry 2E4G, blue). The distance between flavin and K79 is 9 Å.

Although pHis modification has been reported in several enzymes,^39,40 including succinyl Co-A synthetase, a key enzyme in the TCA cycle, this is the first report of such post-translational modification in VHPO family enzymes. To evaluate the role of post-translationally modified H445 during NapH3 catalysis, we carried out enzyme assays using synthetically prepared racemic 4¹⁴ and the H445A variant. Initial rate comparisons revealed that the H445A variant is as active as the native NapH3 (Figure S12), suggesting that phosphorylated H445 does not play a role during the NapH3-catalyzed α-hydroxyketone rearrangement reaction and hints that the substrate binding site is independent of H445. This observation, however, raises the question of why nature performs such post-translational phosphorylation on a residue that, as of yet, has no identifiable role in catalysis.

Phylogenetic Analysis of NapH1 and NapH3.

To generate a broader comparison of NapH1 and NapH3 relative to other VHPO enzymes,⁸ including functionally and structurally validated VHPOs, we constructed a maximum likelihood phylogenetic tree using publicly available sequences (Table S5). Streptomyces VHPOs cluster discretely, independent of known microbial or eukaryotic VHPO enzymes (Figure S13). Within this group, NapH1 and NapH3 cluster according to biochemical function. NapH1 is adjacent to other functionally validated stereoselective chloroperoxidases, while NapH3 clades with MarH2, another nonfunctional streptomycete VHPO. Both NapH3 and MarH2 lack the putative chloramine-forming lysine residue, instead harboring threonine and asparagine residues, respectively (Figure S14). Similarly, this lysine residue is replaced with aspartic acid or cysteine residues in all other structurally characterized VHPOs; however, a function is yet to be identified for those residues (Figure S5). Notably, NapH1 and NapH3 are the only structurally characterized enzymes in the Streptomyces clade of VHPOs, further demonstrating how this work informs a distinct class of VHPO enzymes and offers mechanistic insight into their evolution as enantioselective biocatalysts.

DISCUSSION

Vanadium-dependent haloperoxidases (VHPOs) catalyze halogenation reactions with chloride, bromide, or iodide ions harnessing catalytic vanadate and aqueous H₂O₂ for oxidation. The relatively economic process makes VHPOs ideal starting points for industrial catalyst designs. However, most characterized VHPOs generate short-lived, highly reactive hypohalites that react nonspecifically with various substrates.⁹ The only exceptions are a handful of VCPOs from Streptomyces that catalyze stereospecific chlorofunctionalization reactions on naphthoquinone-based meroterpeneoids.¹¹ Here, we present the first structure of such a VCPO, NapH1, and its homologue, NapH3, the latter of which catalyzes an α-hydroxyketone rearrangement reaction. The two enzymes combined are the VHPOs required for the biosynthesis of napyradiomycin A1 (7), while a third homologue, NapH4, is the VCPO that converts napyradiomycin A1 to B1 (7 → 9) (Figure 1). The structural and biochemical characterization of NapH1 provides a preliminary understanding of how halogenating stereospecificity is achieved, while the structural and functional analysis of NapH3 reveals the reason for its loss of haloperoxidase activity.

In NapH1, K324 in the vanadate binding site is critical for chlorofunctionalization. Substituting K324 with arginine or methionine eliminates the chlorofunctionalization activity but retains the ability to oxidize bromide. Consequently, K324 was shown to facilitate late stage halofunctionalization, adding to the inventory of known amine-mediated stereospecific halogenation reactions. We previously demonstrated that chloramines are effective in chemically mimicking VCPOs for merochlorin biosynthesis.⁴¹ Additionally, a lysine-serine diad near a halogen-oxidizing cofactor was also found in various flavin-dependent chlorinases, as first demonstrated for RebH.^42,43 This flavin-dependent chlorinase uses an active site lysine to form a long-lived, presumably covalently bound lysine-chloramine intermediate, where the distance between lysine and flavin is 9.0 Å (Figure 4C). On the contrary, the halogen-oxidizing vanadate is 5.1 Å from K324 of NapH1 (Figure 4A). The shorter distance between the cofactor and lysine makes transferring a Cl⁺ intermediate to K324 even more plausible in NapH1. On the basis of our NapH1 mutagenesis results and precedents in flavin-dependent halogenases, we propose that K324 in NapH1 may form a chloramine intermediate, thereby shuttling Cl⁺ from the vanadate binding site to the substrate in a stereospecific manner (Figure 5). Furthermore, this lysine residue is conserved in stereospecific VCPOs from Streptomyces (Figure S14) but not in other structurally characterized VHPOs with promiscuous activity, suggesting the lysine residue is essential for the stereospecific halofunctionalization.

Figure 5.

Proposed reaction scheme for stereospecific VHPOs and flavin-dependent halogenases (FDHs). The halogenation reaction catalyzed by stereospecific VHPOs and FDHs proceeds via formation of a hypohalous acid that is shuttled by a conserved lysine residue (presumably via formation of a chloramine intermediate) to the substrate binding site where the halogen atom is stereospecifically incorporated into the substrate. With VHPOs, H₂O₂ oxidizes the halide via formation of an active vanadate oxidizing species, whereas molecular oxygen (O₂) oxidizes the halide via formation of a flavin-C4a-hydroperoxide in FDHs.

In addition to the unique lysine residue, the solvent inaccessible vanadate coordinating site differs from other structurally characterized VHPOs in fungi or algae where the vanadate binding site is exposed to solvent (Figure S6). The only other VHPO with a walled in vanadate binding site is the bacterial VHPO from Zobellia galactanivorans (ZgVHPO),³⁵ in which a partially disordered loop between helices 8 and 9 blocks the access of the solvent to the active site (Figure S6G). Although the function of the vanadate binding site in ZgVHPO is unclear, we hypothesize that the vanadate binding site in NapH1 is essential for stereospecific reactivity by protecting the active Cl⁺ intermediate from reacting with water.

Thus far, our structural investigation has shed light on the mechanism by which NapH1 may control the activated halogen species. Though we have not yet been able to determine a structure of NapH1 or NapH3 with bound substrates or products, the enclosed active sites in NapH1 and NapH3, as well as the proposed chloramine intermediate, indicate substrates bind discretely in the vicinity of vHis in NapH1 and τ-pHis in NapH3. The active site cavity in the proximity of vHis and τ-pHis suggests abundant space near the active site, which allows substrate binding and substrate scope expansion (Figure 4A,B). Using this structural information, we can now engineer VHPOs with enhanced substrate selectivity and use these biocatalysts to install halogen atoms onto an extended scope of substrates in a regio- and stereospecific manner.

In summary, with our structural and biochemical studies of NapH1 and NapH3 enzymes from the napyradiomycin A1 biosynthetic pathway, we have provided insights into the long-standing question in Streptomyces VHPOs of how regio- and stereospecific halogenation is accomplished with highly reactive hypohalous acid. Our results suggest that nature has tackled this problem in the same way it did in the case of flavin-dependent halogenases (FDHs) by recruiting a lysine residue placed between the cofactor and substrate binding sites. We anticipate that our structural studies of Streptomyces VHPOs will lay the foundation for their potential future use as enantioselective biocatalysts.