HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 49, 47476-47485, December 6, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/49/47476    most recent M206342200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Zerbe, K. Articles by Schlichting, I. Search for Related Content PubMed PubMed Citation Articles by Zerbe, K. Articles by Schlichting, I. Social Bookmarking                      What's this?

Crystal Structure of OxyB, a Cytochrome P450 Implicated in an Oxidative Phenol Coupling Reaction during Vancomycin Biosynthesis^*

Katja Zerbe^§, Olena Pylypenko^§¶, Francesca Vitali, Weiwen Zhang, Severine Rouset, Markus Heck, Jan W. Vrijbloed, Daniel Bischoff^**, Bojan Bister^**, Roderich D. Süssmuth^**, Stefan Pelzer, Wolfgang Wohlleben^§§, John A. Robinson^¶¶, and Ilme Schlichting^¶

From the  Institute of Organic Chemistry, University of Zurich, Winterthurerstrasse 190, 8057 Zurich, Switzerland, the ^¶ Max Planck Institute for Molecular Physiology, Department of Physical Biochemistry, Otto Hahn Str.11, 44227 Dortmund, Germany, the  Max Planck Institute for Medical Research, Department of Biomolecular Mechanisms, Jahnstr. 29, 69120 Heidelberg, Germany, the ^** Institut für Organische Chemie, Eberhard-Karls-Universität Tübingen, Auf der Morgenstelle 18, 72076 Tübingen, Germany, and the  Lehrstuhl für Mikrobiologie/Biotechnologie, Eberhard-Karls-Universität Tübingen, Auf der Morgenstelle 28, 72076 Tübingen, Germany

Received for publication, June 26, 2002, and in revised form, August 29, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

Gene-inactivation studies point to the involvement of OxyB in catalyzing the first oxidative phenol coupling reaction during glycopeptide antibiotic biosynthesis. The oxyB gene has been cloned and sequenced from the vancomycin producer Amycolatopsis orientalis, and the hemoprotein has been produced in Escherichia coli, crystallized, and its structure determined to 1.7-Å resolution. OxyB gave UV-visible spectra characteristic of a P450-like hemoprotein in the low spin ferric state. After reduction to the ferrous state by dithionite or by spinach ferredoxin and ferredoxin reductase, the CO-ligated form gave a 450-nm peak in a UV-difference spectrum. Addition of putative heptapeptide substrates to resting OxyB produced type I changes to the UV spectrum, but no turnover was observed in the presence of ferredoxin and ferredoxin reductase, showing that either the peptides or the reduction system, or both, are insufficient to support a full catalytic cycle. OxyB exhibits the typical P450-fold, with helix L containing the signature sequence FGHGXHXCLG and Cys³⁴⁷ being the proximal axial thiolate ligand of the heme iron. The structural similarity of OxyB is highest to P450nor, P450terp, CYP119, and P450eryF. In OxyB, the F and G helices are rotated out of the active site compared with P450nor, resulting in a much more open active site, consistent with the larger size of the presumed heptapeptide substrate.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

Vancomycin and related glycopeptide antibiotics are clinically important natural products that often provide a last line of defense in the treatment of nosocomial infections by multiple drug-resistant strains of Gram-positive bacteria (1). The biosynthesis of these antibiotics is currently attracting great interest, not least because the knowledge may be important in attempts to produce novel glycopeptides by combinatorial biosynthesis methods.

Vancomycin consists of a cross-linked heptapeptide, which is glycosylated on residue 4. The same cross-linked heptapeptide is found in balhimycin and chloroeremomycin (Fig. 1). The peptide backbone is constrained into an unusual conformation by the presence of three side chain to side chain linkages connecting aromatic groups in residues 2 and 4, residues 4 and 6, and residues 5 and 7. Although rapid progress has been made recently in studies of the biosynthesis of the constituent amino acids (2-6), the production and attachment of sugar residues to the aglycon (7-9), an N-methyltransferase (10), as well as the peptide synthetase (11, 12) that builds a linear heptapeptide precursor, the timing of many steps in the biosynthesis remain ill-defined (see below). This is also true of the side chain cross-links, which arise formally by oxidative phenol coupling reactions. Despite the enormous importance of oxidative phenol coupling reactions in natural product biosynthesis, still very little is known about the enzymes that catalyze such reactions, their structures, and their detailed mechanism(s) of action.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Structures of glycopeptide antibiotics and putative linear heptapeptide intermediates (18-20) (1) in vancomycin biosynthesis.

The nucleotide sequences of some biosynthetic genes for four glycopeptide antibiotics, vancomycin (13), chloroeremomycin (14), balhimycin (15), and teichoplanin (16), as well as the structurally related antiviral agent complestatin (17), have been reported recently. The organization of the biosynthetic genes in each cluster is very similar. This includes in the chloroeremomycin and balhimycin clusters three contiguous open reading frames (ORFs)¹ encoding the peptide synthetases, and a cluster of three genes (oxyA, oxyB, and oxyC) encoding P450-like enzymes that catalyze three oxidative phenol coupling reactions. Gene knockout experiments in the balhimycin producer indicate that the first phenol coupling occurs between residues 4 and 6 and is catalyzed by OxyB (18-20). The second coupling should be between residues 2 and 4 catalyzed by OxyA, and the last coupling is between residues 5 and 7 catalyzed by OxyC (18-20). Methylation at the N terminus of the heptapeptide seems to be a late step in the biosynthesis (10). However, the timing of other steps, e.g. chlorination and release of products from the peptide synthetase, are not yet defined. Thus it is not clear whether or not the free linear heptapeptide 1, either with X = H or Cl, is the substrate for OxyB, or for example, whether the peptide remains attached as a thioester through its C terminus to a peptidyl carrier domain during the phenol coupling reactions. For instance, the P450 enzyme NovI that catalyzes the -hydroxylation of tyrosine during novobiocin biosynthesis, hydroxylates a tyrosine moiety that is coupled as a thioester to a peptide carrier domain (21). A similar route for the production of a -hydroxytyrosine moiety may also occur during vancomycin biosynthesis (6). The nikQ gene in nikkomycin biosynthesis has a similar function. NikQ is a P450 that catalyzes the -hydroxylation of histidine while conjugated to a carrier protein (22). So far, no reports of in vitro activity for the enzymes OxyA-C have been described.

We report here the cloning and sequencing of the three contiguous P450-like genes, oxyA, oxyB, and oxyC, from the vancomycin producer Amycolatopsis orientalis. The enzyme OxyB has been produced in Escherichia coli as a soluble protein whose UV-visible spectral properties are typical of P450 monooxygenases. We also describe the crystal structure of OxyB to 1.7-Å resolution, which is the first reported structure of a P450 protein implicated in an oxidative phenol coupling reaction.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

Bacterial Strains and Cloning Vectors-- A. orientalis DSM40040 chromosomal DNA was used to construct a genomic library in the cosmid SuperCos I (Stragene) in E. coli XL1-blue. Subcloning was in the hosts E. coli DH5 and XL1-Blue with plasmids pUC18 and pBluescript. For production of OxyB with a His₆ tag at the N terminus the gene was cloned in pET14b (Novagen), and for OxyB with a native N terminus pET17b (Novagen).

Library Construction and Screening-- Genomic DNA was isolated from A. orientalis, using methods developed for streptomycetes (23) and partially digested with Sau3AI. Fragments of 30-40 kb were isolated by sucrose gradient ultracentrifugation, and ligated into SuperCosI previously digested with XbaI and BamHI. After packaging (Gigapack III XL, Stratgene) the DNA was used to transfect E. coli XL1-blue.

A part of the gtfE gene (13) from A. orientalis DSM40040 was amplified by PCR using the following primers: GtfFor, 5'-ATCCCCTACTTCTATGGCTTCCAC-3'; GtfRev, 5'-GAAGTGAATCAGCAGGTGCTGTTC-3'. The PCR product (549 bp) was cloned into the EcoRV site of pBluescript. This DNA fragment was used to screen the cosmid library described above. A positive clone was used for subcloning and DNA sequencing of the region upstream of the gtfE gene, which includes cytochrome P450 genes oxyA-C (14).

DNA Sequencing and Analysis-- Sequencing templates were obtained by generating nested deletions with exonuclease III. Sequencing with the dideoxy chain termination method was performed using an Applied Biosystems 377A sequencer. Nucleotide sequences have been deposited at the EMBL data base (accession number AF486630). Here the protein sequence is numbered from the N terminus starting with Met¹-Ser²-Glu³-Asp⁴-Asp⁵ etc.

Expression of oxyB-- The oxyB gene was amplified by PCR using the following primers (restriction sites underlined): OxyBFor, 5'-GCTATCTAGACATATGAGCGAGGACGACCCGCGCCC-3'; OxyBRev, 5'-TGATCAAGCTTAGATCTTCACCAAGCAACCATCAGCTCGGTCAAACCGTACG-3'. The PCR products were digested with XbaI and HindIII, and ligated into XbaI/HindIII-digested pBluescript for nucleotide sequencing. After confirming the correct nucleotide sequence, the gene was transferred as an NdeI/HindIII fragment into pET17b (24) to give pOCI810. For the production of His₆-tagged OxyB (His₆-OxyB), the gene was transferred from pOCI810 as an NdeI/XhoI fragment into pET14b to give pOCI1047.

Production of His₆-OxyB-- For production of His₆-OxyB in E. coli BL21(DE3)pLys-S, TB medium (400 ml) with ampicillin (100 µg/ml) and chloramphenicol (34 µg/ml) was inoculated (4%, v/v) with a preculture (grown overnight at 37 °C) and incubated at 24 °C with shaking at 200 rpm. At A₆₀₀ = 0.5, -aminolevulinic acid (8 mg) was added. The culture was induced at A₆₀₀ = 1.0 with isopropyl-1-thio--D-galactopyranoside (0.1 mM) and growth was at 22 °C. A second portion of -aminolevulinic acid (8 mg) was added 20 h after induction and the cultures were harvested by centrifugation after another 24 h.

The cells were disrupted in PG buffer (50 mM potassium phosphate, pH 7.4, 10% glycerol, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 0.1% -mercaptoethanol, 0.02% azide) by 3 passages through a French pressure cell at 1200 p.s.i. The E. coli cell-free extract containing His₆-OxyB was applied to a nickel-nitrilotriacetic acid column (0.5 × 5 cm, Ni²⁺-nitrilotriacetic acid superflow, Qiagen) pre-equilibrated with buffer A (potassium phosphate, pH 7.4, 50 mM, with KCl (300 mM), and imidazole (20 mM)), at a flow rate of 2 ml/min. After washing (20 column volumes of buffer A) the His₆-tagged protein was eluted in buffer A with 300 mM imidazole and dialyzed against buffer B (Tris-HCl (50 mM), pH 7.5, and EDTA (0.5 mM)). The protein was then chromatographed on Mono Q (10/10, Amersham Biosciences) in buffer B at a flow rate of 2 ml/min, and eluted with a linear gradient of 0-0.4 M KCl over 100 ml with a flow rate of 1.5 ml/min. His₆-OxyB eluted with 0.17 M KCl, and showed a single band by SDS-PAGE, gave an electrospray mass spectrum consistent with the expected mass, and the correct N-terminal sequence by Edman degradation.

Production of Native OxyB-- The production of OxyB with a native N terminus was performed using pOCI810 and the same method described for His₆-OxyB (see above), except that induction with isopropyl-1-thio--D-galactopyranoside was at A₆₀₀ = 1.6. The E. coli cell-free extract was applied to a Q-Sepharose Fast Flow column (1.6 × 14 cm) equilibrated with buffer B, and eluted with a linear gradient of 0-1 M KCl over 100 ml at a flow rate of 3.0 ml/min. After dialysis into buffer B, the protein was then applied to a Mono Q 10/10 column and eluted with a linear gradient from 0 to 0.4 M KCl over 100 ml at a flow rate of 1.5 ml/min. After concentrating, the protein was then applied to a Superdex 75 (HR 10/30, Amersham Biosciences) column pre-equilibrated with buffer C (Tris-HCl (50 mM), pH 7.5, with EDTA (0.5 mM) and KCl (0.15 M)) and eluted at a flow rate of 0.2 ml/min. The native OxyB showed a single band by SDS-PAGE, gave an electrospray mass spectrum consistent with the expected mass, and the correct N-terminal sequence by Edman degradation.

UV-Visual Spectrophotometry-- Spectra were measured with a Varian Cary 3 double-beam spectrophotometer equipped with a thermostated cell holder. After thermal equilibration at 30 °C the baseline was zeroed between 300 and 700 nm and the UV-visible spectra of the substrate-free His₆-OxyB and native OxyB (2.5 µM) were recorded in Tris-HCl (50 mM, pH 8.0). The enzymes were reduced by addition of 2 mg of solid Na₂S₂O₄ and spectra were measured under the same conditions. The imidazole complex was generated by adding imidazole (100 mM) to native OxyB in the same buffer.

CO Difference spectra-- A solution of enzyme (2.6 µM) in Tris-HCl buffer (2 ml, 50 mM, pH 8.0) was divided between two tandem cuvettes (sample and reference) and CO was bubbled through the sample cuvette for a few seconds, then 2 mg of solid Na₂S₂O₄ was added to each cuvette, and the difference spectrum was recorded. The P450 concentration was determined from CO difference UV spectra using dithionite-reduced heme, and an extinction coefficient of 91 mM¹ cm¹ (25).

Native OxyB was also reduced by incubation in buffer B with spinach ferredoxin (10 µM), ferredoxin-NADP⁺ reductase (0.05 unit; Sigma), and NADPH (1 mM). CO was then passed through the solution and the UV-difference spectrum was recorded.

Peptide Binding Assays-- A solution of native OxyB (2.6 µM) in Tris-HCl buffer (2 ml, 50 mM, pH 8.0) was divided between two tandem cuvettes (sample and reference). After thermal equilibration at 30 °C the baseline was zeroed between 300 and 700 nm. Successive aliquots of peptide (1 or 2) dissolved in 10% methanol in H₂O were added to the sample cuvette to give a final peptide concentration in the range 10-1000 µM. An equal volume of 10% methanol in H₂O was also added to the reference cuvette (the final methanol concentration did not exceed 1%). The difference spectrum was then recorded.

Crystallization of His₆-OxyB-- Crystals were obtained at 20 °C using the hanging drop method by mixing 1 µl of protein (30 mg/ml in a solution containing 20 mM KCl, 5 mM dithioerythritol, 20 mM K-Hepes, pH 7.5) and reservoir solutions. The latter contained either 1 M ammonium sulfate, 10 mM KCl, 5 mM dithioerythritol, 100 mM K-Hepes, pH 7.5 (AS crystal form), or 9% PEG 3350, 10 mM KCl, 5 mM dithioerythritol, 100 mM K-Hepes, pH 7.5 (PEG crystal from). The crystals reached their final size (350 × 50 × 20 µm, AS crystal form; 300 × 150 × 50 µm, PEG crystal form) within 8-10 days. Heavy atom-derivatized crystals were prepared by soaking crystals in reservoir solution containing the heavy atom compounds for various periods of time. Prior to flash-cooling in liquid nitrogen the crystals were either rinsed briefly in the final cryoprotectant solution (AS crystal form) or soaked in mother liquor solutions containing increasing concentrations of the cryoprotectant (PEG crystal form). In both cases, the cryoprotectant solution consisted of the respective reservoir solution supplemented with 10% glucose and 10% sucrose. Both crystal forms contain one molecule of His₆-OxyB per asymmetric unit. The AS crystal form has the symmetry of the monoclinic space group C2 with unit cell parameters a = 100.4 Å, b = 60.4 Å, c = 90.3 Å,  = 122.8°, whereas the PEG crystal form has the symmetry of space group P2₁2₁2₁, with unit cell parameters a = 60.4 Å, b = 73.8 Å, c = 99.0 Å.

X-ray Data Collection and Structure Determination-- Crystals were kept at 100 K during measurements. Diffraction data were collected using a rotating anode or synchrotron radiation (beamline ID14-1 at the ESRF, Grenoble, beamline X11 at EMBL c/o DESY, Hamburg) and processed with the XDS suite of programs (26) (see Table I). The His₆-OxyB structure was determined with CNS 1.0 (27) using a combination of multiple isomorphous replacement and molecular replacement methods. Several polyalanine models derived from crystal structures of different cytochrome P450s were used as search models for molecular replacement. Although a solution was found with the P450terp structure (Protein Data Bank accession code 1CPT), the phase information did not allow us to build the protein model. Isomorphous HgCl₂ derivatives were obtained for the AS crystal form (Table I). The heavy atom sites were found in both isomorphous and anomalous difference Patterson maps. The heme-iron position was determined by anomalous difference Fourier methods. The heavy atom parameters were refined and phases were calculated to 2.9-Å resolution. The phases were improved by solvent flipping as implemented in the density modification procedure of CNS 1.0 (27). The density modified map was interpretable and an atomic model was built into the 2.9-Å electron density map using the program O (28). The refinement (including simulated annealing) was monitored by R_free, computed with 5% of the data randomly selected and omitted from the refinement. The resulting partial structure was used as a model for molecular replacement with the PEG crystal form. The phases from the molecular replacement solution were used to locate heavy atom sites in the ethyl/mercury/phosphate derivatives. The heavy atom sites agreed with the peaks in the isomorphous difference Patterson map. The heavy atom parameters were refined and multiple isomorphous replacement phases were calculated. Several rounds of manual rebuilding followed by refinement with subsequent combination of partially built model phases and experimental phases, using the data from both crystal forms, allowed completion of the protein model. At this point all data to 1.7-Å resolution were included in the refinement. During several rounds of refinement and manual rebuilding, solvent molecules were included in the model using the WATERPICK routine of CNS (27). All waters were checked manually and removed if displaying unusual H-bonding geometry. The final model contains 375 residues and 370 water molecules. In the AS crystal form, three parts (Met¹-Asp⁴, Ile³³-Asp³⁸, Arg⁷⁰-Arg⁸²) of the His₆-OxyB structure are too disordered to be built into the electron density (see Table I).

In an attempt to obtain a structure of the OxyB substrate complex, OxyB was co-crystallized with the putative substrate SP1066 (1, X = H) (20). Although the peptide was not visible in either crystal form (AS-native 2, PEG-native 2), the data allowed us to build the region Arg⁷⁰-Arg⁸² in the AS crystal form and the loop Ile³³-Asp³⁸ in the PEG crystal form. The coordinates and structure factor amplitudes have been deposited with the Protein Data Bank (29) (accession codes 1lfk, 1lgf, and 1lg9). The superposition of OxyB with other P450 structures was obtained by DALI (www.ebi.ac.uk/dali/) and analyzed with the program O (28).

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

Cloning of OxyB-- The three putative P450 genes responsible for oxidative phenol coupling reactions during vancomycin biosynthesis (named here oxyA, oxyB, and oxyC in analogy to homologues in the balhimycin cluster (15, 18-20)) were cloned from A. orientalis using a fragment of the previously reported glycosyltransferase gene gtfE as a probe (13). About 6.3 kilobase pairs of contiguous DNA upstream of the cloned gtf gene was fully sequenced (G/C content 66.7%). The deduced organization of ORFs in this region, shown in Fig. 2, was found to be the same as that reported for the balhimycin and chloroeremomycin gene clusters (14, 15). oxyA-C are adjacent genes, encoded on the same DNA strand. The start of oxyA is a valine GTG codon, which is preceded by a typical ribosome binding site sequence. oxyA encodes a protein of 391 residues having 88-90% sequence similarity with the homologous OxyA proteins in the chloroeremomycin producer A. orientalis A82846 (denoted ORF7 (14)) and the balhimycin producer A. mediterranei DSM5908 (15) (Table II). A perfect 18-bp inverted repeat is present in the region between oxyA and oxyB, which may serve as a transcriptional terminator. The assigned start codons of oxyB and oxyC are TTG and ATG, respectively, and each has a typical ribosome binding site sequence 8-11 bp upstream. The predicted ORFs comprise 398 and 406 amino acid residues, respectively. The OxyB and OxyC proteins show 89-94% sequence similarity to their homologous ORFs in the chloroeremomycin and balhimycin gene clusters. In contrast, the OxyB from the vancomycin producer, cloned here, shows only 50-58% similarity to the OxyA and OxyC proteins from all three producers (Table II).

View larger version (8K): [in this window] [in a new window]   Fig. 2.   The organization of ORFs in the DNA isolated from the vancomycin producer A. orientalis. On the left side is the 3'-terminal end of the peptide synthetase-3 and on the right side the 5'-terminal end of the halogenase gene (14). The oxyA, oxyB, and oxyC genes are indicated.

Production of OxyB-- The OxyB protein was produced in E. coli (24) as a fusion protein (called here His₆-OxyB) with a 20-residue N-terminal prepeptide, which includes a His₆ tag and a thrombin cleavage site. The His₆-OxyB was purified to apparent homogeneity by SDS-PAGE using metal-ion affinity chromatography and ion-exchange on Mono Q. In addition, OxyB having a native N terminus (called here native OxyB) was also produced in E. coli, and purified to apparent homogeneity in three steps.

UV-visible Spectroscopy-- His₆-OxyB and native OxyB gave essentially identical UV-visible absorption spectra that are typical of P450 hemoproteins (30). As shown in Fig. 3A, the UV-visible absorption spectrum of His₆-OxyB has a Soret band at 419 nm, and  and  peaks at 537 and 571 nm, respectively, typical for those of low-spin cytochrome P450s. Upon reduction with Na₂S₂O₄ the Soret band shifts to 423 nm, and the  and  peaks to 529 and 560 nm. Subsequent addition of CO resulted in the characteristic shift of the Soret band to 450 nm and the occurrence of a smaller peak at 545 nm. The CO difference spectrum contains a prominent peak at 450 nm (Fig. 3B). These results show that the isolated OxyB is in the oxidized form, and can be reduced with sodium dithionite.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   UV-visible absorption spectra of His6-OxyB in Tris-HCl buffer (50 mM, pH 8.0) at 30 °C. A, in green, the spectrum of His6-OxyB; in blue after reduction with Na2S2O4; and in red after addition of imidazole; B, the CO-difference spectrum of reduced protein showing the characteristic P450 band.

The addition of imidazole to the oxidized form causes a red shift in the Soret band to 426 nm, and  and  peaks at 545 and 571 nm (Fig. 3A), which is typical of P450s with imidazole as axial ligand for the iron(III) atom (30, 31). The OxyB proteins with and without the His tag behave identically, suggesting that the His tag does not perturb the heme environment or provide an imidazole ligand for the iron. Indeed, the crystal structure of OxyB shows clearly a bound water molecule in the sixth coordination position (see below).

Binding Assays-- The active site of OxyB was titrated with imidazole. Fig. 4 shows UV-difference spectra obtained with OxyB and increasing concentrations of imidazole. This type II binding spectrum reflects increased occupancy of the axial ligand site by imidazole. The data show that the binding site is saturable at high imidazole concentrations (>20 mM).

View larger version (12K): [in this window] [in a new window]   Fig. 4.   UV-visible difference spectra of OxyB upon addition of increasing concentrations of imidazole.

Binding assays were also performed with native OxyB and two putative substrates, the peptides 1 with X = H and X = Cl. As shown in Fig. 5, the addition of 1 with X = H (1 with X = Cl behaves similarly) to native OxyB, leads to distinctive type I changes in the UV-visible spectrum. A trough is observed in difference spectra at 420 nm, and a peak at 385 nm, although the latter is partially obscured by the absorption because of the peptide (_max  274 nm). Similar difference spectra are obtained in both Tris and in phosphate buffer.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   UV-visible difference spectra of OxyB upon addition of increasing concentrations of the heptapeptide 1 (X = H) (see "Results and Discussion").

It has been known for a long time that substrates and exogenous ligands can elicit two major types of spectral changes with many different P450s (31). The type I changes shift the Soret band from 418 to 388-394 nm (in the UV-difference spectrum a peak at 385 nm and a trough at 420 nm), and indicate the displacement of the axial ligand from the heme-iron atom, with a change in heme spin state from hexa-coordinated low spin to penta-coordinated high spin. The type II complexes are low spin with their Soret absorption around 417-426 nm (typically a trough at 390-410 nm and a peak at 425-435 nm in difference spectra). The concentration dependence of the spectral changes (Fig. 5), show that the binding site is not saturated even at the highest concentrations of peptide (1) tested here (about 0.5 mM). These data indicate a binding behavior typical of a type I ligand, and a rather weak interaction (K_D > 0.5 mM) of each peptide with the enzyme. Attempts to co-crystallize the peptides with OxyB failed (see "Materials and Methods").

Enzymic Activity-- The ability of commercially available spinach ferredoxin and spinach ferredoxin reductase to supply electrons to OxyB in a NADPH-dependent process was tested. Incubation of native OxyB with these redox proteins and NADPH leads to consumption of NADPH and reduction of the heme iron, as evidenced by the appearance of a typical band at 450 nm in UV-difference spectra in a CO binding assay (data not shown). This 450-nm CO-difference band is about 25% lower in intensity compared with that seen after reduction of the enzyme with dithionite under the same conditions (see Fig. 3B). It is still uncertain, however, whether or not these spinach proteins can deliver a second electron, as needed for a full catalytic cycle of a typical P450 enzyme.

When assays for enzymic activity were performed with the substrates 1, where X = H or Cl, in the presence of spinach ferredoxin and spinach ferredoxin reductase, no turnover to the expected product could be detected by high performance liquid chromatography-MS with either peptide. Possible reasons for the lack of turnover include: (a) that the peptides 1 with X = H or Cl are not the correct substrates, and/or (b) that the spinach-derived redox proteins are unable to support a full catalytic cycle of the enzyme.

The design of assays to detect enzyme activity with OxyB is complicated by the lack of current knowledge concerning its natural substrate, and the identity of any essential partner redox proteins. An oxyB null mutant produced the heptapeptide 1 with X = H (20). A different null mutant of the oxy genes produced the heptapeptide 1 with X = Cl (15, 18). These and other gene knockout experiments (15, 18-20), therefore, suggest that the linear heptapeptide should be fully assembled before the cross-linking steps catalyzed by OxyA-C occur. However, the timing of release of the putative heptapeptide (1, X = H or Cl) from the peptidyl carrier domain of the vancomycin peptide synthetase is unclear. It is possible that the heptapeptide must be attached through its C terminus to a peptide carrier domain to act as a substrate for OxyB. In analogy, the P450 enzyme NovI that catalyzes the -hydroxylation of tyrosine during novobiocin biosynthesis, hydroxylates a tyrosine moiety that is coupled as a thioester to a peptide carrier domain (21). A similar route for the production of -hydroxytyrosine may also occur during vancomycin biosynthesis (6).

The crystal structure described below reveals an active site on the distal side of the heme, which, unlike other known P450 enzymes, is highly exposed to the solvent. This might allow docking of a carrier protein and positioning of a coupled heptapeptide close to the active site heme. Presently, an appropriate heptapeptide linked to a peptide carrier protein is not available to test this hypothesis.

The putative reductants of the OxyA-C hemoproteins in A. orientalis, presumably a ferredoxin-like and a ferredoxin reductase-like protein, have also not been identified, and no genes encoding such redox proteins were identified in four of the glycopeptide antibiotic gene clusters sequenced so far (vancomycin (13), balhimycin (15), teicoplanin (16), and chloroeremomycin (14)). A ferredoxin-like gene (of unknown function) was discovered, however, in the complestatin cluster (17). A cytochrome P450 from the higher plant Berberis stolonifera has been cloned and sequenced, and shown to catalyze a C-O phenol coupling reaction using molecular oxygen as the oxidizing agent with a eukaryotic NADPH-cytochrome P450 reductase (32). This important result establishes a role for both molecular oxygen and external electron transfer proteins in phenol coupling reactions catalyzed by P450 enzymes.

Crystallization and X-ray Structure-- OxyB was crystallized in two different space groups using PEG or ammonium sulfate as precipitant. Although attempts to determine the structure of OxyB by molecular replacement using different P450s as search models were successful with P450terp, the phases were not good enough for refinement. Therefore, the OxyB structure was determined by multiple isomorphous replacement using mercury derivatives. OxyB exhibits the typical P450-fold (33, 34). As can be seen from Table III, the structural similarity is highest to P450nor (35), P450terp (36), CYP119 (37), and P450eryF (38) with root mean square differences of 2.9, 2.8, 2.8, and 3.4 Å, respectively. Analogous to P450cam (39) and P450eryF (38) that have poorly defined N termini, the N terminus of OxyB has high temperature factors.

The structure starts with Asp⁵, the A'-helix is only three residues long. It is interesting that OxyB contains an additional -hairpin (0 in Fig. 6), which is also present in P450nor, although in this case a -strand is replaced by a loop conformation. Other P450s lack this structural element. The main chain atoms of Arg¹³ located in this hairpin are solvent accessible, its side chain is turned toward the inside of the molecule, and the guanidinium group forms hydrogen bonds with the carbonyl oxygen atoms of Leu²⁸⁰ and Thr²⁸¹. This interaction seems to stabilize Thr²⁸¹ in an energetically strained conformation (Thr²⁸¹ is the only residue in the nonpreferred region of the Ramachandran plot) in which its hydroxyl group interacts with the carbonyl of Arg²⁷⁸ and the side chain amide of Asn³¹⁶. Pro²⁸³ and Val²⁸² belong to the region connecting the J'-helix and 14 and form a "wall" of the substrate binding pocket. Moreover, Pro²⁸³ is in van der Waals contact with Asn²⁴⁰ and may be required for positioning of this catalytically important residue.

View larger version (34K): [in this window] [in a new window]   Fig. 6.   A, sequence alignment of OxyB, P450nor, P450eryF, Cyp119, P450terp, and P450cam, based on their structural superposition. Gaps in the alignment are indicated by dots. Secondary structure elements of each sequence are shown in red (-helices) and blue (-strands), names of the -helices and -strands are indicated above the sequences. B, stereo view of ribbon representation of the OxyB structure. -Helices and -sheets are labeled according to A.

Major differences to other known P450 structures occur in the region of the B', F, and G helices (see Fig. 6), which are known to be important in substrate binding (33, 34, 40). There is very weak electron density for most of the region corresponding to the B'-helix in one of the AS crystal forms, whereas in the PEG crystal form the residues between Arg⁶⁵ and Glu⁸³ are missing. The distances between the last well defined residues are 14 and 8 Å, respectively, data which indicate a region of high flexibility in the protein.

Other significant differences with known P450s are in the orientation of the F and G helices. The structural similarity is very high to P450nor, both in terms of fold and lengths of the helices, although not their orientation (see Fig. 7). The root mean square difference between the C atoms of residues 162 to 194 and 163 to 195 in OxyB and P450nor, respectively, is 5.4 Å in the native structures but only 1.0 Å after applying a rigid body transformation. In OxyB, the F and G helices are rotated out of the active site compared with P450nor, resulting in a much more open active site (see Fig. 8), consistent with the larger substrate of OxyB.

View larger version (32K): [in this window] [in a new window]   Fig. 7.   Stereo view comparing the distal (B', F, G, and I helices) and proximal sites (Cys ligand loop) of OxyB (black) and P450nor (1rom) (gray). The figure was generated by manual superposition of the hemes and subsequent application of the resulting orientation and translation matrix to the reminder of the protein atoms. The figure was prepared using MOLSCRIPT (50) and RASTER3D (51, 52).

View larger version (93K): [in this window] [in a new window]   Fig. 8.   Stereo view of the electron density map and the model at the distal heme pocket of OxyB (AS crystal form) showing the conformation around Asn240. The 2Fo  Fc A-weighted electron density map is countoured at 1. Possible hydrogen bonds with a cut-off of 3.2 Å are indicated by dashed lines. The figure was prepared using BOBSCRIPT (50, 53) and RASTER3D (51, 52).

Another feature common to OxyB, P450nor (35), and Cyp119 (37) is that the H and I helices are connected by a loop and not by an extended -hairpin as in other P450s. This region has been implicated in binding the redox partner and shown to make direct contact with the FMN binding domain in the crystal structure of the complex of the heme and FMN domains of P450 BM3 (41). The significance of this similarity between P450nor, which does not require a redox protein partner for electron transfer, and OxyB, for which no gene encoding a redox partner has so far been found, is so far unclear.

The most highly conserved regions in the P450 superfamily are near the heme, and include the I and L helices. The latter contains the P450 signature sequence FGHGXHXCLG, with Cys³⁴⁷ being the proximal axial thiolate heme ligand in OxyB. The I-helix spans the whole molecule and is located above the heme (see Figs. 6B and 7). It contains two catalytically important residues, an acidic residue followed in most cases by a threonine. This region is believed to be involved in the proton relay shuttling protons from the solvent to the heme bound oxygen molecule during catalytic turnover (42-45). In OxyB the acidic residue is Asp²³⁹ followed by Asn²⁴⁰ whose side chain points in the active site. The Met-water molecule (Wat²⁶⁸) bound to the heme iron forms hydrogen bonds with the Asn²⁴⁰ amido group and the carbonyl oxygen of Ala²³⁶ (see Fig. 8). The latter interaction and the conformation of the alanine are similar to the corresponding ones in P450terp and substrate-free P450-BM3, and stabilize the iron-bound water molecule (43, 46). In P450-BM3, substrate binding induces a conformational change in the I-helix resulting in a shift of Ala²⁶⁴ and a concomitant movement of the Met-water to a new position thus allowing binding of oxygen to the heme iron (47). A similar situation can be expected for OxyB.

In OxyB there are two water molecules (Wat³⁹ and Wat⁷¹) located in the groove of the I-helix (Fig. 8). Wat⁷¹ forms hydrogen bonds with the carbonyl oxygen atom of Leu²³⁵ and amide of Asp²³⁹. Wat³⁹ forms hydrogen bonds with the backbone amide of Asn²⁴⁰, the carbonyl oxygen of Ala²³⁶, and the side chain amido group of Asn²⁴⁰. It corresponds to Wat²⁰² in substrate-free P450-BM3 and interrupts the normal hydrogen bonding pattern of the backbone residues of the I-helix. This results in a kink of the I-helix that is believed to be important for oxygen binding and the proton shuttle mechanism (48). The distance between the side chain amide of Asn²⁴⁰ and the carbonyl oxygen of Ala²³⁶ is 3.5 Å. This is longer than the corresponding distance in other P450s, such as P450cam, where Thr²⁵² forms a strong hydrogen bond with the carbonyl oxygen atom of Gly²⁴⁸ (2.4 Å) (49). The network of water molecules involving the Met-water Wat²⁶⁸ begins in the niche made up by the regions between the J'-helix and 1-4, the turn between 4-1 and 4-2 and B'-helix part, and extends to the solvent.

The observation of a water molecule bound to the heme iron is consistent with the UV-visible spectra of OxyB indicating a low spin heme (see below). The iron is in the heme plane, the distance to the thiolate sulfur of Cys³⁴⁷ is 2.3 Å. As in other P450s, the A propionic acid interacts in a bidentate fashion with Arg²⁸⁹ (which interacts also with the carbonyl oxygen atom of Thr²⁸⁷ (2.8 Å), Thr²⁸⁷, and Wat¹⁸); the D propionic acid interacts with the imidazole nitrogens of His⁹⁶ and His³⁴⁵, and with Arg¹⁰⁰.

The comparison of the OxyB structures obtained from the ammonium sulfate and PEG crystal forms shows some interesting differences. The most obvious one is that there is weak electron density for the region of the B'-helix in one of the AS crystal forms (see above), and that the 11-12 loop is more disordered. The F and G helices are shifted somewhat, resulting in a more closed active site in the AS crystal form. These differences also result in changes in the conformations of the side chains of Gln²³², Met⁸⁹, and Leu²³⁵ that line the heme cavity. Despite these differences, the water-accessible surface of the heme, which is ~30 Ų, does not change significantly. The differences between the two crystal forms may reflect a functionally important flexibility of the OxyB molecule that is required for substrate binding and concomitant displacement of the Met-water molecule, analogous to the situation in P450-BM3 (47).

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

The hemoprotein OxyB implicated in the first oxidative phenol coupling reaction during vancomycin biosynthesis has been cloned and overproduced in E. coli, both with a native N terminus and as a fusion with an N-terminal hexahistidine tag. The native and the His₆-tagged proteins gave UV-visible absorption spectra characteristic of P450-like hemoproteins, with the heme in the low spin ferric state. After reduction with dithionite the ferrous form is produced, which gives a peak at 450 nm in CO-difference UV-visual spectra. The resting state can also be reduced with spinach ferredoxin and ferredoxin reductase, because the resulting ferrous form again gives the characteristic 450-nm band in CO-difference spectra. This shows that native OxyB can accept one electron from these redox proteins. The addition of putative heptapeptide substrates to ferric OxyB leads to typical type I changes in the UV-visible spectrum. A trough is observed in difference spectra at 420 nm, and an increased absorption at 385 nm, although the latter is partially obscured by the absorption because of the peptide (_max  274 nm). These data indicate a weak binding interaction between the peptides and OxyB, consistent with the displacement of the distal axial water ligand of the heme iron and a change in spin state from hexa-coordinated low spin to penta-coordinated high spin. The putative heptapeptide substrates, however, are not turned over into new products upon incubation with OxyB, spinach ferredoxin, and ferredoxin reductase. The reasons for this lack of turnover are not known, but indicate that the substrates and/or partner redox proteins used are insufficient for a full catalytic cycle of this P450 enzyme.

The His₆-OxyB protein was crystallized in two different space groups using PEG or ammonium sulfate as precipitants. The structures were solved to 1.7-Å resolution. Attempts to co-crystallize the hemoprotein with the putative heptapeptide substrate SP1066 (20) (1, X = H) failed to give visible electron density for bound peptide. OxyB exhibits the typical P450-fold, with highest structural similarity to P450nor, P450terp, CYP119, and P450eryF.

The most highly conserved regions in the P450 superfamily are near the heme, and include the I and L helices. The latter contains the P450 signature sequence FGHGXHXCLG, with Cys³⁴⁷ being the proximal axial thiolate heme ligand in OxyB. The observation of a water molecule bound to the heme iron is consistent with the UV-visual spectra of OxyB indicating a low spin heme. The iron is in the heme plane, the distance to the thiolate sulfur of Cys-347 is 2.3 Å.

The structural similarity is very high to P450nor, both in terms of fold and lengths of the F and G helices, although not their orientation (see Fig. 7). In OxyB, the F and G helices are rotated out of the active site compared with P450nor, resulting in a much more open active site, consistent with the larger presumed substrate of OxyB. In OxyB, P450nor, and Cyp119 the H and I helices are connected by a loop and not by an extended -hairpin as in other P450s. This region has been implicated in binding the redox partner and shown to make direct contact with the FMN binding domain in the crystal structure of the complex of the heme and FMN domains of P450 BM3.

The I-helix spans the whole molecule and is located above the heme (see Figs. 6B and 7). It contains two catalytically important residues, an acidic residue followed in most P450s by a threonine. This region is believed to be involved in the proton relay shuttling protons from the solvent to a heme-bound oxygen molecule during catalytic turnover. In OxyB the acidic residue is Asp²³⁹ followed by Asn²⁴⁰ whose side chain points in the active site (Fig. 8). The differences observed between the two crystal forms of OxyB may reflect a functionally important flexibility of the OxyB molecule that is required for substrate binding and concomitant displacement of the Met-water molecule, analogous to the situation in P450-BM3 (47).

	ACKNOWLEDGEMENTS

We thank Annelies Meier, Georg Holtermann, and Elisabeth Hartmann for expert technical assistance. We (O. P. and I. S.) also thank the staff at EMBL, c/o DESY, and the European Synchrotron Radiation Facility for support during data collection, and Alexey Rak and Roger S. Goody for continuous encouragement and support.

	FOOTNOTES

^* This work was supported in part by the Deutsche Forschungsgemeinschaft and European Union program "European Community- Access to Research Infrastructure Action of the Improving Human Potential Program to the EMBL Hamburg Outstation, Contract HPRI-CT-1999-00017."The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the DDBJ/GenBank^TM/EBI Data Bank with accession number(s) AF486630.

The atomic coordinates and the structure factors (code 1lfk, 1lgf, and 1lg9) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/).

^§ Both authors contributed equally to this work.

^§§ Supported by European Union MEGA-TOP Grant QLK3CT- 1990-00650.

^¶¶ Supported by European Union Grant MEGA-TOP QLK3-CT-1990-00650 and Bundesamt für Bildung und Wissenschaft Grant 99.0241. To whom correspondence may be addressed: Organisch-chemisches Institut, Universität Zürich, Winterthurerstr. 190, 8057 Zürich, Switzerland. Tel.: 41-0-1-635-4242; Fax: 41-0-1-635-6833; E-mail: robinson@oci.unizh.ch.

Supported by the "Richard und Anne-Liese Gielen-Leyendecker Stiftung and the Deutsche Forschungsgemeinschaft." To whom correspondence may be addressed: Max Planck Institute for Medical Research, Dept. of Biomolecular Mechanisms, Jahnstr. 29, 69120 Heidelberg, Germany. Tel.: 49-0-231-133-2738; Fax: 49-0-231-133-2797; E-mail: ilme.schlichting@mpimf-heidelberg.mpg.de.

Published, JBC Papers in Press, August 30, 2002, DOI 10.1074/jbc.M206342200

	ABBREVIATIONS

The abbreviations used are: ORF, open reading frame; PEG, polyethylene glycol.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES