Structure of OxyAtei: completing our picture of the glycopeptide antibiotic producing Cytochrome P450 cascade Kristina Haslinger and Max J. Cryle Max Planck Institute for Medical Research, Heidelberg, Germany

Correspondence M. J. Cryle, EMBL Australia, Monash University, Clayton, Victoria 3800; The Department of Biochemistry and Molecular Biology and ARC Centre of Excellence in Advanced Molecular Imaging, Monash University, Clayton, Victoria 3800, Australia Fax: +61 3 9905 4302 Tel: +61 3 9905 0771 E-mail: [email protected] K. Haslinger, Department of Chemical Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139, USA Fax: +1 617 324 3127 Tel: +1 617 258 8037 E-mail: Kristina.Haslinger@mpimf-heidelberg. mpg.de; [email protected] (Received 10 January 2016, revised 20 January 2016, accepted 25 January 2016, available online 28 January 2016) doi:10.1002/1873-3468.12081 Edited by Richard Cogdell

Cyclization of glycopeptide antibiotic precursors occurs in either three or four steps catalyzed by Cytochrome P450 enzymes. Three of these enzymes have been structurally characterized to date with the second enzyme along the pathway, OxyA, escaping structural analysis. We are now able to present the structure of OxyAtei involved in teicoplanin biosynthesis – the same enzyme recently shown to be the first active OxyA homolog. In spite of the hydrophobic character of the teicoplanin precursor, the polar active site of OxyAtei and its affinity for certain azole inhibitors hint at its preference for substrates with polar decorations. Keywords: Cytochrome P450; glycopeptide antibiotics; nonribosomal peptide synthesis; phenolic coupling; teicoplanin Highlights • X-ray structure of a D-O-E ring forming Cytochrome P450 in glycopeptide biosynthesis • The structure agrees with the canonical fold of peptide-oxidizing P450 enzymes • The active site is polar, which is conserved among D-O-E ring forming enzymes • Highly conserved aromatic amino acids cluster in the active site • Binding of azole inhibitors reflects a preference for binding of polar compounds.

Oxidative tailoring of peptides is an important modification during the maturation of natural products [1]. Of all the oxidative enzymes involved in such processes, Cytochrome P450 monooxygenases (P450s) are one of the most diverse families both in terms of the reactions they catalyze and the substrates they accept [2]. One complex example for P450s acting on peptides is the three- or four-step oxidative cross-linking of aromatic side chains during the biosynthesis of

glycopeptide antibiotics (GPAs), a group of clinically relevant compounds active against gram-positive bacteria. Herein, aryl and aryl-ether bonds are introduced by individual P450s in a highly regiospecific manner and in a strict reaction order. This pathway has been elucidated through in vivo gene disruption studies on the balhimycin- (vancomycin-like, Type-I GPA) and A47934-producing pathways (teicoplanin-like, Type-IV GPA, Fig. 1), thus enabling the assignment of each

Abbreviations Aindex, absorption; GPA, glycopeptide antibiotic; NRPS, nonribosomal peptide synthetase; P450, Cytochrome P450 monooxygenase; PCP, peptidyl carrier protein; suffix –tei or –van, belonging to the teicoplanin- or vancomycin-producing pathway, respectively.

FEBS Letters 590 (2016) 571–581 ª 2016 Federation of European Biochemical Societies

571

Structure of OxyAtei

K. Haslinger and M. J. Cryle

Fig. 1. Cytochromes P450 involved in GPA biosynthesis shown with the teicoplanin aglycone. OxyD (purple, PDB-ID: 3MGX; b-hydroxylase not present in all GPA pathways); OxyB (red, PDB-ID: 1LGF; first phenolic coupling enzyme); OxyE (orange, PDB-ID: 3O1A; second or third phenolic coupling enzyme); OxyC (blue, PDB-ID: 1UED; last aryl coupling enzyme).

cyclization reaction to the respective P450s [3,4]. The conserved P450s OxyA, OxyB and OxyC were found to act in a sequential manner on the heptapeptide precursor in the order OxyB-OxyA-OxyC, during which time the peptide is rigidified and obtains the specific three-dimensional shape that is crucial for the antibiotic activity of mature GPAs. Furthermore, in Type-III and IV GPAs a fourth (nonessential) cross-link is formed by the P450 OxyE around the same time as OxyA is active [5]. Some GPA gene clusters encode an additional P450 (OxyD) [6–9] that is associated with the b-hydroxylation of tyrosine as building blocks for the peptide precursor [10–12]. However, while OxyD has been shown to act on its substrates prior to their incorporation into the peptide [11,12], the peptide tailoring P450s are recruited to the main peptide assembling enzyme complex, the nonribosomal peptide synthetase (NRPS) itself after peptide formation [13]. P450 recruitment to the NRPS-bound peptide occurs via a specialized domain, the X-domain, which is conserved among all GPA NRPSs [14]. Due to the importance of the GPAs and the reliance on their fermentative production, significant research has been conducted into the biosynthesis of their peptide backbone in vitro, with particular emphasis on the P450 reaction mechanism [15], the orchestration of the P450 cascade [14,16,17], the activity of the first peptide 572

oxidizing enzyme OxyB [15,18–22] and the molecular structures of the Oxy enzymes [9,22–24]. The first X-ray crystal structure of OxyB from Amycolatopsis orientalis revealed features that set it apart from other P450 enzymes [9]: the a-helices (A-L) that form the canonical P450 fold with an overall spherical shape around the central, thiolate-ligated heme cofactor appear to leave a large, open cavity on the distal side of the heme. This opening is caused by the comparatively short length of the F- and G-helices that form the ceiling of the active site cavity and their outward rotation [9]. Such arrangements had not been observed and were only later described for P450s with large substrates such as macrolides [25–27] or carrier protein bound amino acids [28,29] and fatty acids [30] indicating their importance for the accommodation of bulky substrates. The structures of OxyC [23] and OxyE [24,31] then showed that this active site topology is characteristic for the group of P450 monooxygenases interacting with NRPS, albeit with minor differences. The latest, and possibly most informative structure revealed the interaction site of these P450s with the recruitment domain of the NRPS (OxyBtei/Xdomain complex): a highly conserved motif in the P450 F-helix (PRDD) mediates the direct interaction with the X-domain and likely allows the distinction between peptide tailoring and precursor supplying

FEBS Letters 590 (2016) 571–581 ª 2016 Federation of European Biochemical Societies

K. Haslinger and M. J. Cryle

P450 enzymes [14]. With these structures uncovered, an important remaining goal was the structural characterization of the second phenolic coupling enzyme in the peptide-tailoring cascade, OxyA. In this study, we now report the X-ray crystal structure of OxyA from Actinoplanes teichomyceticus, OxyAtei, and thus complete the gallery of P450 structures involved in GPA aglycone formation. Furthermore, we performed binding studies of azole inhibitors to probe the P450 active site, which have revealed important differences to the previously studied P450, OxyBtei.

Structure of OxyAtei

protein solution in the front cuvette in a stepwise manner with the same amount of buffer or DMSO added to the reference cuvette and after a 2 min equilibration period spectra were collected. The resulting amplitudes of the difference spectra were extrapolated (DA = AmaxAmin) and plotted against the ligand concentration. The resulting data points were fitted to a one-site binding model (Eqns 1,2) to determine the maximal amplitudes (DAmax) and dissociation constants (KD). The maximal amplitudes were further compared to the initial heme absorption at 418 nm to report change in spin state shift in percent. y ¼ DAmax  ½lig=ðKD þ ½ligÞ; ðimidazoleÞ or

Methods

y ¼ ðDAmax  DA0 Þ  ð½lig  0:5  ðKD þ ½lig  ½OxyA þ sqrtð4  KD  ½lig þ ðKD  ½lig

Cloning of constructs, expression and purification Based on the pET151D::oxyA plasmid containing the synthetic tcp18 gene (Uniprot protein ID: Q6ZZI8) published in [14] a second expression vector (pET151D::oxyAII) was generated by round-the-horn site directed mutagenesis using the primer set: fwd 50 AGTAGCATGTTCGAGGAGATCA ACGTT and rev 50 TCCCTGAAAATACAGGTTTTC to introduce a Gly-Ser-Ser linker between the Tobacco Etch Virus (TEV) protease recognition site and the first helix of OxyAtei. Sequences were confirmed using standard T7 promotor and terminator primers. Both plasmids allow gene expression under the control of a T7 promoter in frame with a TEV-cleavable N-terminal hexahistidine tag. OxyAtei was expressed from the pET151D::oxyA and pET151D::oxyAII vectors and purified as described previously for OxyBtei [22]. In brief, OxyAtei was expressed in the E. coli KRX strain (Promega, Mannheim, Germany) grown in terrific broth media and purified via a four-step protocol: metal affinity chromatography, tag removal by TEV protease, anion exchange (ResourceQ, GE Healthcare, Munich, Germany) and final size exclusion chromatography (Superose12, 10/300 GL, GE Healthcare, Munich, Germany) (yield for OxyAtei:170 nmol/L culture, yield for OxyAtei linker variant: 160 nmol/L); protein identity was confirmed by peptide mass fingerprinting.

UV-visible spectroscopy Imidazole (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in water (100 mM stock), adjusted to pH 8.0 with HCl and stored at rt. Other azole inhibitors (Abblis Chemicals LLC, Houston, TX, USA) were freshly dissolved in DMSO (1 mM stocks). OxyAtei was diluted to 2.5 lM in 1 mL 50 mM Tris/HCl pH 8.0 buffer and split into two cuvettes. The cuvettes were placed into a Jasco V-650 double beam spectrophotometer and thermally equilibrated to 30 °C. The instrument was blanked and the initial baseline was collected between k = 350 and 600 nm. Ligand was added to the

þ ½OxyAÞ2 ÞÞÞ=½OxyA þ DA0 Þ; ðother inhibitorsÞ with DA0 ¼ 0 and ½OxyA ¼ 2:5lm

ð1Þ

ð2Þ

X-ray crystallography Optimized OxyAtei crystals were obtained in a 2 lL hanging drop vapor diffusion setup (6 mg/mL OxyAtei mixed at a 1:1 ratio with 0.1 M Mes/NaOH pH 6.5, 14.4% (w/v) PEG 8000, 20% (v/v) glycerol, 0.16 M calcium acetate; 293 K). Box-like crystals appeared within 1 week (170 lm 9 60 lm 9 60 lm), were harvested in cryo loops without further cryogenic protection and flash cooled in liquid nitrogen. Throughout storage and data collection, the crystals were kept at 100 K. Data collection was performed at the Swiss Light Source (Paul Scherrer Institute, Villigen, Switzerland),  A native data set beamline X10SA, wavelength k = 1.072 A. was collected and processed with the XDS program suite [32] including XDS, XSCALE and XDSCONV. Phases were obtained from molecular replacement with the structure of StaF (unpublished data) by PHASER [33]. The initial model  and manually built was refined with REFMAC [34] (2.1 A) in COOT [35] in iterative cycles until optimal geometry was achieved. The quality and validity of the final structures was assessed by MOLPROBITY [36]. Superposition of polypeptides was performed by SSM [37] implemented in COOT and structural figures were generated using PyMol (Schr€ odinger LLC). PDB search for structurally related entries was performed with Dali [38] and structural comparisons with DaliLite [39]. Sequence alignments were performed with ClustalOmega [40] and sequence logos were generated using Weblogo [41].

Results First crystal structure of a D-O-E ring forming Oxy enzyme, OxyAtei In previous studies we have shown that the predicted gene tcp18 [42,43] in the gene cluster associated with

FEBS Letters 590 (2016) 571–581 ª 2016 Federation of European Biochemical Societies

573

Structure of OxyAtei

K. Haslinger and M. J. Cryle

teicoplanin biosynthesis from Actinoplanes teichomyceticus codes for the D-O-E ring forming P450 monooxygenase OxyAtei, by demonstrating its activity in vitro [14,44]. Since this was the first active monooxygenase of its kind, OxyAtei became an important target for structural analysis and was subjected to crystallization trials. Initial hits were optimized in a hanging drop vapor diffusion setup. Native data sets were collected under cryogenic conditions from single crystals and the highest resolution data set was processed and used for phase determination by molecular replacement. Iterative model building and refinement led to the structure deposited in the protein data bank (PDB) under PDB-ID: 5HH3 (Fig. 2, Table 1). The protein crystallized in space group P212121 with two molecules in the asymmetric unit; both bear a prosthetic heme group with full occupancy (Fig. S1). The two protomers appear to be linked via a coordinative bond between the flexible N terminus of protomer B and the heme iron of protomer A (position 6, distal ligand, Fig. 2C), while the N terminus of protomer A appears to be unbound and mobile. The observed dimerization was highly unexpected since previous solution state experiments had not indicated that dimers are formed [14]: thus, the protein arrangement in the crystal is likely an artifact of the crystallization procedure. Since the coordination only involves a few flexible residues at the N terminus of protomer B that are residuals of the purification tag and since the structured cores of both protomers are highly congruent  [Root mean square deviation (RMSD) of Ca = 1.7 A, Fig. 2B], we herein focus on the structure of protomer A. OxyAtei adopts the common P450 fold comprising one mainly a-helical and one b-sheet rich half that accommodate the heme in the center of the molecule (Fig. 2A) [45]. The prominent I-helix (lime) spans the core of the protein and forms a four helix bundle together with the D-, E- and L-helices (dark gray) below the capping F- and G-helices (green). Furthermore, OxyAtei contains two b-turns (b-2 and b-3) and a five-stranded b-sheet (b-1) (light green). In addition to these canonical secondary structure elements, one further helix is observed at the N terminus (A0 , dark green). This helix lies parallel to the five-stranded b-sheet and thus narrows the opening above the active site compared to other Oxy structures [9,22–24]. The structured cores display strong structural similarity to other P450 proteins in the PDB as identified by Dali search [38] for both molecules present in the asymmetric unit (Table S1). In particular, the great similarity to other Oxy enzymes involved in GPA production (OxyEtei, OxyBvan, OxyBtei and OxyCvan) shows that in spite of the dimerization, OxyAtei dis574

plays the characteristic features of this group of enzymes – namely the rather short F- and G-helices and a large, open active site [9], which allows for the accommodation of the heptapeptide substrate (Fig. 3B). This unique feature clearly sets this group of enzymes apart from the structurally related non-Oxy P450 enzymes identified by Dali search for both molecules present in the asymmetric unit (e.g. CYP105-P1 [25]/-AS1 [46]/-N1 [47], MycG [27], P450nor [48] and OleP [49]). Compared to the OxyB and OxyC structures, in OxyAtei these helices are pulled down toward

Fig. 2. X-ray crystal structure of OxyAtei. Overall structure with secondary structure elements highlighted and labeled (helices A’ to L, b-1 to b-3) (A); overlay of protomer A (green) and B (blue) with focus on the observed differences in helix orientations (B); active site of protomer A with bound N terminus of protomer B (blue = nitrogen, red = oxygen, brown = iron, yellow heme carbon, orange = protomer B carbon atoms; blue mesh = mFc-dFo map, r = 3) (C) and comparison of active site residue conformations in protomers A (green) and B (blue) (D).

FEBS Letters 590 (2016) 571–581 ª 2016 Federation of European Biochemical Societies

K. Haslinger and M. J. Cryle

Structure of OxyAtei

Table 1. Statistics of data collection and refinement for the OxyAtei crystal structure. Data Collection Space Group Cell Dimensions a, b, c ( A) Molecules/asymmetric unit Resolution ( A)a a Rmeas I/rIa CC1/2a Completeness (%)a Redundancy Wilson B-factor ( A)b Refinement Unique Reflectionsa Resolution in Refinement Rwork/Rfree (%)b,c No. of atoms Protein (protomer A/B) Heme (protomer A/B) Water Glycerol/acetate B-factors overall Protein (protomer A/B) Heme (protomer A/B) Water Glycerol/acetate RMSD Bond lengths ( A) Bond angles (°) Geometry statisticsd MolProbity score PDB-ID

P212121 (19) 58.5 103.1 153.0 2 48.8-2.1 8.3 (74.4) 20.0 (4.3) 99.9 (93.2) 100 (99.9) 13.1 46.0 54 785 (7001) 48.8–2.1 19.8/22.9 (24.0/25.8) 3021/3105 43/43 193 54/12 54.1 64.3/45.7 37.3/30.0 45.6 69.9/75.9 0.009 1.3 97.4/0.3e/95.3 1.1 5HH3

a

Numbers in parentheses correspond to the highest resolution shell (2.2–2.1  A). b Rwork = ∑ ||Fo||Fc||/∑|Fo|, calculated from the working reflection set; Rfree calculated in the same manner using the 5% test set reflections. c Numbers in parentheses correspond to the highest resolution bin (2.15–2.10  A). d Calculated by MOLPROBITY [36]: percentage of the protein residues in most favored and disallowed regions of the Ramachandran plot and percentage of favored rotamers. e Residues in disallowed region: F382 (clearly defined density).

the heme, whereas in the OxyE structure these helices adopt an even more closed conformation (Fig. S2). Active site The active site of OxyAtei is lined by the I-helix, the last strand of the b-1 sheet and the B-C connecting loop with the F and G-helices, b-3 and the A’- helix forming the ceiling of the central cavity. In protomer B a short B’-helix is formed that projects upwards perpendicular to the heme plane and forms

direct and water-mediated hydrogen bonds with b-1 and the A’-helix. The following loop then turns toward the I-helix and descends toward the C-helix thus forming the far boundary of the active site (Fig. 2B, blue). Within the I-helix, the catalytically important residues Asp235 and Gln236 are located in close proximity to the heme. Comparison of the side chains in the active sites of the two protomers reveals that all residues in the rigid areas of the active site adopt nearly identical conformations (Fig. 2D); whereas those residues in the more flexible or even unstructured region (close to B- and C-helices) adopt different conformations. In particular, the hydrophobic network only present in the active site of protomer A (green residues in Fig. 2D) and the outward rotation of the side chain of His283 by 107 ° compared to protomer B are likely directly caused by the coordination of the N terminus of protomer B in the active site. However, the overall conformational identity indicates that the coordination of the N terminus does neither distort the structure of the donor protomer nor the active site of the acceptor. The residues lining the upper part of the central cavity of OxyAtei are polar, as is also observed in other Oxy structures [9,22–24]. However, the residues in the direct surroundings of the heme are less hydrophobic than observed for the C-O-D and AB ring forming Oxy enzymes, OxyB and OxyC [9,22,23]. While these enzymes largely share identical residues lining the active site, in particular the b-sheet in the active site of OxyAtei displays some amino acid changes that affect the polarity of the active site: the proline residues in positions 279 and 282 are replaced by an aspartic acid and a threonine, respectively, and methionine 86 is changed into a threonine. In the I-helix Leu235 is replaced by a glycine. At the same time, three aromatic amino acids are introduced at the far end of the active site, namely W81 and F84 in the B/C-region and F228 in the C-terminal end of the I-helix (Fig. 3A). These specific amino acid replacements are largely conserved among OxyA sequences (Fig. 3C, Fig. S3) except for the P282T variation: about half of the OxyA-like sequences identified by NCBI BLASTp [50] and confirmed to be associated with GPA gene clusters, display leucine or isoleucine residues in this position (Fig. 3C). With the current knowledge, however, we cannot correlate the identity of residue 282 to the substrate specificity of the OxyA homologs. Furthermore, it is possible that this amino acid substitution is rather of structural importance: while OxyB, OxyC, and OxyE enyzmes share a conserved PXP motif in this position leading to a loop conformation of the surrounding stretch of amino acids, the T/LHT/P sequence in OxyA proteins allows the formation of a

FEBS Letters 590 (2016) 571–581 ª 2016 Federation of European Biochemical Societies

575

Structure of OxyAtei

K. Haslinger and M. J. Cryle

Fig. 3. Comparison of the OxyAtei active site to OxyBtei (OxyA (green), OxyB (beige, PDB-ID: 4TVF) noncarbon atoms colored as in Fig. 2) (A); OxyA peptide substrate (B); Sequence logos of selected active site regions based on all annotated GPA-modifying P450 sequences (OxyA residues discussed in the text are highlighted with boxes: hydrophobic residues in blue and polar residues in green; residue ranges refer to OxyAtei) (C).

linear strand that is part of the b-1 sheet. The strict conservation of all amino acid exchanges indicates a specific requirement for the described active site architecture: a long fourth strand in the b-1 sheet, polar residues in the heme environment and aromatic amino acids neighboring the flexible B/C region. In particular, the occupied active site of protomer A (Fig. 2D, green) shows that the three aromatic amino acids form a network that may involve p-interactions and could easily contribute to the coordination of a substrate peptide comprising aromatic amino acids. In an attempt to eliminate the heme coordination observed in the original OxyAtei construct, new constructs of the OxyAtei protein were designed with modifications in the N-terminal region. Here, the linker between the TEV recognition site and the native N terminus was shortened and changed into a flexible Gly-Ser-Ser linker. The protein was prepared in an identical manner and appeared well behaved, although we did not obtain any crystals, thus indicating that the dimerization observed in our crystal structure is necessary for crystal packing. Probing the active site of OxyAtei With the active site of OxyAtei shown to be more polar than other Oxys, we set out to test the effect of bind576

ing of azole inhibitors of varying size and hydrophobicity (Fig. 4A): such inhibitors have previously been shown to bind to OxyBtei with similar affinities yet with varying effect on the heme coordination state [22]. Azole inhibitors are powerful heme ligands that are able to displace the resting state water from the distal heme position and to form a near irreversible, coordinative N-Fe bond themselves [51]. This effect can be further complemented by interactions with active site lining residues leading to high apparent affinities [52]. Initially, we tested the binding of the small ligand imidazole to OxyAtei in a UV-visible spectroscopic titration study. Similar to OxyBtei, a type-II spectral response with a red shifted Soret peak is evoked, best observed in the difference spectrum as a minimum at 420 nm and a maximum at 435 nm. While the fitted maximal amplitudes in the OxyAtei spectra are similar to OxyBtei, the affinities differ by two orders of magnitude: for OxyAtei the apparent dissociation constant of imidazole is in the medium micromolar range whereas for OxyBtei it is in the low millimolar range (Fig. 4B/C, Table 2). Titrations of OxyAtei with the other azole inhibitors (Fig. 4A, 1–3) evoked type-II spectral shifts with dissociation constants in the high nanomolar range (Fig. 4D, Table 2). Interestingly, OxyAtei displays similar affinity to all inhibitors, however, with the largest amplitudes

FEBS Letters 590 (2016) 571–581 ª 2016 Federation of European Biochemical Societies

K. Haslinger and M. J. Cryle

Structure of OxyAtei

Table 2. Inhibitor binding to OxyAtei. OxyAtei DAmax  SEa Imidazole Ketoconazole Miconazole Clotrimazole

0.14 0.12 0.12 0.07

   

0.001 0.001 0.002 0.001

KD  SEa 54.0 0.81 0.32 0.37

   

0.1 lM 0.05 lM 0.06 lM 0.05 lM

Spin state shift 44.8% 36.3% 38.8% 22.5%

DAmax and KD derived from the one-site binding model (Eqns 1,2 in methods). SE, standard error of the regression (SE(Pi) = sqrt(SS/ DF) 9 Cov(i,i)), with Pi: i-th adjustable (nonconstant) parameter; SS, sum of squared residuals; DF, degrees of freedom, Cov(i,i): i-th diagonal element of covariance matrix). a

Discussion

Fig. 4. UV-visible spectroscopic binding studies of OxyAtei binding to azole inhibitors. Inhibitors tested: ketoconazole (1), miconazole (2) and clotrimazole (3) (A), spectral response of OxyAtei upon imidazole binding (B), extrapolated amplitudes of imidazole (C) and azole inhibitor binding (D).

observed for the multiply chlorinated and rather flexible ketoconazole and miconazole. Furthermore, it displays decreased amplitudes upon clotrimazole titration – the compound previously shown to be the best ligand for OxyBtei [22]. This rather unexpected difference in spectral amplitude is a very interesting observation and cannot be easily rationalized by comparing both Oxy structures to the recent structure of the clotrimazole-bound P450 OleP (PDB-ID: 4XE3 [49]). Although the aromatic ring of the inhibitor, which lies in parallel to the heme plane, is very close to D279 (only present in OxyA), no severe clashes would arise in the OxyAtei active site for both binding conformations of clotrimazole observed in the OleP structure. Thus, it is more likely that this difference in binding behavior of OxyBtei and OxyAtei originates from the general architecture of their active sites.

We have determined the molecular structure of OxyAtei, the crucial D-O-E ring-forming P450 enzyme involved in teicoplanin biosynthesis. This structure not only completes the collection of peptide tailoring P450s involved in GPA biosynthesis but also reveals striking differences compared to the previously determined structures of OxyB, C and E. In particular the central cavity of OxyA is narrowed by the presence of an additional N terminal helix, the A’-helix and the orientation of the F- and G-helices toward the I-helix. Since this closed conformation of the F- and G-helices is more prominent in protomer A with the active site occupied by the N terminus of the protomer B, it is highly likely that binding of the N terminus induced a closing of the capping helices in a similar manner as described for ligand-bound P450s, yet with less dramatic conformational changes [26,49,53]. In the literature, there is one example of a P450 forming a homodimer in the crystal in a similar manner as OxyAtei, the structure of P450sky (PBD: 4LOE) [29]. In this case, the dimerization also did not cause any distortions in the structure when compared to an unbound P450sky structure and a following study showed that the position of the N terminus within the active site was even highly similar to that of an NRPS-bound azole inhibitor in complex with P450sky (PDB-ID: 4PXH) [54]. Comparison of the OxyAtei structure to the P450sky structures shows that the entry site of the N terminus of protomer B into the P450 active site is in a similar location as the entry site of the NRPScargo or the coordinated tag in the P450sky structures – in the B/C region, close to the G-helix (Fig. 5A). Thus, the bound N terminus may reflect the binding of a substrate to OxyAtei. This suggested binding mode would furthermore be in good agreement with the only structure of a GPA-modifying P450 in complex with an NRPS domain, the OxyBtei/X-domain structure

FEBS Letters 590 (2016) 571–581 ª 2016 Federation of European Biochemical Societies

577

Structure of OxyAtei

K. Haslinger and M. J. Cryle

(PDB-ID: 4TX3) [14]: the immediate proximity of the X-domain to the adjacent peptide attachment site within the peptidyl carrier protein domain of the NRPS (PCP domain, conserved, post-translationally phosphopantetheinylated serine residue), requires the substrate peptide to enter the P450 active site from approximately this position and angle (Fig. 5B). Interestingly, the OxyAtei structure displays an accumulation of aromatic residues in this area that are in close proximity to the bound N terminus and may even participate in a network of p-interactions (Fig. 2D). Since the substrate peptide is mainly composed of aromatic residues this aromatic network within the active site may be important in properly orienting the substrate for OxyA catalysis. In our sequence analysis of all known OxyA homologs, we found that these residues are strictly conserved yet different compared to the other Oxy enzymes (Fig. 3C). In the P450 with highest sequence and structure similarity to OxyA enzymes, OxyE, these positions are occupied by hydrophobic amino acids. This suggests an important role of these residues in substrate binding. A more general comparison of the available Oxy structures shows that in OxyEtei, the active site is even further occluded than in OxyAtei by the closing of the F-and G-helices, the fixed conformation of the loop connecting the B- and C-helices and the C-terminal bsheet that both reach across the active site to form hydrogen bonds with loops in the b1 sheet. In both the OxyAtei and the OxyEtei structures, the C-terminal

loop is characterized by a proline-rich turn that is interspersed with other amino acids in OxyEtei (PGRPAP) thereby allowing a deeper projection into the active site and favoring hydrogen bonding. Since OxyEtei catalyzes the cross-linking of the N-terminal amino acids of the peptide precursor (F-O-G ring), which are furthest away from the peptide attachment site on the NRPS, only minimal penetration of the substrate into the active site is needed. OxyAtei, catalyzing the D-O-E ring formation likely requires the second least depth of peptide entry. Thus, the observed degrees of active site opening with OxyE, being the most occluded, OxyA the second and OxyB and OxyC being most open may be traced back to the way substrates are accommodated in the active site. Another common feature of OxyA and OxyE is that their Ihelix is shifted away from the heme toward the D- and E-helices so that the immediate heme cavity is slightly larger than in the other structures. The relevance of this feature for peptide binding awaits the determination of substrate-bound Oxy structure, which has so far eluded characterization. Furthermore, compared to the other Oxys, OxyAtei displays striking differences in the amino acid composition in the immediate heme binding pocket that introduce additional polarity and structural changes to the active site. We set out to investigate the degree of conservation of these residues among the Oxys of all annotated GPA clusters and found that these positions are highly conserved with characteristic sequences for each Oxy enzyme (Fig. 3C). Although all Oxy enzymes

Fig. 5. Comparison of OxyAtei to the P450sky/PCP7sky complex (PDB-ID: 4PXH) (A) and suggested model for the OxyA recruitment to the NRPS (B) (docking model of the X-domain/OxyBtei complex (PDB-ID: 4PX3), OxyAtei, and PCP7tei (PDB-ID: 2MR8).

578

FEBS Letters 590 (2016) 571–581 ª 2016 Federation of European Biochemical Societies

K. Haslinger and M. J. Cryle

catalyze cross-linking of heptapeptides, the characteristics of their individual substrates are quite diverse depending on their cross-linking state: the degrees of rigidity and sterical hindrance increase with each ring formation and may thus require specific active site motifs that enable the productive binding of substrates for catalysis. Our UV-visible binding studies of azole inhibitors showed that the nature of the active site of OxyAtei has an effect on inhibitor binding: while the more open, hydrophobic active site of OxyBtei enables interactions with rigid, hydrophobic inhibitors, the more closed, polar and possibly more rigid active site of OxyAtei favors binding of more flexible and multiply chlorinated inhibitors. This indicates that the Oxy enzymes, while maintaining the typical P450 fold and active site composition, each bear their characteristic features that are likely important for substrate recognition and specificity.

Structure of OxyAtei

4

5

6

7

Acknowledgements The authors are grateful to Lina Sieverling for practical assistance and Melanie M€ uller for mass spectral analysis; to Anton Meinhart, Thomas Barends and Miroslaw Tarnawski (MPI-Hd) for assistance with data processing; and to Christopher Roome (MPI-Hd) for IT support. Diffraction data were collected at the Swiss Light Source, beamline X10SA, Paul Scherrer Institute, Villigen, Switzerland. We thank the Heidelberg team for data collection and the PXII staff for their support in setting up the beamline. M.J.C. is grateful for the support of the Deutsche Forschungsgemeinschaft (Emmy-Noether Program, CR 392/1–1), Monash University and the EMBL Australia program.

8

9

10

Author contributions KH and MJC designed the study and wrote the manuscript. KH performed experiments.

11

References 1 Cryle MJ, Brieke C and Haslinger K (2013) Oxidative transformations of amino acids and peptides catalysed by Cytochromes P450. In Amino Acids, Peptides and Proteins (Farkas E and Ryadnov M, eds), pp. 1–36. The Royal Society of Chemistry, Cambridge, UK. 2 Ortiz de Montellano PR (ed.) (2015) Cytochrome P450 – Structure, Mechanism, and Biochemistry. Springer, Cham, CH. 3 Bischoff D, Pelzer S, Holtzel A, Nicholson GJ, Stockert S, Wohlleben W, Jung G and Sussmuth RD (2001) The

12

13

FEBS Letters 590 (2016) 571–581 ª 2016 Federation of European Biochemical Societies

biosynthesis of vancomycin-type glycopeptide antibiotics – new insights into the cyclization steps. Angew Chemie-Int Ed 40, 1693–1696. Bischoff D, Pelzer S, Bister B, Nicholson GJ, Stockert S, Schirle M, Wohlleben W, Jung G and S€ ussmuth RD (2001) The biosynthesis of vancomycin-type glycopeptide antibiotics – the order of the cyclization steps. Angew Chemie – Int Ed 40, 4688–4691. Hadatsch B, Butz D, Schmiederer T, Steudle J, Wohlleben W, S€ ussmuth R and Stegmann E (2007) The biosynthesis of teicoplanin-type glycopeptide antibiotics: assignment of P450 mono-oxygenases to side chain cyclizations of glycopeptide A47934. Chem Biol 14, 1078–1089. Solenberg PJ, Matsushima P, Stack DR, Wilkie SC, Thompson RC and Baltz RH (1997) Production of hybrid glycopeptide antibiotics in vitro and in Streptomyces toyocaensis. Chem Biol 4, 195–202. vanWageningen AM, Kirkpatrick PN, Williams DH, Harris BR, Kershaw JK, Lennard NJ, Jones M, Jones SJ and Solenberg PJ (1998) Sequencing and analysis of genes involved in the biosynthesis of a vancomycin group antibiotic. Chem Biol 5, 155–162. Pelzer S, S€ ußmuth R, Heckmann D, Recktenwald J, Huber P, Jung G and Wohlleben W (1999) Identification and analysis of the balhimycin biosynthetic gene cluster and its use for manipulating glycopeptide biosynthesis in Amycolatopsis mediterranei DSM5908. Antimicrob Agents Chemother 43, 1565–1573. Zerbe K, Pylypenko O, Vitali F, Zhang W, Rouset S, Heck M, Vrijbloed JW, Bischoff D, Bister B, S€ ussmuth RD et al. (2002) Crystal structure of OxyB, a cytochrome P450 implicated in an oxidative phenol coupling reaction during vancomycin biosynthesis. J Biol Chem 277, 47476–47485. Puk O, Huber P, Bischoff D, Recktenwald J, Jung G, S€ ussmuth RD, van Pee KH, Wohlleben W, Pelzer S, S€ ußmuth RD et al. (2002) Glycopeptide biosynthesis in Amycolatopsis mediterranei DSM5908: function of a halogenase and a haloperoxidase/perhydrolase. Chem Biol 9, 225–235. Puk O, Bischoff D, Kittel C, Pelzer S, Weist S, Stegmann E, S€ ussmuth RD and Wohlleben W (2004) Biosynthesis of chloro-b-hydroxytyrosine, a nonproteinogenic amino acid of the peptidic backbone of glycopeptide antibiotics. J Bacteriol 186, 6093–6100. Cryle MJ, Meinhart A and Schlichting I (2010) Structural characterization of OxyD, a cytochrome P450 involved in b-hydroxytyrosine formation in vancomycin biosynthesis. J Biol Chem 285, 24562–24574. Bischoff D, Bister B, Bertazzo M, Pfeifer V, Stegmann E, Nicholson GJ, Keller S, Pelzer S, Wohlleben W and S€ ussmuth RD (2005) The biosynthesis of vancomycintype glycopeptide antibiotics – a model for oxidative

579

Structure of OxyAtei

14

15

16

17

18

19

20

21

22

23

24

580

K. Haslinger and M. J. Cryle

side-chain cross-linking by oxygenases coupled to the action of peptide synthetases. ChemBioChem 6, 267–272. Haslinger K, Peschke M, Brieke C, Maximowitsch E and Cryle MJ (2015) X-domain of peptide synthetases recruits oxygenases crucial for glycopeptide biosynthesis. Nature 521, 105–109. Geib N, Woithe K, Zerbe K, Li DB and Robinson JA (2008) New insights into the first oxidative phenol coupling reaction during vancomycin biosynthesis. Bioorganic Med Chem Lett 18, 3081–3084. Zerbe K, Woithe K, Li DB, Vitali F, Bigler L and Robinson JA (2004) An oxidative phenol coupling reaction catalyzed by OxyB, a cytochrome P450 from the vancomycin-producing microorganism. Angew Chemie – Int Ed 43, 6709–6713. Woithe K, Geib N, Zerbe K, Dong BL, Heck M, Fournier-Rousset S, Meyer O, Vitali F, Matoba N, Abou-Hadeed K et al. (2007) Oxidative phenol coupling reactions catalyzed by OxyB: a cytochrome P450 from the vancomycin producing organism. Implications for vancomycin biosynthesis. J Am Chem Soc 129, 6887–6895. Woithe K, Geib N, Meyer O, W€ ortz T, Zerbe K and Robinson JA (2008) Exploring the substrate specificity of OxyB, a phenol coupling P450 enzyme involved in vancomycin biosynthesis. Org Biomol Chem 6, 2861– 2867. Schmartz PC, W€ olfel K, Zerbe K, Gad E, El-Tamany ES, Ibrahim HK, Abou-Hadeed K, Robinson JA, Wolfel K, Zerbe K et al. (2012) Substituent effects on the phenol coupling reaction catalyzed by the vancomycin biosynthetic P450 enzyme OxyB. Angew Chemie-Int Ed 51, 11468–11472. Schmartz PC, Zerbe K, Abou-Hadeed K and Robinson JA (2014) Bis-chlorination of a hexapeptide-PCP conjugate by the halogenase involved in vancomycin biosynthesis. Org Biomol Chem 12, 5574–5577. Brieke C and Cryle MJ (2014) A facile Fmoc solid phase synthesis strategy to access epimerization-prone biosynthetic intermediates of glycopeptide antibiotics. Org Lett 16, 2454–2457. Haslinger K, Maximowitsch E, Brieke C, Koch A and Cryle MJ (2014) Cytochrome P450 OxyBtei catalyzes the first phenolic coupling step in teicoplanin biosynthesis. ChemBioChem 15, 2719–2728. Pylypenko O, Vitali F, Zerbe K, Robinson JA and Schlichting I (2003) Crystal structure of OxyC, a cytochrome P450 implicated in an oxidative C-C coupling reaction during vancomycin biosynthesis. J Biol Chem 278, 46727–46733. Cryle MJ, Staaden J and Schlichting I (2011) Structural characterization of CYP165D3, a cytochrome P450 involved in phenolic coupling in teicoplanin biosynthesis. Arch Biochem Biophys 507, 163–173.

25 Xu L-H, Fushinobu S, Ikeda H, Wakagi T and Shoun H (2009) Crystal structures of cytochrome P450 105P1 from Streptomyces avermitilis: conformational flexibility and histidine ligation state. J Bacteriol 191, 1211–1219. 26 Xu L-H, Fushinobu S, Takamatsu S, Wakagi T, Ikeda H and Shoun H (2010) Regio- and stereospecificity of filipin hydroxylation sites revealed by crystal structures of cytochrome P450 105P1 and 105D6 from Streptomyces avermitilis. J Biol Chem 285, 16844– 16853. 27 Li S, Tietz DR, Rutaganira FU, Kells PM, Anzai Y, Kato F, Pochapsky TC, Sherman DH and Podust LM (2012) Substrate recognition by the multifunctional cytochrome P450 MycG in mycinamicin hydroxylation and epoxidation reactions. J Biol Chem 287, 37880–37890. 28 Cryle MJ, Bell SG and Schlichting I (2010) Structural and biochemical characterization of the cytochrome P450 CypX (CYP134A1) from Bacillus subtilis: a cyclol-leucyl-l-leucyl dipeptide oxidase. Biochemistry 49, 7282–7296. 29 Uhlmann S, S€ ussmuth RD and Cryle MJ (2013) Cytochrome P450 sky interacts directly with the nonribosomal peptide synthetase to generate three amino acid precursors in skyllamycin biosynthesis. ACS Chem Biol 8, 2586–2596. 30 Cryle MJ and Schlichting I (2008) Structural insights from a P450 carrier protein complex reveal how specificity is achieved in the P450(BioI) ACP complex. Proc Natl Acad Sci USA 105, 15696–15701. 31 Li Z, Rupasinghe SG, Schuler MA and Nair SK (2011) Crystal structure of a phenol-coupling P450 monooxygenase involved in teicoplanin biosynthesis. Proteins-Struct Funct Bioinf 79, 1728–1738. 32 Kabsch W (2010) XDS. Acta Crystallogr D Biol Crystallogr 66, 125–132. 33 McCoy AJ, Grosse-Kunstleve RW, Adams PD, Winn MD, Storoni LC and Read RJ (2007) Phaser crystallographic software. J Appl Crystallogr 40, 658–674. 34 Murshudov GN, Vagin AA and Dodson EJ (1997) Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr D Biol Crystallogr 53, 240–255. 35 Emsley P, Lohkamp B, Scott WG and Cowtan K (2010) Features and development of Coot. Acta Crystallogr D Biol Crystallogr 66, 486–501. 36 Chen VB, Arendall WB, Headd JJ, Keedy DA, Immormino RM, Kapral GJ, Murray LW, Richardson JS and Richardson DC (2010) MolProbity: all-atom structure validation for macromolecular crystallography. Acta Crystallogr D Biol Crystallogr 66, 12–21. 37 Krissinel E and Henrick K (2004) Secondary-structure matching (SSM), a new tool for fast protein structure

FEBS Letters 590 (2016) 571–581 ª 2016 Federation of European Biochemical Societies

K. Haslinger and M. J. Cryle

38

39

40

41

42

43

44

45

46

47

48

49

alignment in three dimensions. Acta Crystallogr D Biol Crystallogr 60, 2256–2268. Holm L and Rosenstr€ om P (2010) Dali server: conservation mapping in 3D. Nucleic Acids Res 38, W545–W549. Holm L and Park J (2000) DaliLite workbench for protein structure comparison. Bioinformatics 16, 566–567. Sievers F, Wilm A, Dineen D, Gibson TJ, Karplus K, Li W, Lopez R, McWilliam H, Remmert M, S€ oding J et al. (2011) Fast, scalable generation of high-quality protein multiple sequence alignments using Clustal Omega. Mol Syst Biol 7, 539. Crooks GE, Hon G, Chandonia J-M and Brenner SE (2004) WebLogo: a sequence logo generator. Genome Res 14, 1188–1190. Sosio M, Bianchi A, Bossi E and Donadio S (2000) Teicoplanin biosynthesis genes in Actinoplanes teichomyceticus. Antonie van Leeuwenhoek, Int J Gen Mol Microbiol 78, 379–384. Sosio M, Kloosterman H, Bianchi A, de Vreugd P, Dijkhuizen L and Donadio S (2004) Organization of the teicoplanin gene cluster in Actinoplanes teichomyceticus. Microbiology 150, 95–102. Brieke C, Peschke M, Haslinger K and Cryle MJ (2015) Sequential in vitro cyclization by cytochrome P450 enzymes of glycopeptide antibiotic precursors bearing the X-domain from nonribosomal peptide biosynthesis. Angew Chemie – Int Ed 54, 15715–15719. Poulos TL, Finzel BC, Gunsalus IC, Wagner GC and  crystal structure of Kraut J (1985) The 2.6-A Pseudomonas putida cytochrome P-450. J Biol Chem 260, 16122–16130. McLean KJ, Hans M, Meijrink B, van Scheppingen WB, Vollebregt A, Tee KL, van der Laan J-M, Leys D, Munro AW and van den Berg MA (2015) Singlestep fermentative production of the cholesterollowering drug pravastatin via reprogramming of Penicillium chrysogenum. Proc Natl Acad Sci USA 112, 2847–2852. Lim Y-R, Hong M-K, Kim J-K, Doan TTN, Kim DH, Yun C-H, Chun Y-J, Kang L-W and Kim D (2012) Crystal structure of cytochrome P450 CYP105N1 from Streptomyces coelicolor, an oxidase in the coelibactin siderophore biosynthetic pathway. Arch Biochem Biophys 528, 111–117. Kudo T, Takaya N, Park SY, Shiro Y and Shoun H (2001) A positively charged cluster formed in the heme-distal pocket of cytochrome P450nor is essential for interaction with NADH. J Biol Chem 276, 5020–5026. Montemiglio LC, Parisi G, Scaglione A, Sciara G, Savino C and Vallone B (2016) Functional analysis and

Structure of OxyAtei

50

51

52

53

54

crystallographic structure of clotrimazole bound OleP, a cytochrome P450 epoxidase from Streptomyces antibioticus involved in oleandomycin biosynthesis. Biochim Biophys Acta 1860, 465–475. Altschul SF, Gish W, Miller W, Myers EW and Lipman DJ (1990) Basic local alignment search tool. J Mol Biol 215, 403–410. Jefcoate CRR (1978) Measurement of substrate and inhibitor binding to microsomal cytochrorne P-450 by optical-difference spectroscopy. In Methods in Enzymology - Biomembranes - Part C: Biological Oxidations (Fleischer S and Packer L, eds), pp. 258– 279. Elvesier, Amsterdam, NL. Correia MA and Hollenberg PF (2015) Inhibition of Cytochrome P450 Enzymes. In Cytochrome P450 – Structure, Mechanism and Biochemistry. (Ortiz de Montellano PR, ed), pp. 177–259. Springer, Cham, CH. Yu X, Cojocaru V and Wade RC (2013) Conformational diversity and ligand tunnels of mammalian cytochrome P450s. Biotechnol Appl Biochem 60, 134–145. Haslinger K, Brieke C, Uhlmann S, Sieverling L, S€ ussmuth RD and Cryle MJ (2014) The structure of a transient complex of a nonribosomal peptide synthetase and a cytochrome P450 monooxygenase. Angew Chemie – Int Ed 53, 8518–8522.

Supporting information Additional supporting information may be found in the online version of this article at the publisher’s web site: Fig. S1. Asymmetric unit of the OxyAtei crystals (A) with a zoomed in view of the two heme moieties in the active sites of the protomers (B). Fig. S2. Overlay of OxyAtei with the other known structures of GPA producing Oxy enzymes. OxyA colored as in Fig. 2, others in beige (OxyBvan PDB 1LG9 and 1LGF; first phenolic coupling enzyme from vancomycin biosynthesis); OxyCvan (PDB: 1UED; last aryl coupling enzyme from vancomycin biosynthesis); OxyE (PDB: 3O1A; second or third phenolic coupling enzyme from teicoplanin biosynthesis); and OxyBtei (PDB: 4TVF; first phenolic coupling enzyme from teicoplanin biosynthesis). Fig. S3. Multiple sequence alignment of the OxyA proteins experimentally characterized to date with the secondary structure of OxyAtei indicated as cartoon. Table S1 Most closely related structures of OxyAtei identified by Dali search [1] with the RMSD of Ca given for both protomers (A/B).

FEBS Letters 590 (2016) 571–581 ª 2016 Federation of European Biochemical Societies

581