ORF6 from the clavulanic acid gene cluster of Streptomyces clavuligerus has ornithine acetyltransferase activity Nadia J. Kershaw 1 , Heather J. McNaughton 1 , Kirsty S. Hewitson 1 , Helena Herna ´ ndez 2 , John Griffin 3 , Claire Hughes 3 , Philip Greaves 3 , Barry Barton 3 , Carol V. Robinson 2 and Christopher J. Schofield 1 1 Oxford Centre for Molecular Sciences and The Dyson Perrins Laboratory, UK; 2 Oxford Centre for Molecular Sciences, Central Chemistry, Oxford, UK; 3 GlaxoSmithKline Pharmaceuticals, Worthing, West Sussex, UK The clinically used beta-lactamase inhibitor clavulanic acid is produced by fermentation of Streptomyces clavuligerus.The orf6 gene of the clavulanic acid biosynthetic gene cluster in S. clavuligerus encodes a protein that shows sequence homology to ornithine acetyltransferase (OAT), the fifth enzyme of the arginine biosynthetic pathway. Orf6 was overexpressed in Escherichia coli (at  15% of total soluble protein by SDS/PAGE analysis) indicating it was not toxic to the host cells. The recombinant protein was purified (to > 95% purity) by a one-step technique. Like other OATs it was synthesized as a precursor protein which underwent autocatalytic internal cleavage in E. coli to generate a and b subunits. Cleavage was shown to occur between the alanine and threonine residues in a KGXGMXXPX–(M/L)AT (M/L)L motif conserved within all identified OAT sequences. Gel filtration and native electrophoresis analyses implied that the ORF6 protein was an a 2 b 2 heterotetramer and direct evidence for this came from mass spectrometric analyses. Although anomalous migration of the b subunit was observed by standard SDS/PAGE analysis, which indicated the presence of two bands (as previously observed for other OATs), mass spectrometric analyses did not reveal any evidence for post-translational modification of the b subunit. Extended denaturation with SDS before PAGE resulted in observation of a single major b subunit band. Purified ORF6 was able to catalyse the reversible transfer of an acetyl group from N-acetylornithine to glutamate, but not the formation of N-acetylglutamate from glutamate and acetyl-coenzyme A, nor (detectably) the hydrolysis of N-acetylornithine. Mass spectrometry also revealed the reaction proceeds via acetylation of the b subunit. Keywords: ornithine acetyltransferase; clavulanic acid; N-terminal nucleophile hydrolase; arginine biosynthesis. Streptomyces clavuligerus produces a number of b-lactams, including clavulanic acid (Scheme 1, 1), which is a potent inhibitor of serine b-lactamases and is clinically used in combination with penicillin antibiotics [1,2]. Whilst a synthesis of clavulanic acid has been achieved, the known routes are low yielding and produce racemic material [2,3]. Thus, it is produced commercially by fermentation of S. clavuligerus and as a result its biosynthesis has been of considerable interest, particularly with respect to the optimization of fermentation titres. The biosynthetic pathway to clavulanic acid has been partially elucidated (Scheme 1) [1,2]. It begins with the condensation of arginine and glyceraldehyde 3-phosphate, catalysed by ORF2, to produce 2-carboxyethyl-arginine [4]. It has been shown [5] that arginine is a later metabolic intermediate than ornithine as, when the pathway from ornithine to arginine is blocked, ornithine cannot be incorporated into clavulanic acid. 2-Carboxyethyl-arginine is cyclized to give the first formed b-lactam, deoxyguanidi- noproclavaminic acid, via an ATP mediated ring closure catalysed by b-lactam synthetase (BLS/ORF3) [6–8]. Hydroxylation by clavaminic acid synthase (CAS/ORF5), followed by hydrolysis of the guanidino side-chain catalysed by proclavaminate amidino hydrolase (PAH/ORF4) yields proclavaminic acid [9], which undergoes two further oxida- tion steps, again catalysed by CAS, to produce bicyclic clavaminic acid [10]. This is converted via an unknown process involving epimerization at the C-3 and C-5 centres to give (3R,5R)-clavaldehyde which is reduced, in a reaction catalysed by an NADPH reductase (CAD/ORF9), to clavulanic acid (reviewed in [1]). In S. clavuligerus the genes for b-lactam biosynthesis are grouped into a ÔsuperclusterÕ [11] in which the operons for clavulanic acid and cephalosporin biosynthesis are adjacent. The cephamycin gene cluster has been entirely sequenced [12]. Including resistance, regulatory and transport genes, 11 genes have so far been reported in the clavulanic acid gene cluster. Furthermore, the early clavam biosynthetic genes are duplicated in a paralogous cluster responsible for the biosynthesis of other clavams [13,14]. Thus, only functions for the gene products of orfs2–5 in the early stages of the pathway and the final stage (orf 9) of the clavulanic acid gene cluster have been assigned (see above). A BLAST search and sequence alignments indicate that ORF6 displays 40% identity with ornithine acetyltransfer- ase (OAT) from Thermus thermophilus and 30–36% identity with OATs from other species [15,16]. OAT is one of the enzymes involved in the biosynthesis of arginine from glutamate, in which ornithine is a key intermediate (Scheme 2) [17]. Ornithine is produced from glutamate via Correspondence to C. J. Schofield, Oxford Centre for Molecular Sciences and The Dyson Perrins Laboratory, South Parks Road, Oxford, OX1 3QY, UK. Fax: + 44 1865275654, Tel.: +44 1865275625, E-mail: christopher.schofield@chem.ox.ac.uk Abbreviations: OAT, ornithine acetyltransferase; CAS, clavaminic acid synthase. (Received 16 November 2001, revised 22 February 2002, accepted 25 February 2002) Eur. J. Biochem. 269, 2052–2059 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02853.x four N-acetylated intermediates, beginning with N-acetyl- glutamate and ending with N-acetylornithine. In most prokaryotes (with known exceptions of Enterobacteriaceae and Sulfolobus solfataricus [17,18]) ornithine formation is catalysed by OAT (also know as ARGJ), which transfers an acetyl group from N-acetylornithine to glutamate. Hence, once the pathway is initiated from glutamate and acetyl- CoA by acetylglutamate synthase (ARGA), the acetyl group can be recycled, thus feeding the first step of the pathway. Some OATs are bifunctional, being capable of forming N-acetylglutamate from acetyl-CoA and glutamate [19–21]. Here we report studies on the orf6 gene product (ORF6). They demonstrate that ORF6 has OAT activity and provide mass spectrometric evidence that ORF6, and by implication other OATs, exists as an a 2 b 2 heterotetramer and that catalysis proceeds via acetylation of the b subunit. EXPERIMENTAL PROCEDURES DNA manipulations were carried out by standard proto- cols [22]. All chemicals were of reagent grade and obtained from Sigma-Aldrich Co. Ltd. unless otherwise stated. Restriction enzymes and IMPACT-CN System TM were purchased from New England BioLabs Inc. Oligonucleo- tides were synthesized by SigmaGenosys Ltd. Protein concentrations were determined by the method of Bradford [23]. PCR The Orf6 gene was amplified by PCR from wild-type S. clavuligerus genomic DNA (from GlaxoSmithKline) and directly cloned into the PCR-Script TM vector (Stratagene). The primers used were: forward primer, 5¢-ACGCT CATATGTCCGACAGCACACCGAAGACG-3¢; reverse primer, 5¢-CCATGTCCCTCCTGCCCTCGTCACCTTG CAT-3¢. Orf6 was subsequently cloned into pET24a(+) using BamHI/NdeI restriction sites. It was also cloned into the pTYB12 and pTYB11 vectors of the IMPACT-CN System TM using NdeI/EcoRI, and EcoRI/SapI sites, respec- tively, to generate an intein-chitin binding domain fusion vector. Expression and purification of ORF6 in E. coli The pTYB12/orf6 and pTYB11/orf6 plasmids were used to transform E. coli BL21 (DE3) cells and grown at 30 °Cin 2TY media containing ampicillin at 100 lgÆmL )1 . When the D 600 reached 0.5–0.8, the temperature was lowered to 15 °C and protein expression was induced by the addition of 0.3 m M isopropyl thio-b- D -galactoside. Following overnight incubation, the cells were harvested by centrifugation at 14 333 g for 20 min at 4 °C. The cells were resuspended in 50 m M Tris/HCl, pH 7.5, and lysed by sonication. The lysate was cleared by centrifugation for 10 min in a bench- top centrifuge at 11 000 g and loaded onto a column of chitin beads pre-equilibrated with column buffer (20 m M Tris/HCl pH 8.0, 0.1 m M EDTA, 2 M NaCl). The column was washed with 15 column volumes of column buffer containing 0.1% (v/v) Triton X-100. Following incubation with column buffer containing 50 m M dithiothreitol for 16 h at 23 °C to initiate cleavage of the intein tag protein, the ORF6 was subsequently eluted with column buffer. ORF6 was judged to be > 95% pure by SDS/PAGE analysis (17.5%). All protein samples were desalted, concentrated to  4mgÆmL )1 andstoredat)80 °C. Gel filtration chromatography The native molecular mass of purified ORF6 was estimated by gel filtration using a Superdex 200 HR column equilibrated with 100 m M Tris/HCl, pH 7.5. The column was calibrated with ribonuclease A (13.7 kDa), chymotrypsinogen A (25 kDa), ovalbumin (43 kDa) and bovine serum albumin (67 kDa) (Gel Filtration Calibra- tion Kit, Amersham Pharmacia Biotech.). An elution volume parameter (K av ) was calculated for each of the calibration proteins and a calibration curve constructed. By calculating K av for ORF6 the native molecular mass was established. Enzyme assays The ORF6 assay contained 100 m M Tris/HCl, pH 7.5, 6.0 m M N-acetylornithine, 6.0 m M glutamate (or other putative substrate) and 30 lg of enzyme in a final volume Scheme 1. Biosynthetic pathway leading to clavulanic acid. BLS, beta- lactam synthetase (ORF3); PAH, proclavaminate amidino hydrolase (ORF4); CAS, clavaminate synthase (ORF5); CAD, clavaldehyde dehydrogenase (ORF9). The ÔunnaturalÕ CAS-catalysed hydroxylation of L -N-acetylarginine to (2S,3R)-3-hydroxy-N-acetylarginine is shown boxed. Scheme 2. The role of OAT (ARGJ) in the biosynthesis of arginine. ORF6 carries out the same reaction. The proposed acyl-enzyme in- termediate is shown boxed. Ó FEBS 2002 ORF6 from the clavulanic acid gene cluster (Eur. J. Biochem. 269) 2053 of 0.1 mL. Incubation was at 37 °C for 20 min. Ornithine acetyltransferase activity was measured by modification of the procedure of Chinard [24] which employs ninhydrin to determine the presence of ornithine. Following incubation, derivatization was accomplished by the addition of 0.4 mL MilliQ water and 0.5 mL ninhydrin reagent [1.2 M citric acid/1.5% (w/v) ninhydrin in 2-methoxy- ethanol, 1 : 2]. The mixture was then vortexed for 15 s, andheatedat95°C for 12 min. The absorbance at 470 nm was measured and the amount of ornithine produced was determined by reference to a standard curve of 0–1 lmol ornithine. ORF6 kinetic studies were carried out in the presence of 1.0–6.0 m M N-acetylornithine and 6.0 m M glutamate. K m and the specific activity values were calculated using SIGMAPLOT (SPSS inc.). The pH profile was determined using assay conditions as above except with Tris/HCl buffer at pH 7.5–9.0, Mops buffer at pH 7.0 and Mes buffer at pH 6.0–6.5, followed by ninhydrin derivatization. For 1 H NMR (500 MHz) assays, substrate concentra- tions were doubled, and the amount of enzyme increased to 45 lg. The samples were incubated at 37 °Cfor2h, lyophilized and resuspended in D 2 O. The 1 HNMR (500 MHz) spectrum was recorded immediately. N- L -Acetylarginine for incubation with CAS was pro- duced by the standard assay, containing 6.0 m M N-acetyl- ornithine and 6.0 m M arginine, but the incubation time was increased to 2 h. Following the incubation, ORF6 was heat denatured at 95 °C for 5 min and precipitated protein removed by centrifugation. To 50 lL of the supernatant was added 40 lgofCAS,10m M FeSO 4 and 10 m M a-ketoglutarate, in a final volume of 0.1 mL. This was incubated at 30 °C for 10 min. The products were detected by HPLC, using a reverse-phase C 18 octadecylsilane column (150 · 4.6 mm), monitoring absorbance at 218 nm. An isocratic gradient of 10% methanol was used at a flow rate of 1 mLÆmin )1 . N-Acetylarginine and 3-hydroxy-N-acety- larginine had retention volumes of 3.2 and 2.9 mL, respec- tively. Gel electrophoresis Both SDS/PAGE and native PAGE (17.5%) were per- formed using standard techniques with a Bio-Rad Mini Protean II kit. Samples were prepared by heating for 5 min at 95 °Cwith2· SDS/PAGE sample loading buffer [0.635 M Tris/HCl, 0.01% bromophenol blue (w/v), 10% SDS (w/ v), 50% glycerol (w/v), 1% 2-mercaptoethanol (v/v)]. Extended denaturation was achieved by heating with 2 · SDS/PAGE loading buffer at 95 °C for 30 min. Mass spectrometric studies Electrospray ionization mass spectra were recorded on a Micromass BioQ II-ZS triple quadrupole mass spectro- meter. Samples (10 lL) were introduced into the electro- spray source via a loop injector as a solution with a final protein concentration of 5 pmolÆlL )1 in water/acetonitrile (1 : 1, v/v) containing 0.2% formic acid. For the nanoflow ESI mass spectrometry work, ORF6 samples were prepared at a concentration of 20 l M , in both 100 m M and 20 m M ammonium acetate, pH 7.0. For the substrate selectivity experiments, samples of 20 l M ORF6 with a 20-fold molar excess of substrate in 20 m M ammo- nium acetate, pH 8.0 were left at room temperature for 1–2 h. After this time, an aliquot was removed and partially denatured by addition of an equal volume of acetonitrile. Data were acquired using a quadrupole time-of-flight mass spectrometer (QTOF I, Micromass UK Ltd, Altrin- cham, UK). Positive ion spectra were recorded to compare samples at different ionic strength and negative ion spectra were used for the N-acetyl donors study. Caesium iodide was used to calibrate the instrument over the mass range 100–10 000 m/z (manual pusher set at 180 ls) with an acquisition step of 5 s. Samples were loaded into borosili- cate capillaries, 1.0 mm outer diameter · 0.5 mm inner diameter (Clark Electrical Instruments, Reading, UK), which were drawn down to a fine taper and coated with gold in-house. The capillary tip was cut manually under a stereomicroscope to give the required diameter and flow. A nitrogen backing gas line was used to initiate and maintain a flow from the capillary. Nitrogen at room temperature was also used as drying gas, and the ESI source was not heated. RESULTS AND DISCUSSION Expression and purification Previously it has been reported that high-level expression of OATs is problematic, with expression above a critical level becoming deleterious for E. coli hosts [25,26]. Expression of orf6 using the pET24a(+) vector resulted in production of an insoluble 42-kDa protein. However, orf6 was overex- pressed, using both the pTYB12 and pTYB11 vectors, in E. coli BL21 (DE3) cells as approximately 15% of the total soluble protein by SDS/PAGE analysis. The pTYB vectors allow for affinity purification of the target protein with subsequent thiol-mediated cleavage of the tag in a single step, utilizing a commercially avaliable modified intein protein-splicing protocol. This single step purification using the IMPACT-CN TM chitin bead column gave ORF6 protein samples of > 95% purity by SDS/PAGE analysis. The pTYB12 vector introduced an additional three amino acids (alanine, glycine and histidine) to the N-terminus of ORF6 compared to that derived from the pTYB11 vector. Subsequent activity analysis comparisons indicated that the additional three amino acids had no affect on the functional properties of ORF6. Almost identical CD traces were obtained for the N-terminally extended and wild-type proteins, indicating that the extra N-terminal residues did not affect the gross structure of ORF6. Autoproteolytic processing of ORF6 SDS/PAGE analysis of the purified ORF6 showed the protein as two distinct bands, with approximate masses of 19 and 25 kDa (Fig. 1). From its gene sequence, ORF6 is predicted to have a molecular mass of 42 kDa, and hence ORF6 appeared to have undergone proteolysis, forming two subunits. Only a trace band was sometimes present at  42 kDa consistent with an efficient autocatalytic cleavage in E. coli, which does not itself contain OAT [17]. The N-terminal amino-acid sequence of the 19-kDa band (from pTYB11) was shown to be MSDSTPKTPR, identical to that expected for the N-terminus of ORF6. This was 2054 N. J. Kershaw et al. (Eur. J. Biochem. 269) Ó FEBS 2002 consequently designated as the a subunit, the b subunit being the larger C-terminal portion. The b subunit had the N-terminal sequence TLLTFFATDA, which is consistent with cleavage occuring between alanine 180 and threonine 181 of the KGVGMLEPDMATLL motif. Processing of ORF6 into a and b subunits is consistent with its assignment as an OAT. All 17 known (or putative) OATs contain the motif KGXGMXXPX–(M/L)AT(M/ L)L, with a predicted post-translational cleavage site between the alanine and threonine residues [27]. The alcohol of the ÔunmaskedÕ threonine is believed to act as a nucleophile which is acylated during the catalytic cycle of OAT. Thus, OAT is a member of the family of N-terminal nucleophilic enzymes [28,29]. Experimental evidence has been reported for this cleavage site in OAT from three thermophilic organisms Methanoccocus jannaschii, Thermo- toga neapolitana and Bacillus stearothermophilus [25], which were shown to undergo cleavage to form a and b subunits following expression in E. coli.OATfromSaccharomyces cerevisiae, has also been shown to cleave at this point and evidence provided for an autocatalytic rather than a protease-mediated process [30]. By standard SDS/PAGE analyses, the b subunit of ORF6 was seen to consist of major (lower) and minor (upper) bands with apparent molecular masses of  25 kDa. Both b subunit bands gave identical N-terminal sequences (TLLTFFATDA). Previously it has been repor- ted that the b subunit of OATs from thermophiles run anomalously by SDS/PAGE analysis [25] and the b subunit of OAT from yeast has been observed as a double band [30]. To investigate this, denatured ORF6 was analysed by electrospray ionization mass spectrometry, which indicated that only a single b subunit species was present, with no evidence for a post-translational modification as previously suggested for the anomalous migration [25]. The mass spectrum gave an a subunit mass of 18 816.2 Da and a b subunit mass of 22 811.7 Da, which corresponds almost exactly with the predicted masses of 18 815.4 and 22 810.3, respectively, for a cleavage site between alanine 180 and threonine 181. It seems likely that the ÔanomalousÕ obser- vations for ORF6 result from the SDS/PAGE technique and may reflect aggregation/incomplete denaturation. Con- sistent with this, extension of the time the ORF6 protein was heated with SDS before electrophoresis resulted in apparent conversion of the lower b subunit band to the upper band. Native PAGE of the ORF6 protein suggested that it exists as an ab heterodimer under these conditions. However by gel filtration chromatography the molecular mass was determined to be  84 kDa suggesting an a 2 b 2 stoichiometry. Presumably the interactions between the two ab units are such that they are sufficiently weak to be disrupted by electrophoresis conditions. It is possible that yeast OAT, which has been reported to be an ab dimer, can also exist as a heterotetramer as for ORF6 and other OATs. To provide direct evidence for the formation of a hetero- tetramer we investigated the use of nanoflow ESI mass spectrometry for samples under native solution conditions. At a protein concentration of 20 l M in 100 m M ammonium acetate solution at pH 7.0, the major species observed was the a 2 b 2 heterotetramer (experimental mass: 83 308.3 Da, calculated mass: 83 251.6 Da) consistent with the gel filtration results. Charge states from the ab heterodimer and monomers were prominent (Fig. 2A). Homodimers were not detected at any significant level. At a lower concentration of ammonium acetate (20 m M ), the intensity of the heterotetramer charge states decreased relative to the monomers and the heterodimer was the major oligomer observed (Fig. 2B). Fig. 1. SDS/PAGE gel of protein purification. Lane 1, molecular mass markers; lane 2, cell lysate; lane 3, purified ORF6. Note the presence of two bands for the b subunit in the latter. Fig. 2. Comparison of positive ion nanoflow ESI mass spectra of ORF6 in (A) 100 m M and (B) 20 m M ammonium acetate at pH 7.0 showing the distribution of oligomeric species. Peak labels refer to the lower charge states for the a and b monomers. Oligomeric species are indicated. Experimental masses (n ¼ 6): a,18815.6±0.2Da;b,22810.4± 0.1 Da; heterodimer (ab), 41 629.1 ± 0.6 Da; heterotetramer (a 2 b 2 ), 83 308.3 ± 11.0 Da. Ó FEBS 2002 ORF6 from the clavulanic acid gene cluster (Eur. J. Biochem. 269) 2055 Kinetics and mechanism of ORF6 OATs catalyse the transfer of an acetyl group from L -N-acetylornithine to L -glutamate to form L -ornithine and L -N-acetylglutamate. The presence of ornithine can be detected by derivatization with acidic ninhydrin followed by UV spectroscopic analysis, observing the absorption at 470 nm [24]. Using this assay it was possible to observe the formation of ornithine from N-acetylornithine and gluta- mate as catalysed by ORF6 (Fig. 3). The K m value for L -N-acetylornithine at 37 °C and pH 8.0 was estimated to be 3.6 m M , with a specific activity of 87 nmolÆmin )1 Æmg )1 . This K m value is similar to K m values for other OATs, which include 0.5 m M for T. aquaticus [18] and 9.6 m M for M. jannashi [25]. Maximal enzyme activity was between pH 7.5–8.0, which is similar to pH profiles for other OATs [25,31] (Fig. 4). To confirm the identity of the ORF6 assay products and to develop an assay that did not rely upon ornithine detection, the crude reaction mixtures were characterized by 1 H NMR (500 MHz) spectroscopy after termination by heat treatment. Spectra were obtained for both forward (i.e. forming ornithine) and reverse reactions in D 2 Oafter quenching. Transfer of the acetyl group was shown to be freely reversible. The equilibrium position (after a 2-h incubation at 37 °C) for the ( L -N-acetylornithine/ L -gluta- mate)/(ornithine/ L -N-acetylglutamate mixture) was found to be  3 : 2, respectively, by integration of the a-acetyl CH 3 group of the N-acetylated compounds (Fig. 5). The OAT from Chlamydomonas reinhardti is reported to have L -N-acetylornithine hydrolase activity [32]. However, incubation of ORF6 with L -N-acetylornithine in the absence of a potential acetyl acceptor, led to no detectable obser- vation (i.e. < 5%) of L -ornithine by the 1 HNMRand ninhydrin assays. Presumably, incubation of L -N-acetylor- nithine results in reversible acetylation of the active site threonine 181 [25] and deacetylation by water cannot compete with that by L -N-acetylornithine. Some OATs (e.g. that from Neisseria gonorrhoeae, Bacillus stearothermophilus, S. cerevisae and B. subtilis) are capable of forming N-acetylglutamate from glutamate and acetyl-CoA [19,21,25,27]. Because the ninhydrin assay monitors ornithine production it was not possible to use this assay with acetyl-CoA and hence the reaction was monitored by 1 H NMR. Within the sensitivity of this assay (< 5% conversion on 1.0 mg scale) it was not possible to detect any N-acetylglutamate production, and therefore ORF6 appears to behave as a monofunctional OAT. There are no obvious differences between the amino-acid sequences that could account for the differ- ence between monofunctional and bifunctional OATs. However, the sequences of the monofunctional class appear to have a slightly shorter N-terminus [33]. ORF6, which belongs to the monofunctional class, shares this property. Previously, with OAT from B. stearothermophilus and T. neapolitana, radioactive studies have indicated that there is a covalent acetylation of the enzyme during catalysis [25]. By analogy with other N-terminal nucleophile enzymes, the nucleophilic site of the enzyme is the alcohol side-chain of threonine 181, the N-terminal residue of the b subunit. Direct evidence for the proposed mechanism came from ESI mass spectrometry. For ORF6, using negative ion nanoflow ESI mass spectrometry, three acetyl donors were examined ( L -N-acetylornithine, L -N-acetylarginine and L -N-acetylglu- tamate). Negative ion spectra were recorded due to a lower observed level of salt adducts compared to that present in the corresponding positive ion spectra. Incubation of ORF6 with N-acetylglutamate followed by denaturation of the protein indicated that approximately 37% of the b subunit had undergone acetylation (Fig. 6). Some acetylation was also observed for N-acetylornithine although at a lower level (< 20%) than N-acetylglutamate (data not shown). Under the same conditions, no acetylation of the b monomer was evident when N-acetylarginine was employed as the sub- strate. Fig. 3. UV spectrum of ninhydrin derivatized reaction mixture at time T ¼ 0andT ¼ 30 min. Fig. 4. pH profile of ORF6. Fig. 5. 1 H NMR spectrum of incubation of N-acetylornithine (NAO) and various acetyl group acceptors with ORF6. The acetyl group of NAO appears at 1.98 p.p.m. and is at lower field in all cases. The acetyl groups for the acetylated acceptors appear between 1.98 and 1.96 p.p.m. 2056 N. J. Kershaw et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Substrate selectivity of ORF6 A series of alternative acetyl acceptors were also assayed with ORF6, using L -N-acetylornithine as the acetyl donor. D -Glutamate was not a substrate and ORF6 selected L -glutamate as a substrate from racemic D , L -glutamate. This indicates that ORF6 (and probably other OATs) could be used as an alternative to the widely used amino-acid acylases [34], in particular for the resolution of racemic mixtures of certain amino acids with side chains containing polar groups. The side-chain selectivity of the acetyl-acceptor was probed, assaying L -aspartate and D -aspartate, L -a-amin- oadipic acid, L -leucine, and L -alanine. With these substrates, no ornithine was produced, indicating that the acetyl group was not being transferred. However L -arginine, L -glutamine and L -lysine did act as acetyl acceptors. After 20 min incubation, the relative amounts of ornithine produced using L -arginine and L -glutamine compared to L -glutamate were 24% and 16%, respectively, as determined by the ninhydrin assay. Lysine could not be measured as it interfered with the ornithine derivatization conditions. Activity was shown for all three by the 1 H NMR assay, indicating the equilibrium position for the N-acetylorni- thine:ornithine mixture was approximately 2 : 1 as deter- mined by integration of the hydrogens of the a-acetyl CH 3 group of the N-acetylated compounds (Fig. 5). The obser- vation that arginine acts as a substrate is particularly interesting because L -N-acetylarginine is a good substrate for CAS [10], being hydroxylated to give (2S,3R)-3- hydroxy-N-acetylarginine (Scheme 1). L -N-acetylornithine is a much poorer substrate for CAS [10]. A combination of ORF6andCASmightbeusedtoprepare/ferment 3-hydroxy derivatives of arginine and ornithine that might be useful as chiral starting materials for pharmaceutical production. Incubation of ORF6 with N-acetylornithine and arginine to produce N-acetylarginine, followed by incubation with CAS and appropriate cofactors resulted in the formation of 3-hydroxy-N-acetylarginine, as detected by HPLC analysis. The observation that arginine/N-acetylarg- inine can act as an acetyl acceptor/donor raises the question as to how selectivity is achieved in vivo. Kinetic studies have shown that OAT from B. stearo- thermophilus operates via a ping-pong bi-bi mechanism in which the enzyme only accepts the glutamate after the ornithine has left the active site [25]. The selectivity and mass spectrometric results are consistent with this and suggest ORF6 has a side-chain binding site that can accommodate the side-chains of ornithine, glutamate, lysine and arginine, but not, for example, that of a-aminoadipate. It is interesting to note that the side chain can contain basic, acidic or neutral functional groups. Clearly X-ray crystal- lography of ORF6 will be interesting, as there is currently no 3D structural information available on any OAT. Role of ORF6 in clavulanic acid biosynthesis These results clearly demonstrate that ORF6 catalyses the formation of L -N-acetylglutamate from L -N-acetylornithine and L -glutamate, a key step in arginine biosynthesis, and are the first experimental evidence for the functional assignment of ORF6 from the clavulanic acid super-cluster. The first step of the biosynthetic pathway to clavulanic acid (Scheme 1, 1) involves the condensation of arginine with a C-3 metabolite, glyceraldehyde 3-phosphate [4]. The role of ORF6 may therefore be to increase the arginine available for clavulanic acid production. Deletion mutant studies show that when ORF6 production is blocked, clavulanic acid production is completely halted in starch-asparagine based medium, yet when grown in soy based medium clavulanic acid production is approximately 40% of the wild-type [13]. Jensen et al. have provided an explanation for this observation by revealing that paralogous genes, not directly responsible for clavulanic acid biosynthesis, are only expressed in the soy based medium [13]. Another OAT in the arginine biosynthesis cluster of S. clavuligerus (ARGJ) has also been identified [35], suggesting that the OAT role of ORF6 is directed towards increased intracellular concen- trations of arginine for clavam biosynthesis rather than primary metabolism. Because ORF6 is not the only OAT in S. clavuligerus it may seem surprising that no clavulanic acid production was observed for the orf6 deletion mutant in starch/asparagine media [13]. Deletion mutants for other enzymes required for arginine biosynthesis have been reported to affect clavulanic acid production [36]. Thus, elimination of ARGC (an enzyme involved in synthesis of ornithine from glutamate in the arginine biosynthetic pathway) has been shown to interfere specifically with the production of clavulanic acid [36], albeit under different growth conditions. Fig. 6. Acetylation of ORF6 b monomer monitored by negative ion nanoflow ESI mass spectrometry. Spectra have been transformed on to a mass scale. The average masses from three spectra are shown. The experimental mass difference between the acetylated and nonacetylated forms was 41.8 ± 0.1 Da (42.0 Da calculated difference). (A) 20 l M ORF6 in 20 m M ammonium acetate pH 8.0 (B) 20 l M ORF6 with a 20-fold molar excess of N-acetylglutamate in 20 m M ammonium acetate pH 8. Ó FEBS 2002 ORF6 from the clavulanic acid gene cluster (Eur. J. Biochem. 269) 2057 There is no direct evidence for a role for multienzyme complexes or channelling in clavam biosynthesis. However, the lability (and likely toxicity) of some of the highly reactive intermediates, especially 3R,5R-clavaldehyde, suggests that intermediate channelling would be advantageous. Channel- ing may also explain the selectivity problems posed by the observation that ORF6 can accept arginine/N-acetylargi- nine as substrates and the latter is a substrate for CAS. In the case of other secondary metabolites, e.g. flavonoids [37] there is evidence for channelling and protein–protein interactions between ÔindividualÕ enzymes. ORF6 may represent a link (possibly structural) between primary and secondary arginine metabolism in S. clavuligerus. Finally, because the full sequence of reactions in clavam biosynthesis in S. clavuligerus has not been elucidated, another role for ORF6 cannot be ruled out. ACKNOWLEDGEMENTS OCMS is jointly supported by the Biotechnology and Biological Sciences Research Council, Engineering and Physical Sciences Research Council and Medical Research Council. N. J. K. is supported by an Engineering and Physical Sciences Research Council studentship and HJM by a Biotechnology and Biological Sciences Research Council CASE award. C. V. R. is supported by the Royal Society in a University Research Fellowship. We thank R. T. Aplin for mass spectrometric analyses and A. Willis for amino acid sequencing and Prof. S. E. Jensen for encouragement. REFERENCES 1. Jensen, S.E. & Paradkar, A.S. (1999) Biosynthesis and molecular genetics of clavulanic acid. Anton. Leeuwen. Int. J. 75, 125–133. 2. Baggaley, K.H., Brown, A.G. & Schofield, C.J. (1997) Chemistry and biosynthesis of clavulanic acid and other clavams. Nat. Prod. Report, 4, 309–333. 3. Bently, P.H., Brooks, G., Gilpin, M.L. & Hunt, E. (1979) A new total synthesis of (±)-clavulanic acid. Tetrahedron Lett. 21, 1889– 1890. 4. Khaleeli, N., Li, R. & Townsend, C.A. (1999) Origin of the b-lac- tam carbons in clavulanic acid from an unusual thiamine pyro- phosphate-mediated reaction. J. Am. Chem. Soc. 121, 9223–9224. 5. Valentine, B.P., Bailey, C.R., Doherty, A., Morris, J., Elson, S.W., Baggaley, K.H. & Nicholson, N.H. (1993) Evidence that arginine is a later metabolic intermediate than ornithine in the biosyntheis of clavulanic acid by Streptomyces clavuligerus. J. Chem. Soc., Chem. Commun., 1210–1211. 6. Bachmann,B.O.,Li,R.&Townsend,C.A.(1998)beta-Lactam synthetase: a new biosynthetic enzyme. Proc. Natl Acad. Sci. USA 95, 9082–9086. 7. McNaughton, H.J., Thirkettle, J.E., Zhang, Z., Schofield, C.J. & Jensen, S.E. (1998) b-Lactam synthase: implications for b-lac- tamase evolution. J. Chem. Soc., Chem. Commun., 2325–2326. 8. Miller, M.T., Bachmann, B.O., Townsend, C.A. & Rosenzweig, A.C. (2001) Structure of b-lactam synthetase reveals how to synthesize antibiotics instead of asparagine. Nat. Struct. Biol. 8, 684–689. 9. Elson, S.W., Baggaley, K.H., Davidson, M., Fulston, M., Nicholson, N.H. & Risbridger, G.D. (1993) The identification of three new biosynthetic intermediates and one further enzyme in the clavulanic acid pathway. J. Chem. Soc., Chem. Commun., 1212–1214. 10. Lloyd,M.D.,Merritt,K.D.,Lee,V.,Sewell,T.J.,Wha-Son,B., Baldwin, J.E. & Schofield, C.J. (1999) Product-substrate engineering by bacteria: studies on clavaminate synthase, a tri- functional dioxygenase. Tetrahedron 55, 10201–10220. 11. Ward, J.M. & Hodgson, J.E. (1993) The biosynthetic genes for clavulanic acid and cephamycin production occur as a super- cluster in 3 streptomyces. FEMS Microbiol. Lett. 110, 239–242. 12. Alexander, D.C. & Jensen, S.E. (1998) Investigation of the Streptomyces clavuligerus cephamycin C gene cluster and its reg- ulation by the CcaR protein. J. Bacteriol. 180, 4068–4079. 13. Jensen,S.E.,Elder,K.J.,Aidoo,K.A.&Paradkar,A.S.(2000) Enzymes catalysing the early steps of clavulanic acid biosynthesis are encoded by two sets of paralogous genes in Streptomyces clavuligerus. Antimicrob. Agents. Ch. 44, 720–726. 14. Mosher, R.H., Paradkar, A.S., Anders, C., Barton, B. & Jensen, S.E. (1999) Genes specific for the biosynthesis of clavam meta- bolites antipodal to clavulanic acid are clustered with the gene for clavaminate synthase 1 in Streptomyces clavuligerus. Antimicrob. Agents. Ch. 43, 1215–1224. 15. Hodgson, J.E., Fosberry, A., Rawlinson, N.S., Ross, H.N.M., Neal, R.J., Arnell, J.C., Earl, A.J. & Lawlor, E.J. (1995) Clavu- lanic acid biosynthesis in Streptomyces clavuligerus:genecloning and characterization. Gene 166, 49–55. 16. Jensen, S.E., Alexander, D.C., Paradkar, A.S. & Aidoo, K.A. (1993) Extending the b-lactam biosynthetic gene cluster in Strep- tomyces clavuligerus.InIndustrial Microorganisms: Basic and Applied Molecular Genetics (Baltz, R.H., Hegeman, G.D. & Skatrud, P.L., eds), pp. 169–176. American Society for Microbi- ology, Washington, DC. 17. Cunin, R., Glansdorff, N., Pie ´ rard,A.&Stalon,V.(1986)Bio- synthesis and metabolism of arginine in bacteria. Microbiol. Rev. 50, 314–352. 18. Van de Casteele, M., Demarez, M., Legrain, C., Glansdorff, N. & Pie ´ rard, A. (1990) Pathways of arginine biosynthesis in extreme thermophilic archaeo- and eubacteria. J. General Microbiol. 136, 117–1183. 19. Martin, P.R. & Mulks, M.H. (1992) Sequence anaylsis and complementation studies of the argJ gene encoding ornithine acetyltransferase from Neisseria gonorrhoeae. J. Bacteriol. 174, 2694–2701. 20. Mountain, A., McChesney, J., Smith, M.C. & Baumberg, S. (1986) Gene sequence encoding early enzymes of arginine synth- esis within a cluster in Bacillus subtilis, as revealed by cloning in Escherichia coli. J. Bacteriol. 165, 1026–1028. 21. Sakanyan, V., Charlier, D., Legrain, C., Kochikyan, A., Mett, I., Pie ´ rard, A. & Glansdorff, N. (1993) Primary structure, partial- purification and regulation of key enzymes of the acetyl cycle of arginine biosynthesis in Bacillus stearothermophilus – dual function of ornithine acetyltransferase. J. Gen. Microbiol. 139, 393–402. 22. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, Cold Spring Harbour Laboratory Press, Cold Spring Harbour, New York. 23. Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248–254. 24. Chinard, F.P. (1952) Photometric estimation of proline and ornithine. J. Biol. Chem. 199, 91–95. 25. Marc, F., Weigel, P., Legrain, C., Almeras, Y., Santrot, M., Glansdorff, N. & Sakanyan, Y. (2000) Characterisation and kinetic mechanism of mono- and bifunctional ornithine acetyl- transferases from thermophilic microorganisms. Eur. J. Biochem. 267, 5217–5226. 26. Savchenko, A., Weigel, P., Dimova, D., Lecocq, M. & Sakanyan, V. (1998) The Bacillus stearothermophilus argCJBD operon har- bours a strong promoter as evaluated in Esherichia coli cells. Gene 212, 167–177. 27. Crabeel, M., Abadjieva, A., Hilven, P., Desimpelaere, J. & Soetens, O. (1997) Characterisation of the Saccharomyces cerevi- siae ARG7 gene encoding ornithine acetyltransferase, an enzyme also endowed with acetylglutamate synthase activity. Eur. J. Bio- chem. 250, 232–241. 2058 N. J. Kershaw et al. (Eur. J. Biochem. 269) Ó FEBS 2002 28. Brannigan, J.A., Dodson, G., Duggleby, H.J., Moody, P.C.E., Smith, J.L., Tomchick, D.R. & Murzin, A.G. (1995) A protein catalytic framework with an N-terminal nucleophile is capable of self-activation. Nature 378, 416–419. 29. Dodson, G. & Wlodawer, A. (1998) Catalytic triads and their relatives. Trends Biochem. Sci. 23, 347–352. 30. Abadjieva, A., Hilven, P., Pauwels, K. & Crabeel, M. (2000) The yeast ARG7 gene product is autocatalysed to two subunit peptides yielding active ornithine acetyltransferase. J. Biol. Chem. 275, 11361–11367. 31. Liu, Y.S., Heeswijck, R.V., Hoj, P. & Hooengraad, N. (1995) Purification and characterisation of ornithine acetyltransferase from Saccharomyces cerevisiae. Eur. J. Biochem. 228, 291–296. 32. Staub, M. & Denes, G. (1966) Mechanism of arginine biosynthesis in Chlamydomonas reinhardti I. Purification and properties of ornithine acetyltransferase. Biochem. Biophys. Acta 128, 82–91. 33. Baetens, M., Legrain, C., Boyen, A. & Glansdorff, N. (1998) Genes and enzymes of the acetyl cycle of arginine biosynthesis in the extreme thermophilic bacterium Thermus thermophilus HB27. Microbiol. 144, 479–492. 34. Chenault, H.K., Dahmer, J. & Whitesides, G.M. (1989) Kinetic resolution of unnatural and rarely occuring amino-acids – enan- tioselective hydrolysis of N-acyl amino-acids catalysed by acylase- I. J. Am. Chem. Soc. 111, 6354–6364. 35. Rodriguez-Garcia, A., de la Fuente, A., Perez-Redondo, R., Martin, J.F. & Liras, P. (2000) Characterization and expression of the arginine biosynthesis gene cluster of Streptomyces clavuligerus. J. Mol. Microb. Biotechnol. 2, 543–550. 36. Ludovice, M., Martin, J.F., Carrachas, P. & Liras, P. (1992) Characterisation of the Streptomyces clavuligerus argC gene encoding N-acetylglutamyl-phosphate reductase – expression in Streptomyces lividans and effect on clavulanic acid production. J. Bacteriol. 174, 4606–4613. 37. Burbulis, I.E. & Winkel-Shirley, B. (1999) Interactions among enzymes of the Arabidopsis flavonoid biosynthetic pathway. Proc. Natl Acad. Sci. USA 96, 12929–12934. Ó FEBS 2002 ORF6 from the clavulanic acid gene cluster (Eur. J. Biochem. 269) 2059 . ORF6 from the clavulanic acid gene cluster of Streptomyces clavuligerus has ornithine acetyltransferase activity Nadia J. Kershaw 1 , Heather J UK The clinically used beta-lactamase inhibitor clavulanic acid is produced by fermentation of Streptomyces clavuligerus .The orf6 gene of the clavulanic acid