The kinetic properties of various R258 mutants of deacetoxycephalosporin C synthase Hwei-Jen Lee 1 , Young-Fung Dai 1 , Chia-Yang Shiau 2 , Christopher J. Schofield 3 and Matthew D. Lloyd 4 1 Department of Biochemistry, National Defense Medical Centre, Taipei, Taiwan, ROC; 2 Institute of Medical Science, National Defense Medical Centre, Taipei, Taiwan, ROC; 3 The Oxford Centre for Molecular Sciences and Dyson Perrins Laboratory, South Parks Road, Oxford, UK; 4 Department of Pharmacy and Pharmacology, University of Bath, Claverton Down, Bath, UK Site-directed mutagenesis was used to investigate the control of 2-oxoacid cosubstrate selectivity by deacetoxycephalo- sporin C synthase. The wild-type enzyme has a requirement for 2-oxoglutarate and cannot efficiently use hydrophobic 2-oxoacids (e.g. 2-oxohexanoic acid, 2-oxo-4-methyl-penta- noic acid) as the cosubstrate. The following mutant enzymes were produced: R258A, R258L, R258F, R258H and R258K. All of the mutants have broadened cosubstrate selectivity and were able to utilize hydrophobic 2-oxoacids. The efficiency of 2-oxoglutarate utilization by all mutants was decreased as compared to the wild-type enzyme, and in some cases activity was abolished with the natural cosubstrate. Keywords: b-lactam biosynthesis; cephem; chemical cosub- strate rescue; nonhaem iron(II) oxygenase; 2-oxoglutarate. Deacetoxycephalosporin C synthase (DAOCS; Swiss-Prot P18548) catalyses a key step in the cephamycin C biosyn- thetic pathway in Streptomyces clavuligerus, i.e. the ring- expansion of penicillin N (1) to deacetoxycephalosporin C (DAOC, 2) [1–7] (Scheme 1). A sequence-related oxygenase, deacetylcephalosporin C synthase (DACS; Swiss-Prot 42220), catalyses the subsequent hydroxylation of the exocy- clic methyl group of (2) to give deacetylcephalosporin C (DAC, 3). The DAC product is then converted by a series of reactions, including 7-hydroxylation, into cephamycin C (4). Oxidative reactions in this pathway are catalysed by nonhaem iron(II) and 2-oxoglutarate (2-OG)-dependent oxygenases [4]. Two of these enzymes, DAOCS and DACS, are part of a sequence-related subgroup of enzymes [8], which also include deacetoxy/deacetylcephalosporin C syn- thase (the Cephalosporium acremonium bifunctional protein catalysing the both ring-expansion and hydroxylation reactions; Swiss-Prot P11935) and isopenicillin N synthase (which is not a 2-OG-dependent oxygenase, but catalyses the formation of the bicyclic penicillin nucleus; S. clavuli- gerus enzyme Swiss-Prot P10621). Understanding of the catalytic mechanism of DAOCS [9–11] has been advanced by a combination of biochemical and X-ray crystallographic studies [1,12–17]. These studies have given insights into how the oxidation of 2-oxoglutarate (the cosubstrate) and the penicillin substrate are coupled. R258 and S260 are located in the 2-OG binding pocket and bind the 5-carboxylate of the cosubstrate by electrostatic interactions. Mutation of R258 to glutamine results in loss of coupling between penicillin oxidation and 2-OG conver- sion [17]. Moreover, the activity of the R258Q mutant could be restored by the use of ÔunnaturalÕ 2-oxoacids (which are not utilized by the wild-type enzyme) as cosubstrates (Ôchemical cosubstrate rescueÕ) [17]. A similar change in cosubstrate selectivity upon mutation of the analogous arginine residue has been observed with the 2-OG-depend- ent oxygenase, phytanoyl-CoA 2-hydroxylase [18,19]. This paper reports the biochemical properties of other DAOCS mutants, in which R258 has been replaced with uncharged or positively charged amino acids. Materials and methods Materials Chemicals were obtained from the Sigma-Aldrich Chemical Co. or Merck and were of at least analytical grade or higher. Reagents were also supplied by: Amersham Biosciences (protein chromatography systems and columns); Bohringer- Mannheim (ATP); MBI (1 kb and 100 bp DNA gel markers); Bio-Rad (Kunkel mutagenesis reagents); New England Bio-Laboratories (enzymes for molecular biology); Novagen (pET vectors); Gibco/BRL (mutagenesis primers); Phenomenex (HPLC columns). Correspondence to H J. Lee, Department of Biochemistry, National Defense Medical Centre, No. 161, Sec. 6, Minchuan East Road, Neihu 114, Taipei, Taiwan, Republic of China. Fax: + 886 2 87921544, Tel.: + 886 2 87910832, E-mail: hjlee@ndmctsgh.edu.tw; and M. D. Lloyd, Department of Pharmacy and Pharmacology, University of Bath, Claverton Down, Bath BA2 7AY. Tel.: + 44 1225 386786, E-mail: M.D.Lloyd@bath.ac.uk Abbreviations: DACS, deacetylcephalosporin C synthase; DAOCS, deacetoxycephalosporin C synthase; ESI-MS, electrospray ionization mass spectrometry; G-7-ADCA, phenylacetyl-7-aminodeacetoxy cephalosporanic acid; 2-OG, 2-oxoglutarate; 2-OH, 2-oxohexanoate; 2-OMP, 2-oxo-4-methyl-pentanoate. (Received 14 November 2002, revised 29 January 2003, accepted 4 February 2003) Eur. J. Biochem. 270, 1301–1307 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03500.x Site-directed mutagenesis Oligonucleotide-directed mutagenesis of R258 mutants was performed using the Kunkel method [20,21]. Single-stranded DNA in the pET24a plasmid (pHL1 [1]) containing the DAOCS gene was used as a template for mutant produc- tion, with the appropriate primers (Table 1). The presence of the desired mutations was confirmed by automated DNA sequencing (ABI 3100 sequencer). Protein purification Recombinant DAOCS and mutant proteins were expressed in Escherichia coli BL21(DE3) as described previously at 37 °C after induction by 0.4 m M isopropyl thio-b- D -gal- actoside. Wild-type and mutant enzymes were purified as described [1,16] using a HiPrep 16/10 Q XL anion-exchange column, which was eluted with 0.16–0.26 M NaCl gradient. Fractions (10 mL) containing DAOCS (as judged by SDS/ PAGE analyses) were further purified using a Sephacryl S-100 column (XK26 column, 2.6 · 86 cm, 386 mL). Protein oligomerization was assessed as reported [1] using gel filtration chromatography with an S-200 10/30 column equilibrated in gel filtration buffer. Secondary structures of the purified wild-type and mutant proteins were compared by circular dichroism analyses. The molecular masses of all mutants were determined by ESI-MS (R258A: calculated 34470.6 Da, observed 34469 ± 5 Da; R258L: calculated 34512.6 Da, observed 34514 ± 5 Da; R258F: calculated 34546.7 Da; observed 34548 ± 4 Da; R258H: calculated 34536.6 Da, observed 34538 ± 5 Da; R258K: calculated 34527.7 Da, observed 34528 ± 6 Da). Activity assays Production of deacetoxycephems was assayed using the reported HPLC or holed-plate assays [1] in a final volume of 100 lL. Reaction mixtures contained 50 m M Hepes/NaOH, pH 7.5, 2 m M tris(carboxyethyl)phosphine, ascorbate, FeSO 4 and 2-oxoacid, and 10 m M penicillin G or ampicillin as ÔprimeÕ substrate. Approximately 0.05 mg of enzyme was used in each assay. E. coli · 580 was used as the indicator organism in the holed-plate assay [22]. HPLC separation of the products was performed by isocratic elution with 25 m M ammonium bicarbonate in 16% (v/v) methanol on a Hypersil C4 column (250 · 4.6 mm). Retention volumes for the G-7-ADCA and cephalexin products (from penicillin G and ampicillin, respectively) were 21.6 and 11.2 mL, respectively. 2-OG conversion assays were conducted as described previously [13], and were corrected for 2-OG conversion in the absence of penicillin substrate. Apparent kinetic parameters for the penicillin substrate (± SD) were determined by HPLC assay [1]. Rates were determined in at least triplicate for at least six concentrations of penicillin G or ampicillin at a single defined concentration of the 2-oxoacid cosubstrate (2 m M ). Results and discussion Purification and characterization of mutants Wild-type DAOCS and all mutants were purified to at least 85% purity (as judged by SDS/PAGE and gel filtration analyses) using a combination of anion-exchange and gel filtration chromatographies. The DNA sequences of the R258 mutants and the ESI-MS analyses showed that only Table 1. Primer sequences for production of the R258 mutants of DAOCS by the Kunkel method [20,21]. Mutation Primer sequence R258K 5¢-ACTGGAGGTCTTGCTGCTGC-3¢ R258H 5¢-ACTGGAGGTGTGGCTGCTGC-3¢ R258A 5¢-ACTGGAGGTCGCGCTGCTGC-3¢ R258Q [17] 5¢-ACTGGAGGTCTGGCTGCTGC-3¢ R258L 5¢-ACTGGAGGTCAGGCTGCTGC-3¢ R258F 5¢-ACTGGAGGTGAAGCTGCTGC-3¢ Scheme 1. Conversion of penicillin N1 to cephamycin C4. 2-OG ¼ 2-oxo-glutarate; R ¼ D -d-(a-aminoadipoyl)-; * methyl group incorporated into cephem ring by DAOCS. 1302 H J. Lee et al.(Eur. J. Biochem. 270) Ó FEBS 2003 the desired mutations were present. Circular dichroism analyses indicated that no gross changes in the secondary structure of the mutant enzymes had occurred compared to wild type DAOCS. Activity assays for 2-oxoglutarate and penicillin oxidation Previously it has been reported that wild-type DAOCS is only able to utilize a few 2-oxoacids as cosubstrates [17], with the natural 2-oxoacid, 2-oxoglutarate, giving the highest activity. Significant activity was also observed with 2-oxohexanedioate (2-oxoadipate) (64%) and 2-oxopro- panoate (pyruvate) (5%). The R258Q mutation reduced activity with 2-oxoglutarate and 2-oxopropanoate, but allowed utilization of other 2-oxoacids. Thus, all mutants were assayed for their ability to catalyse 2-OG conversion and penicillin oxidation in the presence of various 2-oxoacids (Table 2, Scheme 2). Cephem produc- tion (G-7-ADCA or cephalexin from penicillin G or ampicillin, respectively) was monitored by HPLC and holed-plate bioassays. The latter assay can give useful confirmatory data, as this provides evidence that the enzyme is producing a cephem antibiotic product. However, a linear response is only observed over a limited range of product concentrations and the response also depends on the antibiotic potency of the cephem product [23]. Thus, the results from the holed-plate assay should be interpreted with care. The effect of mutations appears to be similar regardless of whether the prime substrate was penicillin G or ampicillin. Utilization of 2-oxoglutarate by the R258 mutants The level of 2-OG conversion in the presence of a penicillin substrate was reduced with all the R258 mutants (Table 2, assay 1). The R258K mutant retained the highest levels of activity in the presence of penicillin and ampicillin, presumably due to maintenance of a charge interaction between the lysine side-chain and 2-OG. The R258H mutant retained little activity with 2-OG, which was surprising in light of the results with the R258K mutant. This probably arises due to the R258H side-chain being deprotonated. The R258A, R258L and R258F mutants almost completely abolished activity with 2-OG, presuma- bly due to unfavourable charge/hydrophobic interactions with the cosubstrate 5-carboxylate. Penicillin G and ampicillin oxidation in the presence of 2-oxoglutarate All mutants displayed reduced penicillin conversion in the presence of 2-OG (Table 2, assays 2 and 3). In some cases Table 2. Activity comparison of wild type and R258 mutants in the presence of alternative 2-oxoacids using 10 m M penicillin G or 10 m M ampicillin as substrate. Results are normalized to wild-type enzyme with 2-oxoglutarate as cosubstrate. The specific activities were 190 nmolÆmin )1 Æmg )1 for the 2-oxoglutarate conversion assay [7], 102 nmolÆmin )1 Æmg )1 for penicillin conversion by HPLC assay [1] and 72 nmolÆmin )1 Æmg )1 for the holed-plate assay [1]. Standard deviations for these assays were approximately 15% and 10%, respectively, based on triplicate assays. Assay Cosubstrate DAOCS enzyme Wild-type R258K R258H R258Q R258A R258L R258F Penicillin G substrate (10 m M ) 1. 2-OG conversion (CO 2 production) (a) 2-OG 100 66 6 30 19 26 < 5 2. HPLC assay (G-7-ADCA production) (a) 2-OG 100 28 12 18 < 2 16 9 (b) 2-OH < 2 23 23 46 53 74 57 (c) 2-OMP < 2 50 51 105 85 110 93 3. Holed-plate assay (G-7-ADCA production) (a) 2-OG 100 7 < 2 < 2 < 2 < 2 < 2 (b) 2-OH < 2 3 6 25 25 26 54 (c) 2-OMP < 2 7 24 72 49 76 69 Ampicillin substrate (10 m M ) 1. 2-OG conversion (CO 2 production) (a) 2-OG 47 28 < 5 49 12 < 5 6 2. HPLC assay (cephalexin production) (a) 2-OG 52 10 4 9 6 3 5 (b) 2-OH < 2 11 38 38 36 45 56 (c) 2-OMP < 2 17 48 55 55 58 75 3. Holed-plate assay (cephalexin production) (a) 2-OG 41 7 < 2 < 2 < 2 < 2 < 2 (b) 2-OH < 2 10 15 37 31 41 58 (c) 2-OMP < 2 10 6 37 11 34 43 Scheme 2. Conversion of penicillin G to G-7-ADCA and ampicillin to cephalexin by DAOCS. Ó FEBS 2003 Co-substrate selectivity of DAOCS (Eur. J. Biochem. 270) 1303 reduced penicillin oxidation is simply a consequence of the less efficient utilization of 2-OG as a cosubstrate. However, reduced cephem production may also be due to uncoupling of the two reactions, i.e. the stoichiometry is less than 1 : 1. Comparison of the results for assays 1a and 2a imply that the level of 2-OG conversion is greater than penicillin G conversion by the R258A mutant, i.e. some uncoupling is occurring. When ampicillin is used as the prime substrate uncoupling is clearly present in all mutants except for R258H, which does not catalyse conversion of 2-OG. The difference in the levels of coupling between 2-OG and penicillin G or ampicillin oxidation is probably a reflected in the apparent K m values for these substrates. The interpretation of these results is complicated because the levels of uncoupling will be variable depending on the mutant and the nature of the cosubstrate. Uncoupled conversion has been shown to be responsible for auto- oxidation of other oxygenases [24], and this may give rise to enzyme inactivation. This uncoupling may be a consequence of steric interactions between the penicillin substrate and the hydrophobic side-chain of the coproduct (the carboxylic acid that is analogous to succinate in the wild-type reaction), thus causing misorientation of the penicillin to the high-energy ferryl [Fe(IV)¼O/Fe(III)–O Æ ] intermediate [16,25]. Crystal- lographic analysis of DAOCS complexed with iron(II), succinate and CO 2 demonstrated that the 4-carboxylate of succinate may be bound by the side-chain of R160 or R162 in the absence of any substrate [12]. This structure may reflect catalysis during uncoupled reaction cycles (i.e. 2-OG conversion in the absence of a penicillin substrate) in the wild-type enzyme. Functional analysis of R160 and R162 by site-directed mutagenesis suggests that these two residues bind the nucleus of the penicillin substrate [16]. Rescue of enzyme activity of R258 mutants by alternative cosubstrates In the case of wild-type DAOCS, 2-OG cannot be replaced by either 2-OH or 2-OMP (Table 2, assays 2b and 2c). Because it was not possible to measure cosubstrate conver- sion (assay 1) the possibility of uncoupled 2-oxoacid conversion in the wild-type enzyme cannot be excluded. All of the R258 mutants were able to catalyse cephem production from both penicillin G and ampicillin in the presence of aliphatic 2-oxoacids (2-OH and 2-OMP), showing that broadening of cosubstrate selectivity occurs when the side-chain of R258 is substituted by a positively charged, polar [17] or hydrophobic side-chain. Significant activity in the presence of hydrophobic cosubstrates was unexpected for the R258K and R258H mutants because the presence of a positive charge in the active site would be expected to suppress cosubstrate binding. In the case of the R258K mutant this may be allowed because of disordering of the lysine side-chain, as was previously observed for the R258Q mutant (Fig. 1) [17]. InthecaseoftheR258Hmutantitismorelikelythatthe Fig. 1. Structure comparison of the active site of wild-type DAOCS (RCSB PDB accession no. 1rxg) complexed with iron (II) and 2-oxoglutarate and the R258Q mutant (RCSB PDB accession no. 1hjf) complexed with iron (II) and 2-oxo-4-methyl-pentanoate (2-OMP). Wild-type residues are represented by the red structure, mutant residues by the green structure. This figure was produced using VIEWERPRO (http://www.accelrys.com/ viewer). 1304 H J. Lee et al.(Eur. J. Biochem. 270) Ó FEBS 2003 imidazole side-chain is predominately in an uncharged state or that the positive charge is no longer proximal to the cosubstrate side-chain. This result is consistent with loss of activity in this mutant when 2-OG is used as the cosubstrate. Higher levels of penicillin oxidation were observed when 2-OMP was used as a cosubstrate compared to 2-OH. The results with the R258F mutant are particularly impressive, as this single point mutation appears to almost completely change the DAOCS cosubstrate selectivity from 2-OG to hydrophobic cosubstrates. Kinetic parameters of R258 mutants Kinetic parameters for the mutants were determined using the HPLC assay with either penicillin G or ampicillin as substrates, in the presence of 2-OG, 2-OH or 2-OMP, as appropriate. Due to the low activities of most mutants with 2-OG, only the R258K mutant was subjected to kinetic analysis. Similar trends were observed for both mutants and cosubstrates irrespective of whether penicillin G or ampi- cillin was the ÔprimeÕ substrate (Tables 3 and 4), and the results are consistent with those presented in Table 2. However, care is required when interpreting these results for several reasons. Firstly, only one concentration of cosub- strate was used in each experiment and so apparent kinetic parameters are reported. Secondly, the concentration of the 2-oxoglutarate cosubstrate has been shown to be a critical factor in determining the activity of the wild-type enzyme, which displays complex biphasic kinetics [26]. It is unclear if a similar phenomenon is observed for rescuing cosubstrates with mutant enzymes. Mutants can be grouped into those that affect apparent K m values, those that affect k cat values and those that affect both K m and k cat values. In the case of the R258K mutant with 2-OG, a modest decrease in k cat is observed whilst K m is considerably increased for both penicillin substrates. When 2-OH is used as cosubstrate a decrease in k cat is observed for the conversion of penicillin G in all mutants, except for the R258F mutant. This decrease is absent when Table 4. Kinetic parameters of R258 mutants using ampicillin as ÔprimeÕ substrate. Parameters were determined by HPLC assay with a fixed concentration of 2 m M cosubstrate. Enzyme Fixed cosubstrate (2 m M ) k cat (s )1 ) K m (m M ) k cat /K m ( M )1 Æs )1 ) WT 2-OG 0.014 ± 0.001 2.6 ± 0.24 5.4 R258K 2-OG 0.011 ± 0.002 13 ± 0.66 0.8 R258K 2-OH 0.009 ± 0.0009 24 ± 2.3 0.4 R258H 2-OH 0.007 ± 0.0005 4.2 ± 0.54 1.7 R258Q 2-OH 0.017 ± 0.003 6.6 ± 0.35 2.6 R258A 2-OH 0.008 ± 0.0009 4.2 ± 0.20 1.8 R258L 2-OH 0.012 ± 0.001 2.6 ± 0.50 4.6 R258F 2-OH 0.012 ± 0.002 1.9 ± 0.28 6.3 R258K 2-OMP 0.013 ± 0.002 15 ± 3.2 0.9 R258H 2-OMP 0.0053 ± 0.0002 4.1 ± 0.44 1.3 R258Q 2-OMP 0.014 ± 0.001 3.3 ± 0.44 4.2 R258A 2-OMP 0.011 ± 0.001 4.5 ± 0.43 2.4 R258L 2-OMP 0.0086 ± 0.00005 1.3 ± 0.01 6.6 R258F 2-OMP 0.011 ± 0.00005 0.79 ± 0.09 14 Table 3. Kinetic parameters of R258 mutants using penicillin G as substrate. Parameters were determined by the HPLC assay with a fixed concentration of 2 m M cosubstrate. Enzyme Fixed cosubstrate (2 m M ) k cat (s )1 ) K m (m M ) k cat /K m ( M )1 Æs )1 ) WT 2-OG 0.021 ± 0.002 1.1 ± 0.21 19 R258K 2-OG 0.010 ± 0.001 6.4 ± 0.75 1.6 R258K 2-OH 0.001 ± 0.0005 4.7 ± 0.37 2.1 R258H 2-OH 0.013 ± 0.0004 3.0 ± 0.29 4.3 R258Q 2-OH 0.010 ± 0.0001 0.79 ± 0.15 13 R258A 2-OH 0.013 ± 0.0003 0.79 ± 0.11 17 R258L 2-OH 0.019 ± 0.002 1.7 ± 0.14 11 R258F 2-OH 0.024 ± 0.0002 1.6 ± 0.01 15 R258K 2-OMP 0.020 ± 0.004 7.0 ± 0.85 2.9 R258H 2-OMP 0.014 ± 0.002 1.5 ± 0.04 9.3 R258Q 2-OMP 0.021 ± 0.001 1.5 ± 0.18 14 R258A 2-OMP 0.020 ± 0.002 1.8 ± 0.27 11 R258L 2-OMP 0.020 ± 0.002 1.2 ± 0.22 17 R258F 2-OMP 0.025 ± 0.004 0.84 ± 0.16 30 Ó FEBS 2003 Co-substrate selectivity of DAOCS (Eur. J. Biochem. 270) 1305 2-OMP is used as cosubstrate. The K m values for penicillins are significantly increased in mutants with side-chains that can be positively charged (R258K and R258H), whilst mutants with hydrophobic side-chains, such as R258F, appear to have similar or slightly lower K m values. Similar trends were observed when ampicillin was used as the prime substrate, although the magnitude of the changes in kinetic parameters do not always correspond to those observed with penicillin G. Notably, a modest increase in catalytic efficiency (as judged by k cat /K m values) occurs for the R258F mutant when 2-OMP is the cosubstrate with both penicillin G and ampicillin. Role of arginine residues in controlling cosubstrate selectivity The results highlight the essential nature of an arginine residue in the 2-oxoglutarate binding site in DAOCS and related oxygenases in controlling cosubstrate selectivity. The key factors determining which alternative cosubstrates can efficiently rescue oxygenase activity appear to be steric constraints upon the size of the cosubstrate side-chain, and the flexibility (or absence) of the substituted side-chain. 2-Oxoacids with up to six carbons or equivalent appear to be efficient cosubstrates, whilst larger 2-oxoacids (e.g. 2-oxo-octanoate) are not [17,19]. Efficient coupling of cosubstrate conversion with penicillin oxidation is also crucial, and this also appears to be partially controlled by the size of the cosubstrate side-chain. Some oxygenases, e.g. prolyl 4-hydroxylase [27] appar- ently have an active site lysine residue in place of arginine, but still maintain control of cosubstrate binding. How this is achieved is unclear, but it may be due to orientation of several residue side-chains to accomplish the selective charge–charge interaction. The presence of an additional histidine residue in the prolyl 4-hydroxylase binding site may be required to maintain the lysine side-chain in a positively charged state. Thus, although the combination of an arginine residue with a serine (or remote threonine [28]) residue appears to be an important mechanism for control- ling cosubstrate selectivity in many oxygenases, it does not appear to be universal. Acknowledgements We thank Dr R. T. Aplin for mass spectrometric analyses and Dr B. Odell for NMR analyses. The Biotechnology and Biological Sciences Research Council, Engineering and Physical Sciences Research Coun- cil, Medical Research Council, the Wellcome Trust, the European Union and the National Science Council, Republic of China are thanked for financial assistance. References 1. Lloyd, M.D., Lee, H J., Harlos, K., Zhang, Z.H., Baldwin, J.E., Schofield, C.J., Charnock, J.M., Garner, C.D., Hara, T., Ter- wisscha van Scheltinga, A.C., Valega ˚ rd, K., Viklund, J.A.C., Hajdu, J., Andersson, I., Danielsson, A ˚ . & Bhikhabhai, R. (1999) Studies on the active site of deacetoxycephalosporin C synthase. J. Mol. Biol. 287, 943–960. 2. Schofield, C.J., Baldwin, J.E., Byford, M.F., Clifton, I.J., Hajdu, J. & Roach, P.L. (1997) Proteins of the penicillin biosynthesis pathway. Curr. Opin. Struct. Biol. 7, 857–864. 3. Baldwin, J.E. & Schofield, C.J. (1992) The Chemistry of b-Lactams, Blackie Academic and Professional, London, UK. 4. Prescott, A.G. & Lloyd, M.D. (2000) The iron(II) and 2-oxoacid- dependent dioxygenases and their role in metabolism. Nat. Prod. Report 17, 367–383. 5. Jensen, S.E. (1986) Biosynthesis of cephalosporins. CRC Crit. Rev. Biotechnol. 3, 277–301. 6. Martin, J.F. (1998) New aspects of genes and enzymes for b- lactam antibiotic biosynthesis. Appl. Microbiol. Biotechn 50, 1–15. 7. Baldwin, J.E. & Abraham, E. (1988) The biosynthesis of peni- cillins and cephalosporins. Nat. Prod. Report 5, 129–145. 8. Prescott, A.G. (1993) A dilemma of dioxygenases (or where bio- chemistry and molecular biology fail to meet). J. Exp. Bot. 44, 849–861. 9. Townsend, C.A., Theis, A.B., Neese, A.S., Barrabee, E.B. & Poland, D. (1985) Stereochemical fate of chiral-methyl valine in the ring expansion of penicillin N to deacetoxycephalosporin C. J. Am. Chem. Soc. 107, 4760–4767. 10. Pang, C.P., White, R.L., Abraham, E.P., Crout, D.H.G., Lutstorf, M., Morgan, P.J. & Derome, A.E. (1984) Stereochemistry of the incorporation of valine methyl groups into methylene groups in cephalosporin C. Biochem. J. 222, 777–788. 11. Baldwin, J.E., Adlington, R.M., Kang, T.W., Lee, E. & Schofield, C.J. (1988) The ring expansion of penicillins to cephems: a possible biomimetic process. Tetrahedron 44, 5953–5957. 12. Lee, H J., Lloyd, M.D., Harlos, K., Clifton, I.J., Baldwin, J.E. & Schofield, C.J. (2001) Kinetic and crystallographic studies on deacetoxycephalosporin C synthase (DAOCS). J. Mol. Biol. 308, 937–948. 13. Lee, H J., Lloyd, M.D., Harlos, K. & Schofield, C.J. (2000) The effect of cysteine mutations on the activity of recombinant dea- cetoxycephalosporin C synthase from S. clavuligerus. Biochem. Biophys. Res. Commun. 267, 445–448. 14. Lee, H J., Schofield, C.J. & Lloyd, M.D. (2002) Active site mutations of recombinant deacetoxycephalosporin C synthase. Biochem. Biophys. Res. Commun. 292, 66–70. 15. Valega ˚ rd, K., Terwisscha van Scheltinga, A.C., Lloyd, M.D., Hara, T., Ramaswamy, S., Perrakis, A., Thompson, A., Lee, H J., Baldwin, J.E., Schofield, C.J., Hajdu, J. & Andersson, I. (1998) Structure of a cephalosporin synthase. Nature 394, 805–809. 16. Lipscomb, S.J., Lee, H J., Mukherji, M., Baldwin, J.E., Schofield, C.J. & Lloyd, M.D. (2002) The role of arginine residues in sub- strate binding and catalysis by deacetoxycephalosporin C syn- thase. Eur. J. Biochem. 269, 2735–2739. 17. Lee, H J., Lloyd, M.D., Clifton, I.J., Harlos, K., Dubus, A., Baldwin,J.E.,Frere,J M.&Schofield,C.J.(2001)Probingthe cosubstrate selectivity of deacetoxycephalosporin C synthase: The role of arginine-258. J. Biol. Chem. 276, 18290–18295. 18. Mukherji, M., Chien, W., Kershaw, N.J., Clifton, I.J., Schofield, C.J., Wierzbicki, A.S. & Lloyd, M.D. (2001) Structure-function analysis of phytanoyl-CoA 2-hydroxylase mutations causing Refsum’s Disease. Hum. Mol. Genet. 10, 1971–1982. 19. Mukherji, M., Kershaw, N.J., MacKinnon, C.H., Clifton, I.J., Schofield, C.J., Wierzbicki, A.S. & Lloyd, M.D. (2001) ÔChemical co-substrate rescueÕ of phytanoyl-CoA 2-hydroxylase mutants causing Refsum’s Disease. J. Chem. Soc. Chem. Commun., 972– 973. 20. Kunkel, T.A. (1985) Rapid and efficient site-specific mutagenesis without phenotypic selection. Proc. Natl Acad. Sci. USA 82, 488– 492. 21. Kunkel, T.A., Roberts, J.D. & Zakour, R.A. (1987) Rapid and efficient site-specific mutagenesis without phenotypic selection. Methods Enzymol. 154, 367–382. 22. Baldwin, J.E., Adlington, R.M., Coates, J.B., Crabbe, M.J.C., Crouch, N.P., Keeping, J.W., Knight, G.C., Schofield, C.J., Ting, 1306 H J. Lee et al.(Eur. J. Biochem. 270) Ó FEBS 2003 H.H., Vallejo, C.A., Thorniley, M. & Abraham, E.P. (1987) Purification and initial characterization of an enzyme with dea- cetoxycephalosporin C synthetase and hydroxylase activities. Biochem. J. 245, 831–841. 23. Shibata, N., Lloyd, M.D., Baldwin, J.E. & Schofield, C.J. (1996) Adipoyl-6-APA is a substrate for deacetoxycephalosporin C syn- thase (DAOCS). Bioorg.Med.Chem.Lett.6, 1579–1584. 24. Liu, A., Ho, R.Y.N., Que, L., Ryle, M.J., Phinney, B.S. & Hau- singer, R.P. (2001) Alternative reactivity of an a-ketoglutarate- dependent iron (II) oxygenase: enzyme self hydroxylation. J. Am. Chem. Soc. 123, 5126–5127. 25. Hamilton, C.S., Yasuhara, A., Baldwin, J.E., Lloyd, M.D. & Rutledge, P.J. (2001) Contrasting fates for 6-a-methylpenicillin N upon oxidation by deacetoxycephalosporin C synthase (DAOCS) and deacetoxy/deacetylcephalosporin C synthase (DAOC/ DACS). Bioorg. Med. Chem. Lett. 11, 2511–2514. 26. Dubus, A., Lloyd, M.D., Lee, H J., Schofield, C.J., Baldwin, J.E. & Frere, J M. (2001) Substrate selectivity studies on deacetoxycephalosporin C synthase using a direct spectro- photometric assay. Cellular and Molecular Life Sciences 58, 835– 843. 27. Myllyharju, J. & Kivirikko, K.I. (1997) Characterization of the iron and 2-oxoglutarate-binding sites of prolyl 4-hydroxylase. EMBO J. 16, 1173–1180. 28. Zhang,Z H.,Ren,J.,Stammers,D.K.,Baldwin,J.E.,Harlos,K. & Schofield, C.J. (2000) Structural origins of the selectivity of the trifunctional oxygenase clavaminic acid synthase. Nat. Struct. Biol. 7, 127–133. Ó FEBS 2003 Co-substrate selectivity of DAOCS (Eur. J. Biochem. 270) 1307 . 5¢-ACTGGAGGTGTGGCTGCTGC-3¢ R258A 5¢-ACTGGAGGTCGCGCTGCTGC-3¢ R258Q [17] 5¢-ACTGGAGGTCTGGCTGCTGC-3¢ R258L 5¢-ACTGGAGGTCAGGCTGCTGC-3¢ R258F 5¢-ACTGGAGGTGAAGCTGCTGC-3¢ Scheme. sequences for production of the R258 mutants of DAOCS by the Kunkel method [20,21]. Mutation Primer sequence R258K 5¢-ACTGGAGGTCTTGCTGCTGC-3¢ R258H 5¢-ACTGGAGGTGTGGCTGCTGC-3¢ R258A