Adventitious reactions of alkene monooxygenase reveal common reaction pathways and component interactions among bacterial hydrocarbon oxygenases William L. J. Fosdike 1 , Thomas J. Smith 1,2 and Howard Dalton 1 1 Department of Biological Sciences, University of Warwick, Coventry, UK 2 Biomedical Research Centre, Sheffield Hallam University, UK Alkene monooxygenase (AMO) (EC 1.14.13.69), of Rhodococcus rhodochrous (formerly Nocardia corallina) B-276 belongs to a family of soluble multicomponent oxygenases that possess a binuclear iron active centre [1–5]. This group of enzymes includes AMOs from two other bacterial sources [6–8], the soluble methane monooxygenases (sMMOs) (EC 1.14.13.25) produced by certain methanotrophic bacteria [9] and a range of oxygenases that perform aromatic ring hydroxylations with a variety of specificities [3,10]. All enzymes for which data are available have been found to comprise at least three components: (1) multisubunit binuclear iron active centre-containing terminal oxygenase, where oxygen activation and substrate oxygenation occur; (2) an NAD(P)H-dependent reductase that sup- plies electrons to the active centre of the terminal Keywords alkene monooxygenase; alkyne; component interactions; peroxide shunt reaction; turnover-dependent inhibition Correspondence H. Dalton, Department of Biological Sciences, University of Warwick, Coventry CV4 7AL, UK Fax: +44 24 76523568 Tel: +44 24 76523552 E-mail: h.dalton@warwick.ac.uk Website: http://www.bio.warwick.ac.uk/res/ frame.asp?id=4; http://www.shu.ac.uk/schools/sci/biomed/ bmrc/tomsmith.htm (Received 24 January 2005, revised 14 March 2005, accepted 21 March 2005) doi:10.1111/j.1742-4658.2005.04675.x Alkene monooxygenase (AMO) from Rhodococcus rhodochrous (formerly Nocardia corallina) B-276 belongs to a family of multicomponent nonheme binuclear iron-centre oxygenases that includes the soluble methane mono- oxygenases (sMMOs) found in some methane-oxidizing bacteria. The enzymes catalyse the insertion of oxygen into organic substrates (mostly hydrocarbons) at the expense of O 2 and NAD(P)H. AMO is remarkable in its ability to oxidize low molecular-mass alkenes to their corresponding epoxides with high enantiomeric excess. sMMO and other well-character- ized homologues of AMO exhibit two adventitious activities: (1) turnover- dependent inhibition by alkynes and (2) activation by hydrogen peroxide in lieu of oxygen and NAD(P)H (the peroxide shunt reaction). Previous stud- ies of the AMO had failed to detect these activities and opened the possi- bility that the mechanism of AMO might be fundamentally different from that of its homologues. Thanks to improvements in the protocols for culti- vation of R. rhodochrous B-276 and purification and assay of AMO, it has been possible to detect and characterize turnover-dependent inhibition of AMO by propyne and ethyne and activation of the enzyme by hydrogen peroxide. These results indicate a similar mechanism to that found in sMMO and also, unexpectedly, that the enantiomeric excess of the chiral epoxypropane product is significantly reduced during the peroxide shunt reaction. Inhibition of the oxygen⁄ NADH-activated reaction, but not the peroxide shunt, by covalent modification of positively charged groups revealed an additional similarity to sMMO and may indicate very similar patterns of intersubunit interactions and ⁄ or electron transfer in both enzyme complexes. Abbreviations AMO, alkene monooxygenase; sMMO, soluble methane monooxygenase; sulfo-NHS-acetate, sulfosuccinimidyl acetate. FEBS Journal 272 (2005) 2661–2669 ª 2005 FEBS 2661 oxygenase; and (3) a small component, known as the coupling or gating protein, which is also required for full activity [3–5]. The terminal oxygenase components of soluble methane monooxygenases [9,11], aromatic monooxygenases [10] and the AMO of Xanthobacter sp. strain Py2 [12] all have an (abc) 2 quaternary struc- ture, whereas the AMOs from R. rhodochrous B-276 [13] and Mycobacterium sp. [8] lack the c subunit. In addition, the aromatic monooxygenases and AMOs from Xanthobacter Py2 and Mycobacterium sp. also possess a Rieske iron-sulfur protein that appears to pass electrons between the reductase and terminal oxygenase [12,14–16]. The binuclear iron-centre monooxygenases are char- acterized by close biochemical and, apparently, struc- tural similarities. There is complete conservation of the protein ligands to the binuclear iron site (four glu- tamyl and two histidyl residues) and all catalyse similar hydrocarbon monooxygenation reactions at the expense of NADH and dioxygen [3–5]. The enzymes show a gamut of substrate ranges and regio- and stereo-selectivities. For instance, the substrate range of AMO from R. rhodochrous B-276 is restricted almost exclusively to alkenes [17], whereas sMMOs have a remarkably wide range of substrates that includes alka- nes, alkenes and aromatic compounds [18]. Consistent with their different enzymatic activities, the terminal oxygenase components of AMOs are termed epoxygen- ases as their products are epoxides; the equivalent components of the other enzymes add hydroxyl groups to their natural substrates and are therefore known as hydroxylases. There are also important differences in the enantiopurity of products obtained from oxygen- ation of alkene substrates: AMO from R. rhodochrous B-276 epoxygenates propene to R-epoxypropane with high enantiomeric excess (83%) [1,13], whereas sMMO produces a nearly racemic mixture of products with the same substrate (S.E. Slade, T.J. Smith and H. Dalton, unpublished observations). In addition to these oxygenation reactions, several well characterized binuclear iron-centre monooxygen- ases have been found to exhibit two adventitious reac- tions: turnover-dependent inhibition by alkynes and the so-called peroxide shunt reaction. Terminal alkynes such as ethyne have been shown to act as suicide sub- strates of sMMO [19], soluble butane monooxygenase [20], several aromatic monooxygenases [21,22] and the AMO from Xanthobacter sp. Py2 [12,23], presumably by oxygenation to ketenes that then covalently modify and inactivate the enzymes [19]. Irreversible inhibition of the heme active-site monooxygenases of the cyto- chrome P450 family, which are not related to the binu- clear iron centre-containing enzymes, may also proceed via similar ketene intermediates [24]. The peroxide shunt reaction allows the terminal oxygenase compo- nent to perform oxygenation reactions in the absence of other protein components and NAD(P)H, if the oxidant is hydrogen peroxide rather than dioxygen. The peroxide shunt has been observed in sMMO [25,26] and toluene 2-monooxygenase [27], as well as unrelated monooxygenases of the cytochrome P450 family [28]. In sMMO it has been shown that the whole-complex (i.e. NADH-dependent) and peroxide- activated activities can be functionally separated by covalently modifying positively charged groups on the hydroxylase, which inhibits the whole-complex reaction but not the peroxide shunt [29]. Neither of these adventitious activities has to our knowledge previously been reported in the rhodococcal AMO. The rhodococcal AMO was previously found to be inhibited by propyne, but the inhibition appeared to be competitive because K m for propene oxygen- ation was increased by the inhibitor but V max was unchanged [1]. This, together with the lack of any pub- lished account of the peroxide shunt reaction in the rhodococcal AMO, opened the possibility of funda- mental differences in reaction mechanism and ⁄ or sub- strate selectivity between AMO and its homologues. In addition to their possible mechanistic significance, these inferences had potential implications for com- mercial application of the enzyme. An AMO that was not destroyed by alkynes would be tolerant of alkyne contamination of the alkene feedstock, whereas one that did not exhibit the peroxide shunt reaction could not be economically employed in a cell-free system without a means for regeneration of NADH. In the context of a whole-cell biocatalyst that would bypass problems with coenzyme regeneration, we began the present study of AMO by investigating the effect of alkynes on AMO-dependent growth of R. rhodochrous B-276. Subsequently, by refining the purification, inhi- bition and activity assay protocols we have been able to produce large amounts of pure high-activity AMO and to perform a more thorough investigation of its interaction with alkynes and hydrogen peroxide than has previously been possible. Results Alkynes strongly inhibit growth of R. rhodochrous B-276 on propene The effect of alkynes on whole-cell systems expressing AMO was investigated by monitoring the growth of flask cultures of R. rhodochrous B-276 in the presence of a range of alkynes (Fig. 1). A fivefold molar excess Adventitious reactions of alkene monooxygenase W. L. J. Fosdike et al. 2662 FEBS Journal 272 (2005) 2661–2669 ª 2005 FEBS of propene over the alkyne was used as the carbon and energy source to ensure that cell growth was dependent on AMO. The observed complete inhibition of growth in the presence of propyne and but-1-yne was qualita- tively a much more severe effect than was expected from the relatively low level (£ 70%) of competitive inhibition that the previous data implied [1] and suggested either that these alkenes were inhibiting AMO to a greater extent than had been observed in the purified enzyme experiments or that some other essential metabolic pro- cess was inhibited by alkynes. Even more remarkably, ethyne (which barely inhibited purified AMO in the previous study [1]), caused significantly delayed growth and reduced growth rate relative to the control. Preincubation of AMO with alkyne and NADH allowed turnover-dependent inhibition to be observed In order to explain the unexpectedly large effect of alkynes on the growth of AMO-expressing R. rhodo- chrous B-276 cells on propene, the effect of alkynes on the activity of purified AMO was re-examined. Previously, the inhibition properties of AMO were investigated in reactions in which propene and the alkyne were present simultaneously and were added before NADH, which is essential for turnover of the AMO complex [1,13]. Hence, although the effect of propyne in causing an increase in K m with respect to propene without significant change in V max was consistent with reversible competitive inhibition, turn- over-dependent and -independent events could not be distinguished. Consequently, the nature of the inhibition could not be unambiguously assigned. To resolve this uncertainty, purified AMO components and NADH were preincubated aerobically in the pres- ence of ethyne or propyne (5% v ⁄ v in the headspace) and subsequent assay of the propene oxygenation activity after removal of the alkyne clearly demonstra- ted irreversible inhibition at 80% (Table 1). It was clear that this inhibition was predominantly turnover- dependent because omission of NADH during the pre- incubation phase completely abolished it. Conversely, when a much higher concentration of ethyne (50% v ⁄ v, corresponding to an increase in calculated liquid- phase concentration from 1.8 mm to 18 mm) and pro- pene were added at the same time, before NADH, no inhibition was observed during a 10-min assay. A sim- ilar assay using propyne (35% v ⁄ v, corresponding to an increase in calculated liquid-phase concentration from 3.1 mm to 22 mm) in place of ethyne resulted in only a 50% reduction in epoxide formation over a 10-min assay, relative to the control in which nitrogen was substituted for the alkyne. The fact that the pres- ence of propene protects against inhibition by the alky- nes is consistent with the alkynes’ acting as suicide substrates which compete for the same active site on the enzyme as the natural substrate propene. Residual AMO activity was measured as a function of the time between the start of the reaction (addition of NADH) and removal of the alkyne by flushing with nitrogen. First-order decay of enzyme activity was observed, from which first order decay constants for the inactivation of the enzyme by propyne and ethyne under these conditions could be calculated (Fig. 2). It is likely that the linear part of the graph in Fig. 2 does not cross the ordinate at the position corresponding to the uninhibited enzyme activity because the measured reaction times do not include the time taken to remove the alkyne during the flushing process and are there- fore uniformly underestimated. The positive deviation of the latest data points from the extrapolated linear Fig. 1. Effect of alkynes on the growth of R. rhodochrous B-276 using propene as the growth substrate. Cultures were incubated aerobically in the presence of propene plus nitrogen (solid symbols, solid line), ethyne (solid symbols, dotted line), propyne (open sym- bols, solid line) and but-1-yne (open symbols, dotted line). Table 1. Turnover-dependent inhibition of AMO by alkynes. Preincubation before assay a Percent activity b None 100 ± 10 Ethyne 96 ± 12 Propyne 85 ± 8 Ethyne + NADH 18 ± 4 Propyne + NADH 21 ± 2 a Reactions contained 8 lM of epoxygenase and 12 lM each of reductase and GST-coupling protein fusion in a total volume of 100 lL; the headspace concentration of alkyne was 11% (v ⁄ v) and the reaction was preincubated for 10 min before removal of the alkyne and assaying with propene as the substrate, in the presence of O 2 and NADH, as described in the Experimental procedures. b Specific activities are given as percentages of the uninhibited activity of 192 nmolÆmin )1 Æ(mg of epoxygenase) )1 . W. L. J. Fosdike et al. Adventitious reactions of alkene monooxygenase FEBS Journal 272 (2005) 2661–2669 ª 2005 FEBS 2663 portion possibly reflects the exhaustion of alkyne later in the reaction, if a number of turnovers are required per inactivation event. Improved epoxygenase preparations allowed detection of the AMO-catalysed peroxide shunt reaction Investigation of the effect of hydrogen peroxide on AMO was facilitated by the improved protocols for growth of R. rhodochrous B-276 and purification of the binuclear iron site-containing epoxygenase component described in the Experimental procedures. Use of pro- pene as the sole carbon and energy source ensured a high level of AMO expression by making cell growth dependent on AMO. The subsequent purification pro- tocol yielded > 95% pure epoxygenase with a specific activity of at least 145 nmolÆmin )1 Æ[mg of protein] )1 , which was more than twice as active as previous pre- parations produced from cells grown in rich medium [1,13]. In addition, the use of smaller assay reaction volumes (0.1 mL rather than 0.5 mL used previously) enabled extensive investigations to be carried out at high epoxygenase concentrations. Thanks to these improvements, the peroxide shunt reaction was found to be detectable and easily quantifiable at 40 lm epoxygenase after only 3 min reaction time, using 500 mm hydrogen peroxide and propene at 37% (v ⁄ v) in the headspace gas. The specific activity of the hydrogen peroxide-activated reaction under these con- ditions was 32 nmolÆmin )1 Æ[mg of epoxygenase] )1 , i.e. 22% of the NADH-dependent reaction catalysed by the whole AMO complex. The amount of epoxypro- pane produced was unchanged when the reaction was purged of oxygen by flushing with oxygen-free nitro- gen for 5 min before addition of the hydrogen per- oxide, showing that the reaction did not require molecular oxygen. Formation of product was depend- ent on the presence of hydrogen peroxide and, more importantly, active epoxygenase. Omission or prior heat-denaturation (100 °C, 5 min) of the epoxygenase completely abolished formation of epoxypropane under otherwise identical conditions. The rate of epoxypropane formation via the peroxide shunt reaction was linear with epoxygenase concentra- tion between 20 and 60 lm (data not shown). When a reaction time of 3 min was used, the reaction rate was linear with hydrogen peroxide concentration up to 1.0 m, the highest concentration tested, suggesting that the K m for hydrogen peroxide was considerably greater than 1 m. This contrasts with the lower value of K m for hydrogen peroxide of 66 mm estimated from experi- ments with the corresponding hydroxylase component of sMMO [25]. When longer reaction times were used, the average rate did not increase beyond about 500 mm hydrogen peroxide (Fig. 3), probably because of pro- gressive inactivation of the enzyme by higher concentra- tions of hydrogen peroxide in a manner that was also observed with sMMO [25]. The AMO epoxygenase- catalysed peroxide shunt reaction showed moderate inhibition by the coupling protein component of AMO (29 ± 8% inhibition relative to an activity of 30 ± 2 nmolÆmin )1 Æ[mg of epoxygenase] )1 at a coupling protein ⁄ epoxygenase molar ratio of 3 : 1), again similar to the result obtained with sMMO [25]. Fig. 2. Kinetics of inactivation of AMO by (A) propyne and (B) ethyne. Alkynes were added to the headspace at 6.25% (v ⁄ v), which was calculated to give equilibrium aqueous-phase concentrations of propyne and ethyne of 3.9 and 2.2 m M, respectively. Residual AMO activity was measured after removal of the alkyne using pro- pene as the substrate and are derived from single-timepoint meas- urements of the product after 10 min reaction time; 100% activity corresponded to 153 nmolÆmin )1 Æ(mg of epoxygenase) )1 . Error bars show standard deviation from three experiments. First order decay constants during the exponential decay periods were 0.083 min )1 and 0.13 min )1 for propyne and ethyne, respectively. The zero-time measurement is derived from enzyme that was not exposed to the alkyne. Adventitious reactions of alkene monooxygenase W. L. J. Fosdike et al. 2664 FEBS Journal 272 (2005) 2661–2669 ª 2005 FEBS The peroxide shunt reaction in AMO showed low stereoselectivity even in the presence of the coupling protein When epoxypropane produced via the whole-complex reaction was subjected to chiral analysis, the R-enantio- mer predominated with an enantiomeric excess >80%, consistent with the 83% previously observed [1]. In contrast, epoxypropane produced by the AMO epoxy- genase via the peroxide shunt reaction, whilst still showing a predominance of the R-enantiomer, had an enantiomeric excess of only 25%. In sMMO from Methylosinus trichosporium OB3b [30], it has been observed that the protein B compo- nent, equivalent to the coupling protein of AMO, influences the regioselectivity of the enzyme. If the coupling protein influenced product distribution in AMO also, it was possible that the low stereoselecti- vity observed via the peroxide shunt was due to the absence of the coupling protein. It was found that the presence of the coupling protein (up to a threefold molar excess relative to the epoxygenase) had no signi- ficant effect on the chiral composition of the epoxypro- pane product in the peroxide shunt reaction, although the possibility that the coupling protein was damaged by the high concentration of hydrogen peroxide used in this experiment cannot be excluded. Modification of positively charged residues on the surface of the epoxygenase allowed functional separation of the whole-complex and peroxide shunt activities As treatment of the hydroxylase component of sMMO (equivalent to the epoxygenase of AMO) with reagents specific for positively charged moieties completely inhibited the whole-complex reaction but not the peroxide shunt [29], covalent modification of AMO afforded an additional test of its biochemical similarity to sMMO. When the epoxygenase of AMO was reacted with the primary amine-specific reagent sulfo-NHS-acetate, the whole complex reac- tion was abolished whilst the activity via the per- oxide shunt was unaffected, as in the sMMO system. The control experiment described in the Experimental procedures confirmed that the specific inactivation of the whole-complex reaction was due solely to the effect of the sulfo-NHS-acetate on the epoxygenase component and not due to the effect of any carried over reagent during the assay reaction. A similar functional separation of the peroxide shunt and whole-complex reactions was observed by using the arginine-specific reagent p-hydroxyphenylglyoxal, which did not significantly inhibit the peroxide shunt reac- tion but resulted in progressive inactivation of the whole-complex reaction (data not shown). These results suggested that accessible positively charged residues were necessary for interactions between the enzyme components but not per se for substrate oxy- genation at the active site. Consistent with the hypo- thesis that protein–protein interactions between the AMO components require positive charges on the epoxygenase, chemical modification of the coupling protein with sulfo-NHS-acetate or p-hydroxyphenyl- glyoxal had no effect on its activity in the whole- complex propene oxygenation reaction (data not shown). Discussion The observations that the rhodococcal AMO, like other binuclear iron-centre monooxygenases, exhibits turnover-dependent inhibition by alkynes and can be activated by hydrogen peroxide support a unified mechanism for oxygenation reactions catalysed by this family of enzymes. The results of Gallagher et al. [1], where it was observed using partially purified AMO that propyne increased the apparent K m with respect to propene but left V max unchanged, can now be reinterpreted as showing that competition between the substrate propene and the inhibitor propyne not only alleviates inhibition at high substrate concentration by preventing propyne from blocking the active site but in so doing also protects the enzyme from the irreversible inhibition that would result from turnover of the alkyne. The relatively small amount of inhibition of growth and AMO activity seen when ethyne was pre- sent at the same time as propene are consistent with the conclusion that the two-carbon ethyne competes Fig. 3. Kinetics of peroxygenation of propene catalysed by the AMO epoxygenase. Activity was measured by quantifying epoxy- propane formation as described in the Experimental procedures, using 40 l M epoxygenase and reaction times of 3 min (solid line), 10 min (dotted line) and 15 min (dashed line). Error bars show standard deviations from three experiments. W. L. J. Fosdike et al. Adventitious reactions of alkene monooxygenase FEBS Journal 272 (2005) 2661–2669 ª 2005 FEBS 2665 less well with propene for the active site than the three-carbon propyne. The approximately twofold dif- ference in solubility of the two alkynes would not be expected to account for the gross difference in inacti- vation that was observed in competition with the pro- pene substrate. As it is now clear that the adventitious activities of AMO are broadly similar to those of other binuclear iron-centre monooxygenases, the range of activities exhibited by this family of enzymes can be explained by differences in substrate binding and the ability of the highly oxygenating species at the active site to acti- vate high energy C–H bonds such as those in benzene, methane and other alkanes. The fact that AMO can be activated via hydrogen peroxide is consistent with the presence of a binuclear iron–peroxo intermediate sim- ilar to that found during the catalytic cycle of sMMO [31,32]. Whether, as previously proposed [33], AMO is unable to oxidize methane because it lacks the charac- teristic and probably diferryl (Fe IV 2 ) intermediate Q observed in sMMO must await future studies. Whatever the structural differences that underlie the difference in active-site reactivity of the binuclear iron- centre monooxygenases, it appears to be relatively subtle as far as the immediate environment of the binuclear iron site is concerned. Not only are the ligat- ing residues perfectly conserved throughout the group [3] but also the recent structure of the hydroxylase component of toluene ⁄ o-xylene monooxygenase has indicated that the only appreciable difference between the active-centre ligation environments of this enzyme and sMMO, which share only 24% identity in their a-subunits, are relatively minor differences in active site hydrogen bonding patterns and the coordinating nitrogen atom of one histidine ligand [34]. The results obtained with the reagents specific for positively charged groups reveal an additional level of similarity between AMO and sMMO that allows functional separation of the whole-complex and per- oxide shunt reaction in both enzyme systems. In both systems, dioxygen activation and ⁄ or functional interactions between the enzyme components require accessible positively charged moieties on the terminal oxygenase, whereas the actual process of oxygen insertion into the substrate does not. There may be at least two positively charged moieties, perhaps unidentified conserved residues, that are independ- ently necessary for electron transfer and⁄ or intersub- unit interactions in each enzyme because primary amine- and arginine-specific reagents have similar effects in both systems. The diminished enantiomeric excess of epoxypro- pane production observed when AMO was activated via the peroxide shunt reaction is reminiscent of the altered substrate specificity and product distribution seen in the peroxide shunt reaction in sMMO [25,30]. In sMMO from Methylosinus trichosporium OB3b, the coupling protein component has been shown to influence the regioselectivity of oxygenation reactions [30] and mutagenesis and modelling studies have suggested that the coupling protein interacts directly with the hydroxylase active site and controls substrate entry [35]. The binding site for the coup- ling protein of AMO may be similarly positioned on the epoxygenase. However, unless the coupling pro- tein is damaged by the high concentration of hydro- gen peroxide used, presence of the coupling protein is clearly not the sole determinant of enantioselective catalysis in AMO because addition of the coupling protein to the AMO epoxygenase-catalysed peroxide shunt reaction reduced the overall reaction rate but did not increase the enantiomeric excess of the prod- uct. Whilst the operation of the peroxide shunt reaction in AMO presents an opportunity for devel- opment of an AMO-based biocatalyst that is inde- pendent of reduced coenzyme, the low product enantiomeric excess obtained via the peroxide shunt presents an important new question about the origin of enantioselectivity in this enzyme and a challenge to novel applications of AMO for chiral synthesis. Experimental procedures Bacterial strains and growth conditions R. rhodochrous B-276 was cultivated at 30 °C in nitrate min- imal salts liquid medium or on nitrate minimal salts agar [36], using propene as the sole carbon and energy source. Batch cultures for analysis of the effect of alkynes on growth were performed in 1-L cultures in 2-L conical flasks (Quick- fit, Fisher, Loughborough, UK), sealed with Subaseal Ò rub- ber seals (W. Freeman, Barnsley, UK). After inoculation, 600 mL of headspace gas was removed and replaced by 500 mL of propene, plus 100 mL of the appropriate alkyne or nitrogen, as stated for each experiment. Growth was monitored over a 5-day incubation at 30 °C with shaking (180 r.p.m.) by measuring the OD 600 of samples removed through the seal with a hypodermic syringe. Large-scale con- tinuous growth of R. rhodochrous B-276 for purification of the AMO epoxygenase and reductase component was achieved by using a 2000 Series fermentor (LH Engineering, Stoke Poges, UK) with a working volume of 4 L and a dilu- tion rate of 0.035 h )1 . The culture was gassed with a 1 : 10 (v ⁄ v) propene ⁄ air mixture at a flow rate of 1 LÆh À1 , agitated at an imepellor speed of 450 r.p.m. and maintained at pH 7.0. The Escherichia coli strain used to produce the Adventitious reactions of alkene monooxygenase W. L. J. Fosdike et al. 2666 FEBS Journal 272 (2005) 2661–2669 ª 2005 FEBS recombinant glutathione S-transferase (GST)-coupling pro- tein fusion was cultivated as described previously [37]. Purification of the AMO components All protein purification procedures were conducted at 0–4 °C. For purification of the AMO epoxygenase (i.e. ter- minal oxygenase) and reductase components, cells from 20 L of propene-grown R. rhodochrous B-276 culture at an OD 600 of 12–15 were harvested at 3000 g using a Westfalia continuous centrifuge (Northvale, NJ, USA), washed by centrifugation (10 000 g, 10 min) and resuspension in 10 mm MgSO 4 ,5%(v⁄ v) glycerol, 25 mm Mops pH 7.5 and resuspended in a minimal volume of the same buffer containing benzamidine (1 mm), dithiothreitol (1 mm) and a total glycerol concentration of 15% (v ⁄ v) (buffer B). Deoxyribonuclease I (Sigma, Gillingham, UK) was added to 20 lgÆmL )1 and the cells were broken by passing twice through a high-pressure cell disrupter (172 MPa; Constant Systems, Warwick, UK), after which the cell-free extract was centrifuged (48 000 g, 1 h) and the supernatant (the soluble extract) was removed. The soluble extract was applied to a DE52 anion exchange column (Whatman, Maidstone, UK; 5 cm · 15 cm), previously equilibrated with buffer B. After wash- ing with buffer B, the column was eluted with a step gradi- ent of 150 and 250 mm NaCl in buffer B. The reductase was purified from the 250-mm NaCl fraction as described previously [13]. The 150-mm NaCl fraction was concentra- ted to a volume of 200 mL using a 900 cm 2 surface-area spiral-wound ultrafiltration cartridge (Amicon, Stonehouse, UK) with a molecular size cut-off of 10 kDa. MgSO 4 was added to a final concentration of 0.5 m and then the epoxygenase-containing solution was applied to a Phenyl Sepharose high-performance column (Amersham-Pharma- cia, Little Chalfont, UK; 2.6 · 12 cm) previously equili- brated with buffer D (25 mm Mops pH 7.5 plus benzamidine and dithiothreitol, each at 1 mm) containing 0.5 m MgSO 4 . Proteins were eluted with a linear gradient of 0.5–0 m MgSO 4 in buffer D. Fractions containing epoxygenase activity, which eluted at 0 mm MgSO 4 , were adjusted to an MgSO 4 concentration of 10 mm and concen- trated using a 30 kDa molecular size cut-off membrane (Amicon). The phenyl Sepharose column was washed with 0.7% (w ⁄ v) sodium cholate after use to remove residual protein and prevent loss of epoxygenase yield during subse- quent preparations. The concentrated epoxygenase was applied to a Mono Q anion exchange column (Amersham- Pharmacia; 1 · 10 cm) that had been equilibrated with 25 mm Mops pH 7.5 containing 15% (v ⁄ v) glycerol and 10 mm MgSO 4 and then proteins were eluted with a lin- ear gradient of 0–400 mm NaCl in the same buffer. The pure epoxygenase eluted at 300 mm NaCl. The coupling protein used throughout this study, which was a recombinant GST-coupling protein fusion that is fully active with the GST tag attached, was produced in Escherichia coli and purified by affinity chromatography [37]. AMO assays and inhibition studies AMO assays and alkyne inhibition reactions were per- formed in 100-lL reaction volumes in 1.7-mL crimp-seal glass vials. For measurement of activity via the whole- complex AMO reaction, the three AMO components were mixed with 25 mm Mops pH 7.5 containing 10 mm MgSO 4 to give 8 lm epoxygenase, 12 lm reductase and 12 lm coupling protein. The vial was then sealed and 0.7 mL of the headspace gas was removed and replaced with 0.7 mL of propene, after which the vial was preincubated at 30 °C for 30 s. The reaction was initiated by adding NADH (to 100 lm) and the vial was shaken (180 r.p.m.) at 30 °C for a further 3 min, unless otherwise stated, before the epoxypro- pane formed was quantified by gas chromatography of 0.5-mL gas phase samples, as described previously [1]. Activity assays of the individual AMO components during purification were performed in the presence of an excess of the other two components. Peroxide shunt assays were per- formed in a similar manner to the whole-complex reactions except that reductase and (except where stated otherwise) the coupling protein were omitted and the reaction was started by adding hydrogen peroxide instead of NADH. The reaction time and concentrations of protein compo- nents and hydrogen peroxide are stated for each experi- ment. Inhibition of the AMO complex by alkynes was achieved as follows. Epoxygenase (8 lm), reductase (12 lm) and coupling protein (12 lm) were mixed and the reaction vial was sealed. Headspace gas was removed and replaced with an equal volume of the alkyne, to give the alkyne concentration stated for each experiment. Concen- trations of alkynes in the aqueous phase were calculated using Henry’s law and Henry’s constants of 0.068 mÆatm )1 [38] and 0.037 mÆatm )1 (http://www.mpch-mainz.mpg.de/ $sander/res/henry.html), respectively, for propyne and eth- yne. The vial was preincubated with shaking at 30 °C for 30 s and then, except where stated otherwise, NADH was added to 100 lm. Turnover-dependent inhibition was then allowed to proceed under the same incubation conditions, for 10 min unless stated otherwise. Unreacted alkyne was removed by flushing with nitrogen for 5 min; the vial was then flushed with air and the remaining AMO activity was assayed by replacing 0.7 mL of the headspace gas with 0.7 mL of propene and incubating at 30 °C for a fur- ther 10 min, before removing 0.5 mL of the headspace gas for quantitation of the epoxypropane product by gas chro- matography as above. Residual propene oxidation activity was measured after 10 min reaction with the alkyne, at a range of concentrations, in the presence of NADH (100 lm). The decay of enzyme activity as a function of W. L. J. Fosdike et al. Adventitious reactions of alkene monooxygenase FEBS Journal 272 (2005) 2661–2669 ª 2005 FEBS 2667 time was analysed by plotting the natural logarithm of the residual activity against time, to yield the apparent first- order decay constant, k_app, from the slope of the linear portion of the graph [39]. Chiral analysis Reactions for analysis of chiral composition were scaled up to 1-mL volume and incubated at 30 °C until the epoxypro- pane concentration was about 2 mm. Epoxypropane was extracted using 200–300 lL of diethyl ether, which was dried using molecular sieve and analysed by means of capil- lary GC using a Phillips 4500 GC apparatus fitted with a Chiraldex (20 m · 0.25 mm) a-cyclodextrin trifluoroacetate column (Advanced Separation Technologies, Whippany, NJ, USA). The injection volume was 1 lL, the split ratio was 100 : 1 and the carrier gas (nitrogen) flow rate was 1.3 mLÆmin )1 . The injector and flame ionization detector were maintained at 150 °C and the column temperature was 30 °C. Modification of positively charged groups Covalent modification of the AMO epoxygenase and coup- ling protein was effected using protocols based on those used with sMMO [29], as follows. Accessible primary amine groups were modified by reacting the epoxygenase or coupling protein (10 mgÆmL )1 ) with a 13-fold molar excess of sulfosuccinimidyl acetate (sulfo-NHS-acetate; Pierce, Rockford, IL, USA) at room temperature for 30 min. Un- reacted sulfo-NHS-acetate was removed by buffer exchange via three cycles of ultrafiltration [using a Microcon centrifu- gal concentrator (Amicon)] and dilution. In the case of the epoxygenase, a 30 kDa cut-off membrane was used and each time the sample was diluted with a volume equal to the original sample volume of 25 mm Mops pH 7.5 con- taining 15% (v ⁄ v) glycerol. In the case of the coupling pro- tein, the membrane had a 10 kDa cut-off and glycerol was omitted from the dilution buffer. The reaction between the epoxygenase and sulfo-NHS-acetate abolished the activity of the epoxygenase in the AMO whole-complex reaction and so a control was performed to confirm that the observed inhibition was due to reaction of the sulfo-NHS- acetate with the epoxygenase and not due to reaction of carried over reagent with the other AMO components. Here, the epoxygenase was not added to the reaction until after the sulfo-NHS-acetate had been removed by buffer exchange. No inhibition of the reaction was observed, con- firming that the buffer exchange procedure was effective in removing the sulfo-NHS-acetate and showing that inhibi- tion required contact between the epoxygenase and sulfo- NHS-acetate. Modification of accessible arginyl side-chain guanidinium groups was effected by reacting the epoxyge- nase or coupling protein (10 mgÆmL )1 ) with p-hydroxyphe- nylglyoxal (12 mm) for 20 min at room temperature, after which the reaction was quenched by adding arginine to 60 mm. References 1 Gallagher SC, Cammack R & Dalton H (1997) Alkene monooxygenase from Nocardia corallina B-276 is a member of the class of dinuclear iron proteins capable of stereospecific epoxygenation reactions. Eur J Biochem 247, 635–641. 2 Saeki H & Furuhashi K (1994) Cloning and characteri- sation of the Nocardia corallina B-276 gene cluster encoding alkene monooxygenase. J Ferment Bioeng 78, 399–406. 3 Leahy JG, Batchelor PJ & Morcomb SM (2003) Evolu- tion of the soluble diiron monooxygenases. FEMS Microbiol Rev 27, 449–479. 4 Kotani T, Yamamoto T, Yurimoto H, Sakai Y & Kato N (2003) Propane monooxygenase and NAD + -depend- ent secondary alcohol dehydrogenase in propane metabolism Gordonia sp. strain TY-5. J Bacteriol 185, 7120–7128. 5 Sluis MK, Sayavedra-Soto LA & Arp DJ (2002) Mole- cular analysis of the soluble butane monooxygenase from ‘Pseudomonas butanovora’. Microbiology 148, 3617–3629. 6 Zhou NY, Jenkins A, Chion CKNCK & Leak DJ (1999) The alkene monooxygenase from Xanthobacter strain Py2 is closely related to aromatic monooxy- genases and catalyzes aromatic monohydroxylation of benzene, toluene, and phenol. Appl Environ Microbiol 65, 1589–1595. 7 Hartmans S, Weber FJ, Somhorst DPN & de Bont JAM (1991) Alkene monooxygenase from Mycobacter- ium – a multicomponent enzyme. J Gen Microbiol 137, 2555–2560. 8 Coleman NV & Spain JC (2003) Distribution of the coenzyme M pathway of epoxide metabolism among ethene- and vinyl chloride-degrading Mycobacterium strains. Appl Environ Microbiol 69, 6041–6046. 9 Murrell JC, Gilbert B & McDonald IR (2000) Molecu- lar biology and regulation of methane monooxygenase. Arch Microbiol 173, 325–332. 10 Johnson GR & Olsen RH (1995) Nucleotide-sequence analysis of genes encoding a toluene benzene-2-monooxy- genase from Pseudomonas sp. strain JS150. Appl Environ Microbiol 61, 3336–3346. 11 Lipscomb JD (1994) Biochemistry of the soluble methane monooxygenase. Annu Rev Microbiol 48, 371–399. 12 Small FJ & Ensign SA (1997) Alkene monooxygenase from Xanthobacter strain Py2 – purification and charac- terization of a four-component system central to the bacterial metabolism of aliphatic alkenes. J Biol Chem 272, 24913–24920. Adventitious reactions of alkene monooxygenase W. L. J. Fosdike et al. 2668 FEBS Journal 272 (2005) 2661–2669 ª 2005 FEBS 13 Miura A & Dalton H (1995) Purification and characteri- sation of the alkene monooxygenase from Nocardia corallina B-276. Biosci Biotechn Biochem 59, 853–859. 14 Pikus JD, Studts JM, Achim C, Kauffmann KE, Mu ¨ nck E, Steffan RJ, McClay K & Fox BG (1996) Recombi- nant toluene 4-monooxygenase: catalytic and Mo ¨ ss- bauer studies of the purified diiron and Rieske components of a four-protein complex. Biochemistry 35, 9106–9119. 15 Byrne AM, Kukor JJ & Olsen RH (1995) Sequence analy- sis of the gene cluster encoding toluene-3-monooxygenase from Pseudomonas pickettii PKO1. Gene 154, 65–70. 16 Bertoni G, Martino M, Galli E & Barbieri P (1998) Analysis of the gene cluster encoding toluene ⁄ o-xylene monooxygenase from Pseudomonas stutzeri OX1. Appl Environ Microbiol 64, 3626–3632. 17 Furuhashi K (1992) Biological routes to optically active epoxides. In Chirality in Industry (Collins, A, Sheldrake, G & Crosby, J, eds), pp. 167–186. John Wiley & Sons, London. 18 Smith TJ & Dalton H (2004) Biocatalysis by methane monooxygenase and its implications for the petroleum industry. In Petroleum Biotechnology, Developments and Perspectives (Vazquez-Duhalt, R & Quintero-Ramirez, R, eds). Elsevier, Amsterdam. 19 Prior SD & Dalton H (1985) Acetylene as a suicide sub- strate and active site probe for methane monooxygenase from Methylococcus capsulatus (Bath). FEMS Microbiol Lett 29, 105–109. 20 Hamamura N, Storfa RT, Semprini L & Arp DJ (1999) Diversity in butane monooxygenases among butane- grown bacteria. Appl Environ Microbiol 65, 4586–4593. 21 Yeager CM, Bottomley PJ, Arp DJ & Hyman MR (1999) Inactivation of toluene 2-monooxygenase in Bur- kholderia cepacia G4 by alkynes. Appl Environ Microbiol 65, 632–639. 22 Keener WK, Watwood ME, Schaller KD, Walton MR, Partin JK, Smith WA & Clingenpeel SR (2001) Use of selective inhibitors and chromogenic substrates to differ- entiate bacteria based on toluene oxygenase activity. J Microbiol Methods 46, 171–185. 23 Ensign SA, Hyman MR & Arp DJ (1992) Cometabolic degradation of chlorinated alkenes by alkene monooxy- genase in a propylene-grown Xanthobacter strain. Appl Environ Microbiol 58, 3038–3046. 24 Foroozeh M, Primrose G, Guo Z, Bell LC, Alworth WL & Guengerich FP (1997) Aryl acetylenes as mechanism- based inhibitors of cytochrome P450-dependent mono- oxygenase enzymes. Chem Res Toxicol 10, 91–102. 25 Jiang Y, Wilkins PC & Dalton H (1993) Activation of the hydroxylase of sMMO from Methylococcus capsula- tus (Bath) by hydrogen peroxide. Biochim Biophys Acta 1163, 105–112. 26 Andersson KK, Froland WA, Lee S-K & Lipscomb JD (1991) Dioxygen independent oxygenation of hydrocarbons by methane monooxygenase hydroxylase component. New J Chem 15, 411–415. 27 Newman LM & Wackett LP (1995) Purification and characterization of toluene 2-monooxygenase from Bur- kholderia cepacia G4. Biochemistry 34, 14066–14076. 28 White RE & Coon MJ (1980) Oxygen activation by cytochrome P-450. Annu Rev Biochem 49, 315–356. 29 Balendra S, Lesieur. C, Smith. TJ & Dalton H (2002) Positively charged amino acids are essential for electron transfer and protein–protein interactions in the soluble methane monooxygenase complex from Methylococcus capsulatus (Bath). Biochemistry 41, 2571–2579. 30 Froland WA, Andersson KK, Lee SK, Liu Y & Lipscomb JD (1992) Methane monooxygenase compo- nent B and the reductase alter the regioselectivity of the hydroxylase component-catalysed reactions. J Biol Chem 267, 17588–17597. 31 Lee S-K, Nesheim JC & Lipscomb JD (1993) Transient intermediates of the methane monooxygenase catalytic cycle. J Biol Chem 268, 21569–21577. 32 Liu KE, Valentine AM, Qiu D, Edmondson DE, Apple- man EH, Spiro TG & Lippard SJ (1995) Characterisa- tion of a diiron (III) peroxo intermediate in the catalytic cycle of methane monooxygenase hydroxylase from Methylococcus capsulatus (Bath). J Am Chem Soc 117, 4997–4998. 33 Valentine AM, Stahl SS & Lippard SJ (1999) Mechanis- tic studies of reduced methane monooxygenase hydroxy- lase with dioxygen and substrates. J Am Chem Soc 121, 3876–3887. 34 Sazinski MH, Bard J, Di Donato A & Lippard SJ (2004) Crystal structure of the toluene ⁄ o-xylene mono- oxygenase hydroxylase from Pseudomonas stutzeri OX1. J Biol Chem 279, 30600–30610. 35 Brazeau B, Wallar BJ & Lipscomb JD (2003) Effector proteins from P450 cam and methane monooxygenase: lessons in tuning nature’s powerful reagents. Biochim Biophys Res Commun 312, 143–148. 36 Dalton H & Whittenbury R (1976) The acetylene reduc- tion technique as an assay for nitrogenase activity in the methane oxidizing bacterium Methylococcus capsulatus (Bath). Arch Microbiol 109, 147–151. 37 Smith TJ, Lloyd JS, Gallagher SC, Fosdike WLJ, Murrell JC & Dalton H (1999) Heterologous expression of alkene monooxygenase from Rhodococcus rhodo- chrous B-276. Eur J Biochem 260, 446–452. 38 Simpson LB & Lovell FP (1962) Solubility of methyl, ethyl, and vinyl acetylene in several solvents. J Chem Eng Data 7, 498–500. 39 Cornish-Bowden A (1995) Fundamentals of Enzyme Kinetics, 2nd edn. Portland Press, London. W. L. J. Fosdike et al. Adventitious reactions of alkene monooxygenase FEBS Journal 272 (2005) 2661–2669 ª 2005 FEBS 2669 . Adventitious reactions of alkene monooxygenase reveal common reaction pathways and component interactions among bacterial hydrocarbon oxygenases William. of flask cultures of R. rhodochrous B-276 in the presence of a range of alkynes (Fig. 1). A fivefold molar excess Adventitious reactions of alkene monooxygenase