Probing Electron Transfer in Flavocytochrome P ... - Wiley Online Library

Eur. J. Biochem. 239, 403-409 (1996) 0 FEBS 1996

Probing electron transfer in flavocytochrome P-450 BM3 and its component domains Andrew W. MUNRO’, Simon DAFF’, John R. COGGINS’, J. Gordon LINDSAY’ and Stephen K. CHAPMAN’ ’ Division of Biochemistry and Molecular Biology, University of Glasgow, Glasgow, UK Department of Chemistry, University of Edinburgh, Edinburgh, UK (Received 11 December 1995/25 April 1996)

-

EJB 95 2026/4

Rapid events in the processes of electron transfer and substrate binding to cytochrome P-450 BM3 from Bacillus megaterium and its constituent haem-containing and flavin-containing domains have been investigated using stopped-flow spectrophotometry. The formation of a blue semiquinone flavin form occurs during the NADPH-dependent reduction of the flavin domain and a species with a similar absorption maximum is also seen during reduction of the holoenzyme by NADPH. EPR spectroscopy confirms the formation of the flavin semiquinone. The formation of this semiquinone is transient during fatty acid monooxygenation by the holoenzyme, but in the presence of excess NADPH the species reforms once fatty acid is exhausted. Electron transfers through the reductase domain are too rapid to limit the fatty acid monooxygenation reaction. The substrate-binding-induced haem iron spin-state shift also occurs much faster than the k,,, at 25°C. The rate of first electron transfer to the haem domain is also rapid; but it is of the order of 5-10-times larger than the k,,, for the enzyme (dependent on the fatty acid used). Given that two successive electron transfers to haem iron are required for the oxygenation reaction, these rates are likely to exert some control over the rate of fatty acid oxygenation reactions. The presence of large amounts of NADPH also results in decreased rates of electron transfer from flavin to haem iron. In the difference spectrum of the active fatty acid hydroxylase, features indicative of a high-spin iron haem accumulate. These are in accordance with the presence of large amounts of an Fe’+-product bound enzyme during turnover and indicate that product release may also contribute to rate limitation. Taken together, these data suggest that the catalytic rate is not determined by the accumulation of a single intermediate in the reaction scheme, but rather that it is controlled in a series of steps.

Keywords: cytochrome P-450; stopped-flow kinetics ; EPR; electron transfer.

The cytochrome P-450 monooxygenases (P-450) are a ubiquitous superfamily of haem enzymes which catalyse insertion of oxygen into an enormous variety of both physiological and nonphysiological organic substrates [l -31. P-450 generally fall into one of two broad classes. Class I P-450 (bacterial/mitochondrial) are three component systems comprised of an NAD(P)H-binding flavoprotein reductase, a small iron-sulfur protein and the P-450, which is membrane bound in eukaryotic forms [4].Class I1 P450 (microsomal) are two component systems comprising an FAD-containing and FMN-containing NADPH-cytochrome P450 reductase and the P-450. This class is found almost exclusively in eukaryotes, where both components are membrane bound [4]. There are many prokaryotic class I P-450, the best characterised being the P-450cum camphor hydroxylase from Pseudomonus putida. P-450cum (CUP 101) was the first P-450 for which an atomic structure was determined [5]. More recently, a unique prokaryotic class I1 flavocytochrome P-450 from Bacillus megaterium has been characterised [6, 71. P-450 BM3 (CUP 102) has evolved from the fusion of genes encoding a fatty acid hydroxylase P-450 and a eukaryotic-like P-450-reductase to create a Correspondence t o A. W. Munro, Division of Biochemistry and Molecular Biology, The Davidson Building, University of Glasgow, Glas-

&ow, G12 SQQ, UK Fax: +44 141 330 4620. Abbreviation. P-450, cytochrome P-450 monooxygenase.

novel gene (cypl02) encoding a 119-kDa flavocytochrome [ S ] . P-450 BM3 is soluble and its haemoprotein domain has higher amino acid sequence similarity to eukaryotic family IV P-450 than to any bacterial P-450 [S].Sub-genes encoding the haem and flavin domains of P-450 BM3 have been generated by the PCR and overexpressed in Escherichia coli, and all polypeptides have been purified to homogeneity [9, 101. The haem domain was crystallised and the atomic structure was determined recently - providing a good structural model for class 11 P-450 [ll]. The unique construction of P-450 BM3 makes it attractive not only as a structural model, but also as a tool for studying the reaction kinetics and electron transfer processes in a single polypeptide class I1 P-450 system. For P-450cam (CYPIOI), camphor binding induces a lowto-high-spin shift of the haem iron, causing an increase in redox potential of the ferric haem iron and facilitating its one-electron reduction by the iron-sulfur protein putidaredoxin. A similar mechanism is thought to control the reduction of other P-450 enzymes [12-141. The next step in the ordered P-450 catalytic cycle is binding of oxygen to the ferrous haem. The further oneelectron reduction of the ferrous-dioxygen species leads to the formation of a reactive high-valent iron-oxo species which mediates substrate oxygenation [15]. Electron transfer to haem is thought to be the major determinant of rate control in P-450cum. However, it is not certain whether first or second electron transfer is slowest [16-18]. In eukaryotic class I1 systems, the second electron may be delivered from either P-450 reductase

404

Munro et al. ( E m J. Biochem. 239)

seen to increase significantly in cultures of pBM23 transformants grown to stationary phase, as has been reported previously [lo]. Approximately 30 g wet cell pellets were the starting points for polypeptide purification. Following cell breakage using a French press (two passes at 64 MPa), polypeptides were purified to homogeneity by successive steps of ammonium sulfate precipitation, ion-exchange chromatography on DEAE-Sephacel and either affinity chromatography on 2',5'-ADP-Sepharose (for intact P-450 BM3 and its reductase domain) or by affinity for Bio-Gel HTP (DNA grade) hydroxyapatite (for the haem domain of P-450 BM3) as previously described [lo]. A final gel-filtration purification step (Sephacryl S-300 HR) was used for removal of minor contaminants or degradation products, as required. Spectroscopy, protein and enzyme assays. All ultravioletvisible spectroscopy was performed on a Shimadzu 2101 spectrophotometer (Shimadzu corporation). Protein concentrations were determined using the methods of Bradford [30]and by the bicinchoninic acid technique [31] with BSA as standard. Cytochrome P-450 concentrations were estimated by the method of Omura and Sato [32] using an E of 91 mM -'cm-' at 450 nm for the reduced plus CO adduct. Steady-state fatty-acid-dependent NADPH oxidation was measured at 25°C in 20 mM Mops, MATERIALS AND METHODS pH 7.4, plus 100 mM KCl, using an E of 6.2 mM-' cm-' at E. coli strains, plasmid and bacteriophage vectors. E. coli 340nm for the oxidation of NADPH. The assay system constrains TG1 (sclpE, ksdA5, thi, d(luc-proAB), F' [ truD36, tained 20 nM P-450 BM3, 0.5 mM fatty acid and 0.2 mM proAB', lacl", lacZAMlS]) [23] and XL-1 Blue (supE44, NADPH. Oxygenated fatty acid products were separated by rehsdR17, recAl, gyrA46, thi, relA1, lac-, F' [proAB', lacI", verse-phase HPLC as previously described [33]. Steady-state cyIucZAMlS, TnlO(tet')]) 124) were used for overexpression of tochrome c reductase activity was determined using an c of the wild-type (c~p102)and mutant genes encoding cytochrome 21 mM-' cm-' at 550 nm for the reduced-minus-oxidised cytoP-450 BM3 (CYPI 02) and the PCR-generated sub-genes encod- chrome c in the same buffer, as previously described [lo]. Rate ing its constituent reductase and haem domains. Construction of constants shown are representative of at least five separate deterplasmids used for expression of P-450 BM3 and its constituent mi nations. Stopped-flow measurements of transient absorbance changes domains in E. coli has been described previously [lo]. Briefly, pBM23 consists of a 5-kb segment of B. megaterium chromo- associated with reduction of flavins (459 nm and 47.5 nm), binding of fatty acid substrates (419 nm and 390 nm), reduction of somal DNA (containing the cyp102 gene) cloned as a XbuIEcoRI restriction fragment into vector pUCl19 [25] and expres- cytochrome c (550 nm) and binding of carbon monoxide to resion is from the Baeil1u.r promoter. Plasmid pBM27 was used duced haem iron (450nm) were measured using an Applied for expression of the flavoprotein reductase domain of P-450 Photophysics SF. 17 MV stopped-flow kinetics spectrophotomeBM3 (initiating methionine and residues 473 to end) and con- ter. Reactions were performed at 25 "C (unless otherwise stated) sists of an approximately 1.8-kb PCR fragment of cypl02 cut in 20 mM Mops, pH 7.4, plus 100 mM KCI. Substrate binding with BumHI (to generate cohesive ends from restriction enzyme (419 nm and 390 nm) was measured in the presence of 500 pM sites engineered at 5' ends of PCR primers) and cloned into fatty acids, flavin reduction (459 nm and 475 nm) and flavin vector ptuc85 under an inducible tac promoter [26]. Plasmid semiquinone formation (600 nm) were measured in the presence pBM20 was used for the expression of the haem-containing do- of 500 pM NADPH, and cytochrome c reduction (550 nm) was main of P-450 BM3 (residues 1-472) and consists of an ap- measured in the presence of 1 pM cytochrome c. Flavin-to-haem proximately 1.5-kb PCR fragment of cyplO2 digested with electron transfer was measured in carbon-monoxide-saturated EcoRI and RainHT (to generate cohesive ends as with pBM27) solutions which had previously been degassed and bubbled for and cloned into pUC118 [25] under the luc promoter. FMN- 10 min with oxygen-free nitrogen. Rates were measured at deficient mutants G570D and W574D of P-450 BM3 holoen- 450 nm in the presence of 500 pM fatty acid. Enzyme concentrazyme have been described previously 1271. tions were from 20 pM to 200 pM (flavin-to-haem electron Molecular biology techniques. DNA manipulations, bacte- transfer) and from 1 pM to 10 pM (other assays). Analysis of rial transformations and other molecular techniques were per- stopped-flow data were performed using the SF.17MV software formed by standard methods 1281. and Origin (Microcal), both of which use non-linear leastExpression and purification of intact cytochrome P-450 squares regression analysis. Traces were fitted to single or BM3 and its constituent domains. E. coli transformants carry- multiple exponentials, as appropriate. ing plasmids encoding wild-type P-450 BM3 and its constituent Rapid quenched-flow EPR experiments for analysis of the flavoprotein reductase domain were grown overnight to high cell flavin semiquinone content of cytochrome P-450 BM3 during density in Terrific Broth plus antibiotic (ampicillin) [29]. Isopro- steady-state monooxygenation of arachidonic acid or reduction pyl-thio-P-D-galactopyranoside inducer (final concentration of cytochrome c were performed in 20 mM Mops, pH 7.4, con25 mg/ml) was added to facilitate expression from plasmids taining 100 mM KC1. For each experiment, a 1-ml reaction mixpBM27 (reductase domain) and pBM20 (haem-containing do- ture containing 500 pM arachidonate or 500 pM oxidised cytomain). Intact P-450 BM3 was expressed from plasmid pBM23 chrome c. 20 pM cytochrome P-450 BM3 and 500 pM NADPH under the control of its own promoter and without addition of was generated by rapid mixing, forced through a variable length exogenous inducer. Expression levels of intact P-450 BM3 were of HPLC steel tubing before being sprayed into a bath of isopen-

or cytochrome b, 1191. Here, the slow interaction of the membranous P-450 and P-450 reductase/cytochrome b, to form a catalytic complex is the primary determinant of reaction rate. Various researchers have genetically engineered fusion proteins between microsomal P-450 and P-450 reductase [20, 211. Recently, Sakaki et al. [22] studied a protein fusion between rat P450 1Al and yeast NADPH-cytochrome P-450 reductase, and concluded that product release had replaced flavin to haem electron transfer as the rate-limiting step in this fusion protein's catalytic cycle. Flavocytochrome P-450 BM3 has an extremely high catalytic rate, the main reason for which is likely to be the fusion of the redox partners in close proximity, allowing rapid flavin-tohaem electron transfer. In this study, we have utilised large quantities of purified P-450 BM3 and domains for spectroscopic and stopped-flow spectrophotometric assays to determine the velocity of various electron-transfer and substrate-binding steps in its catalytic cycle. These are measured by transient and stable absorbance changes associated with alterations in reduction/oxidation states of the bound haem and tlavin chromophores, and with attachment of substrates and haem ligands.

Munro et al. (ELMJ. Biochem. 239)

405

Fig. 1. Electron transfer scheme for the catalytic cycle of flavocytochrome P-450 BM3. The productive catalytic cycles of electron transfer through the three coenzymes of the enzyme are shown. In cycle 1, two electrons (represented as black dots) derived from the transfer of NADPH hydride are passed to FAD to form the hydroquinone (FAD"). Two successive single-electron transfers to FMN regenerates oxidised FAD. In cycle 2, the FMN is reduced to the semiquinone form (FMN') by transfer of an electron from FAD hydroquinone or semiquinone (FAD'). This FMN semiquinone then transfers its electron to the haem. In cycle 3, oxidised (ferric) P-450 haem (P4.50) binds its fatty acid substrate (S) and can then be reduced by an electron derived from the FMN semiquinone to form the femus substrate-bound species ('P450[S]). Oxygen is then bound and a second electron is transferred from FMN semiquinone ("P450[S][02]). Subsequent (rapid) steps are not clearly understood, but are known to

involve the formation of a transient high valency 0x0-iron (active oxygen) species which performs the monooxygenation reaction. Hydroxylated fatty acid (SOH) and water are then released to regenerate the oxidised P 4 5 0 . The point at which cytochrome c is reduced is indicated in cycle 2.

tane (Probalo), maintained at - 140°C by a liquid-nitrogencooled cryostat. The apparatus was provided by Dr John Ingledew, University of St Andrews. To vary the time base for the experiment, different lengths of steel tube were used, to allow different degrees of aging. The frozen reaction mixture was then packed into a an EPR tube and stored in liquid nitrogen. Samples analysed were those frozen after 20 ms and 50 ms of turnover. Samples of NADPH-reduced cytochrome P-450 BM3 holoenzyme and reductase domain (20 pM) were also frozen 30 s after the addition of excess (200 pM) NADPH, by freezing Eppendorf tubes containing the mixtures in liquid nitrogen. These samples were also packed into EPR tubes and stored in liquid nitrogen. All EPR measurements were performed on a Bruker instrument. Integrals were calculated for the FMN semiquinone signal at 184°C. Materials. Molecular biology reagents were purchased from Boehringer or United States Biochemicals. DEAE-Sephacel was purchased from Pharmacia-LKB. All other reagents and enzymes were obtained from Sigma.

RESULTS Steady-state kinetic measurements of substrate turnover catalysed by P-450 BM3 and its reductase domain. Intact cytochrome P-450 BM3 catalyses the oxygenation of a variety of fatty acid substrates, including lauric acid and the polyunsaturated fatty acid arachidonate. At 25 "C NADPH-dependent fatty acid monooxygenation occurs at approximately 15.5 s - ' for lauric acid, and 24.5 s-' for arachidonic acid (Table 1).The catalytic cycle involved is complex, with the involvement of the binding of fatty acid, then oxygen to the haemoprotein, two successive electron-transfer steps to the haemoprotein from the reduced flavoenzyme partner, formation of the active oxygen specics which catalyses the insertion of oxygen inlo Lhe fatty acid, other (rapid) electronic rearrangements in the haem and release of product (Fig. 1). P-450 BM3 and its reductase domain both catalyse rapid reduction of cytochrome c, electrons passing from FMN onto the oxidised electron acceptor. At 25 "C, the k,,, values for cytochrome c reduction by P-450 BM3 and its reductase domain were virtually identical (Table 1). These rates also reflect complex processes, including the binding of NADPH, hydride-ion

Table 1. Steady-state kinetic parameters for substrate turnover by cytochrome P-450 BM3 and its reductase domain. Rate constants were determined at 25"C, as described in Materials and Methods. The

fatty acid used is given in parenthesis. Reaction

Rate constant for P-450 BM3

reductase domain

S-'

Cytochrome c reduction 43.8 ? 1.2 41.0 i- 0.9 Fatty acid hydroxylation 15.5 _t 0.5 (laurate) 21.9 ? 1.2 (myristate) 24.5 -+ 1.3 (arachidonate) ~

transfer to FAD and dissociation of NADP, interflavin electron transfer (FAD to FMN), binding of oxidised cytochrome c and transfer of a single electron from the FMN to the cytochrome c to facilitate reduction (Fig. 1 ) .

Stopped-flow measurement of substrate binding to P-450 BM3 and its P-450 domain. The binding of fatty acid substrates to P-450 BM3 induces a haem ferric (Fe") iron low-to-highspin shift. The haem Soret peak shift from approximately 419 nm to approximately 390 nm accompanies the binding of all known fatty acid substrates. Measurements of the rates of absorbance change on the binding of lauric acid were made at 390 nm, 419 nm and 406 nm, the latter being an isosbestic point for the low-spin and high-spin forms of P-450 BM3 holoenzyme, and for its P-450 domain. Rate constants in excess of 700 s-' (near the limit of resolution of the instrument) were measured for the increase in absorbance at 390 nm and for the decrease in absorbance at 419 nm, rcgardless of thc fatty acid used. The rates were essentially identical for both holoenzyme and P-450 domain (Table 2). As expected, the change in absorbance at 406 nm was negligible (less than 0.002) in both cases.

Stopped-flow measurements of reduction of flavins in P-450 BM3 and its reductase domain. The reduction of flavins in P450 BM3 holoenzyme and in the reductase domain were measured at 459 nm and at 475 nm, the former wavelength being

406

Munro et al. ( E M J . Biochern. 239)

Table 2. Stopped-tlow kinetic parameters for cytochrome P-450 BM3 and its reductase domain. Rate constants were measured under saturating conditions, as described in Materials and Methods. All values

0.401

,

.

,

.

.

,

,

'

'

1

I

OX

were determined at 25 "C, except where indicated. For flavin reduction two values represent the fast and slow phases, respectively, from biphasic traces. Reaction

Rate constant for P-450 BM3

P-450 domain

Reductase domain

>800

-

s ' -~

~

Fatty acid binding >SO0 Flavin reduction 758 4 ? 5.9 and 117.6 t 2.4 Semiquinone 620.0 ? 8.3 formation

~

-

770.5 1 6.7 and 235.0 i 4.0 452.2 t 7.5 (25°C) 226.9 t 6.4 (5" C)

119.1 i- 0.7

(2°C) Flavin-to-haem electron transfer

223.4 i- 4.1 (myristate) 130.0 ? 2.4 (Inurate)

-

located at approximately the peak of the longer wavelength absorption band for the f-lavins in the enzymes and the latter at a wavelength at which there was no longer a contribution from the haem Soret band to the flavin absorbance in P-450 BM3 holoenzyme. Rate constants measured were essentially identical at the two wavelengths for the reduction of both P-450 BM3 holoenzyme and its reductase domain, using NADPH as the reductant. In the presence of excess NADPH, a biphasic curve was described at 25°C with a fast phase in excess of 700 s-' and a slow phase of approximately 130 s-' (Table 2). At lower temperatures the velocity of the fast phase can be decreased, to approximately 500 s-' at 5°C and to less than 450 s-' at 2°C. The rate constants are vastly in excess of that for steady-state P-450 BM3 fatty acid monooxygenase activity, or, indeed, for the reduction of cytochrome c and other artificial electron acceptors (which depend only on electron transfer through the flavins of the reductase domain). The rate constants measured reflect NADPH binding as well as hydride transfer to FAD and electron transfer to FMN.

Flavin semiquinone formation during reduction of flavins in P-450 BM3 and its reductase domain. At sufficiently high concentrations of reductase domain, the addition of excess NADPH results in the development of a clearly visible blue colour in the enzyme, strongly indicating the presence of a neutral semiquinone species [34]. Examination of the visible spectrum of oxidised and NADPH-reduced BM3 reductase domain reveals not only large decreases in absorbance in the 380-550-nm rcgion as the flavins are reduced, but also a smaller increase in absorbance in the 560-630-nm region (peak absorbance at approximately 600 nm) which would correspond to a blue semiqiiinone species (Fig. 2). While addition of a powerful reductant such as sodium dithionite (E" = -527 mV) faciltates the complete four electron reduction of BM3 reductase, this evidently does not occur with the weaker physiological reductant NADPH (E" = -324 mV). EPR studies confirm the presence of the semiquinone in NADPH-reduced freeze-quenched samples of P450 BM3 and BM3 reductase domain (Fig. 3). These enzyme

1 396.1

500.0

600.0

646.9

Wavelength (nm.)

Fig. 2. Formation and decay of the blue flavin semiquinone species during reduction and oxidation of P-450 BM3 reductase domain. Oxidised (OX) P-450 BM3 reductase (4.5 pM) in 20 mM Mops containing 100 mM KCI was reduced (RED) with an excess (200 pM) NADPH

and allowed to reoxidise aerobically. Absorption spectra were recorded at the time intervals indicated in minutes. The semiquinone species is evident immediately upon reduction (600-nm band - SQ) but disappears at an early stage in the reoxidation. The semiquinone disappears by the first (10 min) time point, at which stage the major flavin absorbance band (at 459 nm) is less than 30% recovered. Further regeneration of the major absorption band is seen over subsequent time intervals as the reductase slowly uses aerobic oxygen as an electron acceptor. The spectrum becomes virtually indistinguishable from that of the original oxidised form after approximately 2.5 h.

samples catalysed virtually identical rates of cytochrome c reduction (indicating near-identical FMN content) and were both estimated at 15 pM. Quantitation of the EPR signals showed that flavin semiquinone at approximately 15 pM was present in both the reduced samples, indicating that only one of the two flavins was i n a semiquinone form, the other being fully reduced. By analogy with eukaryotic cytochrome P-450 reductase, the semiquinone is likely to reside on the FMN [35-371. The seiniquinone signal was absent from P-450 BM3 samples frozen during active oxygenation of arachidonate or laurate. When the NADPH-reduced reductase domain is allowed to oxidise in aerobic conditions, the semiquinone absorbance band disappears within 2-3 min of the complete oxidation of NADPH, while the full oxidation of the flavins requires a much longer period. This indicates that the first electron is lost from the semiquinone species during aerobic oxidation (Fig. 2). Under anaerobic conditions (following extensive flushing of a sealed cuvette with oxygen-free nitrogen) the semiquinone absorbance is stable for more than 1 h. Under aerobic conditions, the semiquinone appears less stable in P-450 BM3 holoenzyme than in the reduc-

Munro er al. (ELMJ. Biochem. 239)

--

0.33

0.35

0.35

0.33

Magnetic Field Strength (T)

Fig. 3. Detection of flavin semiquinone in NADPH-reduced P-450 BM3 and BM3 reductase by EPR. Cytochrome P-450 BM3 and BM3 reductase domain (30pM) were reduced with an excess (150pM) of NADPH and frozen after 30 s, as described in Materials and Methods. EPR measurements were performed as described in Materials and Methods. Signals indicating neutral (blue) semiquinone species are evident in both samples. Quantitation of these signals indicates that there is approximately 15 pM flavin semiquinone present in both samples; i.e. that only one of the two flavins is in the semiquinone form. Integrals were calculated for the FMN semiquinone at -184°C relative to a 1 mM Cu (11) standard.

150

-/

50

0

v

1Ff 0

0 50

100

150

200

250

[P450-BM3] (pM)

Fig. 4. Michaelis-Menten curve for the reduction of cytochrome c by P-450 BM3 holoenzyme. The rates of reduction of cytochrome c (1 pM) by P-450 BM3 were measured at 550 nm at 25°C in 20 mM Mops, pH 7.4, containing 100 mM KCI, using stopped-flow. The concentration of enzyme was varied from 20 pM to 200 pM and NADPH was present at a final concentration of 400 pM. First-order rate constants were derived from the averages of numerous traces at the different enzyme concentrations.

tase domain, possibly due to slow leakage of electrons to the haem. Rate constants for semiquinone formation in BM3 reductase are shown in Table 2. The rate constants fell with decreasing temperature, but all fitted well to single exponential curves. It is possible that the rate of formation of the semiquinone species corresponds to the second phase seen in the flavin reductions described above. As expected, no semiquinone formation was detected during the reduction of FMN-deficient P-450 BM3 mutants G570D and W574D [27], where the complete reduction of FAD is achieved by hydride transfer from NADPH and there is no rapid exit route for the electrons on FAD.

Stopped-flow measurement of electron transfer to cytochrome c. The transfer of electrons to cytochrome c through the

407

reductase domain of cytochrome P-450 BM3 is monitored by the increase in absorbance at 550 nm. The rate of cytochrome c reduction was measured under pseudo-first-order conditions using 1 pM cytochrome c and various P-450 BM3 concentrations (20-200 pM). Traces fitted best to single exponential curves. At 25"C, the rate of cytochrome c reduction was seen to depend greatly on thc concentration of the electron donor, and to vary from approximately 31 s-' at 20 pM P-450BM3 up to approximately 110 s-' at 200 pM P-450 BM3 (Table 2). Similar results were obtained using BM3 reductase domain. Using stopped-flow measurements, the K,, for cytochrome c was estimated as 100t 3 pM and the rate constant for electron transfer, kec,was 187 5 3 s-' (Fig. 4). This rate is higher than that obtained from steady-state analysis (Table 1) and may reflect that transfer of the first electron from reduced P-450 BM3 to cytochrome c occurs faster than subsequent transfers. The data indicate that electron transfer to cytochrome c is the rate-limiting step in its reduction by P-450 BM3.

Stopped-flow measurement of electron transfer between the flavin and haem domains of P-450 BM3 holoenzyme. A lower limit for the rate constants of transfer of the first electron between the haem and flavin domains of cytochrome P-450 BM3 is most easily obtained by the measurement of the rate of accumulation of the ferrous iron-carbon monoxide adduct (absorbing at approximately 450 nm) which provides the diagnostic recognition assay for this class of haemoprotein. Previous studies with P-4.50 BM3 indicate that the association of carbon monoxide with this enzyme is extremely rapid (k,,,, of approximately 400X10" M-' s-l at 20°C for the palmitate-bound enzyme), although the binding of of the natural ligand, dioxygen, may occur considerably faster [38]. Mixing of CO-saturated solutions of fatty acid-bound P-450 BM3 (2 pM) with CO-saturated solutions of NADPH provided monophasic curves of absorbance increase at 450 nm. The rate of P-450 formation was rapid, but showed dependence on the fatty acid used, being markedly faster for rnyristic acid than for the poorer substrate laurate (Table 2). Mixing of CO-saturated NADPH-reduced P-450 BM3 holoenzyme in one syringe with CO-saturated fatty acid solution in the other gave similar data, but rates were seen to decrease significantly as the concentration of NADPH was increased above approximately 200 pM, indicating that the presence of excess reductant may inhibit the reactivity of P-450 BM3 by influencing the rate of flavin to haem electron transfer.

Difference spectrum of P-450 BM3 during active turnover of fatty acids. The difference spectrum of cytochrome P-450 BM3 was recorded during steady-state turnover of lauric and arachidonic acids. Excess substrates were incubated with P-450 BM3 and difference spectra in the visible region were recorded at 20°C during the steady-state monooxygenation of both laurate and arachidonate. A deep trough at approximately 419 nm and a peak at 390 nm are the major features seen in these difference spectra. These are typical of a high-spin ferric haem iron and indicate that a predominant species present during steady-state has a fatty acid molecule attached at the haem iron active site.

DISCUSSION Dissection of the catalytic cycle of cytochrome P-450 BM3 reveals that the measurable biphasic rates of flavin reduction occur too fast to represent rate-limiting steps in the processes of fatty acid monooxygenation or reduction of artificial electron acceptors, such as cytochrome c. This is in agreement with the results of Klein and Fulco [39] who found that cytochrome c

408

Munro et al. (Eur: J . Biochem. 239)

reduction could take place simultaneously with fatty acid hydroxylation catalysed by P-450 BM3, without effect on the rate of the latter process. The formation of a semiquinone species occurs during the NADPH-dependent reduction of P-450 BM3 holoenzyme and of its reductase domain, and the rate of formation of this species at 600 nm may correspond to the second phase of flavin reduction seen at 459 nm or 475 nm. Freezequench EPR studies under steady-state conditions show that the semiquinone does not accumulate during acive fatty acid oxygenation or cytochrome c reduction by P-450 BM3, indicating that electron transfer through this species is rapid. A large number of intermediates are known to exist in the class I1 P-450 electron-transfer chain (NADPH cytochrome P450 reductase and P-450 haemoprotein) for which cytochrome P-450 BM3 represents an experimentally tractable model. Eukaryotic P-450 systems are usually composed of one or more membranous components and generally have catalytic rates 100fold (or more) slower than that of P-450 BM3. In many of these forms, the rate-limiting step is the productive collision of individual enzymes to form an enzyrnically active heterodimer. The individually expressed domains of cytochrome P-450 BM3 interact rather poorly by comparison with the rapid electrontransfer process catalysed by the holoenzyme, and it is also the case here that molecular dissection of P-450 BM3 changes the rate-limiting step to one of domain interaction. However, it is unlikely that the same step is rate-limiting in the catalytic process of the holoenzyme. Since the P-450 and reductase domains are fused together in close proximity; it may well be the case that they no longer retain high affinity for each other following physical separation [40]. That cytochrome c reduction occurs simultaneously with fatty acid hydroxylation and without effect on the the rate of the latter [391 is strongly suggestive that the FMN-dependent reduction of the exogenous cytochrome occurs markedly faster than the rate-limiting step in the normal inonooxygenase reaction cycle of P-450 BM3 and, hence, that steps preceding reduction of FMN are not those which limit the fatty acid hydroxylation reaction. Thus, we expect that NADPH binding and electron transfer to FAD, and the one-electron transfer from FAD to FMN to occur at least as fast as the rate of reduction of cytochrome c (up to approximately 190 s-' at 25°C). Stopped-flow analysis proves that flavin reduction in P-450 BM3 is an extremely rapid process which cannot limit the catalytic cycle. The reduction of FAD in FMN-deficient mutant P-450 BM3 eiizyines G570D and W574D is also extremely rapid (data not shown). These mutants cannot reduce cytochrome c to any appreciable extent, but they retain the ability to reduce other electron acceptors (such as potassium ferricyanide and methyl viologen) at levels up to approximately 35% of that of the wildtype enzyme 1271. These data support the hypothesis that the direction of electron flow in cytochrome P-450 BM3 is from NADPH --+ FAD FMN -+ haem, as previously shown for mammalian P-450 reductase 1361. Our data are in accordance with those of Klein and Fulco [39] who provided steady-state kinetic evidence that electron transfer through flavins was too rapid to limit fatty acid hydroxylation. The semiquinone form is clearly visualised in the reductase domain of cytochrome P-450 BM3 - the enzyme turning from yellow to blue when reduced by an excess of NADPH. No such peak was visible during the reduction of FMN-deficient mutants G570D and W574D, indicating that the FMN provides the coloured semiquinone seen. EPR studies show that only one flavin (the FMN) in in the semiquinone state. For P-450 BM3 in the presence of fatty acid and NADPH, there is a rapid formation of semiquinone inside the first few milliseconds, but this disappears rapidly and does not return until the completion of fatty

-

acid monooxygenation (providing NADPH is in excess). It may be the case that the FMN semiquinone is the species responsible for electron transfer to haem iron and that the electron-donation step occurs too rapidly for the species to accumulate during steady-state fatty acid hydroxylation. Many (soluble) microbial cytochrome P-450 systems have catalytic rates many fold higher than those of their eukaryotic counterparts. It is interesting that the catalytic rate of P-450 BM3 is very similar to that of the P-450cam system, perhaps suggesting that there may be a common rate-limiting step in the reactions of these three-component and one-component systems. However, while it is generally accepted that an electron-donation step is used to control the catalytic rate of P-450cam [IS], it is obvious that the first electron transfer to P-450 BM3 haem occurs more rapidly than the steady-state rate. The first electron transfer to the haem iron appears to be dependent on the fatty acid used and this may be a major feature explaining the 3-4fold difference in the turnover rates of various fatty acid substrates. With myristic acid, a transfer rate of approximately 200 s ' was measured, similar to that recently reported [41], although the rate was rather lower with the poorer substrate laurate. Given that the overall rate of catalysis is governed by the sum of the inverse rates for each of the individual processes involved, it would appear that first electron transfer must play some role in controlling turnover. Since two successive electron transfers to haem are required for fatty acid oxygenation, the influence of flavin to haem electron transfer on rate-control becomes more apparent. We are currently investigating the effect of high concentration of NADPH in controlling flavin to haem electron transfer. Likewise, while the initial reduction of FAD is extremely rapid, the slower second phase of flavin reduction may also exert some control. The formation of transient species post-binding of oxygen to ferrous P-450 are likely to be too rapid to limit the reaction. Similarly, the rate constants for the binding of both oxygen and carbon monoxide are very high and should not limit the rates determined to any significant extent. However, the accumulation of a high-spin ferric species in the difference spectrum of active P-450 BM3 suggests that release of hydroxylated fatty acid product may be a factor in rate control. The presence of the absorbance difference typical of the the ferric species suggests that product release may occur somewhat more slowly than the transfer of the second electron to the haem. On the basis of studies in other systems, the oxyferrous P-450 would be expected to exhibit absorbance characteristics rather different from the ferric form and from those observed in the difference spectrum [22]. Since there is little uncoupling of NADPH oxidation to lauric acid hydroxylation by P-450 BM3, it is unlikely that auto-oxidation of the oxyferrous form has occurred. Thus, the species observed during steady-state turnover is likely to be ferric haem with product (hydroxylated lauric acid) bound. This is not an altogether surprising finding, given that certain substrates can undergo multiple rounds of hydroxylation by P-450 BM3. This suggests that products retain affinity for the active site and can also diminish the catalytic rate. A similar conclusion regarding influence of product release on catalytic activity was arrived at by Sakaki et al. [22] in their study on the properties of a rat P-450 1Allyeast NADPH-cytochrome P-450 reductase fusion protein. The conclusion from these studies on the catalytic cycle of P-450 BM3 is that the enzyme appears to be well evolved, with control over catalytic rate exerted by a number of distinct steps in the catalytic cycle. The authors wish to thank the Biotechnology and Biological Sciences Research Council (UK) for financial support for these studies. A.

Munro et al. (Eu,: J. Biochem. 239) W. Munro is a Royal Society of Edinburgh Caledonian Research Fellow. Thanks are also due to Professor Armand Fulco and Dr Michael Klein of the Department of Biological Chemistry and Laboratory of Structural Biology, University of California, Los Angeles, for provision of the clones encoding FMN-deficient mutants of P-450 BM3.

REFERENCES 1. Nelson, D. R., Kamataki, T., Waxman, D., Guengerich, F. P., Estabrook, R. W., Feyereisen, R., Gonzalez, F. J., Coon, M. J., Gunsalus, I. C., Gotoh, O., Okuda, K. & Nebert, D. W. (1993) DNA Cell Biol. 12, 1-51. 2. Guengerich, F. P. (1991) J. Biol. Chem. 266, 10019-10022. 3. Asperger, 0. & Kleber, H.-P. (1991) in Microbial and plant cytochromes P-450: biochemical characteristics, genetic engineering andpractical implications (Ruckpaul, K. & Rein, H., eds) pp. 153, Taylor and Francis, London. 4. Degtyarenko, K. N. & Archakov, A. 1. (1993) FEBS Lett. 332, 1-8. 5. Poulos, T. L., Finzel, B. C. & Howard, A. J. (1987) J. Mol. Bid. 195, 687-700. 6. Narhi, L. 0. & Fulco, A. J. (1986) J. Biol. Chem. 261, 7160-7169. 7. Narhi, L. 0. & Fulco, A. J. (1987) J . Bid. Chem. 262, 6683-6690. 8. Ruettinger, R. T., Wen, L.-P. & Fulco, A. J. (1989) J. Bid. Chem. 264, 10987- 10995. 9. Li, H., Darwish, K. & Poulos, T. L. (1991) J. Biol. Chem. 266, 11909- 119 14. 10. Miles, J. S., Munro, A. W., Rospendowski, B. N., Smith, W. E., McKnight, J. & Thomson, A. J. (1992) Biochem. J. 288, 503509. 11. Ravichandran, K. G., Boddupalli, S. S., Hasemann, C. A., Peterson, J. A. & Deisenhofer, J. (1993) Science 261, 731-736. 12. Remmer, H., Schenkman, R. W., Estabrook, H., Sasame, H., Gillette, J. R., Narasimhulu, D., Cooper, Y. & Rosenthal, 0. (1966) Mol. Pharmacol. 2, 187-190. 13. Schenkman, J. B., Remmer, H. & Estabrook, R. W. (1967) Mol. Pharmacol. 3, 113-123. 14. Sligar, S. G. (1976) Biochemistrjj 15, 5399-5406. 15. Mueller, E. J., Loida, P. J. & Sligar, S. G. (1995) Cytochrome P450. Structure, mechanism and biochemistry, 2nd edn, (Ortiz de Montellano, P. R., ed.) pp. 83-124, Plenum Press, New York. 16. Brewer, C. B. & Peterson, J. A. (1988) J. Biol. Chem. 263, 791798. 17. Pederson, T. C., Austin, R. H. & Gunsalus, I. C. (1977) in Micros o nm and drug oxidations (Ullrich, V., ed.) pp. 275-283, Pergamoii Press, New York.

409

18. Martinis, S. A,, Ropp, J. D., Sligar, S. G. & Gunsalus, 1. C. (1991) Microbial and plant cytochromes P-450: biochemiccil chnracteristics, genetic engineering and practical implications (Ruckpaul, K. & Rein, H., eds) pp. 54-86, Taylor and Francis, London. 19. Peterson, J. A. & Prough, R. A. (1986) Cytochrome P-450. Structure, mechanism and hiochemisfry (Ortiz de Montellano, P. R., ed.) pp. 89-117, Plenum Press, New York. 20. Murakami, H., Yabusaki, Y., Sakaki, T., Shibata, M. & Ohkawa, H. (1987) DNA (NY) 6, 189-197. 22. Fisher, C. W., Shet, M. S., Caudle, D. L., Martin-Wixtrom, C. A. & Estabrook, R. W. (1992) Proc. Nut1 Acad. Sci. USA 89, 1081710821. 22. Sakaki, T., Kominami, S., Takemori, S., Ohkawa, H., Akiyoshi-Shibata, M. & Yabusaki, Y. (1994) Biochemistry 33, 4933-4939. 23. Gibson, T. J. (1984) PhD thesis, University of Cambridge, England. 24. Bullock, W. O., Fernandez, J. M. & Short, J. M. (1987) Biotechniques 5, 376-379. 25. Vieira, J. & Messing, J. (1987) Methods Enzymol. 153, 3-1 1. 26. Marsh, P. (1986) Nucleic Acids Res. 14, 3603. 27. Klein, M. L. & Fulco, A. J. (1993) J . Bid. Chem. 268, 7553-7562. 28. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular cloning: a laboratory manual, 2nd edn, Cold Spring Harbor Laboratory Press, Cold Spring Harbor NY. 29. Tartof, K. D. & Hobbs, C. A. (1987) Bethesda Res. Lab. Focus 9, 12. 30. Bradford, M. M. (1976) An d . Biochem. 72, 248-254. 31. Smith, P. K., Krohn, R. I., Hermanson, G. T., Mallia, A. K., Gartner, F. H., Provenzano, M. D., Fujimoto, E. K., Goeke, N. M., Olson, B. J. & Klenk, D. C. (1985) Anal. Biochem. 150, 76-85. 32. Omura, T. & Sato, R. (1964) J. Bid. Chem. 239, 2379-2385. 33. Okita, R. T., Clark, J. E., Rice Okita, J. & Masters, B. S. S. (1991) Methods Enzymol. 206, 432-441. 34. Muller, F., Hemmerich, P., Ehrenberg, A,, Palmer, V. & Massey, V. (1970) Eu,: J. Biochem. 14, 185-190. 35. Kurzban, G. P., Howarth, J., Palmer, G. & Strobel, H. W. (1990) J. Biol. Chem. 265, 12272-12279. 36. Verinillion, J. L., Ballou, D. P., Massey, V. & Coon, M. J. (1981) J . Biol. Chem. 256, 266-277. 37. Otvos, J. D., Krum, D. P. & Masters, B. S. S. (1986) Biochemistry 25, 7220-7228. 38. Sevrioukova, I. F. & Peterson, J. A. (1995) Arch. Biochem. Biophys. 317, 397-404. 39. Klein, M. L. & Fulco, A. J. (1994) Biochim. Biophys. Acta 1201, 245-250. 40. Munro, A. W., Lindsay, J. G., Coggins, J. R., Kelly, S. M. & Price, N. C. (1994) FEBS Lett. 343, 70-74. 41. Govindaraj, S. & Poulos, T. L. (1995) Biochemistry 34, 11 221 11 226.