Direct Observation of a Novel Perturbed Oxyferrous Catalytic Intermediate during Reduced Putidaredoxin-initiated Turnover of Cytochrome P-450-CAM

The single turnover of (1R)(+)-camphor-bound oxyferrous cytochrome P450-CAM with one equivalent of dithionite-reduced putidaredoxin (Pdx) was monitored for the appearance of transient intermediates at 3 °C by double mixing rapid scanning stopped-flow spectroscopy. With excess camphor, three successive species were observed after generating oxyferrous P450-CAM and reacting versus reduced Pdx: a perturbed oxyferrous derivative, a species that was a mixture of high and low spin Fe(III), and high spin ferric camphor-bound enzyme. The rates of the first two steps, ∼140 and ∼85 s-1, were assigned to formation of the perturbed oxyferrous intermediate and to electron transfer from reduced Pdx, respectively. In the presence of stoichiometric substrate, three phases with similar rates were seen even though the final state is low spin ferric P450-CAM. This is consistent with substrate being hydroxylated during the reaction. The single turnover reaction initiated by adding dioxygen to a preformed reduced P450-CAM·Pdx complex with excess camphor also led to phases with similar rates. It is proposed that formation of the perturbed oxyferrous intermediate reflects alteration of H-bonding to the proximal Cys, increasing the reduction potential of the oxyferrous state and triggering electron transfer from reduced Pdx. This species may be a direct spectral signature of the effector role of Pdx on P450-CAM reactivity (i.e. during catalysis). The substrate-free oxyferrous enzyme also reacted readily with reduced Pdx, showing that the inability of substrate-free P450-CAM to accept electrons from reduced Pdx and function as an NADH oxidase is completely due to the incapacity of reduced Pdx to deliver the first but not the second electron.

The cytochrome P-450 family of heme-containing mono-oxygenases is involved in the metabolism of xenobiotics and in the production of physiologically important molecules (1). A defining characteristic of the P-450 family is the proximal thiolate-ligated heme. P450-CAM (P450-CAM, CYP101) 3 from Pseudomonas putida catalyzes the hydroxylation of (1R)(ϩ)-camphor to form (1R)(ϩ)-5-exo-hydroxycamphor (Reaction 1). The electrons required for the reaction flow from NADH to the flavoprotein, putidaredoxin reductase, then to the iron-sulfur (2Fe/2S) protein, putidaredoxin (Pdx), and finally to P450-CAM. In the widely quoted P450 reaction cycle shown in Fig. 1 (1), substrate binding converts the ferric low spin resting state (1) to the ferric high spin form (2). Electron transfer from reduced Pdx gives high spin deoxyferrous P-450 (3). Dioxygen binding yields the oxy-ferrous adduct, a ferrous-O 2 /ferric superoxide resonance hybrid (4a 7 4b). CO binding to 3 generates the ferrous-CO derivative (5). Addition of the second electron from reduced Pdx has been proposed to yield a ferric peroxo species (6a), protonation of which gives the hydroperoxo state (6b). Protonation of the distal oxygen produces water and compound I, a ferryl porphyrin -cation radical (7). This highly oxidizing intermediate abstracts a hydrogen atom from the substrate onto the ferryl oxygen to produce a substrate radical. Radical recombination yields the oxygenated product, and the addition of H 2 O reforms 1.
Intermediates 6 and 7 are apparently too reactive to build up during turnover, and thus, have never been seen during catalysis. Nonetheless, 6a and 6b have been detected by Davydov et al. (2,3) in low temperature EPR experiments in which oxyferrous P450-CAM (4b) is reduced by hydrated electrons generated by ␥-radiation. Egawa et al. (4) have detected P-450 compound I (7) by reacting substrate-free ferric P450-CAM (1) with m-chloroperoxybenzoic acid to yield a species that is spectrally similar to the compound I of another thiolate-ligated heme enzyme, Caldariomyces fumago chloroperoxidase (5). Sligar and coworkers (6) have reported similar results with thermostable Cyp119. Shünemann et al. (7,8) have observed a ferryl heme species with substrate-free P450-CAM in reactions with peroxyacetic acid using rapid freeze quenching techniques, but instead of a classical P450-CAM compound I, they reported formation of compound II plus a tyrosine radical, analogous to the compound ES species observed when cytochrome c peroxidase reacts with hydrogen peroxide (1). Spolitak et al. (9) have shown that this compound ES-like species evolves from compound I in an acid-dependent step, presumably by abstracting an electron (or hydrogen atom) from a nearby tyrosine residue. The cryogenic crystal structure (10) of oxyferrous P450-CAM exposed to x-rays under conditions known to produce hydrated electrons may possibly be that of P-450 compound I (7), although other species such as 1 cannot be entirely ruled out.
Pdx not only delivers electrons to P450-CAM, it is also an effector of catalysis (11)(12)(13)(14). In this regard, Pochapsky et al. (15) have proposed that Pdx helps prevent uncoupling of electron transfer from product formation. It has been shown that in catalysis, putidaredoxin reductase does not form a ternary complex (putidaredoxin reductase/Pdx/P450-CAM) (16). Therefore, single turnover experiments to probe the oxygenation mechanism using Pdx and P450-CAM in the absence of putidaredoxin reductase are likely to be relevant to the natural catalytic events. In resonance Raman (17,18), IR (19), NMR (20,21), and EPR studies (11,22), Pdx binding has been shown to cause protein conformational changes that distort the P450-CAM heme structure and, especially, the Fe-S bond. In particular, Tosha et al. (20) proposed that Pdx binding leads to structural changes that facilitate oxygen activation, and Unno et al. (18) suggested that the structural changes may correlate with H-bonding to the proximal Cys. Similarly, Nagano et al. (19) proposed that Pdx binding promotes electron donation from the proximal Cys to the iron-bound O 2 to facilitate O-O bond cleavage. The role of hydrogen bonding in stabilizing heme-thiolate coordination in P450-CAM has been examined by Yoshioka et al. (23) and in a thiolate-ligated model system by Suzuki et al. (24).
In this report, we have used double mixing rapid scanning stoppedflow spectroscopy to monitor the reaction of oxyferrous P450-CAM with reduced Pdx in the presence or absence of substrate. The effect of oxidized Pdx on the stability of oxyferrous P450-CAM has also been studied. This study builds upon the highly demanding single mixing/ single wavelength kinetics study of Brewer and Peterson (25), as well as the earlier work of Pederson et al. (26) and Hui Bon Hoa et al. (27), and provides considerable additional detail about the reaction between oxyferrous P-450 and reduced Pdx, including the effector roles of Pdx with P450-CAM. Two recent studies provide useful background for the reactions of reduced P-450 with dioxygen. Tosha et al. (28) recently investigated the roles of amino acids at the putative Pdx⅐P450-CAM interface, building on previous work by Unno et al. (29). Zhang et al. (30) have studied the analogous reduction of oxyferrous mammalian P-450 -2B4 by its flavoprotein P-450 reductase.

EXPERIMENTAL PROCEDURES
Materials and Sample Preparation-(1R)(ϩ)-Camphor was purchased from Sigma, sodium dithionite was from Aldrich, and buffer chemicals were from Fisher. 5-exo-Hydroxy-(1R)-camphor was synthesized from (1R)-camphor following the procedure reported by Li et al. (31) and was purified by silica gel column chromatography. P450-CAM and Pdx were individually overexpressed in Escherichia coli and purified as reported (32). The ferric camphor-bound P450-CAM and oxidized Pdx concentrations were determined using ⑀ 391 ϭ 102 mM Ϫ1 cm Ϫ1 and ⑀ 455 ϭ 10.4 mM Ϫ1 cm Ϫ1 , respectively (33). The Pdx used had a ratio of A 325 /A 280 Ն 0.63 to establish purity.
For stopped-flow experiments, P450-CAM and Pdx in 50 mM potassium phosphate, pH 7.4, 100 mM KCl, with or without 500 M camphor, as required, were placed in tonometers either as separate solutions or mixed. Substrate-free P450-CAM was prepared by Sephadex G-25 chromatography in 50 mM MOPS, pH 7.0, to remove camphor, followed by a second G-25 column in 50 mM potassium phosphate, pH 7.4, 100 mM KCl (34); stoichiometric camphor-bound P450-CAM was prepared by passing the protein through the latter G-25 column and then titrating back with just enough camphor to form the high spin ferric state. Protein solutions were made anaerobic by multiple cycles of alternately evacuating and flushing with oxygen-free argon at 25°C over a period of 30 min. Reduced P450-CAM and Pdx were prepared by titration with microliter increments of a concentrated sodium dithionite stock solution until the ferric states were fully reduced as judged spectrally.
Stopped-flow Experiments-Rapid scan stopped-flow experiments were carried out at 3-4°C on a Hi-Tech Ltd. SF-61 DX2 instrument equipped for anaerobic work and with a Hi-Tech MG-6000 rapid scan diode array detector. Dead times were determined to be 1.5 ms. In single and double mixing experiments, numerous scans were recorded; the data displayed are representative. For double mixing experiments, a delay time of 100 ms was optimal for complete oxyferrous enzyme formation with essentially no autoxidation. The kinetic data were fit to various reaction models by the KinetAsyst program (Hi-Tech Ltd.) or by singular value decomposition and multiple component analysis to derive spectra of intermediates using the Specfit program from Spectrum Software Associates.
Product Binding-The K d value for binding 5-exo-hydroxy-camphor to ferric substrate-free P450-CAM was determined by Hill analysis of ligand titration data obtained by the addition of microliter aliquots of a concentrated stock solution of substrate to the substrate-free enzyme.

RESULTS AND DISCUSSION
Formation and Stability of Oxyferrous P450-CAM-The reaction of dioxygen at 4°C with camphor-bound reduced P450-CAM occurs with a rate constant of ϳ10 6 M Ϫ1 s Ϫ1 (data not shown). Thus, at a dioxygen concentration of 130 M (aerobic buffer prepared at 20°C (35) mixed with anaerobic buffer), the rate would be about 130 s Ϫ1 (t1 ⁄ 2 ϭ ϳ5 ms). The UV-visible absorption spectrum of oxyferrous P450-CAM ( Fig. 2) has Soret and visible region absorption peaks at 418 and 556 nm, respectively (419 and 554 nm at Ϫ30°C in 65% ethylene glycol/phosphate buffer, pH 7.4) (36). After formation of the oxyferrous species, essentially no spectral changes were observed for up to 60 s, as expected from the reported t1 ⁄ 2 for autoxidation of 25 min at this temperature (11). Substrate-free oxyferrous P450-CAM was also rapidly produced in a similar manner and likewise had UV-visible absorption peaks at 418 and 556 nm (data not shown). The t1 ⁄ 2 for autoxidation of substrate-free oxyferrous P450-CAM was determined to be 90 s at 4°C, pH 7.4. In the double mixing experiments, the oxyferrous P450-CAM was allowed to form for ϳ100 ms and then was mixed with the next reagent. Consequently, more than 99% of the P450-CAM was in the oxyferrous form at the time of the second mix.
The Reaction of Camphor-bound Oxyferrous P450-CAM with Oxidized Pdx-In principle, oxidized Pdx should be incapable of reacting with oxyferrous cytochrome P-450 because, being in the oxidized state, it cannot provide an electron to reduce the oxyferrous enzyme. However, Lipscomb et al. (11) demonstrated nearly 30 years ago that oxidized Pdx could mediate the formation of up to one-half an equivalent of product from substrate-bound oxyferrous P450-CAM. The mechanism of this process has yet to be established. To quantify the kinetics of this effect, we examined the reaction of a 2-fold excess of oxidized Pdx with oxyferrous P450-CAM, first in the presence of a large excess of camphor and then in the presence of only a 2-fold excess of substrate.
Dithionite-reduced ferrous P450-CAM in the presence of ϳ100-fold excess of camphor was mixed, initially against aerobic buffer to create oxyferrous P450-CAM and then against a 2-fold excess of oxidized Pdx ( Fig. 2A). The first trace is the spectrum of camphor-bound oxyferrous P450-CAM plus that of oxidized Pdx. Over the next 60 s, spectral changes yielded the spectrum of camphor-bound high spin ferric P450-CAM with peaks at 392, 539, and 561 nm plus minor contributions from Pdx. The t1 ⁄ 2 for this reaction is 10 s, 150-fold less than that for the formation of ferric camphor-bound P450-CAM from the oxyferrous enzyme (t1 ⁄ 2 ϭ 25 min) in the absence of oxidized Pdx.
The same experiment was repeated with oxyferrous P450-CAM prepared in the presence of only a 2-fold excess of substrate (Fig. 2B). The first spectral scan again shows the spectrum of oxyferrous P450-CAM. The last trace is the spectrum of a mixture of high and low spin state REACTION 1 Perturbed Oxyferrous P-450 Catalytic Intermediate DECEMBER 23, 2005 • VOLUME 280 • NUMBER 51 forms of ferric P450-CAM caused by the presence of the slight excess of camphor, plus contributions from Pdx. Normally, at these concentrations, one would expect that P450-CAM would be fully saturated with camphor and be high spin. However, the presence of Pdx weakens the binding of camphor (11). The t1 ⁄ 2 for breakdown of oxyferrous P450-CAM under these conditions was estimated to be 12 s. These two experiments ( Fig. 2) reveal that the rate of conversion of oxyferrous to ferric P450-CAM is Ͼ100-fold enhanced (smaller t1 ⁄ 2 ) in the presence of oxidized Pdx. The rate enhancement is essentially the same whether the experiment is done in the presence of a large or small excess of camphor. This quantifies the dramatic extent to which oxidized Pdx decreases the t1 ⁄ 2 for formation of the ferric enzyme that Lipscomb et al. (11) had qualitatively reported.
The Reaction of Camphor-bound Oxyferrous P450-CAM with Reduced Pdx in the Presence of Excess Camphor-This double mixing experiment involves the formation of camphor-bound oxyferrous P450-CAM in the first mixing step in which reduced P450-CAM is mixed with oxygenated buffer (Fig. 1, 3 3 4). After reacting for 100 ms, the second mixing step in the stopped-flow instrument introduces dithionite-reduced Pdx to initiate several steps converting 4 to 2 (Fig. 1). The camphor concentration (500 M) was in large excess of the P450-CAM concentration (10 M). Fig. 3 shows spectra recorded during the first 1000 ms of the reaction. The inset is an expanded view of the Soret absorption region. The first scan at 1.5 ms after the second mix is identified principally as the UV-visible absorption spectrum of oxyferrous P450-CAM, with a Soret absorption peak at 418 nm and a peak in the visible region at 554 nm. An initial increase in the intensity and a slight blue shift of the Soret absorption peak of the oxyferrous enzyme from 418 to ϳ414 nm ensues during the first 17 ms, with an isosbestic point at 421 nm, and this is followed over the next 41 ms by a decrease in intensity and a further blue shift in peak position to a broad "double hump." This process occurs with an isosbestic point at 405 nm. As will be discussed below, the rate of the second phase correlates with the rate reported by Brewer and Peterson (25) for the electron transfer from reduced Pdx to oxyferrous P450-CAM coupled with the formation of product. From this, we conclude that the protein is in the ferric state by the end of phase 2, and we thereby assign the broad double hump spectrum to a mixture of low spin and high spin ferric P450-CAM species. The third and slowest (k Ͻ 5 s Ϫ1 ) phase seen in Fig. 3 occurs between 61 and 1000 ms. In this step, the Soret band is evolving with an isosbestic point at 408 nm into what is clearly high spin camphor-bound ferric P450-CAM. Singular value decomposition and global analysis of the data in Fig. 3 are shown in Fig. 4, which illustrates a fully resolved quite intense perturbed Soret peak at 413.5 nm (species B). The single peak in the visible region has decreased slightly in intensity and shifted from 554 to ϳ548 nm. We refer to this species evolving in phase I as perturbed oxyferrous P450-CAM. This intermediate converts to C, the double-humped spectrum, which slowly converts to D, high spin camphor-bound ferric P450-CAM (Fig. 4).
Because the perturbed oxyferrous enzyme Soret absorption peak is at nearly the same wavelength as that of ferric P450-CAM oxygen-donor ligand complexes (37), we were initially concerned that the species at the end of the first phase might in fact be the product complex formed by binding of the alcohol oxygen of 5-exo-hydroxycamphor to ferric P450-CAM. However, as seen in the inset to Fig. 4, although the spectrum of 5-exo-hydroxy-camphor-bound ferric P450-CAM also has a Soret absorption peak at 413 nm, it has two distinct peaks in the visible region at 566 and 534 nm rather than the single peak seen for the perturbed oxyferrous species. Thus, despite the near coincidence in the position of their Soret absorption peaks, the overall absorption spectrum of 5-exo-hydroxycamphor-bound ferric P450-CAM is quite different from that of the perturbed oxyferrous P450-CAM⅐Pdx. These results demonstrate that the species formed at the end of phase 1 is not the product complex.
To obtain more accurate and resolved kinetic data to use in determining rate constants for the first two phases, the experiment was repeated using single wavelength detection. Fig. 5 shows the spectral changes that occur at 418 nm for the first 100 ms. Data from this time period covers the first two phases of the reaction and could be fitted with a sum of two exponential functions. The first phase is characterized by a rate of ϳ140 Ϯ 14 s Ϫ1 , and the second phase is characterized by a rate of ϳ85 Ϯ 8 s Ϫ1 . Essentially identical rate constants were obtained at several different wavelengths (data not shown). The first rate corresponds to the conversion of oxyferrous P450-CAM to the perturbed oxyferrous intermediate.
To assign the second rate constant, we turn to the work of Brewer and Peterson (25), who examined the reaction of reduced Pdx with oxyferrous P450-CAM under similar conditions to those described herein, but they used single mixing/single wavelength (420 nm) stopped-flow absorption spectroscopy, because of the limitations of the instrumentation available at that time. Only a single reaction phase with a rate of 60 -100 s Ϫ1 was observed (from our data this would be expected, because the conversion of oxyferrous to the perturbed oxyferrous is essentially isosbestic at 420 nm). They assigned the rate to the electron transfer from reduced Pdx to oxyferrous P450-CAM on the basis of their careful parallel measurements of both the rate of electron transfer from reduced Pdx observed by freeze-quench EPR spectroscopy and the rate of product (5-exo-hydroxycamphor) formation, determined from chemical quench experiments. We therefore conclude that the rate of the second phase observed in Figs. 3 and 5 (ϳ85 s Ϫ1 ) likely corresponds to the rate of electron input from reduced Pdx to the perturbed oxyfer-    DECEMBER 23, 2005 • VOLUME 280 • NUMBER 51 rous intermediate and the formation of product. Additional support for this conclusion comes from the appearance during this phase of absorbance at ϳ640 nm that is characteristic of high spin ferric P-450 (38).

Perturbed Oxyferrous P-450 Catalytic Intermediate
The Reaction of Oxyferrous P450-CAM with Reduced Pdx in the Presence of Stoichiometric Camphor-We repeated the experiment with oxyferrous P450-CAM prepared with essentially only one equivalent of camphor bound to verify that camphor was being hydroxylated, as well as to separate out steps involved in releasing product and rebinding camphor. Camphor binds to ferric P450-CAM very tightly (K d ϭ ϳ0.5 M). This makes it possible to prepare a fully substrate-bound enzyme sample with essentially stoichiometrically added substrate. Reaction of the oxyferrous state of such a sample with reduced Pdx depletes the single equivalent of camphor, so that at the end of the experiment, the low spin ferric state (substrate-free) has formed. Fig. 6 displays UVvisible absorbance spectra recorded during the reaction. The formation of low spin substrate-free ferric P450-CAM is confirmed by the appearance of distinct alpha and beta absorption bands in the visible region at 536 and 564 nm that are characteristic of low spin ferric P450-CAM (536 and 569 nm) (37). The lack of any high spin ferric enzyme in the final spectrum is consistent with a single turnover that is essentially 100% coupled, i.e. for every two electrons added, one product molecule is generated without formation of hydrogen peroxide.
Although the Soret absorption maxima for camphor-bound oxyferrous P450-CAM (418 nm) and substrate-free ferric low spin P450-CAM (417 nm) are at essentially the same wavelength, the distinctions between the kinetic phases are easily seen. Monitoring the reaction at 390 nm under single wavelength conditions allows the three phases to be followed (Fig. 6, inset). From the single wavelength data, rate constants for the first two phases of ϳ140 Ϯ 14 and ϳ70 Ϯ 7 s Ϫ1 were calculated. Essentially identical rate constants were obtained at several different wavelengths (data not shown). These values are very similar to the rates determined for the reaction in the presence of excess camphor (see above). The first two phases can also be readily distinguished in a blow-up of the Soret region; the first phase has an isosbestic point at 427 nm, whereas the second phase isosbestic point is at 423 nm (data not shown).
The Reaction of Substrate-free Oxyferrous P450-CAM with Stoichiometric Reduced Pdx-Having examined the reaction of oxyferrous P450-CAM with reduced Pdx in the presence of excess as well as stoi-chiometric camphor, we next examined the reaction in the absence of substrate. Fig. 7 displays spectra recorded during the reaction of substrate-free oxyferrous P450-CAM, produced in the first mix, with equimolar dithionite-reduced Pdx. The first scan is that of substratefree oxyferrous P450-CAM. The last is of substrate-free low spin ferric P450-CAM with peaks at 416.5, 533, and 568 nm. This process occurs in a single phase with a half-life (t1 ⁄ 2 ) of 31 ms under these conditions. This reaction with reduced Pdx is ϳ400-fold faster than that with oxidized Pdx (Fig. 2), suggesting that reduced Pdx is reducing the oxyferrous complex. Because substrate is absent from the reaction pocket, the rapid formation of low spin ferric P450-CAM almost certainly results in the formation of H 2 O 2 .
Substrate binding to low spin ferric P450-CAM converts the heme iron to high spin and alters the reduction potential from Ϫ330 to Ϫ173 mV versus NHE (normal hydrogen electrode) (1). Thus, substrate binding greatly facilitates electron transfer from reduced Pdx (E°ϭ Ϫ196 mV) (1) to the ferric P450-CAM heme to generate the five-coordinate high spin deoxyferrous state. Therefore, in the absence of substrate, Pdx does not readily reduce P450-CAM; as a result, enzymatic NADH consumption is dependent upon substrate binding to P450-CAM to make the potential more favorable. This provides a mechanism for preventing waste of reducing equivalents and formation of reactive oxygen species in the absence of substrate. The present experiment shows that reduced Pdx is readily able to provide the second electron to the dithionitereduced dioxygen-bound ferrous enzyme, when no substrate is present. This result unmistakably demonstrates that the inability of P450-CAM to accept electrons from reduced Pdx and reduce dioxygen, presumably to hydrogen peroxide, in the absence of substrate, i.e. to function as an NADH oxidase, is due to the inability of reduced Pdx to deliver the first electron to the substrate-free ferric enzyme, not the second electron.

Reaction of a Preformed Ferrous P450-CAM⅐Pdx Complex with O 2 in the Presence of Excess
Camphor-Another way to examine the reaction of oxyferrous P450-CAM and reduced Pdx is to preform a complex of the two reduced proteins under anaerobic conditions in the presence of substrate and then react this mixture with dioxygen in a single mixing experiment. As in the previous experiments done in the presence of substrate, this rapidly brings together all four components of the reaction: the two reduced proteins, the substrate, and dioxygen, so that the resulting spectral changes can be monitored. The obvious difference is  The ferrous P450-CAM⅐Pdx complex was prepared by combining an equimolar amount of camphor-bound ferric P450-CAM (in the presence of excess camphor) and ferric Pdx in one tonometer and reducing the complex with dithionite. 4 The resulting complex was mixed against buffer that had been saturated with 100% O 2 at 20°C and then cooled. Fig. 8 shows overlaid traces from two shots to show the overall events in the reaction. The first shot was recorded to 300 ms (to emphasize the early events), and the second shot was recorded to 30 s to demonstrate the formation of oxidized P450-CAM. Nearly the same spectral pattern is seen in this single mixing experiment as in the double mixing experiment (Fig. 3), in which preformed oxyferrous P450-CAM in the presence of excess camphor was combined with reduced Pdx in the second mixing operation. The time course of the reaction was also monitored under single wavelength conditions (Fig. 8B). A new process with a larger rate constant is seen (250 Ϯ 25 s Ϫ1 ) that corresponds to the second order formation of the oxyferrous intermediate (k ϭ ϳ4 ϫ 10 5 M Ϫ1 s Ϫ1 ) (the oxyferrous species was preformed in the double mixing experiment). This is followed by two phases (170 Ϯ 17 and 92 Ϯ 9 s Ϫ1 ) that are very similar to those determined in the previous experiments (see above) and a slower phase (ϳ0.25 s Ϫ1 ) that is currently not chemically characterized. The first phase was too fast to quantify from the diode array data, but the rate constant could be determined from data collected at single wavelengths (Fig. 8B).
As indicated above, after the formation of the oxyferrous species in the single mixing experiment (Fig. 8), the spectral sequence closely matches that seen in the double mixing experiment in the presence of excess camphor (Fig. 3). This shows that the perturbed oxyferrous species is not simply due to the binding of reduced Pdx, because in the single mixing experiments reduced Pdx is already bound to reduced P450-CAM.
Single Turnover Kinetics-From the experimental data just presented, two self-consistent sets of rates have been calculated (Figs. 5, 6, inset, and 8): the conversion of oxyferrous to perturbed oxyferrous P450-CAM (ϳ150 s Ϫ1 ) and the next step attributed to electron transfer from reduced Pdx (ϳ80 s Ϫ1 ) to form the double-humped spectrum (when excess camphor is present). Essentially, the same two rates were seen whether the reaction was triggered by the final addition of reduced Pdx or by dioxygen. The first rate for formation of the perturbed oxyferrous P450-CAM⅐Pdx complex was not seen in the previous single wavelength/single mix stopped-flow experiments, probably because of the choice of wavelengths (25)(26)(27). The second rate, which is assigned to the electron transfer from reduced Pdx to oxyferrous P450-CAM in the complex, is comparable with that reported by Brewer and Peterson (25) for the rate of electron transfer from reduced Pdx to oxyferrous P450-CAM under similar conditions (between 60 and 100 s Ϫ1 , depending on the concentration of reduced Pdx).
The observation of essentially the same rate for formation of the perturbed oxyferrous species (ϳ150 s Ϫ1 ) regardless of whether the complex of the two proteins was preformed or reduced Pdx was mixed with oxyferrous P450-CAM implies that the formation of the complex between the two proteins is very rapid. We estimate the rate of complex formation must be at least 500 s Ϫ1 , which at ϳ10 M implies a second order rate constant of 5 ϫ 10 7 M Ϫ1 s Ϫ1 , nearly diffusion-controlled.
The formation of the perturbed oxyferrous state and subsequent electron transfer steps, occur with rate constants that are greater than the turnover rate 5 observed for camphor hydroxylation under multiple turnover conditions. Clearly these two processes occur rapidly enough 4 The absorption spectrum of the camphor-bound deoxyferrous P450-CAM⅐reduced Pdx complex has a Soret peak at 409 nm and a peak in the visible region at 546 nm and is similar to the spectrum of camphor-bound deoxyferrous P450-CAM (peaks at 408 and 542 nm) (37). The lack of a distinct absorption peak at 315 nm showed that there was no excess dithionite. 5 The V max at 25°C is 40 s Ϫ1 (38); the value at 4°C is likely to be less than 10 s Ϫ1 .   ) that is considerably slower than the turnover rate. Bearing in mind that to achieve the fastest P450-CAM turnover rate requires a substantial excess of reduced Pdx (39), not the stoichiometric level used herein, the slowness of the third rate observed under single turnover conditions with stoichiometric reduced Pdx suggests that such a process (formation of ferric P-450⅐Pdx⅐camphor complex) is not part of the normal catalytic process. With excess reduced Pdx present under multiple turnover conditions and ready to bind to ferric P-450 to initiate the next round of catalysis, the third phase of the turnover process would presumably occur much more rapidly. Functional Implications of the Formation of the Perturbed Oxyferrous Intermediate-These experiments have focused on a reaction involving four components: two reduced proteins, camphor, and dioxygen. Once these four ingredients have been combined, the next species observed, the perturbed oxyferrous P450-CAM⅐Pdx complex, is the same (Figs. 3 and 8) regardless of whether reduced Pdx or dioxygen is added last. Fig.  9 shows that the spectra obtained under these two conditions for the perturbed oxyferrous species are nearly indistinguishable. The requirement for the presence of all four components also shows that the phase attributed to the formation of "perturbed" oxyferrous P-450 is not simply due to reduced Pdx binding.
Tosha et al. (20) have proposed that the binding of reduced Pdx to ferrous-CO P450-CAM leads to structural changes that, in the related oxyferrous enzyme, could facilitate dioxygen activation. Similarly, Nagano et al. (19) have examined the interaction of reduced Pdx with the ferrous-CO enzyme and have argued that in the oxyferrous enzyme, Pdx binding would enhance electron donation from the proximal Cys to assist O-O bond cleavage to generate P-450 compound I. The key role of the proximal Cys as an electron-releasing factor ("push") to facilitate O-O bond cleavage to generate compound I (Fig. 1, 6b 3 7) in the mechanism of dioxygen activation by P-450 has been extensively discussed (1,38,40,41). Numerous spectroscopic studies show that the interaction of reduced Pdx with low spin ferrous-CO and -NO P450-CAM adducts cause spectral changes attributable to structural rearrangements in the vicinity of the heme iron center (11,(17)(18)(19)(20)(21)(22). For example, Unno et al. (18) have suggested that the spectral changes they observed upon interaction of reduced Pdx with the ferrous-NO enzyme may correlate with changes in H-bonding to the proximal cysteine ligand.
The present study is the first to investigate the reaction of reduced Pdx with oxyferrous wild type P450-CAM in real time with rapid scanning UV-visible absorption spectroscopy. This has afforded the observation of the transient perturbed oxyferrous species that precedes electron transfer; this could be interpreted as an activating event that sets up the dioxygen-bound heme iron for electron transfer from reduced Pdx and subsequent O-O bond cleavage. Building on the extensive previous work examining the interaction of reduced Pdx with ferrous-CO and -NO states of P450-CAM just discussed, we propose that the transient spectral perturbation of oxyferrous P450-CAM reported in this study reflects a modification of the H-bonding to the proximal Cys. The role of H-bonding to the proximal Cys by peptide amide N-H groups has been examined by Ogliaro et al. (42) using density functional theory; these studies concluded that H-bonding strengthens the iron-sulfur bond while enhancing the porphyrin -cation radical character of compound I at the expense of a diminished thiyl radical component. Experimental investigations have also revealed a link between H-bonding to the thiolate sulfur in P-450 and mono-oxygenase activity (23,24,43).
The spectral properties of thiolate-ligated heme systems such as P-450, chloroperoxidase, and model heme complexes have been extensively investigated (38,44). In particular, Roach et al. (45) examined how the properties of the thiolate sulfur donor influence the UV-visible absorption spectra in the high spin ferric state; they found that complexes with aromatic thiolates having electron withdrawing substituents had a Soret absorption peak at ϳ390 nm, as in high spin ferric P450-CAM, whereas adducts with aromatic thiolates containing electron donating substituents had their Soret peak at ϳ400 nm, as in high spin ferric chloroperoxidase. This suggests that in this state, a blueshifted Soret peak correlates with a less electron-rich thiolate sulfur donor atom. Consistent with the trends seen in the ferric state spectra (45), oxyferrous P450-CAM has its Soret absorption peak at 418 nm, blue-shifted from that of oxyferrous chloroperoxidase at 428 nm (36). The further shift to 414 nm in the perturbed oxyferrous P-450 reported here is thereby suggestive of transient formation of a less electron-rich thiolate sulfur donor atom. Similar spectral trends have been reported for His-ligated oxyferrous heme complexes (46). Because thiolate-ligated hemes are more difficult to reduce than non-thiolate-ligated complexes, weakening of the thiolate sulfur donor strength to oxyferrous P-450 could facilitate reduction of that species to generate the next intermediate in the P-450 reaction cycle. Modulation of the H-bonding to the thiolate sulfur proximal ligand is a reasonable way to modify ligand donor properties. Finally, we note that the reaction of reduced Pdx with oxyferrous D251N P450-CAM produces an altered oxyferrous species with a red-shifted Soret peak and very slow turnover kinetics (47,48). This observation is consistent with our proposal that the blue shift in the Soret peak position of the oxyferrous state correlates with the ease of reduction and, therefore, the rate of turnover of the enzyme. In addition, Nagano et al. (49) and Tosha et al. (50) have prepared and structurally characterized the L358P mutant of P450-CAM, which has altered hydrogen bonding to the proximal thiolate sulfur. The spectral properties of this mutant are similar to those induced by Pdx binding.
In summary, we have observed transient formation of a perturbed oxyferrous intermediate following assembly of the four components: two reduced proteins, substrate, and dioxygen that are needed to form the next P-450 intermediate that follows the oxyferrous state, presumably the ferric peroxo species. The spectral properties of the perturbed oxyferrous state suggest that it has a less electron-rich thiolate ligand, i.e. it is more easily reduced. Modification of the H-bonding manifold to the thiolate sulfur would be the simplest way to achieve this effect. Generation of a less electron-rich thiolate would, in turn, serve to facilitate electron transfer from reduced Pdx to the oxyferrous adduct to produce the peroxo heme 6a in Fig. 1. The perturbed oxyferrous species may be a direct spectral signature of the effector role of Pdx on P450-CAM reactivity (i.e. during catalysis). The results reported herein also seem to provide an answer to one of the many conundrums of the P-450 mechanism. The efficient delivery of an electron to oxyferrous P-450 (Fig. 1, 4 3 6a) would be favored by an electron withdrawing component, whereas the O-O bond cleavage of the resultant one-electron reduced oxyferrous state (Fig. 1, 6a and 6b) would be facilitated by a strong electron releasing factor (push) (1,38,40,41). The present results suggest that initial interaction of reduced Pdx tones down the donor properties of the thiolate ligand in the oxyferrous enzyme to allow for more facile electron transfer, after which the donor strength of the thiolate might return, perhaps after protonation of the peroxoferric state (Fig. 1, 6a 3 6b) neutralizes some of the negative charge built up at the heme center.