NMR Study on the Structural Changes of Cytochrome P450cam upon the Complex Formation with Putidaredoxin

We investigated putidaredoxin-induced structural changes in carbonmonoxy P450cam by using NMR spectroscopy. The resonance from the β-proton of the axial cysteine was upfield shifted by 0.12 ppm upon the putidaredoxin binding, indicating that the axial cysteine approaches to the heme-iron by about 0.1 Å. The approach of the axial cysteine to the heme-iron would enhance the electronic donation from the axial thiolate to the heme-iron, resulting in the enhanced heterolysis of the dioxygen bond. In addition to the structural perturbation on the axial ligand, the structural changes in the substrate and ligand binding site were observed. The resonances from the 5-exo- and 9-methyl-protons of d-camphor, which were newly identified in this study, were upfield shifted by 1.28 and 0.20 ppm, respectively, implying that d-camphor moves to the heme-iron by 0.15–0.7 Å. Based on the radical rebound mechanism, the approach of d-camphor to the heme-iron could promote the oxygen transfer reaction. On the other hand, the downfield shift of the resonance from the γ-methyl group of Thr-252 reflects the movement of the side chain away from the heme-iron by ∼0.25 Å. Because Thr-252 regulates the heterolysis of the dioxygen bond, the positional rearrangement of Thr-252 might assist the scission of the dioxygen bond. We, therefore, conclude that putidaredoxin induces the specific heme environmental changes of P450cam, which would facilitate the oxygen activation and the oxygen transfer reaction.

Cytochrome P450 (P450) 1 is a heme-containing monooxygenase, which catalyzes the hydroxylation reactions of a wide variety of natural and unnatural substrates such as steroids, fatty acids, hydrocarbons, and xenobiotics (1). The P450 catalysis reaction requires two electrons from NADH or NADPH through the redox-linked proteins. In microsomal P450s, the electron transfer (ET) from NADPH to P450s is mediated by NADPH-P450 reductase containing FMN and FAD as the redox centers. In contrast, the electron from NADH or NADPH is sequentially transferred to ferredoxin reductase, ferredoxin, and finally P450 in mitochondrial and bacterial systems. The ET reaction between P450 and its redox partner is essential for the subsequent activation of molecular oxygen and the monooxygenation reaction.
To understand the mechanism of the ET reaction in P450 and its redox partner system, the ET reaction between cytochrome P450cam (P450cam) from Pseudomonas putida, one of the bacterial P450s, and its redox partner, putidaredoxin (Pdx), has intensively been investigated. P450cam catalyzes the regio-and stereo-specific hydroxylation of its substrate, d-camphor (1, 2) by accepting two electrons from NADH. The ET from NADH to P450cam is sequentially mediated by a flavin group of putidaredoxin reductase and a [2Fe-2S] center of putidaredoxin (Pdx). With the availability of the P450cam (3) and Pdx (4,5) structures, many investigators have performed the kinetic (6 -9), mutational (10 -14), and theoretical (15,16) studies to clarify the molecular mechanism for the ET reaction in the P450cam⅐Pdx system. These studies demonstrated that Pdx interacts with the proximal surface of P450cam through the electrostatic interaction (12,15,17) and suggested that Arg-112 at the putative Pdx binding site forms the ET pathway in the P450cam⅐Pdx complex (12,16).
It was also revealed that Pdx acts as the specific electron donor for the turnover reaction of P450cam (18). Lipscomb et al. (18) carried out the mixing of oxy-P450cam with several electron donors, including the non-physiological electron donors as well as the physiological redox partner, Pdx, and examined the formation of the hydroxylation product. Low potential ironsulfur protein such as spinach ferredoxin and bovine adrenodoxin can donate an electron to ferric P450cam (the first ET) but yield no hydroxylation products. On the other hand, the addition of rat liver cytochrome b 5 (cyt b 5 ) and bacterial rubredoxins to oxy-P450cam can yield the hydroxylation product, whereas these non-physiological electron donors cannot transfer the first electron to ferric P450cam. Thus, the specific complex formation between P450cam and Pdx is required for the turnover reaction.
We have studied the mechanism of the specific complex formation between P450cam and Pdx (19,20) and clarified that the P450cam⅐Pdx system exhibits unique features compared with other protein⅐protein ET complexes. Aoki et al. (19) have reported the negative values for the changes in the enthalpy and entropy for the formation of the P450cam⅐Pdx complex determined from the isothermal titration and suggested that the recognition mechanism between P450cam and Pdx is similar to that of the antigen⅐antibody complex formation. A subsequent study performed by Furukawa and Morishima (20) examined the osmotic pressure dependence of the association between P450cam and Pdx and revealed that water molecules participate in the complex formation and cause the negative entropy change upon the complexation of P450cam with Pdx.
Other important feature of the P450cam⅐Pdx complex is that Pdx induces conformational changes of P450cam upon the complex formation (21)(22)(23)(24)(25)(26)(27)(28). EPR and resonance Raman studies clearly exhibited that the complex formation with Pdx converts the spin-state of ferric P450cam from the high to low spin state (21,24). Associated with the change of the spin-state, the heme-iron axial ligand (Fe-S) stretching mode is upshifted by 3 cm Ϫ1 (21). In the ferrous-CO form, the Fe-CO and FeC-O stretching modes indicate 2 cm Ϫ1 upshift and 8 cm Ϫ1 downshift, respectively, by the binding of reduced Pdx, respectively (29). Recent multidimensional NMR study on the complex of P450cam with Pdx showed that the binding of Pdx structurally perturbs the several regions involving the substrate access channel in P450cam (28).
These structural changes of P450cam upon the binding of Pdx are supposed to be essential for the enzymatic activity of P450cam. Shimada et al. (30,31) found that the mutations at the putative Pdx binding site, Arg-109 or Arg-112, inhibit the conformational changes in P450cam and suppress the hydroxylation activity to 1-500 M/min/M heme corresponding to 1/1000 to 1/3 of that for the wild-type enzyme (12,31). Although the Pdx-induced structural changes of P450cam have been supposed to be crucial for the P450cam catalysis, it is still unclear how the conformational changes in the active site of P450cam promote the turnover reaction. It is strongly desirable to identify the detailed structural changes of P450cam upon the binding of Pdx.
To clarify the conformational changes by the binding of Pdx to P450cam and discuss the functional significance on the catalytic oxygenation reaction, we characterized the Pdx-induced conformational changes of P450cam by using NMR spectroscopy. In the diamagnetic ferrous-CO form of P450cam, the ring-current of the porphyrin separates several resonances of the protons near the heme-iron from the crowded signal region between 0 and 10 ppm (32)(33)(34). In fact, Mouro et al. (35) reported that the ␤-proton of the axial Cys in carbonmonoxy P450cam indicates the resolved peak at Ϫ2.76 ppm. However, most of the ring-current-shifted signals of the ferrous-CO form of P450cam have not yet been assigned. Therefore, we first assigned several ring-current-shifted NMR signals by comparing wild-type and mutants of P450cam in the absence and presence of the substrate and substrate analogues. With the help of the two-dimensional NOESY spectrum, we have successfully assigned the signals arising from d-camphor and Thr-252. Based on these signal assignments, we examined the structural effects of the Pdx binding on the heme environmental structure. As a control experiment for the Pdx-induced structural changes, we also measured the NMR spectrum for carbonmonoxy P450cam in the presence of cyt b 5 , whose binding site in P450cam is supposed to overlap with that of Pdx. The detailed information on the structural changes of the heme pocket upon the binding of Pdx allowed us to discuss the functional significance of the Pdx-induced structural changes in P450cam.

EXPERIMENTAL PROCEDURES
Expression and Purification of Proteins-Wild-type P450cam was expressed in a strain of Escherichia coli, BL21, as an inclusion body. Following the procedures described in our previous reports (36, 37), we performed the heme reconstitution and the purification of P450cam.
P450cam mutants (T252A and T252G) were expressed in an E. coli strain, JM109, and purified as previously reported (38). Purified proteins with the R/Z value (A 392 /A 280 ) greater than 1.5 were employed in this study. The purified samples were dissolved in 50 mM potassium phosphate buffer at pH 7.4 containing 50 mM KCl and 1 mM d-camphor and stocked at Ϫ70°C. The substrate-free sample was prepared by passing the substrate bound P450cam through a G-25 column equilibrated with 50 mM Tris-HCl buffer at pH 7.4. The buffer of the eluted sample was exchanged to 50 mM potassium phosphate buffer at pH 7.4 containing 50 mM KCl.
Construction of the expression vector for the Asp-38 3 Asn mutant of Pdx (D38N) was reported by our group (14). Wild-type Pdx and the D38N mutant were expressed in an E. coli strain, RR1, and were purified by the method of Gunsalus and Wagner (39) with a minor modification (40). The preparation of Pdx used in this study had a A 325 nm to A 280 nm ratio of at least 0.60. The sample was dissolved in 50 mM potassium phosphate buffer at pH 7.4 containing 50 mM KCl and 80 l/100 ml ␤-mercaptoethanol. ␤-Mercaptoethanol was removed from the solution before the NMR measurements to avoid the interference of the complex formation with P450cam.
Rat liver cyt b 5 , whose C-terminal membrane binding domain was deleted, was used in this study. The expression and purification of cyt b 5 were carried out by using a method described previously (41). Purified proteins with the R/Z value (A 412 /A 280 ) greater than 5.5 were employed. The purified samples were dissolved in 50 mM potassium phosphate buffer at pH 7.4 containing 50 mM KCl and 1-mM d-camphor.
NMR Measurements-One-dimensional 1 H NMR measurements were performed on a BRUKER Avance DRX500 spectrometer at 40°C. Chemical shifts were referenced to the residual water signal that was calibrated against tetramethylsilane. The WET and PRESAT pulse sequences were used to minimize the water signal (42). One-dimensional NOE experiments with 30 -100 ms of the mixing times were performed using the standard pulse sequence (43). Two-dimensional NOESY, COSY and TOCSY spectra were recorded with a BRUKER Avance DRX600 equipped with a cryo-cooled probe at 25°C by using the standard pulse sequence (43)(44)(45).
NMR samples were prepared by the following procedures. Substratebound P450cam was dissolved in 10% D 2 O or 100% D 2 O containing 50 mM potassium phosphate, 50 mM KCl, and 1-mM d-camphor at pH (pD) 7.4. Substrate-free sample was dissolved in 10% D 2 O or 100% D 2 O containing 50 mM potassium phosphate, 50 mM KCl at pH (pD) 7.4. The d-camphor analogue-bound form of P450cam was prepared by use of the same buffer as the substrate-free sample but saturated with one of the d-camphor analogues: norcamphor, adamantanone, or 3-bromocamphor. Concentrations of P450cam in the NMR samples were 0.2 mM and 0.5-1.0 mM for the one-dimensional NMR and one-dimensional NOE/ two-dimensional NOESY measurements, respectively. 500 l of samples was transferred to the NMR tubes, which were capped with a rubber septum. After flushing with argon gas, samples were reduced by the injection of a small aliquot of degassed sodium dithionite solution. CO gas was anaerobically introduced into the NMR tube to prepare the ferrous-CO form of P450cam.
Measurement of the Dissociation Constant (K d ) of Reduced Pdx with the Ferrous-CO Form of P450cam-We measured one-dimensional 1 H NMR spectra of the ferrous-CO form of P450cam in the presence of various concentrations of Pdx red from 50 M to 2 mM. The dissociation constant (K d ) of the complex between Pdx red and the ferrous-CO form of P450cam was calculated by the shifts of the NMR signals upon the binding of Pdx red . We confirmed that the rate of equilibrium between Pdx-bound P450cam and Pdx-free P450cam was fast in our experimental condition (see "Results" session). In this fast exchange limit, K d can be determined by Equation 1 (46), where ␦ obs denotes the chemical shift of P450cam in the titration experiment. ␦ max and ␦ 0 represent the chemical shifts of P450cam in the presence and absence of Pdx red , respectively.
[P450] tot and [Pdx] tot correspond to the total concentrations of the ferrous-CO form of P450cam and Pdx red in the sample solution, respectively.

RESULTS
Assignments of NMR Signals for the Ferrous-CO Form of P450cam- Fig. 1 illustrates the NMR spectra for the fer-rous-CO forms of P450cam in the absence and presence of the substrate, d-camphor, and substrate analogues, showing several well resolved and ring-current-shifted resonances in the upfield region (35). The intensities of the signals in Fig. 1 were unchanged in the deuterated solution (data not shown). In these ring-current-shifted signals, the signal a observed at Ϫ2.76 ppm in the presence of d-camphor has already been assigned to the ␤-proton of the axial cysteine, Cys-357 (35). However, other ring-current-shifted signals b, c, and d have not yet been assigned.
The most prominent spectral change by the binding of dcamphor (traces A and E) is the increased intensity of the signal b. Based on the intensity of the signal a, the intensity of the signal b in the absence of the substrate (trace A) corresponds to three protons. The intensity of the signal b was increased to six protons by addition of d-camphor (trace E), indicating that a signal of three protons, which is designated as the signal x, overlapped with the signal b in the presence of d-camphor. It should be noted that the increase of the intensity of the signal b was not observed in the presence of the substrate analogue, norcamphor (trace B) or adamantanone (trace C), whereas the addition of 3-bromocamphor (trace D) increased the intensity of the signal b. These observations suggest that the signal x might be assigned to one of the methyl groups, 8-or 9-methyl groups, attached to the C7 position of d-camphor.
To further examine the assignments of the signal x, we measured the two-dimensional NMR spectra, NOESY, COSY, and TOCSY, of the ferrous-CO form of P450cam in the presence of d-camphor. Unfortunately, however, we could not obtain the high quality COSY and TOCSY spectra due to the large molecular weight of P450cam (M r ϭ 46,000) (data not shown) (47). Thus, the assignment of the signal x was validated by examining the NOESY data in relationship to the crystal structure of carbonmonoxy P450cam. Fig. 2 depicts the upfield region of the NOESY spectra for carbonmonoxy P450cam in the presence and absence of d-camphor at 25°C. Due to the low temperature to avoid the denaturation of the samples during the NOESY measurements, the chemical shifts for the ring-current-shifted signals in the one-dimensional spectra at the top of Fig. 2 were different from those at 40°C (Fig. 1). In the absence of dcamphor, the signals b (Ϫ1.23 ppm), c (Ϫ1.14 ppm), and d (Ϫ1.04 ppm) were observed. On the other hand, in the dcamphor-bound enzyme, the signal at Ϫ0.90 was observed as a resolved signal, and the signals Ϫ1.14 and Ϫ1.24 ppm were observed as shoulder signals due to the overlapping with the signal at Ϫ1.20 ppm. The signal at Ϫ1.23 ppm (signal b) in the substrate-free P450cam gave NOE cross-peaks with the signals at 0.07 (b1), 1.56 (b2), 3.96 (b3), 4.38 (b4), and 4.94 ppm (b5) (red spectrum in Fig. 2), which are the same NOE pattern observed for the shoulder peak at Ϫ1.24 ppm in the d-camphorbound enzyme (black spectrum in Fig. 2). This indicates that the signal at Ϫ1.24 ppm in the d-camphor-bound enzyme corresponds to the signal b. The signal at Ϫ1.14 ppm (signal c) in the substrate-free enzyme exhibited the NOE cross-peaks (c1, c2, c3, and c4) as found for the signal at Ϫ1.14 ppm in the d-camphor-bound enzyme, showing that the signal at Ϫ1.14 ppm in the d-camphor-bound enzyme is derived from the protons that give the signal c in the substrate-free form. The signal at Ϫ1.04 ppm (signal d in red spectrum of Fig. 2) in substratefree P450cam displayed strong NOE cross-peaks with the signal from the ␣-meso proton of the heme (9.74 ppm) (d4) and the signals at 1.33 (d1), 4.01 (d2), and 4.19 ppm (d3), and the same NOE pattern was also observed for the signal at Ϫ0.90 ppm in the d-camphor-bound enzyme. These results clearly confirm that the remaining signal at Ϫ1.20 ppm (signal x in the black spectrum) is derived from the substrate, d-camphor. It should be noted here that the signal x shows NOEs to the 8-methyl-(3.43 ppm) and ␦-mesoprotons (10.17 ppm) of the heme (black spectrum in Fig. 2). Because both of the 8-methyl-and ␦-mesoprotons of the heme are located near the 9-methyl group of d-camphor in the crystal structure of the ferrous-CO form of P450cam (48) (Fig. 3), we can unambiguously assign the signal x to the 9-methyl group of d-camphor.
The two-dimensional NOESY spectrum of the ferrous-CO form of P450cam also gave us information about the signal d.
The signal d displayed cross-peaks with the signal x, 9-methyl group of d-camphor (Ϫ1.20 ppm) and ␣-meso proton of the heme (9.74 ppm) (black spectrum in Fig. 2), and its complete disappearance in the Thr-252 3 Ala mutant (trace F of Fig. 1) suggests that the signal d originates from the side chain of Thr-252. The signal d was also missing in the Thr-252 3 Gly mutant (data not shown), which also supports this assignment. To verify the assignment of the signal d, we compared the NOESY spectrum for d-camphor-bound carbonmonoxy wild- type P450cam with that of the T252A mutant. As shown in Fig.  2 (green spectrum), the pattern of the NOE connectivities for the signals at Ϫ1.24 and Ϫ1.14 ppm in the T252A mutant indicates that these signals correspond to the signals b and c in the wild-type enzyme, respectively. The signal at Ϫ1.34 ppm in the T252A mutant gave NOE cross-peaks with the 8-methyl-(3.44 ppm) and ␦-meso-(10.19 ppm) protons of the heme (green spectrum in Fig. 2), allowing us to assign the signal at Ϫ1.34 ppm to the 9-methyl group of d-camphor. Thus, the signal d, which is missing in the T252A mutant, can be ascribed to the ␥-methyl group of Thr-252. This assignment is supported by the close location of Thr-252 to the heme-iron shown in the crystal structure of carbonmonoxy P450cam (Fig. 3). The remaining signals, signals b and c, are still unidentified. Although their signal intensities and upfield shifts are suggestive of the methyl groups near the heme plane, we cannot find appropriate protons in the crystal structure. The specific assignments of the ring-current-shifted signals are summarized in Table I. 1 H NMR Spectra of Ferrous-CO P450cam in the Presence and Absence of Pdx red -To investigate the conformational changes around the heme in P450cam upon the binding of Pdx, we measured 1 H NMR spectra of ferrous-CO P450cam in the presence of Pdx red . Fig. 4 represents the 1 H NMR spectra of the ferrous-CO forms of P450cam in the presence of 0, 0.5, 1, 2, and 4 equivalents of Pdx red . With the increase in the concentration of Pdx red , the ring-current-shifted signals were shifted, implying that the binding of Pdx red affects the heme environmental structure of carbonmonoxy P450cam.
Although the Pdx-induced shifts in the signals b and c were less than 0.02 ppm, the signals a, d, and x showed more prominent shifts. The signal d was downshifted by about 0.18 ppm, and more than 0.1 ppm of upshift was observed for the signals x and a. However, the most remarkable positional shift was detected for the signal y, which was missing in the absence of Pdx red (trace A). As displayed in Fig. 4 (trace G), the signal y gave NOEs with the signals d and x. Because the signals d and x were assigned to ␥-methyl of Thr-252 and 9-methyl of dcamphor, respectively, the signal y would be derived from one of the protons in the distal side of the heme. Because the 5-exo proton of d-camphor is located near both of the 9-methyl group and ␥-methyl protons of Thr-252, it is likely that the signal y originates from the 5-exo proton of d-camphor. The enhanced upfield shift of the signal y also corresponds to the close proximity of the 5-exo proton to the heme-iron.
To confirm that these positional shifts of the resonances are associated with the complex formation of P450cam with Pdx, the dissociation constant for the complex was estimated by the chemical shifts of the signals. Because no significant line broadening indicates that the equilibrium between free and bound proteins is fast (46), Equation 1 for the dissociation constant, K d , under the fast exchange process is available for the P450cam⅐Pdx system. By using Equation 1, the observed chemical shifts for the signals a, d, and x were plotted as a function of the concentration of Pdx red (Fig. 5). The experimental data points were fitted by Equation 1 with K d ϭ 134, 138,  and 149 M for the signals a, d, and x,   respectively (Table I). As shown in Fig. 5, the titration curves were well fitted by Equation 1, assuming a 1:1 complex, and the K d values obtained from the signals a, d, and x were almost indistinguishable, indicating that the changes of the NMR signals are induced by the 1:1 complex formation between P450cam and Pdx red .
To estimate the limiting shift of the signal y, which is not well resolved in the absence of Pdx red , we extrapolated the titration curve for the signal y to the concentration of Pdx red ϭ 0 with the determined K d value (140 M) (Fig. 5, dotted line). The chemical shift of the signal y in the absence of Pdx red was estimated to be Ϫ1.18 ppm. The estimated position of the signal y would be consistent with the broad shoulder around Ϫ1.1 ppm on the NMR spectrum in the absence of Pdx (trace A of Fig. 4).
The K d value obtained from the NMR signal changes is, however, 10 times larger than the value (13 M) estimated from the CO stretching frequency at room temperature (31). One of the reasons for the difference would be the different temperature for the measurements. A previous study reported that K d for the complex formation of the ferric form of P450cam with oxidized Pdx increases with the elevation of the temperature (7). To get highly resolved signals, the NMR spectra were measured at higher temperature (40°C) than the measurements for the CO stretching frequency, which would lead to the larger dissociation constant in the NMR measurements.
It is also quite interesting that these Pdx-induced spectral changes were not detected in the absence of the substrate, d-camphor (data not shown). Even in the presence of 5 equivalents of Pdx red , the spectral pattern between Ϫ0.5 and Ϫ3.0 ppm was not seriously perturbed. The Pdx-induced shifts for both of the signals a and d, the ␤-proton of Cys-357, and the ␥-methyl group of Thr-252, were less than 0.05 ppm. The Pdxinduced chemical shifts are, therefore, specific for substratebound P450cam.
NMR Spectral Changes upon the Binding of the Pdx Mutant-Previous reports indicated that the mutations of Arg-112 at the putative Pdx binding site inhibit the conformational changes upon the binding of Pdx and decrease the ET rate from Pdx (12,31), which permit us to suggest that the conformational changes of P450cam upon the binding of Pdx are crucial for the ET from Pdx to P450cam. To correlate the spectral changes with the ET activity, we followed the spectral changes by addition of a Pdx mutant, Asp-38 3 Asn (D38N), whose ET activity is nearly abolished (13,14). Fig. 6 displays the NMR spectra of carbonmonoxy P450cam in the presence of 2.5 and 6 equivalents of reduced D38N Pdx. As shown in the NMR spectra, no significant spectral changes were found in the region between Ϫ0.5 and Ϫ3.0 ppm. Previously, our group determined the Michaelis constants for D38N Pdx to P450cam in the P450cam catalytic cycle at 22°C, which was reduced to 1/16 of that for wild-type Pdx (14). Under the condition we used here, 30% (60 M) of P450cam would form the P450cam⅐Pdx (D38N) complex in the presence of 6 equivalents of the D38N mutant, which should be clearly detected on the NMR spectrum. The absence of the spectral changes in P450cam with D38N Pdx is not due to the low affinity of the mutant, but to the loss of the ability of Pdx to induce the conformational changes in P450cam.
NMR Spectra in the Presence of Cyt b 5 -Pdx-induced structural changes in P450cam have been supposed to be crucial for the P450cam catalysis (21,23,24,26,31). To delineate the specific role of Pdx in the P450cam-catalyzed oxygenation reaction, we measured NMR spectra of P450cam in the presence of cyt b 5 as a control experiment. From the binding competition experiment, oxidized cyt b 5 and Pdx are believed to bind to the same or similar site on the ferric P450cam surface (49), raising the possibility that cyt b 5 could also induce the structural changes in P450cam.  (50) is widely used for calculating the ring-current contribution. The magnitude of the ring-current shift of the proton near the aromatic ring, rc (ppm), is given by Equation 2 (50,52), where C is a constant related to the properties of the aromatic ring, and and z represent the distances between the aromatic ring and the proton measured perpendicular to the aromatic ring and outward in the aromatic ring, both of which are expressed in units of the radius of the aromatic ring. K(k) and E(k) are complete elliptic integrals of the first and second kind, respectively. The modulus k is expressed by Equation 3.
Among the observed ring-current-shifted signals, the ␤-proton of the axial Cys can serve as a structural marker for the proximal side of the heme. Based on Equation 2 (50, 52), we estimate the possible shifts of the ␤-proton of the axial Cys either along the axis perpendicular to the heme plane (z-axis) or in the plane parallel to the heme plane. The upfield shift of the ␤-proton of the axial Cys (Ϫ2.76 to Ϫ2.88 ppm) can be caused by an approach of ϳ0.05 Å along the z-axis or a 0.15-Å movement to the center of the porphyrin ring in the plane parallel to the heme plane. Therefore, the ␤-proton of the axial Cys would move toward the heme-iron by ϳ0.05-0.15 Å upon the binding of Pdx. Such a structural change of the proximal side in P450cam is qualitatively consistent with the previous results from the resonance Raman spectroscopy (21). The complex formation between ferric P450cam and oxidized Pdx induces the upshift of the heme-axial ligand (Fe-S) stretching mode by 3 cm Ϫ1 (21), implying the movement of the axial thiolate to the heme-iron.
The upshift of the Fe-S stretching mode by 3 cm Ϫ1 is, however, quite small, because the 0.1-Å movement of the thiolate sulfur toward the heme-iron corresponds to the upshift of the heme-axial ligand stretching mode by ϳ50 cm Ϫ1 from the density functional theory calculation for the heme⅐imidazole complex (53). We, therefore, suggest that the positional change of the thiolate sulfur relative to the heme-iron would be smaller than that of the ␤-carbon of the axial Cys, because the change of the dihedral angle between the Fe-S bond and the ␤-carbon of the axial Cys by 10°can result in an approach of 0.1 Å of the ␤-carbon of the axial Cys to the heme-iron without significant movement of the axial Cys. Thus, the movement of the thiolate sulfur relative to the heme-iron might be smaller than 0.1 Å.
The structural changes in the distal side can be also assessed by following the changes in the chemical shifts of the ␥-methyl group of Thr-252 and the protons of d-camphor and using Equation 2. The 9-methyl group of d-camphor (signal x), showing the NMR signal that was upfield-shifted by 0.20 ppm, would approach the heme-iron by ϳ0.15 Å along the z-axis or 0.15 Å in the plane parallel to the heme plane. The upfield shift of the resonance from the 5-exo proton of d-camphor (signal y) (1.28 ppm) can be interpreted with ϳ0.5-Å movement vertical to the heme plane or an approach of Ͼ0.7 Å to the center of the porphyrin ring in the plane parallel to the heme plane. These Pdx-induced structural changes at the heme distal side were also reported by Pochapsky et al. (28). They measured multidimensional NMR spectra of 15 N-and 13 C-labeled P450cam in the absence and presence of Pdx to characterize the Pdx-induced structural changes in P450cam. By the Pdx binding, the resonance from Tyr-96, whose hydroxyl group is hydrogenbonded to the carbonyl oxygen in d-camphor (3), was perturbed. The perturbation of the hydrogen bond between Tyr-96 and d-camphor is consistent with the movement of d-camphor upon the binding of Pdx, we presented here.
In contrast to the approach of d-camphor toward the heme iron, the downfield shift of the resonance from the ␥-methyl group of Thr-252 reflects the movement of the ␥-methyl group away from the heme-iron by ϳ0.15 Å along the z-axis or 0.15 Å in the plane parallel to the heme plane. Pochapsky et al. (28) also observed the movement of the I-helix upon the complexation with Pdx, which would correspond to the positional change in Thr-252. The different positional changes observed between the substrate and ␥-methyl group of Thr-252 suggest that the heme plane of P450cam is tilted by the binding of Pdx as illustrated in Fig. 8. It should be stressed that the tilt of the heme plane upon the binding of Pdx would be mediated by Arg-112. Because Arg-112 in P450cam, which is crucial for the interaction with Asp-38 in Pdx (12,15,16), is hydrogen-bonded to 6-heme propionate (3,54), it is likely that the binding of Pdx to P450cam shifts the side chain of Arg-112, leading to the positional change of one of the propionate groups and the tilting of the heme plane. The positional change of Arg-112 is evident from the significant perturbation of the N⑀H resonance of Arg-112 upon the Pdx binding (28). Furthermore, the crucial contribution of the interaction between Arg-112 in P450cam and Asp-38 in Pdx to the tilting of the heme plane is supported by the absence of the spectral changes upon the complex formation of D38N Pdx with P450cam (Fig. 6).
The tilt of the heme plane upon the binding of Pdx is supported by the shifts of the low frequency Raman lines in the 300 -800 cm Ϫ1 region (23). Although the specific assignments have not yet been reported for P450cam, the low frequency Raman lines would involve the vibrational modes of the peripheral groups of the heme as observed for myoglobin (55). The shifts of the Raman lines in the low frequency region, therefore, likely indicate that the environment surrounding the peripheral groups of the heme is changed by the binding of Pdx.
Correlation of Redox Potential and Chemical Shift of ␤-Proton of Axial Cysteine-The approach of the axial Cys to the heme-iron upon the complexation with Pdx would enhance the electronic donation from the axial Cys to the heme-iron and lower the redox potential of the heme-iron (36,56). We found a clear relationship between the redox potentials and the chemical shifts of the signal a (the ␤-proton of the axial Cys) for P450cam complexed with various substrates (35,57) (Fig. 9). The relationship confirms the lowered redox potential of P450cam by the Pdx binding. As clearly shown in Fig. 8, the signal a was shifted by 0.12 ppm upon the binding of Pdx, which corresponds to the negative shift by ϳ90 mV of the redox potential of the heme-iron in P450cam.
The possible negative shift of the redox potential of P450cam upon the binding of Pdx could be caused by the enhancement of the electronic donation from the axial Cys to the heme-iron as suggested by Unno et al. (21). They reported that the binding of negatively charged Pdx to P450cam would diminish the posi-tive character of the proximal surface of P450cam, resulting in the enhancement of the anionic character of the axial Cys (21) and eventually lowering the redox potential of P450cam. 2 Based on the Marcus equation, the negative shift of 90 mV decelerates the ET rate to only 20% of that calculated without the negative shift. However, the redox potential is not only the determinant factor for the ET rate and other factors, the reorganization energy and the distance between the redox centers, also have the substantial contribution. The mutations of Arg-112 at the putative Pdx binding site drastically inhibit the conformational changes of P450cam upon the binding of Pdx (12,31) and decelerate the ET rate (12), providing evidence that the conformational changes accelerate the ET rate. It is premature to conclude that the binding of Pdx compensates for the reduced redox potential by decreasing the reorganization energy and/or shortening of the distance between the redox 2 Sligar and Gunsulus reported the redox potential of P450cam was not altered by the binding of Pdx (71). The redox potential of P450cam in the complex with Pdx should be re-investigated, although it is difficult to determine it. centers in the ET reaction to P450cam, but we can safely conclude that Pdx red induces the specific conformational changes in the heme environmental structure of P450cam to regulate the reorganization energy and/or the distance between the redox centers for the ET reaction to P450cam.
Significance of the Pdx-induced Conformational Changes on P450cam Catalytic Reaction-The current NMR results demonstrated that the axial Cys moves toward the heme-iron by the Pdx binding, which would enhance the electronic donation from the axial Cys to the heme-iron. The enhancement of the electronic donation from the axial Cys, referred to as the enhanced "push effect," was believed to facilitate the heterolysis of molecular oxygen bound to the heme-iron (58). The P450cam mutant, Leu-358 3 Pro, whose Fe-CO and FeC-O Raman lines indicated up-shift and down-shift, respectively, as found for wild-type P450cam in the presence of Pdx, exhibited the high oxygen activation activity, confirming that the O-O bond scission is accelerated by the enhanced push effect (36,58,59). The Pdx-induced conformational changes at the proximal side of the heme, therefore, substantially contribute to the activation of molecular oxygen on the heme-iron in P450cam.
The Pdx induced conformational changes in the distal side, and the movement of Thr-252 could also facilitate the heterolysis of the dioxygen bond. The water molecule hydrogenbonded to the hydroxyl group of Thr-252 is supposed to play a pivotal role in the O-O bond scission (38,60,61), and the positional perturbation of the water molecule drastically suppresses the O-O bond scission as reported for the T252A mutant. The T252A mutant primarily produces the hydrogen peroxide instead of the hydroxylation product, 5-exo-hydroxycamphor (38,60). The Pdx-induced positional change of Thr-252 would lead to the positional change of the water molecule in the oxygen binding pocket that is favorable for the O-O bond scission.
In addition to the functional significance of the movement of Thr-252 in the P450cam catalysis, the Pdx-induced positional change of d-camphor would be also crucial for the effective turnover. Pochapsky et al. (28) proposed that the primary role of the Pdx-induced structural changes is to prevent the formation of the reaction intermediate without the formation of hydroxylation product by the suppression of the loss of the substrate during the hydroxylation reaction. This idea could also be supported by our current result of the approach of d-camphor to the heme iron upon the Pdx binding. Based on the 19 F NMR study, Crull et al. (62) reported that the reduction of the heme iron from the ferric to ferrous state induces the approach of d-camphor to the heme iron. They suggested that the shorter distance between d-camphor and the heme iron might reflect the higher affinity for d-camphor in the ferrous state than the ferric state (63). Therefore, the Pdx-induced positional change of d-camphor would result in the tight binding of the substrate in the heme pocket, which avoids the loss of the substrate in the hydroxylation reaction.
The approach of d-camphor to the heme-iron would facilitate the oxygen transfer reaction. It is known that an active intermediate of monooxygenation reaction conducted by P450 is a ferryl porphyrin -cation radical, namely compound I. The proximity of the oxygenation site of d-camphor and the reactive oxygen on the heme-iron is crucial for the monooxygenation reaction due to the increased interaction with the oxygen -orbital. The advantages of the conformational changes in the monooxygenation reaction are supported by theoretical studies on the oxygenation mechanism by compound I (64,65). The first step in the proposed oxygenation mechanism of substrate by compound I, the so-called radical rebound mechanism, is the hydrogen atom abstraction from the substrate (66). Yoshizawa et al. (64) theoretically estimated the structural changes in the trajectory runs for the hydrogen atom abstraction from the substrate (64), and reported that, in the transition state of the hydrogen atom abstraction step, the distance between the oxygen atom of compound I and the carbon atom, the oxygenation site, of the substrate is ϳ2.5 Å, which is shorter than that in the initial state (2.9 Å). Because the oxygenation site, C5, of d-camphor is positioned 3.1 Å from the iron-bound oxygen atom in the crystal structure of compound I in P450cam (67), the approach of C5 position of d-camphor to the heme-iron by ϳ0.5 FIG. 9. Correlations between the redox potentials and the chemical shifts of the signal a observed for P450cam bound with various substrates. "⌬ redox potential" and "⌬ chemical shift of the signal a" represent the shifts of the redox potentials and chemical shifts, respectively, of P450cam complexed with various substrates relative to those complexed with d-camphor. The circles represent the correlation between the redox potentials and the chemical shifts of the signal a, and the line is the leastsquare linear fitting. The numbers in the plot indicate the data points for wild-type P450cam in the presence of d-camphor (1), 3-bromocamphor (2), adamantanone (3), d-fenchone (4), and norcamphor (5) and in the absence of the substrate (6). The redox potential data are taken from elsewhere (3,35,57,68,70). Å is consistent with the result from the theoretical calculation. We, therefore, conclude that the approach of the hydroxylation site of the substrate to the reactive oxygen of compound I upon the binding of Pdx facilitates the hydrogen atom abstraction from the substrate, resulting in the preference for the hydrogen atom abstraction of the substrate to the uncoupling reaction and the acceleration of the oxygen transfer reaction. We emphasize that theoreticians could investigate the oxygenation mechanism by compound I using our proposed structure as the initial state to clarify the effect of the Pdx binding on the oxygenation mechanism by compound I.
The essential effects of the relative position of the substrate to the heme-iron on the monooxygenation reaction are also evident by the reaction with norcamphor. The oxygenation site of d-camphor would be located in the close proximity to the reactive oxygen on the heme-iron for the monooxygenation reaction, whereas the oxygenation site of norcamphor is not located near the heme-iron as that of d-camphor (68), and the ratio of the uncoupling reaction producing hydrogen peroxide and water is increased (69). We, therefore, conclude that the Pdx-induced conformational changes facilitate the oxygen transfer reaction to avoid the uncoupling reaction, and the structural changes around the heme would be induced by the specific interaction with the physiological electron donor, Pdx, not by other electron donors.
In summary, we assigned some ring-current-shifted NMR signals derived from d-camphor and Thr-252 by the two-dimensional NOESY and one-dimensional NMR spectra of the substrate free, substrate analogues bound forms of P450cam and the Thr-252 mutant. From the chemical shifts of these NMR signals, we suggested that the binding of Pdx tilts the heme plane of P450cam, leading to the movement of d-camphor and the axial Cys to the heme-iron by 0.1-0.5 Å. The positional perturbation of the axial Cys toward the heme-iron could enhance the axial push effect, which facilitates the heterolysis of the O-O bond. The significant approach of d-camphor to the heme-iron would facilitate the monooxygenation reaction that includes the oxygen transfer reaction from compound I to the oxygenation site of the substrate. We propose that Pdx-induced specific conformational changes of P450cam would facilitate the oxygen activation reaction as well as the ET to oxy-P450cam.