The p67 phox Activation Domain Regulates Electron Flow from NADPH to Flavin in Flavocytochromeb 558 *

An activation domain in p67 phox (residues within 199–210) is essential for cytochromeb 558-dependent activation of NADPH superoxide (O⨪2) generation in a cell-free system (Han, C.-H., Freeman, J. L. R., Lee, T., Motalebi, S. A., and Lambeth, J. D. (1998) J. Biol. Chem. 273, 16663–16668). To determine the steady state reduction flavin in the presence of highly absorbing hemes, 8-nor-8-S-thioacetamido-FAD (“thioacetamido-FAD”) was reconstituted into the flavocytochrome, and the fluorescence of its oxidized form was monitored. Thioacetamido-FAD-reconstituted cytochrome showed lower activity (7% versus 100%) and increased steady state flavin reduction (28 versus <5%) compared with the enzyme reconstituted with native FAD. Omission of p67 phox decreased the percent steady state reduction of the flavin to 4%, but omission of p47 phox had little effect. The activation domain on p67 phox was critical for regulating flavin reduction, since mutations in this region that decreased O⨪2generation also decreased the steady state reduction of flavin. Thus, the activation domain on p67 phox regulates the reductive half-reaction for FAD. This reaction is comprised of the binding of NADPH followed by hydride transfer to the flavin. Kinetic deuterium isotope effects along with K m values permitted calculation of the K d for NADPH. (R)-NADPD but not (S)-NADPD showed kinetic deuterium isotope effects on V and V/K of about 1.9 and 1.5, respectively, demonstrating stereospecificity for theR hydride transfer. The calculated K d for NADPH was 40 μm in the presence of wild type p67 phox and was ∼55 μm using the weakly activating p67 phox (V205A). Thus, the activation domain of p67 phox regulates the reduction of FAD but has only a small effect on NADPH binding, consistent with a dominant effect on hydride/electron transfer from NADPH to FAD.

When the cytochrome is purified from neutrophil plasma membranes, the flavin dissociates and activity is lost, but activity and FAD binding can be restored by incubating the enzyme with phospholipids and FAD (4,6,7).
The flavocytochrome shows no activity in the absence of the cytosolic regulatory proteins p47 phox , p67 phox , and Rac, a small GTP-binding protein (12)(13)(14)(15)(16). Upon cell activation by exposure to bacteria or chemical activators, these proteins assemble on the plasma membrane in a 1:1:1 ratio with flavocytochrome b 558 (17)(18)(19). In resting cells, p47 phox and p67 phox are in a large cytosolic complex (20) along with one or more additional proteins and may translocate en bloc to the membrane. Rac translocates independently of the other cytosolic components (21)(22)(23)(24) but can bind directly p67 phox (25)(26)(27)(28). Details of many of the high affinity protein-protein interactions among cytosolic and some membrane components have been described (29) and involve SH3 domains on p47 phox and p67 phox and target prolinerich sequences in partner proteins.
A key unanswered question has to do with the mechanism by which assembly of the cytosolic proteins with the cytochrome regulates activity. A cell-free model system consisting of the purified flavocytochrome and recombinant cytosolic factor has proven useful (30), and recent evidence points to unique specialized functions for each of the cytosolic regulatory proteins. p47 phox functions as a "regulated adaptor protein," helping to provide binding sites for the other cytosolic factors (31,32). Although it is not essential for cell-free activity, it enhances the affinity of p67 phox and Rac by 2 orders of magnitude. Sitedirected mutagenesis supports a model in which Rac has multivalent interactions, binding simultaneously to p67 phox , membrane, and flavocytochrome through distinct regions on Rac (33)(34)(35). Data are consistent with a model in which Rac, like p47 phox , functions as an adaptor protein, participating in the binding of another essential component(s).
Recent data point to p67 phox as the essential factor that activates electron transfer within the flavocytochrome. We recently identified an "activation domain" within p67 phox that is essential for NADPH oxidase activity (36). Truncation mutants identified this region within residues 199 -210, and a single point mutation at residue 204 completely eliminated NADPH oxidase activity without affecting specific interactions of p67 phox with p47 phox or Rac, or the assembly of the mutant p67 phox within the NADPH oxidase complex. We propose that the activation domain on p67 phox directly activates a particular step in the electron transfer pathway depicted above in Scheme I. The present studies were undertaken to identify the step that is regulated by the activation domain on p67 phox . We provide evidence that the activation domain on p67 phox regulates the reduction of FAD by NADPH but does not affect the binding of NADPH itself, consistent with the regulation of the NADPH 3 FAD hydride/electron transfer reaction.
Preparation of 8-Nor-8-S-thioacetamido-FAD-The starting material, 8-chlororiboflavin was generously provided by Dr. Dale E. Edmondson, Emory University. This was converted to 8-chloro-FAD using partially purified FAD synthetase from Brevibacterium ammoniagenes (7). The appropriate 8-mercapto-FAD was prepared just before use by reaction of 8-chloro-FAD, buffered at pH 8.0, with 5 mM Na 2 S. Excess unreacted sodium sulfide was removed by P-2 column (2 ϫ 70 cm) chromatography. The 8-mercapto-FAD was reacted with excess iodoacetamide for 1 h at 25°C in the dark, and unreacted iodoacetamide was removed by P-2 column chromatography. The resulting analog showed an absorption spectrum identical with that previously published for 8-nor-8-S-thioacetamido-FAD (37), and its concentration was determined using an extinction coefficient of 26 mM Ϫ1 cm Ϫ1 at 475 nm. The fluorescence excitation spectrum corresponded to the absorption spectrum of 8-nor-8-S-thioacetamido-FAD, indicating that this was the fluorescent species. The identity of the FAD analog was confirmed as 8-nor-8-S-thioacetamido FAD by liquid chromatography mass spectrometry (Finnigan-MAT Corp). No mass peaks consistent with sulfoxide and sulfone forms which might be derived from oxidation of the sulfur on the flavin were observed. Absorption and emission spectra were recorded by Hitachi model U-3200 spectrophotometer and Hitachi model F-3000 spectrofluorimeter, respectively.
Preparation of Deuterated NADPH-Deuterated pyridine nucleotides were prepared as in Refs. 38 -41 with some modifications. The cyano form of NADP ϩ was prepared and was then converted into D-NADP ϩ . KCN (1 M) in 5 ml of D 2 O was added to 280 mg of dry NADP ϩ , 0.2 ml of 1 M KOD (in D 2 O) was added, and the solution was incubated for 2 h in the dark. D 2 O (35 ml) containing 11 mmol of KH 2 PO 4 was added, and HCN was removed by gentle bubbling with N 2 in order to convert cyano NADP ϩ into monodeuterated NADP ϩ . The concentration was determined spectrophotometrically using an extinction coefficient of 18 mM Ϫ1 cm Ϫ1 at 260 nm. To prepare the R (A) monodeuterated D-NADPH, a 2-fold molar excess of glucose 6-phosphate was added to 25 ml of D-NADP ϩ along with 50 g of glucose-6phosphate dehydrogenase (Leuconostoc mesenteroides), and reduction was followed spectrophotometrically at 340 nm while maintaining the pH at 7.8 -8.0 with 1 M KOH. To prepare the S (B) form of D-NADPH, a 4-fold molar excess of isocitrate and 5 mg of isocitrate dehydrogenase was added to 25 ml of D-NADP ϩ , and the reduction was monitored and the pH was maintained as above. D-NADPH was purified by fast protein liquid chromatography using a BioScale Q-20 anion exchange column (42) using a linear gradient from 0 to 1 M LiCl, pH 7.8; NADPH elutes at 160 mM LiCl. Fractions showing an A 260 /A 340 ratio of Ͻ2.3 were pooled and lyophilized, and LiCl was removed by washing the dried powder with methyl alcohol. The sample was dried under vacuum and stored at Ϫ20°C until needed. Samples were analyzed by mass spectrometry to confirm isotopic purity.
Preparation of Plasma Membrane, Flavocytochrome b 558 , and Recombinant Cytosolic Proteins-Human neutrophils were isolated from peripheral blood of healthy donors, and plasma membranes were prepared as described (43,44). Purification of flavocytochrome b 558 from solubilized plasma membranes and reconstitution of flavin-depleted cytochrome b 558 with either native FAD or 8-thioacetamido-FAD (heme/ flavin ϭ 2:1, mol ratio) in the presence of both phospholipids (PC/PE/ PI/SM/cholesterol ϭ 4:2:1:3:3, w/w; lipids/protein ϭ 50:1, w/w) and n-octyl glucoside (40 mM) was performed as described (7). Recombinant p47 phox and wild type p67 phox were expressed in Sf9 insect cells and were purified according to Refs. 17 and 45. Rac1 was expressed in DH 5␣ cells as a glutathione S-transferase fusion protein and was purified by binding to glutathione-Sepharose followed by thrombin cleavage (46). Truncated and point-mutated versions of p67 phox developed previously (36) were expressed in Escherichia coli and purified by glutathione-Sepharose affinity chromatography with elution using glutathione (27). Samples were dialyzed to remove free glutathione. Protein concentrations were determined according to Bradford (47).
Assay of Superoxide Generation-NADPH oxidase activity was assayed spectrophotometrically using superoxide dismutase-inhibitable cytochrome c reduction (44), in a Thermomax Kinetic Microplate reader (Molecular Devices, Menlo Park, CA). Rac was preloaded with 5-fold molar excess of GTP␥S for 15 min at room temperature in the absence of MgCl 2 (46). For standard assay conditions, the cell-free reaction mixtures included purified cytochrome b 558 (275 nM) reconstituted with either native FAD or thioacetamido-FAD, 850 nM p47 phox , 900 nM p67 phox , 950 nM Rac1 preloaded with GTP␥S, and 200 -240 M arachidonate in a total volume of 50 l. Four 10-l aliquots of each reaction mixture were transferred to 96-well microassay plates and preincubated for 5 min at 25°C. For each well, 240 l of solution containing 0.2 mM NADPH and 80 M cytochrome c in buffer A (20 mM Tris-HCl, pH 7.4, 50 mM NaCl, 4 mM MgCl 2 , and 1.25 mM EDTA) was added to initiate the reaction. Cytochrome c reduction was quantified by monitoring the absorbance increase at 550 nm using an extinction coefficient of 21 Calculation of Kinetic Data-Reported Michaelis-Menten kinetic parameters were determined using a non-linear least squares fit of the data, programmed in Sigma Plot.
Spectrophotometric and Fluorometric Assays-Heme content was determined by reduced minus oxidized difference spectroscopy at 424 -440 nm using an extinction coefficient of 161 mM Ϫ1 cm Ϫ1 (48). The flavin content of FAD analog-reconstituted cytochrome b 558 was estimated fluorimetrically. Fluorescence spectra were recorded with a Hitachi model F-3000 spectrofluorimeter. Fluorescence changes at 525 nm induced by NADPH-FAD analog oxidoreduction during cell-free NADPH oxidase activation occurred slowly for about 5 min, and the total fluorescence change due to the complete reduction of the FAD analog was measured by adding a few crystals of sodium dithionite. To calculate the percent reduction of the FAD analog at steady state, the fluorescence change at 525 nm attributable to NADPH oxidation was subtracted from that due to oxidoreduction of NADPH and the FAD analog. The time course of heme reduction was derived from the absorbance changes at 558 minus 540 nm, using an extinction coefficient of 21.6 mM Ϫ1 cm Ϫ1 (48).

Reconstitution of Activity with 8-Thioacetamido-FAD-Dur-
ing detergent solubilization and purification, the flavocytochrome b 558 loses flavin so that in its purified form it lacks FAD and has no detectable activity. Activity can be restored using either native FAD (4, 6) or 8-substituted FAD analogs (7). Reconstitution of activity with flavin requires phospholipids, and we previously found that a mixture of PC, PE, PI, SM and cholesterol (4:2:1:3:3, w/w) was most effective in maximally reconstituting activity (7). Under these conditions, maximal activity could be achieved at 0.4 -0.5 flavin per heme (mol/mol), consistent with the previously reported ratio of two hemes/ flavin (7,8). Cytochrome b 558 reconstituted with FAD and lipids exhibited high activity in the cell-free O 2 . generation assay, with turnover numbers from 50 to 75 mol of O 2 . /s/mol of heme (not shown). Fig. 1 shows the effect of 8-thioacetamido-FAD on O 2 . generation. Activity increased linearly up a molar ratio of 0.38 flavins per heme (mol/mol) and remained constant thereafter. The maximum rate was about 7% of that seen with native FAD. These data indicate that 8-thioacetamido-FAD binds tightly to the FAD-binding site in flavocytochrome b 558 but that the bound flavin analog is less efficient than native FAD in transferring electrons from NADPH to superoxide. Steady State Reduction of 8-Thioacetamido-FAD and Heme-The steady state reduction of 8-thioacetamido-FAD by NADPH was assessed fluorophotometrically. Free 8-mercapto-FAD shows very weak fluorescence (7), but reaction of the 8-mercapto-FAD with iodoacetamide results in the production of the fluorescent 8-mercapto-FAD derivative which, unlike native FAD, retains fluorescence when bound to flavocytochrome b 558 (49). As shown in Fig. 2, the flavin analog displays a fluorescence emission maximum at 520 nm, and reduction with dithionite resulted in loss of fluorescence. Cytochrome b 558 was incorporated into phospholipid/octyl glucoside mixed micelles to provide the required phospholipid while minimizing turbidity. The excitation maximum of 8-thioacetamido-FAD bound to the flavocytochrome is 475 nm, 25 nm red-shifted compared with that of native FAD. This longer wavelength excitation maximum minimizes overlap with the NADPH fluorescence excitation spectrum.
Steady state fluorescence of the thioacetamido-FAD reconstituted into flavocytochrome b 558 was analyzed in the presence of the cytosolic regulatory factors p67 phox , p47 phox , and Rac1-GTP␥S as well as the activator sodium dodecyl sulfate. The reduction state of heme was also monitored using the increased absorbance at 558 nm minus the absorbance at the isosbestic point 540 nm. Addition of NADPH (0.10 mM) resulted in the partial reduction of the flavin analog (Fig. 3, panel A). The fluorescence decreased approximately linearly for 5 min, after which there was a much more gradual further bleaching of fluorescence. Superoxide generation under these conditions was approximately linear for more than 10 min, indicating that the leveling off in fluorescence change was not due to a loss in activity of the enzyme. Based on the turnover rate of the enzyme with this flavin analog and the predicted oxidation of NADPH (based on the known turnover number), this second phase of slow fluorescence change is due to the oxidation of NADPH, which shows modest fluorescence at this wavelength. Complete reduction of flavin was achieved by adding a few crystals of sodium dithionite to generate a maximal fluorescence change (arrows in Fig. 3). The steady state reduction levels were calculated based on the percent fluorescence bleaching achieved at 5 min, the approximate intersection of the rapid and slow phases, correcting for the decrease in fluorescence contributed by NADPH oxidation. Based on this calculation, the fraction reduction of flavin after steady state has been achieved is 28 Ϯ 3% (Table I). In contrast to flavin reduction, addition of NADPH produced Ͻ2% steady state reduction of heme based on absorbance changes at 558 nm minus 540 nm ( Table I).
Effect of Cytosolic Regulatory Proteins on the Reduction of Flavin and Heme-The steady state percent reduction of the FAD analog and heme was determined as above in the complete system or in the absence of either p47 phox or p67 phox (Table I). When p47 phox was omitted, there was still significant reduction of flavin (21% compared with 28%). However, when p67 phox was omitted (Table I and Fig. 3, panel B), the flavin was almost completely oxidized. The steady state reduction of flavin correlated with the rate of O 2 . generation under the same conditions (Table I), indicating a functional relationship between flavin reduction and O 2 . generation. In contrast, heme was essentially completely oxidized regardless of the presence of the cytosolic regulatory proteins (Table I).
Role of the Activation Domain in p67 phox in Flavin Reduction-Previous studies (36) imply that an activation domain within residues 199 -210 on p67 phox activates electron transfers within the flavocytochrome. The truncated p67 phox -  failed to support detectable O 2 . generation (Fig. 4, panel A) and resulted in a very low steady state reduction of 8-thioacetamido-FAD (Fig. 4, panel B). Two forms of p67 phox mutated within the activation domain, p67 phox (V204A) and p67 phox (V205A), were evaluated for their effects on activity and steady state reduction of flavin (Fig. 3, panels C and D). The V204A mutation shows essentially no activity, and the V205A form shows low activity. These mutant forms of p67 phox supported very low steady state reduction of FAD (Fig. 4, panel B).
The activation domain is not involved in the interaction with Rac1 or p47 phox , and p67 phox mutated in this region assembles normally within the NADPH oxidase complex (36). Thus, mutation of the activation domain suppresses the reduction of flavin by NADPH in flavocytochrome b 558 . Kinetic Deuterium Isotope Effects on NADPH Oxidase Activity--One attractive hypothesis is that the activation domain on p67 phox regulates either the binding of NADPH itself or another pre-isotopic step (e.g. a conformational change that might juxtapose the pyridine nucleotide and FAD so as to facilitate electron transfer). The K m for a substrate (in this case NADPH) is a complex kinetic term that can differ from the actual binding constant due to the contribution of kinetic terms for post-binding steps. However, a recent treatment of kinetic deuterium isotope effects by Klinman and Matthews (50) led to a simple expression (Equation 1) that allows the direct deter- p67 phox Activation Domain Regulates Electron Flow mination of the K d from K m values plus kinetic deuterium isotope effect data.
The kinetic isotope effect on the rate was first measured at saturating substrate concentration using (monodeuterated) (S)-NADPD versus (R)-NADPD. The R form showed an isotope rate effect of approximately 2, whereas the S form did not show a significant isotope rate effect. Fig. 5, panel A, shows a Lineweaver-Burk plot comparing activity using NADPH and (R)-NADPD. These data demonstrated little if any isotope effect on the K m value but an approximately 2-fold effect on V m . Similar experiments were carried out using p67 phox -(1-210) (this truncated form is referred to as t-p67 phox ) and t-p67 phox (V205A). Truncation has little effect on activity, but the V205A mutation reduces activity to 15-20% of normal. Similar isotope effects were seen regardless of which form of p67 phox was used. Isotope effects on V and V/K in three separate experiments using three different cytochrome or membrane preparations are summarized in Table II. Results were similar using either plasma membranes as a source of cytochrome b 558 or purified cytochrome b 558 reconstituted with native FAD. Purified cytochrome b 558 shows slightly weaker binding of NADPH than does the plasma membrane preparation, but results are otherwise internally consistent. As can be seen, the affinity of NADPH for the enzyme was relatively unaffected by truncation of p67 phox . A small effect of the V205A mutation on the K d (typically 1.2-1.3-fold) was consistently seen in all preparations, but this is insufficient to account for the 3-4-fold decrease in V m seen with this mutation. Thus, the mutation in the activation domain of p67 phox does not have a major effect on the affinity of the enzyme for NADPH. Superoxide Generating Activity Using Pyridine Nucleotide Analogs-The ability to observe a kinetic deuterium isotope effect does not support the idea that NADPH is a "sticky substrate," defined by Cleland (51), as one that reacts to give products as fast or faster than it dissociates from the enzyme. To test further whether product dissociation could be ratedetermining, a series of analogs of NADPH were used as electron-donating substrates, and kinetic parameters were determined as in Fig. 5. Results are summarized in Table III. As shown, for NADPH, deamino-NADPH, and NADH, the V max TABLE I Effects of cytosolic components on NADPH oxidase activity and on steady state reduction of flavin and heme 8-Thioacetamido FAD was reconstituted into purified cytochrome b 558 in the presence of phospholipids as described under "Experimental Procedures." NADPH-dependent superoxide generation was monitored in the presence or absence of p47 phox and p67 phox . Under the same conditions using the complete system, native FAD gave a turnover of 4.88 mol O 2 . /min/nmol heme. The extent of flavin analog reduction (Ef) during turnover is expressed as a percentage of the total fluorescence decrease at 525 nm obtained when dithionite was added, using the following expression, where ⌬F is the fluorescence change observed 5 min after addition of 100 M NADPH, and ⌬F t is the total fluorescence change measured after addition of dithionite. The correction factor 0.26 is due to the estimated contribution to the fluorescence change due to NADPH oxidation, which was determined in a separate experiment from the decreased absorbance at 340 nm corrected for changes at this wavelength due to flavin reduction. The steady state heme reduction level was determined as a percentage of the total dithionite-reducible heme using the reduced minus oxidized absorbance at 558 minus 540 nm.  was essentially identical, whereas the K m varied over a 34-fold range. 2 Thus, for two substrates that bound more weakly than NADPH based on K m values, the V max was not increased as would be predicted if NADPH were behaving as a sticky substrate. DISCUSSION A model has been proposed that attempts to explain individual roles for cytosolic proteins during the protein assembly associated with activation of the respiratory burst (see Introduction). According to this model, it is p67 phox that directly regulates the rate-limiting transfer of electrons within the gp91 phox subunit through its activation domain within the 198 -210 region. In the present study, we have investigated the influence of this region on regulating the rate of specific catalytic steps involved in transferring electrons from NADPH to O 2 . The reductive half-reaction (Reaction 1) and reoxidative half-reaction (Reaction 2) with respect to FAD within gp91 phox are summarized as follows.
We first used steady state kinetics to investigate whether the activation domain in p67 phox stimulates the reductive halfreaction (hypothetical activator 1, above) or the reoxidative half-reaction (hypothetical activator 2, above). If the former were the case then the p67 phox should increase the steady state reduction level of FAD, and mutations should lead to a more oxidized state. The opposite should be true if p67 phox were functioning like activator 2, above. In addition, the heme should become more reduced. Thus, monitoring the steady state reduction of flavin and heme during turnover will distin- 2 The K m value for NADH in our study was approximately 5-6-fold higher than some other published K m values. The reason for this is not clear at this time, although our data seem clear on this point. Differences may have resulted from differences in preparation or purity of cytochrome b 558 and/or cytosolic factors, or other assay-specific differences such as activator concentration, presence of GTP␥S, etc. The two point mutants of p67 phox , V204A and V205A, were constructed using p67 phox -(1-210). The other assay conditions were the same as described in Fig. 3 and Table I guish between these two models.
Two technical problems limit the accuracy and feasibility of monitoring the reduction of native FAD. The extinction coefficient for reduction of FAD is low compared with that of heme, and absorbance changes occur in areas of overlap with the heme. In addition, an earlier study using a high turnover preparation of purified flavocytochrome b 558 showed that the flavin remains mostly oxidized during turnover (52), 3 which makes it difficult to see perturbations of the reduction state. These problems were both overcome by the use of a fluorescent FAD analog, 8-thioacetamido-FAD. In addition to providing a robust fluorescence signal that was not affected by the heme, the FAD analog supported a slower turnover. By using flavocytochrome b 558 reconstituted with 8-thioacetamido-FAD, the steady state reduction of FAD was 28%, compared with reported values of under 10% in the studies of Koshkin et al. (52). These features make it straightforward to determine the effect of wild type and mutated forms of p67 phox on the steady state reduction of the flavin analog.
This approach provides clear data indicating that the flavin reductive half-reaction is regulated by p67 phox and, more specifically, by the activation domain in p67 phox . In the absence of p67 phox , the flavin is nearly fully oxidized, whereas in the presence of p67 phox , the flavin becomes partially reduced. p47 phox had little effect on the reduction state of either flavin or heme. The latter was essentially fully oxidized during all turnover conditions, consistent with previous reports that indicate that reduced heme is reoxidized extremely rapidly by molecular oxygen (48,52). Our data are in agreement with those of Cross and Curnette (53) concerning the role of p67 phox in regulating the reduction of flavin but are not consistent with their suggestion of a role for p47 phox in regulating heme reduction.
The flavin reductive half-reaction can be further subdivided into the following steps, shown in the upper part of Scheme II. The first step is the binding of NADPH to the enzyme and is described by the on and off rate constants k 1 and k 2 . A second step, described by k 3 and its reverse, k 4 , is included for completeness and describes for example a hypothetical conformational or other non-chemical rearrangements that occur prior to flavin reduction itself. The final step (k 5 ), the hydride transfer from NADPH to FAD, is the isotopically sensitive step.
The present studies have utilized deuterium kinetic isotope effects to investigate this reaction. As described above, using Equation 1, it is possible to derive the actual K d for NADPH binding to the enzyme from the K m plus the deuterium isotope effects on K and V/K. As described by Klinman and Matthews (50), K d in this case is defined as the dissociation constant of substrate from all preisotopic complexes. In a case in which binding is the only preisotopic step (i.e. there is no conformational or other change prior to hydride transfer), then K d ϭ k 2 /k 1 , whereas if there is a conformational rearrangement after binding but prior to the isotopic step, then K d includes the kinetic constants for both of the pre-isotopic steps, and is defined in Equation 2.
We observed a small effect (on the order of 20%) of the p67 phox V205A mutation on the K d for NADPH (Table II). Although seen in three experiments, the effect were not of sufficient magnitude to account for the much larger effect of this mutation on the V max . Thus, the K d is affected only minimally, so it is unlikely that any of the preisotopic steps described by k 1 , k 2 , k 3 , or k 4 is significantly regulated by p67 phox . A possible exception is that k 2 and k 1 are both increased by p67 phox , leaving the K d unaffected. However, if k 1 were rate-limiting, then increasing k 1 to the point where it was no longer rate-limiting (or was less rate-limiting) should unmask a larger isotope effect on V max . Thus, these data are most consistent with the regulation by p67 phox of the flavin reduction step, i.e. activation of k 5 or inhibition of k 6 . It was also possible that product dissociation from the prior catalytic cycle might be the step that is regulated by p67 phox . In the limiting case where product dissociation is 3 In contrast, an earlier study (53) reported a 50% steady state reduction of FAD. However, it is not clear whether the partially purified preparation used demonstrated a high turnover rate nor whether artifacts may have arisen due to overlap of the flavin spectrum with the high extinction coefficient hemes. Our studies (Y. Nisimoto and J. D. Lambeth, unpublished data) are in agreement with those of Koshkin et al. (52) and emphasized to us the difficulty of attempting to monitor FAD reduction state in the presence of the highly absorbing heme.

Summary of kinetic isotope effects and K d values for NADPH binding
to the respiratory burst oxidase Superoxide generation was monitored using superoxide dismutaseinhibitable cytochrome c reduction as described under "Experimental Procedures." Kinetic constants were obtained from experiments such as those shown in Fig. 5 and were used to derive D V (the isotope rate effect on V max ) and D (V/K) (the isotope rate effect on V max /K m ). The K d values were calculated using Equation 1. Plasma membrane (PM) was used as the source of the cytochrome in experiments 1 and 2, and purified, FAD-reconstituted cytochrome (CYT) was used in experiment 3. Truncated p67 phox refers to p67 phox -(1-210), and V205A refers to this mutation in the truncated form of p67 phox . p67 phox Activation Domain Regulates Electron Flow the single slow step, this should suppress the isotope rate effect. The ability to observe a significant isotope effect in V and V/K therefore argues against this interpretation. In addition, if product dissociation were rate-limiting, then the products of more weakly binding analogs of NADPH should dissociate more readily from their binding sites, increasing the V max . This is not the case (Table III); the V max is nearly the same for three analogs of NADPH whose K m values vary 40-fold. These data indicate that product release is not rate-limiting nor is it regulated by p67 phox . Thus, the activation domain of p67 phox regulates the reduction of the flavin by NADPH. There is little effect on the affinity of NADPH for forms of the enzyme prior to the hydride transfer step, and data are therefore most consistent with regulation by p67 phox of hydride/electron transfer from NADPH to flavin.