Use of an affinity label to probe the function of the NADPH binding component of the respiratory burst oxidase of human neutrophils.

The respiratory burst oxidase of neutrophils can be activated in a cell-free system in which solubilized membranes, cytosol, and Mg2+ are required and in which sodium dodecyl sulfate is used to convert the dormant oxidase to an active form. The 2',3'-dialdehyde analog of NADPH was used as an affinity label for the cytosolic NADPH binding component of the respiratory burst oxidase from human neutrophils. When treated with this affinity label in the presence of sodium cyanoborohydride to reduce Schiff bases, neutrophil cytosol was shown to lose at least 90% of its activity in the cell-free system. In contrast to normal cytosol, treated cytosol had lost its ability to abolish the lag time required for activation of the oxidase, suggesting that the treated cytosol was no longer able to participate in the rate-limiting activation step. Furthermore, the treated cytosol had lost its ability to convert the oxidase from a form with a high Km to a form with a low Km for NADPH. The ability of dialdehyde-treated cytosol to activate the oxidase could be restored by untreated cytosol with a concentration dependence suggesting that only one kinetically active component of the oxidase was inhibited by treatment with the NADPH analog. Like the dialdehyde-treated cytosol, cytosols from patients with chronic granulomatous disease caused by a deficiency in a cytosolic Mr = 47,000 protein (pp47) fail to participate in the rate-limiting activation step (Curnutte, J. T., Scott, P. J., and Babior, B. M. (1989) J. Clin. Invest. 83, 1236-1240). These chronic granulomatous disease cytosols were nevertheless able to restore limited activity to the dialdehyde-inactivated cytosol in a cell-free activation system. These results are consistent with a model in which (a) the NADPH binding subunit of the oxidase exists in a very slowly dissociating complex with one or more additional cytosolic components, including pp47, and (b) the NADPH binding component of the oxidase controls the affinity of the enzyme for NADPH, either directly or through the binding of additional cytosolic factors.

ated in stimulated neutrophils by a membrane-associated oxidase that catalyzes the one-electron reduction of oxygen to O;, utilizing NADPH as the preferred electron donor according to the reaction: This oxidase is dormant in unstimulated cells, but can be converted to a catalytically active form by exposure of the cell to any of a wide variety of stimuli. The abrupt change in phagocyte oxygen metabolism that follows the activation of this enzyme is termed the respiratory burst (2), and the enzyme itself is called the respiratory burst oxidase.
Despite extensive investigation, the exact mechanism by which the oxidase is activated is incompletely understood. However, the recent development of cell-free oxidase activation systems (3)(4)(5)(6) have provided a system with which the biochemical processes leading to activation can be identified. Experiments using these cell-free systems have demonstrated that activation requires the simultaneous presence of components contributed by cell membranes (either particulate or solubilized (3,7 ) ) and components contributed by the cytosol. In these systems, activation of the oxidase is initiated by the addition either of a detergent (e.g. arachidonate (3)(4)(5)(6) or SDS' (7)) or of protein kinase C plus ATP (8).
Studies validating these cell-free oxidase activation systems have been performed using constituents prepared from the cells of patients with chronic granulomatous disease (CGD), a condition characterized by the inability of phagocytes from affected individuals to generate 0; (2,9,10) owing to a defect in one of the components of the respiratory burst oxidase or its activating apparatus. These studies have demonstrated that in some forms of CGD, the defective component is in the membrane (3,7,11,12), while in others it is in the cytosol (13,14). Cytosol from the latter group of CGD patients is incapable of supporting oxidase activation in the cell-free system (13)(14)(15)(16).
Although cytosol factor activity has been eluted from a gel filtration column in fractions corresponding to an apparent molecular mass of 240 kDa ( E ) , attempts to isolate this activity as a single unitary component have not been successful. Studies of the kinetics of activation of the respiratory burst oxidase in a cell-free system have accounted for this failure by demonstrating a nearly third order relationship between the concentration of cytosol and oxidase activity (14, 17). These findings suggest a mechanism in which at least three kinetically distinct cytosolic factors are required for the activation process (17). This mechanism has recently been confirmed by work showing that ( a ) the cytosol "factor" can be separated into at least four different components (18)

12243
(b) CGD can be caused by a deficiency in either of two of these cytosolic constituents (15,16,18). The functions of these cytosolic constituents and the nature of their participation in the activity of the oxidase, however, have not yet been defined.
Using periodate-oxidized NADPH ("NADPH dialdehyde") as a n affinity label, we have obtained evidence strongly suggesting t h a t in resting neutrophils, one of the cytosolic constituents is the NADPH binding component of the respiratory burst oxidase (19). In the present paper, we describe experiments using cytosol in which the NADPH binding component of the oxidase has been irreversibly inactivated by NADPH dialdehyde. This inactivated cytosol was employed as a reagent to probe the kinetic role that t h e NADPH binding component plays in the activation of the oxidase in a cell-free system. Our studies show that the NADPH binding component participates both in the rate-limiting step in oxidase activation and in the modulation of the affinity of the oxidase for NADPH. Results obtained by mixing the inactivated cytosol with cytosol from patients with cytosol-defective CGD (type 11, Scripps classification (11)) further suggest that in resting neutrophils, the NADPH binding component occurs in the form of a slowly dissociating complex that contains at least one other cytosolic oxidase component, and that this complex participates in oxidase activation as a single kinetic species.

MATERIALS AND METHODS
Reagents were the best grade commercially available and were used without further purification. SDS (electrophoresis purity grade) and sodium deoxycholate (UltrolR grade) were obtained from Bio-Rad and Calbiochem, respectively. Neutrophils and neutrophil subcellular fractions from normal donors were prepared as described previously (18). The protein contents of the subcellular fractions have been previously reported as follows (17): cytosol, 250 +. 42 (S.D.) pg/lO' cell eq; deoxycholate-solubilized membranes, 25.7 & 7.4 Kg/107 cell eq. The cytosol preparations were adjusted prior to aliquoting so that they contained 9 X 10' cell eq/ml.
Cell-free Activation of the Respiratory Burst Oxidase-Activation of the oxidase in the cell-free system was monitored by following the superoxide dismutase-inhibitable reduction of cytochrome c with time at 550 nm in a thermostatted dual beam recording spectrophotometer, using slight modifications of published methods (17). Unless otherwise noted, assay mixtures contained 0.1 mM cytochrome c, 6.5 mM MgC12, 87 mM KCl, 2.6 mM NaCl, 8.7 mM PIPES (pH 7.3), 0.15 mM ATP,' 0.18 mM EGTA, 0.16 mM NADPH, and 0.04 mM SDS. Reactions were performed with 4.17 X lo6 cell eq of deoxycholate-solubilized membranes (0.94 mM final deoxycholate concentration) and 1.33 X 10' cell eq of cytosol per ml of reaction mixture. The reference cuvette also received 45 pg of superoxide dismutase. All components except SDS were mixed in a cuvette and allowed to equilibrate at 25 "C for at least 1 min before activation of the oxidase by the addition of SDS. Following the initiation of the reaction by SDS, the rate of 0; production makes an asymptotic approach to a constant final velocity that represents the rate of 0; production when all the available dormant oxidase has been converted to an active form. This velocity, V,, was calculated from the maximum rate of superoxide dismutase-inhibitable cytochrome c reduction as measured by eye from the strip chart recording and expressed as nanomoles of 0 2 / min/lO' cell eq of membrane.
In the assays in which the concentration of cytosol was varied systematically, incubations were carried out in 96-well plates as previously described (18). Reactions were initiated by the simultaneous addition of SDS to all wells after a 1-min equilibration at room temperature. 0; production was followed by measuring A550 at 5-s intervals with a kinetic microplate reader ( VmXR, Molecular Devices Corp., Menlo Park, CA), calculating the maximum rate of absorbance change with a kinetic software package (Softmax Release 2.0, Molec-*ATP and EGTA are contained in the buffer used to isolate neutrophil subcellular fractions and are carried over into the assay when cytosol is added. Neither substance has any effect on oxidase activity at the concentrations used (12). ular Devices Corp.) that monitored dV,/dt as a function of time. With the exception of the cytosol content, the compositions of the mixtures were as described above, but the reaction volume was 150 ~1 .
For each sample, V, , the activity of the oxidase, was calculated from the difference between the maximum rates of cytochrome c reduction in the sample reaction and in an otherwise identical reaction containing superoxide dismutase (extinction coefficient 20.5 mmol" cm") (18). Reaction orders were calculated by nonlinear least squares regression, fitting the observed data to the equation v,/p = (cytosol content) Order where p represents a computer-derived proportionality constant.
Preparation of NADPH Dialdehyde and Its Incorporation into the Cytosolic NADPH Binding Component-NADPH dialdehyde was synthesized according to a modification of the procedure of Mas and Colman (20). NADP was oxidized with sodium periodate as described. The resulting mixture was adsorbed to a DEAE-agarose column and eluted with a gradient of triethanolamine HCl (pH 8.0,O.Ol to 0.5 M). The NADP dialdehyde was identified as the peak eluting at 0.15 M triethanolamine HC1. The fractions containing this peak were pooled, reduced to NADPH dialdehyde with isocitric acid and isocitric dehydrogenase, and subjected to ion exchange chromatography to isolate the dialdehyde as described in the original method (20). The NADPH dialdehyde content of the final preparation was calculated assuming an extinction coefficient 6340 of 6.22 mM" cm", and aliquots were stored in 0.3 M triethanolamine buffer (pH 8.0) at -70 "C until use. The final preparation of NADPH dialdehyde yielded a single UV-absorbing spot on TLC analysis (19) and contained a single peak (96% purity) absorbing at both 260 and 340 nm on fast protein liquid chromatography analysis with a Mono Q column.
Covalent linkage of the NADPH dialdehyde to the NADPH binding site of the oxidase was performed by incubation of neutrophil cytosol with 0.1 mM NADPH dialdehyde for 30 min at 4 "C, then adding NaCNBH3 (0.5 mM final concentration) and continuing the incubation for an additional 24 h. As described previously (19), this procedure typically results in the loss of more than 90% of the activity of the cytosol as assayed in the cell-free activation system when compared to cytosol incubated for similar periods with buffer alone, or with buffer and NaCNBH3. Following the 24-h incubation, the dialdehyde-treated cytosol (2-10 ml) was dialyzed at 4 "C against 1liter portions of buffer containing 100 mM KCl, 3 mM NaCl, 3.5 mM MgCL, and 10 mM PIPES (pH 7.3), using M , = 3500 exclusion membranes (Spectropor, Spectrum Medical Industries, Los Angeles, CA) and changing the buffer hourly X 4. The cytosol (protein concentration 260 +-19.8 @g/107 cell eq ( n = 4)) was divided into aliquots and stored at -70 "C until use. In the samples used for complementation analysis (see below), more complete inactivation of the NADPH binding component was accomplished by treating the cytosol with 0.2 mM NADPH dialdehyde and 1.0 mM NaCNBH3 according to the foregoing protocol.
Patients-The neutrophils from three patients with Type 11 cytochrome b-positive CGD were used to provide cytosol deficient in defined oxidase components (11, 13). Other work (15,16,18,(21)(22)(23) suggests that these cytosols are probably all deficient in a M, = 47,000 phosphoprotein that has been shown to be closely related to the respiratory burst oxidase; accordingly, they will be designated as pp47-deficient cytosols. Each of the three patients was known to have a severe deficiency of cytosol factor activity (<4% in each case). One of the pp47-deficient type I1 CGD patients (J. C.) has been previously described (13,14,18).

0; Production by Cell-free Systems Containing NADPH
Dialdehyde-treated Cytosol in the Absence and Presence of Untreated Cytosol-Little oxidase activity was generated by a cell-free system containing dialdehyde-treated cytosol (Fig.  L4, O ) , while oxidase activity (V,) in a system containing normal cytosol increased as a power of the cytosol concentration (Fig. lA, 0 In earlier studies, the exponent r, equal to the reaction order for cytosol, was found to be 2.5 (14, 17). This value has been interpreted in terms of a requirement for at least three cyto- solic factors for the expression of oxidase activity in the cellfree system.
Components of the oxidase other than the NADPH binding component are likely to remain active in the dialdehydetreated cytosol. If so, then in an experiment in which small increments of normal cytosol are added to a system containing large amounts of the inactivated cytosol, the observed reaction order for the normal cytosol should be lower than its usual value of 2.5, because the concentrations of active components present in both the normal and the inactivated cytosol (i.e. components unaffected by the dialdehyde) will change very little and accordingly will not contribute to the apparent reaction order. The results of such an experiment are shown in Fig. 1B Kinetic Characterization of the Role Played by the NADPH Binding Subunit of the Oxidase-Kinetic studies have suggested that oxidase activation occurs in two stages: the slow transformation of the fully dormant oxidase from an incompetent to a competent precursor, followed by the rapid conversion of the competent precursor to the catalytically active enzyme. The interval between the addition of activator to the cell-free system and the attainment of the maximum rate of 0; generation (the "lag") is a manifestation of the first stage (7,14,17), while the second stage is in part expressed as an increase in the apparent affinity of the oxidase for NADPH vious studies with the cell-free activation system, we found that oxidase generated at low cytosol concentrations showed a K , for NADPH of ~1 3 0 PM, while at high cytosol concentrations, the K , for the same substrate shifted to =30 PM (14, 17). The pp47-deficient cytosols could mediate the same shift in K, when used to supplement a cell-free system containing limiting amounts of normal cytosol, even though by themselves the defective cytosols supported little oxidase activity at any concentration (14). This finding suggests that the CGD cytosols possessed normal quantities of the cytosolic factor(s) responsible for this shift. In similar experiments, we tested the ability of the dialdehyde-treated cytosol to alter the K , of the low affinity form of the oxidase. The results (Table I ) show that, like oxidase generated using low concentrations of normal cytosol, the residual oxidase formed with dialdehydetreated cytosol alone had a low affinity for NADPH (i.e. a high K,). Furthermore, the low affinity oxidase generated using limiting amounts of normal cytosol was not converted to a high affinity form when the reaction mixture was supple- (1 X lo7 cell eq of treated or untreated cytosol, or buffer) was combined with solubilized neutrophil membrane and all the components of the cell-free oxidase activation assay except NADPH. At t = 3.5 min, SDS (37 p~ during preincubation, 32 p~ during assay) was added, and the mixture was allowed to incubate at 25 "C. At t = 0, 0; production was initiated by the addition of NADPH together with an aliquot of a second cytosol or cytosol equivalent (1 x lo7 cell eq of treated or untreated cytosol or buffer; final volume, 0.75 ml), and 0; generation was measured by following the reduction of cytochrome c at 550 nm as described under "Materials and Methods." The results shown are representative of three experiments performed using different preparations of treated and untreated cytosol. Curve A shows the course of 0; generation when normal cytosol is added first and treated cytosol second. Curves B and C show 0; generation when treated cytosol or buffer, respectively are added first and normal cytosol second. In all experiments, curue A showed no lag in 0; production, while the usual lag was seen in curues B (dialdehydetreated cytosol first) and C (buffer first). The final rate of 0; generation (i.e. V,) was always greater in incubations containing dialdehyde-treated cytosol than in incubations containing buffer, probably reflecting the ability of the unaffected oxidase components in the dialdehyde-treated cytosol to affect the final amount of oxidase generated, despite their failure to abolish the lag. mented with dialdehyde-treated cytosol; the K,,, of the oxidase in the supplemented reaction mixture remained high despite an increase in its V,,, (the latter probably due to other oxidase components contributed by the dialdehyde-inactivated cytosol).

NADPH
Effect on the Lug-When the oxidase in the cell-free system in activated in the presence of substrate, there is an appreciable delay between the addition of the activating agent (SDS in our assays) and the attainment of full 0 2 generation. This lag, a manifestation of the first stage of oxidase activation, can be abolished by activating the enzyme before adding the substrate. (Activation is accomplished by incubating cytosol, membranes, M$+, and SDS for 3.5 min in the absence of substrate.) An earlier study showed that cytosols from pp47deficient CGD patients were unable to support the first stage of activation, because the lag was not eliminated when systems containing these cytosols were preactivated with SDS before adding NADPH (plus normal cytosol, which was needed in these experiments to furnish the systems with the component missing from the defective cytosols) (14). These findings implied that pp47 was required for the first stage of activation (i.e. the conversion of incompetent to competent oxidase precursor). T o evaluate whether dialdehyde-treated cytosol was capable of eliminating the lag, a cell-free system containing the inactivated cytosol was incubated with SDS in the absence of substrate. 0; generation was then initiated by the addition of NADPH together with an aliquot of normal cytosol. The results (Fig. 2 ) demonstrate that the lag is eliminated by preincubation with untreated cytosol, but not by an identical incubation with dialdehyde-treated cytosol or with buffer alone. These findings suggest that the NADPH binding subunit, like pp47, participates in the first stage of oxidase activation.
Competition for Membrane Binding Sites-The foregoing experiments suggest that the cytosolic NADPH binding component interacts directly with the membrane-associated component(s) of the oxidase. If so, it is possible that the dialdehyde-inactivated NADPH binding component would be able to compete with the active component for membrane binding sites. Competition was tested by examining the effect of the dialdehyde-treated cytosol on 0; production by the cell-free system under conditions in which the membrane-associated oxidase components were limiting. The results are shown in Fig. 3. It is seen that when normal cytosol was used, 0, generation increased progressively with cytosol concentration until the cytosol/membrane ratio reached 10 cell eq of cytosol/ cell eq of membrane. The decline in the slope of the curve at ~2 . 0 X lo7 cell eq/ml of cytosol suggests that at this point, the concentration of membrane in the assay begins to limit the amount of oxidase generated at full activation. Therefore, a cytosol concentration of 1.8 X lo7 cell eq/ml was chosen as the starting point to test the ability of the dialdehyde-treated cytosol to compete with untreated cytosol for limited quan- ( n between 3 and 7 for each data point; experiments were performed with two different preparations of cytosol) of the rates of 0; generation with increasing amounts of untreated cytosol. The open circles show the mean k S.E. of the oxidase activity generated when increasing concentrations of NADPH-dialdehyde-treated cytosol are added to 1.8 X lo7 cell eq/ml of untreated cytosol ( n between 3 and 5 for each data point; all experiments were performed with the same preparations of treated and untreated cytosol). Oxidase activity is expressed as a percent of the activity generated by 1.8 X lo7 cell eq/ ml of untreated cytosol in each experiment. 0; production in incubations containing this quantity of untreated cytosol alone were 76.9 k 7.4 nmol/min X (IO7 cell eq of membrane) for the experiments using untreated cytosol only, and 103.5 & 10.6 nmol/min X (10' cell eq of membrane) for the experiments using mixtures of untreated and treated cytosols.

TABLE I1
Complementation between dialdehyde-treated cytosol and pp47-deficient cytosol Assays were performed as described under "Materials and Methods." Incubations contained 2.66 x lo7 cell eq of dialdehyde-treated cytosol alone ("dialdehyde-treatedj or CGD cytosol alone ('I-") or 1.33 X lo7 cell eq each of the dialdehyde-treated and CGD cytosols

( ' I + " ) .
Note that in every case, 0; production by reaction mixtures containing a combination of pp47-deficient cytosol and dialdehydetreated cytosol exceeded 0; production by reaction mixtures containing either cytosol alone (see text). tities of the membrane components of the oxidase. Fig. 3 shows that when the normal cytosol concentration was increased beyond 1.8 X lo7 cell eq/ml, the amount of oxidase a t full activation continued to rise, although in successively smaller increments with each further increment in cytosol concentration. In contrast, the addition of dialdehyde-inactiuated cytosol to reaction mixtures containing normal cytosol a t 1.8 X lo7 cell eq/ml resulted in a progressive fall in activated oxidase, presumably because the inactivated cytosol interfered with the formation of active oxidase complexes by attaching to some of the membrane binding sites to form inactive complexes. The fall in 0; production caused by the inactivated cytosol was not simply an effect of protein, because 0; production was not inhibited when the inactivated cytosol was replaced with a similar amount of ovalbumin (data not shown).
Complementation between Dialdehyde-treated Cytosol and pp47-deficient Cytosols from Patients with CCD-We used complementation studies to ascertain whether pp47 is identical with the dialdehyde-inactivated component. In these studies, test cytosols were mixed in equal volumes and assayed for their ability to restore activity to the cell-free system (Table 11). In each assay, the combination of pp47-deficient and dialdehyde-treated cytosol was able to support oxidase activity at a level greater than seen with either the dialdehydetreated cytosol or the pp47-deficient cytosol when combined with itself (i.e. when used at a volume equal to the combined volumes of the two cytosols in the complementation assay). The rates achieved through complementation, however, represent only a fraction of the oxidase activity that would be produced by reaction mixtures containing normal cytosol only.

DISCUSSION
The kinetics of activation of the respiratory burst oxidase in the cell-free system has been previously explained according to the following model the speed of this conversion is described by k, a first order rate constant.
It seems likely that the oxidase component inactivated by NADPH dialdehyde is the NADPH-binding subunit of the enzyme (19). The kinetic studies have furnished information as to how the NADPH binding component of the oxidase fits into the above model. The inability of the NADPH dialdehyde-treated cytosol to abolish the lag in activation of the oxidase demonstrates that this cytosol lacks a component required for the first stage of the oxidase activation process. Similarly, the failure of cytosols from pp47-deficient neutrophils to abolish the lag indicates that this component is needed for the first stage of the activation process. If the component inactivated by NADPH dialdehyde is different from pp47, the number of cytosolic components needed for the first stage of activation rises to at least 2. Earlier kinetic studies have shown, however, that the cytosol contributes only a single component to the first stage of the activation reaction (14, 17). To reconcile these observations, it is necessary to postulate that the cytosolic components required for this stage of the activation reaction (i.e. pp47 and the dialdehyde-inactivated component, as well as others (18)) must exist in the form of a complex. In the model, this complex corresponds to component S.
Evidence for the existence of such a complex has been provided by earlier studies. Gel filtration of concentrated neutrophil cytosol has yielded a single peak of M , = 240,000 that contains all the components needed to interact with membrane in the cell-free activation assay (12). A cytosolic preparation capable of supporting oxidase activation in the cell-free system has been obtained by affinity chromatography of neutrophil cytosol over 2',5'-ADP-agarose (24) and GTPagarose (15). Although other data clearly indicate the presence of multiple oxidase components in neutrophil cytosol, the ability of each of these purification methods to furnish a single fraction that is capable of supporting oxidase activation suggests that these components exist as a complex.
The complementation studies were carried out to determine the relationship between the dialdehyde-inactivated component and pp47. Interpretation of these studies, however, is not straightforward. Because the reaction order for cytosol is greater than 1, the usual criterion for judging complementation-namely, that the activity in a reaction mixture containing both components be greater than the sum of activities in reaction mixtures containing each component individuallydoes not necessarily apply. Complementation can be identified regardless of reaction order, however, by using the criterion that the activity of a reaction mixture containing the two cytosols at equal volumes u must exceed the activities of each of the reaction mixtures in which the cytosols are used alone, but at a volume of 2u. This criterion is fulfilled by the complementation studies carried out with dialdehyde-treated cytosol and pp47-deficient cytosol, indicating that pp47 is unlikely to be the NADPH binding component of the oxidase.
Oxidase activity in the complementing reaction mixtures, however, was much lower than expected from the activity of reaction mixtures containing normal cytosol at the same concentration (i.e. 2 u ) . We propose that this result may be attributable to a sluggish rate for the dissociation of S into its components. In such a case, the pp47 in the dialdehydeinactivated cytosol is locked up in an inactive S complex and is therefore unavailable to restore activity to the pp47-deficient cytosol. The observed levels of activity, although significantly greater than expected in the absence of complementation, are only a small fraction of the rates expected if the normal oxidase components present in the two defective cytosols were able to interchange freely.
The inability of the dialdehyde-treated cytosol to convert the oxidase from its low affinity to its high affinity form provides evidence that the NADPH binding component of the oxidase participates in that transformation. This finding provides information about the mechanism by which the affinity of the oxidase is regulated. In the resting cell, the NADPH dialdehyde is known to inhibit the oxidase through occupation of a high affinity binding site (19). Thus, the oxidase must either have a second, low affinity NADPH binding site which is not being inhibited under the incubation conditions used in this study, or the affinity of a single NADPH binding site must initially decrease when the NADPH binding protein is translocated to the membrane and then increase again as additional cytosolic factors combine to form the fully active oxidase. If the first hypothesis is correct, the dialdehydetreated cytosol might be unable to alter the affinity of the oxidase for NADPH because all of the available high affinity binding sites are occupied by NADPH dialdehyde. In the second instance, the inhibited NADPH binding protein contributed by the dialdehyde-treated cytosol may sequester the available cytosolic factors that regulate substrate affinity, preventing them from acting on any uninhibited NADPH binding protein.