Endogenous reactive oxygen intermediates activate tyrosine kinases in human neutrophils.

In response to invading microorganisms, neutrophils produce large amounts of superoxide and other reactive oxygen intermediates (ROI) by assembly and activation of a multicomponent enzyme complex, the NADPH oxidase. While fulfilling a microbicidal role, ROI have also been postulated to serve as signaling molecules, because activation of the NADPH oxidase was found to be associated with increased tyrosine phosphorylation (Fialkow, L., Chan, C. K., Grinstein, S., and Downey, G. P.(1993) J. Biol. Chem. 268, 17131-17137). The mechanism whereby ROI induces phosphotyrosine accumulation was investigated using electroporated neutrophils stimulated with guanosine 5′-O-3-thiotriphosphate in order to bypass membrane receptors. In vitro immune complex assays and immunoblotting were used to identify five tyrosine kinases present in human neutrophils. Of these, p56/59, p72, and p77 were activated during production of ROI. Interestingly, the in vitro autophosphorylation activities of p53/56 and p59 were found to decline with ROI production. The mode of regulation of p56/59 was explored in detail. Oxidizing agents were unable to activate p56/59in vitro and, once activated in situ, reducing agents failed to inactivate it, suggesting that the effects of ROI are indirect. Tyrosine phosphorylation of p56/59 paralleled its activation, and dephosphorylation in vitro reversed the stimulation. We therefore conclude that tyrosine phosphorylation is central to the regulation of p56/59 and likely also of p72, which is similarly phosphorylated upon activation of the oxidase. Because ROI have been shown to reduce the activity of tyrosine phosphatases, we suggest that this inhibition allows constitutively active kinases to auto/transphosphorylate on stimulatory tyrosine residues, leading to an increase in their catalytic activity. Enhanced phosphotyrosine accumulation would then result from the combined effects of increased phosphorylation with decreased dephosphorylation.

Neutrophils play a central role in host protection against infection, killing pathogens by a series of rapid and highly regulated responses. These include chemotaxis, phagocytosis, secretion of anti-microbial agents, and generation of reactive oxygen intermediates (ROI) 1 (reviewed by Sha'afi and Molski (1988)). Production of ROI is mediated by a multicomponent enzyme complex, the NADPH oxidase, present in the membranes of neutrophils and other leukocytes (Morel et al., 1991). Functional assembly of the oxidase facilitates the transfer of one electron from cytosolic NADPH to molecular oxygen, producing superoxide. Dismutation of superoxide in turn generates hydrogen peroxide, and both of these molecules can further generate other reactive oxygen intermediates, including hypochlorous acid, hydroxyl radical, and peroxynitrite ion (Halliwell and Gutteridge, 1990). Although the mechanisms whereby NADPH oxidase-derived ROI attack microbial targets are not completely understood, their importance in host defence is highlighted by a rare genetic disorder, chronic granulomatous disease. Patients afflicted with this disorder lack the ability to produce ROI and, as a result, suffer from chronic and recurring infections that can be lethal (Smith and Curnutte, 1991).
Although neutrophils are probably the most efficient source of superoxide, virtually all eukaryotic cells produce ROI, primarily as side products of electron transfer reactions in mitochondria and the endoplasmic reticulum (Halliwell and Gutteridge, 1985). In addition to their microbicidal role in phagocytes, ROI have been suggested to act as signaling molecules in other cells . In principle, ROI constitute good candidate signaling molecules because they are small, rapidly diffusible, and highly reactive. Moreover, both intra-and extracellular concentrations of ROI can be rapidly scavenged by several enzymes, including superoxide dismutase, catalase, and the glutathione peroxidase system, allowing tight control of ROI concentrations and rapid termination of signals. The notion that reactive, small inorganic molecules can function as intracellular signals is supported by the well established role of nitric oxide in the regulation of vascular tone, neurotransmission, and cell-mediated immune responses (Nathan and Xie, 1994). By comparison, much less is known about the role of ROI in signaling, but suggested targets include the transcription factor NFB , tyrosine phosphatases (Hecht andZick, 1992, Fialkow et al., 1994), and phospholipase A 2 (Zor et al., 1993).
Recent observations suggested a role for ROI in neutrophil signal transduction (Fialkow et al., 1993). Neutrophils stimulated to produce ROI were reported to undergo increased tyrosine phosphorylation of several proteins. Exogenous oxidants were able to mimic this response, whereas anti-oxidants could block it. Several lines of evidence suggested that ROI generated by the NADPH oxidase were responsible for the effect, including the finding that the increased tyrosine phosphorylation failed to occur in neutrophils from patients with chronic granulomatous disease. Inasmuch as tyrosine phosphorylation is an important mediator in the regulation of anti-microbial responses (Berkow and Dodson, 1990;Grinstein and Furuya, 1991), ROI may play an important role in the control of auto/ paracrine signaling at sites of inflammation.
The extent of tyrosine phosphorylation is determined by the activity of two competing enzyme families, tyrosine kinases and phosphatases. Earlier in vitro (Hecht and Zick, 1992) and in vivo (Zor et al., 1993) studies have suggested that ROI can inhibit the activity of certain tyrosine phosphatases by oxidation of a conserved cysteine residue within their catalytic domain. Although the inhibition of tyrosine phosphatases may account for the elevated tyrosine phosphorylation induced by ROI, increased activity of tyrosine kinases could conceivably contribute to the response. Indeed, tyrosine kinases have been reported to be activated in lymphocytes by oxidizing agents (Bauskin et al., 1991;Nakamura et al., 1993). For these reasons, we investigated whether endogenous ROI generated by the NADPH oxidase affected the activity of tyrosine kinases in human neutrophils.

EXPERIMENTAL PROCEDURES
Materials-Ficoll 400, dextran T500, protein A-Sepharose, and Sepharose beads were purchased from Pharmacia. Medium RPMI, K-ATP, GTP␥S, NADPH, sodium orthovanadate (NaV), N-acetylcysteine (NAC), diamide, and dithiothreitol were from Sigma. Hydrogen peroxide was from Fisher. Bovine serum albumin and acrylamide were from Boehringer Mannheim. [␥-32 P]ATP and MOPS were from ICN. Truncated T-cell phosphatase was the generous gift of Dr. C. Diltz (Department of Biochemistry, University of Washington, Seattle, WA). Diphenylene iodonium was synthesized in our laboratory as described previously (Collette et al., 1956).
Cell Isolation and Permeabilization-Neutrophils were isolated from fresh heparinized blood of healthy human volunteers by dextran sedimentation, followed by Ficoll-Hypaque gradient centrifugation. Contaminating red cells were removed by NH 4 Cl lysis. Neutrophils were counted using a Model ZM Coulter Counter, resuspended in HEPESbuffered RPMI medium at 10 7 cells/ml, and maintained in this medium at room temperature until use. To minimize proteolysis following extraction of cells in detergent, the cells (10 7 /ml) were pretreated with 5 mM diisopropyl fluorophosphate (Calbiochem) for 30 min at room temperature. For electroporation, 1.5 ϫ 10 7 cells were sedimented and resuspended in 1 ml of ice-cold permeabilization medium. This suspension was transferred to a Bio-Rad Pulser cuvette and permeabilized with two discharges as described previously (Grinstein and Furuya, 1988). The cells were rapidly sedimented and resuspended in fresh, ice-cold permeabilization medium between discharges. Electroporated cells were used immediately. Where indicated, cells were preincubated for 2 min at 37°C in the presence of 2 mM diphenylene iodonium or 2 mM NAC prior to stimulation.
SDS-PAGE and Immunoblotting-SDS-PAGE and immunoblotting were performed essentially as described (Brumell and Grinstein, 1994). Briefly, samples were subjected to electrophoresis in 12% acrylamide gels and blotted onto polyvinylidene fluoride membranes (Millipore). Neutrophil whole cell lysates and immunoprecipitates were blotted with monoclonal anti-phosphotyrosine antibodies (1:5000 dilution) or with anti-sera to the specified tyrosine kinase (1:2000 dilution). Donkey anti-mouse and goat anti-rabbit secondary antibodies coupled to horseradish peroxidase were used at a 1:5000 dilution. Detection was made using the enhanced chemiluminescence system from Amersham Corp. Quantitation of radiograms was performed by densitometry using a Protein Databases Inc. (Huntington Station, NY) model DNA 35 scanner with the Discovery series 1D gel analysis software.
Immunoprecipitation and Immune Complex Kinase Assay-Electroporated cells (1.5 ϫ 10 7 /ml) were solubilized in ice-cold lysis buffer, which contained 1% Nonidet P-40, 150 mM NaCl, 2 mM EDTA, 1 mM NaV, 5 mM NaF, 1 mM phenylmethylsulfonyl flouride, 10 g/m aprotinin, 10 g/ml leupeptin, 1 M pepstatin A, and 50 mM Tris-HCl, pH 8.0. Lysates were centrifuged at 14,000 ϫ g for 5 min to remove unbroken cells and insoluble debris and then precleared with 50 l of Sepharose beads. Antibodies to the indicated tyrosine kinase or to phosphotyrosine were incubated with these lysates for 2 h at 4°C while rotating end over end. Immune complexes were precipitated by addition of 100 l of a 50% slurry of protein A-Sepharose beads, previously blocked with 10% bovine serum albumin in lysis buffer, followed by incubation at 4°C for 2 h. The addition of beads was unnecessary for phosphotyrosine antibodies, which were covalently attached to Sepharose. The immunoprecipitates were washed 4 -6 times and then resuspended in 1 ml of lysis buffer. Equal aliquots of this suspension were used for immunoblotting and for analysis of in vitro kinase activity.
The kinase activity of immune complexes was determined essentially as described . In brief, immunoprecipitates were washed with 1 ml of kinase buffer (5 mM MnCl 2 , 20 mM MOPS, pH 7.0), and autophosphorylating activity was assayed by incubation with 25 l of kinase buffer containing 12.5 Ci of [␥-32 P]ATP and 1 M K-ATP. Where specified, 1 g of rabbit muscle enolase was included as an exogenous substrate. Samples were incubated at 25°C in an Eppendorf Thermomixer, and reactions were stopped by the addition of boiling 2ϫ concentrated Laemmli sample buffer. The samples were subjected to SDS-PAGE, and the gels were stained with Coomassie Blue and dried in gel wrap (Biodesign Inc.). Dried gels were used for direct quantitation of radioactivity with a Molecular Dynamics PhosphorImager using Imagequant software or were subjected to radiography with an intensifying screen.
To study the effect of oxidizing agents on the tyrosine kinase activity of hck, immunoprecipitates of this kinase were isolated from untreated, electroporated cells. After washing, immune complexes were treated for 30 min with 1 mM diamide or 1 mM hydrogen peroxide at 30°C while shaking in a Thermomixer. As above, identical aliquots were used in parallel for immunoblotting and in vitro kinase assay. To study the effect of reducing agents on hck, immune complexes from GTP␥Sstimulated cells were treated with 20 mM NAC or 1 mM dithiothreitol for 30 min at 37°C and processed as above.
To study the role of tyrosine phosphorylation in hck activation, immunoprecipitates obtained from GTP␥S-treated cells were incubated at 37°C for 30 min with or without 2 g/ml of T-cell phosphatase. Aliquots of the beads were used for kinase assays and for immunoblotting with anti-phosphotyrosine antibodies to confirm the effectiveness of dephosphorylation by T-cell phosphatase.
Other Methods-All data are the means Ϯ S.E. or representative of three independent experiments with blood from different donors, with the exception of the in vitro treatment of hck immunoprecipitates with oxidizing and reducing agents, which was performed twice.

NADPH Oxidase-derived ROI Induce Tyrosine Phosphorylation-
The effect of ROI on tyrosine phosphorylation was studied in electropermeabilized cells stimulated with GTP␥S. This approach was chosen for two reasons. First, direct stimulation of GTP-binding proteins bypasses cell surface receptors, circumventing possible direct effects of the latter on the kinases and obviating receptor down-regulation, which can greatly reduce the magnitude and duration of the respiratory burst (reviewed by Klotz and Jesaitis (1994)). Using GTP␥S, activation of the oxidase is sustained, resembling the physiological stimulation elicited by phagocytic stimuli (Grinstein and Furuya, 1991). Second, equilibration of the permeabilized cells with EGTA-containing buffers precludes changes in cytosolic calcium concentration, which might alter tyrosine phosphorylation (Berkow and Dodson, 1990).
As shown in Fig. 1A, the addition of GTP␥S and NADPH to permeabilized cells induced the accumulation of phosphotyrosine on a number of proteins, as determined by immunoblotting (cf. lanes 1 and 4). Treatment of the electroporated cells with GTP␥S and NADPH alone was found to have little effect (lanes 2 and 3). The stimulatory effect of GTP␥S or NADPH was moderated by the presence of active tyrosine phosphatases. This is indicated by the pronounced enhancement in phosphotyrosine accumulation noted when vanadate, a phosphatase inhibitor, was included during stimulation ( Fig. 1A). For this reason, 10 M sodium orthovanadate was included routinely in subsequent assays to minimize dephosphorylation, thereby magnifying the responses. At the concentration used, vanadate itself had negligible effects on tyrosine phosphorylation (see lanes 1 and 2 in Fig. 1C), consistent with earlier findings (Bourgoin and Grinstein, 1992). Moreover, whereas vanadate increased the extent of phosphotyrosine accumulation, the phosphorylated substrates and the time course of phosphorylation were similar in the presence and the absence of the phosphatase inhibitor. As illustrated in Fig. 1B, phosphotyrosine accumulation induced by GTP␥S stimulation was rapid (evident after 1 min) and time-dependent, with a maximal response seen after 10 min. The effect of GTP␥S on tyrosine phosphorylation was entirely dependent on the presence of NADPH. As shown in Fig.  1C (as well as in Fig. 1A), treatment of electroporated cells with GTP␥S had little effect when the nucleotide was omitted (cf. lanes 3 and 4). This finding suggests that generation of superoxide by the NADPH oxidase is required for the increase in tyrosine phosphorylation following stimulation with GTP␥S. In support of this hypothesis, we found that NAC, a powerful anti-oxidant that has been shown to scavenge ROI and increase cytosolic levels of reduced glutathione (Halliwell and Gutteridge, 1985), effectively attenuated the tyrosine phosphorylation produced by GTP␥S in the presence of NADPH. Moreover, diphenylene iodonium, an inhibitor of the flavoprotein component of the NADPH oxidase (Ellis et al., 1988), had a comparable effect (lane 6). These findings are in agreement with those of Fialkow et al. (1993) and indicate that NADPH oxidase-derived ROI promote tyrosine phosphorylation in neutrophils.
Tyrosine Kinases Present in Neutrophils-As an initial step in the study of the mechanism of action of ROI, we determined which of the known tyrosine kinases are present and active in GTP␥S-stimulated 2 neutrophils. Electroporated cells were activated with the nucleotide and immediately solubilized for immunoprecipitation with one of a battery of antibodies to tyrosine kinases. The immune complexes were used for in vitro kinase assays and then subjected to SDS-PAGE and autoradiography. Of the 11 antibodies tested, 5 were found to immunoprecipitate active kinases detectable by their autophosphorylating activity, suggested by the close correspondence of the phosphorylated bands to the known molecular weight of the kinase immunoprecipitated ( Fig. 2A). The active kinases included three src family members; lyn (53 and 56 kDa) 3 , hck (56 and 59 kDa), and fgr (59 kDa). Also included were syk (72 kDa) and btk (77 kDa). In contrast, no significant activity was measurable in src, fyn, yes, blk, lck, and zap-70 immunoprecipitates or when rabbit nonimmune serum ( Fig. 2A, cont.) was used. The presence of lyn, hck, fgr, syk, and btk in neutrophils had been reported previously (Yamanashi et al., 1987;Ziegler et al., 1987;Gutkind and Robbins, 1989;Asahi et al., 1993;Yamada et al., 1993).
In good agreement with the kinase assays of Fig. 2A, the presence of lyn, hck, fgr, syk, and btk in neutrophils was confirmed by immunoblotting whole cell lysates with the same antisera used for precipitation (Fig. 2B). Both the full-length (72 kDa) syk protein as well as its Ϸ40-kDa degradation product were observed upon immunoblotting (Fig. 2B, closed arrowheads). A band of Ϸ65 kDa was also recognized by the syk anti-serum. It is not presently clear whether this polypeptide is related to syk or is merely a fortuitously cross-reacting protein.
It is noteworthy, however, that a band of similar mobility was often found to be phosphorylated in syk immune complex assays (Fig. 3A), suggesting that the 65-kDa polypeptide coimmunoprecipitates and can be phosphorylated by syk.
Modulation of Tyrosine Kinase Activity by ROI-The effect of ROI on neutrophil kinase activity was studied next. To this end, immunoprecipitates were prepared from control and 2 Hereafter, "GTP␥S stimulation" refers to treatment of electroporated neutrophils with 10 M GTP␥S in the presence of 2 mM NADPH and 10 M NaV at 37°C. 3 We observed a phosphoprotein of about 60 -62 kDa in lyn immunoprecipitates. lyn has been extensively studied and only two isoforms have been identified, suggesting that the third band may be a separate protein. Several kinases have been shown to co-precipitate with lyn in other systems, including syk (Sidorenko et al., 1995), btk (Cheng et al., 1994), and a cell cycle regulatory protein, p34cdc2 (Yuan et al., 1995).

FIG. 1. Effect of GTP␥S on tyrosine phosphorylation.
A, NADPH dependence and potentiation by sodium vanadate. Electroporated neutrophils were incubated at 37°C without (Ϫ) or with (ϩ) the following agents for 5 min: 10 M GTP␥S, 2 mM NADPH, and 10 M NaV, as indicated. Cells were then rapidly sedimented, boiled in sample buffer, and subjected to SDS-PAGE. Analysis was performed by immunoblotting with a monoclonal antibody to phosphotyrosine. B, time course of phosphotyrosine accumulation. Electroporated neutrophils were treated without (Ϫ) or with 10 M GTP␥S, 2 mM NADPH, and 10 M NaV for the indicated time (min) and processed as in A. C, dependence of tyrosine phosphorylation on NADPH oxidase-derived ROI. Electroporated neutrophils were treated without (Ϫ) or with (ϩ) 10 M GTP␥S and/or 2 mM NADPH for 5 min at 37°C. Where specified, the cells were treated with 2 mM diphenylene iodonium (DPI) or 2 mM NAC for 2 min at 37°C prior to GTP␥S stimulation. The presence of 10 M NaV during treatment is indicated. The results shown are representative of three separate experiments.
GTP␥S-stimulated cells using antisera to the kinases identified earlier in Fig. 2. As shown in Fig. 3A, the in vitro activity of hck, syk, and btk was noticeably increased following stimulation with GTP␥S. The stimulation of hck was investigated in more detail in Fig. 3B, where immune complexes obtained at various times after the addition of GTP␥S were assayed in the presence of the exogenous substrate enolase. Both the autophosphorylation of hck (closed arrowhead in inset to Fig. 3B) and its ability to phosphorylate enolase (open arrowhead) followed a biphasic course, peaking between 1-5 min and declining thereafter. A similar increase in the ability of syk and btk to phosphorylate enolase was also observed (data not shown). It should be noted that qualitatively similar responses of hck and lyn were seen when NaV was omitted during stimulation of electroporated cells (data not shown).
Although production of ROI stimulated some tyrosine kinases, others were seemingly inhibited. The autophosphorylating abilities of lyn and fgr were diminished (73 Ϯ 14 and 48 Ϯ 16% of control activity, respectively; n ϭ 3) following 1 min of GTP␥S stimulation (Fig. 3A). As for hck, the detailed time course of the effects of ROI on lyn activity was analyzed with enolase as substrate (see Fig. 3C). Interestingly, quantitation of the auto-and enolase-phosphorylating activities of lyn immune complexes revealed a discrepancy. Phosphorylation of the exogenous substrate was markedly increased, whereas autophosphorylation decreased. These findings suggest that nonradioactive phosphate is incorporated into lyn in the cells, prior to immunoprecipitation, precluding subsequent incorporation of radiolabel into these sites. The reduced autophosphorylation is therefore an inaccurate indication of the enzymatic activity of lyn, which is at least transiently stimulated by GTP␥S.
Role of Direct Oxidation in hck Activation by ROI-The mechanism of tyrosine kinase activation by endogenous ROI was studied in detail for hck. This kinase was chosen because it is virtually quiescent in unstimulated cells yet is the most active in immunoprecipitates from activated cells, providing an optimal signal to noise ratio. We first considered the possibility that activation of hck by ROI resulted from direct oxidation of critical residues on the kinase. To test this notion, hck immunoprecipitates obtained from unstimulated cells were treated in vitro with two strong oxidizing agents, diamide and H 2 O 2 . Comparable concentrations of these oxidants have been shown to promote phosphotyrosine accumulation when added to intact cells (Fialkow et al., 1994). Neither diamide nor H 2 O 2 , however, was capable of activating isolated hck in immune complexes (Fig. 4A). Conversely, we found that reducing agents could not reverse the activation of hck immunoprecipitated from lysates of GTP␥S-treated cells. As shown in the rightmost lanes of Fig.  4A, when added directly to the immunoprecipitate neither NAC nor dithiothreitol diminished the activation of hck. These findings contrast the preventive effect of NAC seen when added to permeabilized cells during the respiratory burst, described in Fig. 1B. The inability of oxidants and reducing agents to affect hck autophosphorylation was confirmed by immunoblotting the immunoprecipitates with anti-phosphotyrosine antibodies (Fig.  1B). The kinase was found to be tyrosine-phosphorylated only after stimulation of the cells with GTP␥S, and the phosphotyrosine content was unaffected by oxidizing and reducing agents. Together, these results suggest that ROI do not directly activate hck in GTP␥S-stimulated neutrophils.

FIG. 2. Identification of tyrosine kinases present in neutrophils.
A, immune complex kinase assays were performed in vitro using immunoprecipitates of the tyrosine kinases indicated, prepared from lysates of electroporated neutrophils treated with 10 M GTP␥S, 2 mM NADPH, and 10 M NaV for 2 min. Kinase reactions were stopped, and the material was subjected to SDS-PAGE and autoradiography of the dried gel. The assay was also performed using a rabbit nonimmune serum (cont). B, whole neutrophil lysates were immunoblotted with antisera to the tyrosine kinases indicated. The closed arrowheads point to the tyrosine kinase. The open arrowhead indicates an unidentified protein of Ϸ65 kDa that cross-reacts with the syk antibody.

FIG. 3. Modulation of tyrosine kinase activity by ROI.
A, immune complex kinase assays were performed in vitro using immunoprecipitates of the tyrosine kinases indicated, prepared from lysates of electroporated neutrophils treated without (Ϫ) or with (ϩ) 10 M GTP␥S, 2 mM NADPH, and 10 M NaV for 1 min. B, immunoprecipitates of hck were prepared from lysates of electroporated neutrophils treated without or with 10 M GTP␥S, 2 mM NADPH, and 10 M NaV for the indicated time (min) and subjected to immune complex kinase assays in the presence of enolase. The kinase reactions were stopped, and the samples were subjected to SDS-PAGE followed by autoradiography of the dried gels. A representative experiment is shown in the inset. Bands that correspond to autophosphorylation (closed arrow) and enolase phosphorylation (open arrow) were quantified with a Phosphor-Imager, and the results are presented as the percentage of maximal response in the main panel. C, in vitro autophosphorylation and enolase phosphorylation activities of lyn were determined as in B for hck. The data in B and C are the means Ϯ S.E. of three experiments.

Role of Tyrosine Phosphorylation in the Regulation of Kinase
Activity-Phosphorylation on tyrosine residues has been shown to be an important determinant of the activity of several tyrosine kinases, including those identified in neutrophils. Because tyrosine phosphorylation of hck was detectable when this enzyme was precipitated from stimulated cells (Fig. 4B), we considered the possibility that ROI activate kinases in neutrophils indirectly by mediating their phosphorylation on tyrosine residues. As an initial approach to test this hypothesis, we tested the kinase activity of immune complexes obtained from control and GTP␥S-stimulated cells using anti-phosphotyrosine antibodies. The data in Fig. 5 demonstrate that phosphorylation was greater in precipitates from stimulated cells, suggesting that the relevant kinase activity may be associated with tyrosine phosphorylated proteins. A number of prominent bands displayed increased phosphorylation in vitro (arrowheads in Fig. 5) with molecular masses of approximately 48, 54, 62, 68, 75, and 118 kDa.
The size of some of the phosphoproteins in phosphotyrosine immunoprecipitates correspond to that of the active kinases detailed in Fig. 2. To establish more directly whether the active kinases are tyrosine phosphorylated, immunoprecipitates of lyn, hck, fgr, syk, and btk were prepared from control and GTP␥S-treated cells and probed by immunoblotting with antiphosphotyrosine antibodies. As illustrated in Fig. 6A, endogenous generation of ROI is accompanied by tyrosine phosphorylation of all the kinases studied (indicated with closed arrowheads), with the notable exception of btk, which remained unaffected. The figure also shows that both the intact form of syk as well as its 40-kDa proteolytic fragment (open arrowhead) was phosphorylated on tyrosine.
The correlation between the occurrence of tyrosine phosphorylation and the activation of the tyrosine kinases is further stressed by the similarity of the time courses of both events. In Fig. 6B, the degree of tyrosine phosphorylation was quantified in immunoprecipitates from cells stimulated for varying peri- FIG. 4. Effect of oxidizing/reducing agents on hck activity. hck immunoprecipitates were obtained from lysates of electroporated neutrophils treated without (Ϫ) or with (ϩ) 10 M GTP␥S, 2 mM NADPH, and 10 M NaV (5 min) were incubated at 30°C for 30 min with either 1 mM diamide, 1 mM H 2 O 2 , 20 mM NAC, 1 mM dithiothreitol, or with buffer alone (none), as indicated. An aliquot of the immunoprecipitate was used for in vitro kinase assay (A), and another was used for anti-phosphotyrosine immunoblotting (B). The closed arrows indicate the position of immunoprecipitated hck, whereas the open arrow indicates the position of the exogenous substrate, enolase.
FIG. 5. Phosphotyrosine-associated kinase activity. Anti-phosphotyrosine immunoprecipitates were obtained from lysates of electroporated neutrophils treated without (Ϫ) or with (ϩ) 10 M GTP␥S, 2 mM NADPH, and 10 M NaV (2 min) and used to perform in vitro kinase assays as described under "Experimental Procedures." The closed arrowheads indicate bands that displayed a prominent increase in phosphorylation following stimulation. The results are representative of three separate experiments.
FIG. 6. A, phosphorylation of tyrosine kinases in situ. Anti-phosphotyrosine immunoblotting was performed on immunoprecipitates of the specified tyrosine kinases, obtained from electroporated neutrophils treated without (Ϫ) or with (ϩ) 10 M GTP␥S, 2 mM NADPH, and 10 M NaV for 5 min. The solid arrows indicate the position of the immunoprecipitated kinase. The open arrowhead indicates the 40-kDa proteolytic fragment of syk, also observed in anti-syk immunoblots (see Fig.  2B). B, time course of tyrosine phosphorylation. Cells were stimulated for the times indicated (min), and immunoprecipitates of the indicated kinases were obtained and subjected to SDS-PAGE and anti-phosphotyrosine immunoblotting. The extent of tyrosine phosphorylation was quantified in each case by densistometry and is presented as the percentage of the maximum. The data are the means Ϯ S.E. for three experiments. ods of time with GTP␥S. All four kinases undergo rapid and progressive phosphorylation, which is detectable by 1 min and maximal between 5 and 10 min. This pattern closely resembles the time course of activation of hck determined in Fig. 3C using enolase as the substrate. It therefore appears likely that tyrosine phosphorylation of the kinases regulates their activity.
This notion was directly addressed by treatment of immunoprecipitates with an active tyrosine phosphatase, the truncated T-cell phosphatase. hck was precipitated from GTP␥S-stimulated neutrophils and incubated for 30 min in the presence or the absence of T-cell phosphatase. The effectiveness of the phosphatase was ascertained by immunoblotting the precipitates with anti-phosphotyrosine antibodies. Exposure to T-cell phosphatase led to complete dephosphorylation of hck (indicated by the solid arrowhead in Fig. 7A). Immunoblotting confirmed that equal amounts of hck were present before and after treatment with T-cell phosphatase (data not shown). The autophosphorylating (solid arrowhead) and exogenous kinase activity (open arrowhead) of stimulated and dephosphorylated hck was next compared. Fig. 7B demonstrates that treatment with T-cell phosphatase eliminated the kinase activity of hck stimulated by ROI.

DISCUSSION
In this report, we analyzed the mechanism leading to increased phosphotyrosine accumulation following ROI production in neutrophils. In electroporated cells treated with GTP␥S, we detected an elevated activity of several kinases, measured in vitro. The activation of these kinases was rapid and correlated well with the increase in tyrosine phosphorylation observed under these conditions.
Kinases of three separate families were found to be activated by ROI, as determined by autophosphorylation and phosphorylation of an exogenous substrate, enolase. hck, a member of the src family of tyrosine kinases and highly expressed in granulocytes and macrophages (Ziegler et al., 1987), displayed little activity in untreated cells but was rapidly stimulated following the addition of GTP␥S. syk, which belongs to a separate family of kinases, also displayed increased activity following ROI production. In contrast, the closely related ZAP-70 tyrosine kinase, thought to be important in B-and T-cell receptor signaling (Sefton and Taddie, 1994) was not detectable in active neutrophils using our immune complex kinase assay. btk, a member of the tec family of tyrosine kinases, is expressed in cells of myeloid and lymphoid lineage (Yamada et al., 1993) and was also activated by ROI. To our knowledge, activation of btk in neutrophils had not been reported previously.
Although ROI production led to the activation of some tyrosine kinases, it appeared to have an opposite effect on the activity of others when estimated from autophosphorylation in immune complex kinase assays. Thus, lyn and fgr displayed high activities in untreated, electroporated neutrophils, which decreased following GTP␥S stimulation. However, at least in the case of lyn, the apparent decrease in activity likely reflected occupancy of substrate sites by nonradioactive phosphate, which may have occurred in situ, prior to immunoprecipitation. 3 Indeed, the ability of the enzyme to phosphorylate exogenous substrates was increased following stimulation of the respiratory burst. Therefore, caution must be exercised when equating the autophosphorylating and catalytic activities of tyrosine kinases.
None of the other tyrosine kinases tested were found to be activated following generation of ROI. These included yes, which is reported to be present in neutrophils, where it can be stimulated by granulocyte macrophage colony-stimulating factor (Corey et al., 1993). Clearly, although 11 different antisera were used, our survey was incomplete, because other tyrosine kinases are likely to exist in neutrophils.
The mechanism underlying the activation of the kinases by ROI was explored in some detail using hck as a prototype. Although ROI production in situ appeared to activate hck, oxidizing agents could not mimic this effect when applied to hck immunoprecipitates in vitro. Moreover, reducing agents failed to reverse the activation of hck isolated from GTP␥S-treated cells. We conclude that hck activity is not regulated directly by ROI but rather by some other post-translational modification. Though not tested directly, we suggest by extension that activation of the other kinases is similarly indirect.
Because tyrosine phosphorylation of lyn, hck, fgr, and syk was found to occur upon stimulation by GTP␥S, this posttranslational modification was considered as a possible mechanism of regulation. This notion was evaluated using T-cell phosphatase to dephosphorylate activated hck. This procedure was found to eliminate the activity of the kinase, suggesting that tyrosine phosphorylation mediates the effect of ROI on hck activation.
Tyrosine kinase activity of src family members is thought to be suppressed by phosphorylation of a C-terminal residue, conserved among family members (Cooper, 1988;Liu et al., 1993;Cooper and Howell, 1993). Dephosphorylation of this residue has been shown to increase the activity of src family kinases (Cooper and King, 1987), in apparent conflict with our findings with hck, where complete dephosphorylation of the enzyme led to its inactivation. However, recent evidence has questioned this simple model of regulation. This includes the finding that a T-cell line lacking CD45 (the phosphatase that activates the src family member lck) was found to have higher lck activity even though its inhibitory C-terminal tyrosine residue was hyperphosphorylated (Burns et al. 1994). Although dephosphorylation of the C terminus may be important for the derepression of src family members, a number of unique tyrosine resi- FIG. 7. Effect of tyrosine dephosphorylation on hck activity. hck was immunoprecipitated from lysates of neutrophils treated with 10 M GTP␥S, 2 mM NADPH, and 10 M NaV for 5 min. Immunoprecipitates were then treated for 30 min without (Ϫ) or with (ϩ) recombinant T-cell phosphatase (TC-PTP) at 30°C while shaking. An aliquot of the immunoprecipitated material was subjected to anti-phosphotyrosine immunoblotting (A), and another was used to perform the in vitro kinase assay (B dues have been reported to be phosphorylated upon activation. These include the so-called autophosphorylation site within the kinase domain (Smart et al., 1981;Patchinsky et al., 1982) and sites within the N-terminal domain of some src family members (Souda et al., 1993). Evidence exists that phosphorylation of these residues is essential for kinase activity, possibly by stabilization of the active kinase (Veillette and Fournel, 1990;Mustelin, 1994). Recent evidence has also implicated serine phoshorylation in the regulation of src family members (Winkler et al., 1993;Watts et al., 1993). Whereas the regulation of src family kinases remains incompletely understood, our findings imply that tyrosine phosphorylation is necessary to maintain the activity of hck following activation by endogenous ROI.
The steps that follow ROI generation and lead to kinase phosphorylation are unknown, but some insight is provided by recent reports that (a) critical conserved cysteine residues exist in the catalytic domain of many tyrosine phosphatases (Fisher et al., 1991) and (b) that both exogenous (Zor et al., 1993, Fialkow et al., 1994 as well as endogenous oxidants 4 can inactivate tyrosine phosphatases, likely by targeting their critical sulfhydryl moieties. In view of these considerations, the following scenario can be envisaged. Under basal conditions, the accumulation of tyrosine phophoproteins and the autophosphorylation and stimulation of tyrosine kinases, some of which are constitutively active, are prevented by the offsetting action of tyrosine phosphatases. This delicate balance can be disrupted when ROI diminish the rate of dephosphorylation by reaction with sulfhydryl side chains in the catalytic domain of one or more tyrosine phosphatases. Indeed, in neutrophils, CD45 has been shown to be susceptible to inactivation by oxidants (Fialkow et al., 1994), and other phosphatases present in these cells (e.g. PTP-1C) 5 are likely to be similarly affected.
It is noteworthy that modulation of tyrosine kinase activity has been reported in lymphoid cells exposed to H 2 O 2 . Treatment with the oxidant was found to activate syk but not lyn, resembling our observations in neutrophils. Like most other cells, lymphocytes can potentially generate ROI by electron transfer reactions in mitochondria and the endoplasmic reticulum. However, the magnitude of the oxidative response is far greater in phagocytes, which express high levels of the NADPH oxidase (see the Introduction). In this regard, it is important that in the present experiments activation of phosphorylation was elicited by endogenously generated ROI, lending credence to the physiological significance of the observations. It is possible to envisage that stimulation of the NADPH oxidase, one of the earliest effectors of neutrophils, could promote phosphotyrosine accumulation by the combined inhibition of phosphatases and activation of kinases. This could in turn have important consequences on more slowly developing responses such as phagocytosis and degranulation. In this context, hck has been suggested to have a role in phagocytosis (Lowell et al., 1994), and syk has also been proposed to be essential to the anti-microbial response (Asahi et al., 1993). It is also conceivable that ROI secreted by neutrophils may have paracrine effects, stimulating neighboring quiescent neutrophils or other cells present in the inflammatory milieu, including lymphocytes and macrophages.