Phosphorylation of the Respiratory Burst Oxidase Subunit p47 phox as Determined by Two-dimensional Phosphopeptide Mapping PHOSPHORYLATION BY PROTEIN KINASE C, PROTEIN KINASE A, AND A MITOGEN-ACTIVATED PROTEIN KINASE*

The respiratory burst oxidase is responsible for super- oxide (O 2 .) production by phagocytes and B lymphocytes. This multicomponent enzyme is dormant in resting cells but is activated on exposure of the cells to an appropri- ate stimulus. Upon activation, several serine residues on the cytosolic oxidase subunit p47 phox become phosphorylated. Using two-dimensional tryptic phosphopeptide mapping, we studied the phosphorylation of p47 phox in 32 P i -loaded Epstein-Barr virus-transformed B lympho- blasts expressing wild type p47 phox or any of several p47 phox Ser 3 Ala mutants. We were able to identify the labeled peptides from wild type p47 phox as those containing Ser 303/304 , Ser 315 , Ser 320 , Ser 328 and/or Ser 359/370 , and Ser 345/348 ; no 32 P-labeled Ser 310 -containing peptide was found. When purified p47 phox was phosphorylated in vitro by various protein kinases, varying phosphopep- tide patterns were observed. Protein kinase C phosphorylated all the peptides except the one containing Ser 345/ 348 ; protein kinase A phosphorylated the peptide containing Ser 320 and one or both of the peptides containing Ser 328 and Ser 359/370 ; while mitogen-activated protein kinase phophorylated only the peptide contain- ing Ser 345/348 . These findings suggest that these three kinases play distinct roles electrophoresis chromatography on plate as Phosphopeptides were detected by expos-ing the thin layer plates to Reflexion film (DuPont for 24–72 h at (cid:50) 70 °C with

The respiratory burst oxidase is responsible for superoxide (O 2 . ) production by phagocytes and B lymphocytes.
This multicomponent enzyme is dormant in resting cells but is activated on exposure of the cells to an appropriate stimulus. Upon activation, several serine residues on the cytosolic oxidase subunit p47 phox become phosphorylated. Using two-dimensional tryptic phosphopeptide mapping, we studied the phosphorylation of p47 phox in 32 P i -loaded Epstein-Barr virus-transformed B lymphoblasts expressing wild type p47 phox or any of several p47 phox Ser 3 Ala mutants. We were able to identify the labeled peptides from wild type p47 phox as those containing Ser 303/304 , Ser 315 , Ser 320 , Ser 328 and/or Ser 359/370 , and Ser 345/348 ; no 32 P-labeled Ser 310 -containing peptide was found. When purified p47 phox was phosphorylated in vitro by various protein kinases, varying phosphopeptide patterns were observed. Protein kinase C phosphorylated all the peptides except the one containing Ser 345/ 348; protein kinase A phosphorylated the peptide containing Ser 320 and one or both of the peptides containing Ser 328 and Ser 359/370 ; while mitogen-activated protein kinase phophorylated only the peptide containing Ser 345/348 . These findings suggest that these three kinases play distinct roles in the activation of the respiratory burst oxidase, each of them catalyzing the phosphorylation of a different group of serines in p47 phox .
The respiratory burst oxidase of phagocytes and B lymphocytes catalyzes the reduction of oxygen to superoxide (O 2 . ) at the expense of NADPH (1)(2)(3)(4)(5)(6). In resting cells the enzyme is inactive, and its components are distributed between the cytosol and the membranes of secretory vesicles. When the cells are activated, the cytosolic components migrate to the membranes, where they associate with the membrane-bound components to assemble the catalytically active oxidase (1,5,7). When the oxidase is activated, p47 phox , one of the cytosolic subunits, becomes phosphorylated on several serines (8,9). We recently found that in human neutrophils serines Ser 303 , Ser 304 , Ser 320 , Ser 328 , Ser 345 , Ser 348 , and Ser 359 and/or Ser 370 are phosphorylated and that other serines lying between Ser 303 and Ser 379 could be phosphorylated (10). We further showed that at least one of these serines is absolutely required for oxidase activation in whole cells stimulated with PMA (11). In this study, we report the use of site-directed mutagenesis combined with two-dimensional phosphopeptide mapping to further characterize the phosphorylation of p47 phox in B lymphocytes and to compare the in vitro phosphorylation of purified p47 phox by various serine/threonine-specific protein kinases.
Site-directed Mutagenesis and Transfections-Most of the mutants used for these experiments have been previously reported (11). S303A,S304A was constructed by cloning the WT p47 phox cDNA template into the XbaI/NotI fragment of pBluescript KSII, then mutating by an oligonucleotide-directed technique using the oligonucleotide CG-GATGGCCGCGCGCCGGGGCGGCGC (deviations from the WT sequence are shown in boldface) (11). The elimination of a BssHII site in the mutant was used for screening. S328A-S359A was constructed by the sequential introduction of the single mutations S348A, S345A, S359A, and S328A into the WT clone using templates reported elsewhere (11). S370A,S379A was similarly created by the sequential introduction of the mutations S379A and S370A into the WT clone. S328A-S379A was then constructed by replacing the NaeI/NarI fragment of S370A,S379A with the NaeI/NarI fragment of S328A-S359A. In every case, the mutations were confirmed by dideoxynucleotidebased sequencing (11). The wild type or mutant cDNAs were then excised from Bluescript, cloned into the XbaI/NotI sites of the mammalian expression vector EBOpLPP and transfected into p47 phox -deficient EBV-transformed B lymphocytes as described elsewhere (12). 32 P Labeling of Transfected Cells and p47 phox Purification-Transfected B lymphoblasts were labeled with 32 P i as described previously (11). Briefly, the cells were incubated overnight in phosphate-free medium, then transferred to fresh medium containing 32 P i (0.2 mCi/ml) and incubated for 4 h at 37°C. The cells were then activated for 12.5 min with PMA (1 g/ml/10 8 cells), after which their p47 phox was isolated and purified by immunoaffinity chromatography as described before (10).
In Vitro Phosphorylation of p47 phox -p47 phox was isolated by immunoprecipitation from resting neutrophils exactly as described elsewhere (10,16). Labeling with protein kinase A was performed by incubating a reaction mixture containing 1 g of p47 2 mM EGTA, 0.5 mM NaF, and 0.2 mM ␤-glycerophosphate for 30 min at 37°C. For the protein kinase C reaction, 1 g of immunopurified p47 phox was incubated with 0.5 g of protein kinase C in 20 mM Tris-Cl (pH 7.5), 10 mM MgCl 2 , 0.5 mM CaCl 2 , 1 mM dithiothreitol, 50 M (1 Ci) [␥-32 P]ATP, 5 g/ml diolein, and 50 g/ml phosphatidylserine in a total volume of 50 l. The reaction was carried out for 30 min at 30°C. Phosphorylation by MAP kinase was accomplished by incubating 1 g of p47 phox with 0.5 g of MAP kinase (p42-ERK2) and 50 M (1 Ci) [␥-32 P]ATP under the conditions used for the protein kinase A reaction.
SDS-PAGE and Tryptic Phosphopeptide Analysis-Immunoprecipitated or recombinant 32 P-labeled p47 phox was analyzed by 10% SDS-PAGE and blotted to nitrocellulose using the Laemmli (13) and Towbin et al. (14) systems, respectively. Labeled p47 phox was detected on the immunoblot by a specific anti-peptide antibody (15) and by autoradiography for 1 to 2 h. The band corresponding to p47 phox was digested with trypsin, and the resulting peptides were separated by high voltage electrophoresis and chromatography on a cellulose thin layer plate as described elsewhere (10,17). Phosphopeptides were detected by exposing the thin layer plates to Reflexion film (DuPont NEN) for 24 -72 h at Ϫ70°C with an intensifying screen. For the peptide-containing serines 359 and 370, phosphorylation was analyzed by digesting the blotted p47 phox with Glu-C, separating the resulting peptides by Tris-Tricine SDS-PAGE and subjecting the resulting gel to autoradiography. Phosphorylated peptide 359/370 appears as a radioactive band at Ϸ4 kDa. Each autoradiogram is representative of two or three separate experiments, each carried out with a different transfection.

Identification of Phosphopeptides on the Tryptic Peptide Map
of Phosphorylated p47 phox , Including a Phosphopeptide Containing Ser 315 -During the activation of the respiratory burst oxidase in phagocytes and B lymphocytes, p47 phox becomes extensively phosphorylated. Using CNBr cleavage followed by proteolysis, Tricine gel electrophoresis, and Edman degradation, we showed that the targets of phosphorylation in p47 phox are the serine residues lying between Ser 303 and Ser 379 inclusively, and that among these serines the following are phosphorylated: Ser 303 , Ser 304 , Ser 320 , Ser 328 , Ser 345 , Ser 348 , Ser 359 and/or Ser 370 , and Ser 379 (10,11), the last phosphorylated much less extensively than the rest. The methods employed in that study were elaborate and time-consuming, however, and presented certain limitations: in qualitative analysis because manual Edman degradation is only reliable through the first 10 -15 cycles (18) and in quantitative analysis because of unequal losses of phosphorylated peptides during the workup of the samples. A simpler and more quantifiable method would be the analysis of phosphorylation by two-dimensional tryptic peptide mapping (17). We therefore carried out experiments to identify the phosphopeptides on the tryptic peptide map of 32 P-labeled p47 phox . Our approach was to express p47 phox Ser 3 Ala mutants in EBV-transformed p47 phox -deficient B lymphocytes; to load these lymphocytes with 32 P i and then activate them with PMA to label their p47 phox ; and finally to purify the labeled p47 phox mutants, map them, and look for differences between those maps and the map of 32 P-labeled WT p47 phox . In a tryptic digest of p47 phox , the phosphorylation targets are distributed among several peptides (Table I; trypsin is unable to split Lys-Pro and Arg-Pro bonds) (17). Of these, the peptide containing Ser 379 and Ser 381 would probably be difficult to see because it would contain very little 32 P relative to the other peptides (10,11). The results (Fig. 1, left) showed that the map of WT p47 phox contained six major phosphopeptides (arrows), all of which were seen in each of 15 separate experiments, together with several minor phosphopeptides whose presence in the maps was inconstant. Taking into consideration the serines known to be phosphorylated during oxidase activation (10) and the peptides generated by tryptic digestion of p47 phox (Table I), we made a number of mutants in which two or more serines had been converted to alanines, and used these together with mutants containing a single Ser 3 Ala change to identify the labeled peptides on the two-dimensional map.
Restricting our analysis to the six constant phosphopeptides (Fig. 1), we found that maps of p47 phox mutants containing single Ser 3 Ala mutations were the same as WT maps (Fig. 1, left) except for the map of S320A, which lacked a single spot, and the map of S315A, which lacked two spots (Fig. 1, left). The latter result suggests that at least two peptides were produced, probably because of partial cleavage at the sequence RKR (residues 316 -318). Sequences containing basic residues in tandem are known to be susceptible to partial cleavage (17).
Certain of the peptides of interest contain two serines, and it may be that the elimination of both serines is necessary to eliminate a spot corresponding to such a peptide. In accord with this idea, a single spot was eliminated from the phosphopeptide maps of S303A,S304A and S345A,S348A (Fig. 1, left). The map of the sextuple mutant S328A-S379A lacked two spots: one known to correspond to the peptide Ser 345/345 , and another that by elimination has to represent the two peptides Ser 328 and Ser 359/370 , because the spots corresponding to the remaining phosphopeptides (i.e. Ser 303/304 , Ser 315 , and Ser 320 ) were present on the map. 2 It appears that the phosphorylated peptides Ser 328 and Ser 359/370 coincide on the tryptic peptide map. Finally, these results suggest that Ser 310 is not phosphorylated during oxidase activation.
A diagram of the tryptic peptide map of WT p47 phox giving the identities of the major peptides is shown in Fig. 2. This diagram also shows the location of 32 P i , which proved to be a useful marker for identifying peptides on maps with missing spots. 32 P i lies at a level between peptides Ser 315 and Ser 345/348 , and migrates toward the anode, while the phosphopeptides all migrate toward the cathode. The location of the 32 P i spot relative to the point of application of the sample (marked by a plus sign (ϩ) in the lower left corner of the figure) provides a mobility standard that allows the identification of individual peptides even in the absence of information concerning the overall distribution of spots on the chromatogram, as would occur under conditions in which only one or two peptides are phosphorylated.
Comparison of Phosphopeptides in p47 phox Isolated from Activated Neutrophils and EBV-transformed B Cells-Using tryptic phosphopeptide mapping, we compared the sites of phosphorylation of p47 phox from activated human neutrophils and EBV-transformed normal B lymphocytes. The results (Fig. 3) showed that the same six major phosphopeptides appeared in both maps. These findings suggest that the p47 phox phosphopeptide assignments made through the experiments described above are valid for neutrophils as well as B lymphocytes, and that this method can therefore be employed to study . production by the respiratory burst oxidase. For this purpose, p47 phox immunopurified from resting neutrophils was incubated with [␥-32 P]ATP together with the kinase of interest (protein kinase A, protein kinase C, or MAP kinase), then repurified and analyzed either by CNBr cleavage followed by Tris-Tricine SDS-PAGE (19) or by tryptic peptide mapping. As shown previously with p47 phox that had been phosphorylated in intact cells (10), the sites phosphorylated in vitro by all three kinases were located in the C-terminal CNBr peptide of the labeled p47 phox (Fig. 4). Peptide mapping showed that protein kinase C phosphorylated all the p47 phox peptides except the peptide corresponding to Ser 345/348 (Fig. 5,  top), while protein kinase A, a more selective kinase, phosphorylated only the Ser 320 peptide and the Ser 328 and/or the Ser 359/370 peptides (Fig. 5, middle). When p47 phox phosphorylated by protein kinase C or protein kinase A was cleaved by Glu-C endopeptidase and then analyzed by SDS-PAGE, a labeled fragment appeared at 4 kDa (Fig. 6), indicating that the peptide containing Ser 359/370 was phosphorylated. The status of Ser 328 remained unresolved, although its phosphorylation in PMA-activated neutrophils suggests that it was probably phosphorylated at least by protein kinase C. MAP kinase, used here as an example of a proline-directed kinase, phosphorylated only the Ser 345/348 peptide, as expected from the sequences around the target serines in p47 phox (Fig. 5, bottom). Taken together, these results suggest that the three kinases could play different roles in regulating the activity of the respiratory burst oxidase.
On the phosphopeptide maps of purified p47 phox that had been labeled with a known kinase, major spots were seen that were not present on the maps of p47 phox labeled in whole cells. These spots were disregarded as irrelevant to the physiological labeling pattern of activated p47 phox .

DISCUSSION
Tryptic peptide mapping of p47 phox labeled with 32 P either in intact cells or in a cell-free system provides an efficient way of identifying which of the target peptides are phosphorylated, and in combination with image analysis of radioactivity could yield important information on the relative quantities of phos-phate on various of the serines of the protein. The results obtained by tryptic peptide mapping retain a certain amount of ambiguity, however, because they provide no information as to which of the two serines on a two-serine peptide is (are) phosphorylated. Whether it is important to answer that question will depend on studies correlating structure and function in Ser 3 Ala mutants of p47 phox , although a partial answer is provided by our recent report showing that Ser 3 Ala mutations of individual serines from position 303 to 370 have little effect on oxidase activity (11).
The present results provide some information as to the order of phosphorylation of the target serines on p47 phox . Except for the mutant S315A, whose anomalous properties were discussed above, mutations affecting the serines on a single tryptic peptide caused the loss of at most one spot on the phosphopeptide map. This finding suggests that there is no target serine whose phosphorylation is absolutely dependent on the phosphorylation of a serine on a different peptide, or the phosphorylation of a group of such serines. Rather, it appears that these serines can be phosphorylated in any order.
We previously showed that when the respiratory burst oxidase is activated, serines Ser 303 , Ser 304 , Ser 320 , Ser 328 , Ser 345 , Ser 348 , Ser 359 and/or Ser 370 , and Ser 379 of p47 phox are phosphorylated (10,11). The present studies confirm the earlier results by another method, and in addition have shown that Ser 315 is also phosphorylated, bringing the total number of phosphorylated serines in the C-terminal region of activated p47 phox to 9 or 10. Ser 379 has already been found to play an important role in oxidase activation and p47 phox translocation, and it is likely that protein kinase-mediated phosphorylation of other target serines is equally important. In fact, several lines of evidence already support a role for protein kinase C in oxidase activation. For example, PMA, an activator of several forms of protein kinase C, is a powerful stimulator of O 2 . production in whole cells (20). Purified p47 phox is a good substrate for protein kinase C in vitro (21), while staurosporine, a potent inhibitor of protein kinase C (and other kinases), blocks PMA-induced O 2 .
generation as well as the phosphorylation of p47 phox (22,23). Finally, we show in this study that the phosphopeptide map of p47 phox isolated from PMA-activated neutrophils and EBVtransformed B lymphocytes is identical to the phosphopeptide map of p47 phox phosphorylated in vitro by protein kinase C, except for the absence of the Ser 345/348 peptide from the latter map. These findings suggest that one or more of the PMAresponsive forms of protein kinase C could be a critical mediator of oxidase activation. We showed recently that Ser 345 and Ser 348 are not required for oxidase activation (11), since the S345A,S348A mutant of p47 phox is fully active in EBV-transformed B cells. This finding suggests that, in contrast to pro- FIG. 4. Phosphopeptides produced by CNBr cleavage of p47 phox purified by immunoprecipitation from human neutrophils phosphorylated with protein kinase C, protein kinase A, or MAP kinase. The experiment was carried out as described under "Experimental Procedures" using p47 phox purified from human neutrophils. The arrow shows the location of the C-terminal CNBr fragment of p47 phox . tein kinase C, phosphorylation of p47 phox by proline-directed kinases such as MAP kinase may have little to do with oxidase activation. The role of target serines in the regulation of oxidase activity is currently under investigation in our laboratory.
Phosphorylation of p47 phox was also shown to occur upon addition of dibutyryl cAMP to neutrophil cytoplasts or cytosol, suggesting that p47 phox is also a substrate for protein kinase A (24). Dibutyryl cAMP did not induce O 2 . production, however, indicating that phosphorylation by protein kinase A alone is not sufficient to activate the oxidase. It is possible, in fact, that the the phosphorylation of p47 phox by protein kinase A prevents the assembly of the oxidase, since the elevation of neutrophil cAMP inhibits O 2 . production (25). Our results show that protein kinase A phosphorylates fewer target serines than protein kinase C, phosphorylating only peptides Ser 320 , Ser 359/370 , and possibly Ser 328 (a maximum of four target serines), in contrast to the five peptides (up to seven target serines) phosphorylated by protein kinase C. The protein kinase A targets could be responsible for the negative regulation of p47 phox phosphorylation and oxidase activation by protein kinase A.