Covalent modification of protein kinase C isozymes by the inactivating peptide substrate analog N-biotinyl-Arg-Arg-Arg-Cys-Leu-Arg-Arg-Leu. Evidence that the biotinylated peptide is an active-site affinity label.

We recently reported that the peptide substrate analog Arg-Lys-Arg-Cys-Leu-Arg-Arg-Leu (RKRCLRRL) irreversibly inactivates the protein kinase C (PKC) isozymes alpha, beta, and gamma in a dithiothreitol-sensitive manner by an active site-directed mechanism. We hypothesized that the inactivation mechanism entailed covalent complex formation between the PKC isozyme and the inactivator peptide. In this report, N-biotinylated analogs of RKRCLRRL that inactivate Ca2+-dependent PKC activity were designed and tested for their ability to covalently label PKC isozymes. A purified PKC isozyme mixture (alpha, beta, gamma, epsilon, zeta) was incubated with the N-biotinylated peptides and then subjected to denaturing gel electrophoresis, transferred to nitrocellulose, and probed for avidin-reactive species. The Ca2+-dependent PKC subfamily members PKC-alpha, -beta, and -gamma comigrated at 82 kDa and were distinguished by isozyme-specific immunoprecipitation. N-Biotinyl-RRRCLRRL covalently labeled all of the isozymes examined. When the isozymes were denatured prior to incubation with the N-biotinylated peptides, no labeling was observed. Inactivation of the Ca2+-dependent PKC subfamily by the N-biotinylated peptides was associated with covalent labeling of the 82-kDa PKC subspecies. The concentration dependence curves observed with N-biotinyl-RRRCLRRL were similar for inactivation and covalent labeling. The rank order of potency of three N-biotinylated peptides was the same for the inactivation and covalent labeling. Both the inactivation and covalent labeling were dithiothreitol-sensitive, and they were each subject to protection by MgATP and a peptide substrate analog. The covalent label was mapped to the catalytic domain of PKC by limited proteolysis of the modified enzyme. These results provide evidence that the N-biotinylated inactivator peptides are active-site affinity labels of PKC. The inactivator peptides most likely function by S-thiolating the active-site Cys residue conserved in PKC. This is the first report to demonstrate covalent labeling of PKC by a peptide substrate analog.

The phospholipid-dependent isozyme family protein kinase C (PKC) 1 plays a pivotal role in multifarious physiological processes that include cell growth and differentiation, muscle contraction, neurotransmission, and platelet activation (1,2). Phorbol ester tumor promoters are potent and selective activators of PKC isozymes (3). Overexpression of specific PKC isozymes in mammalian cells has been shown to disrupt growth control mechanisms and in some cases to produce a transformed phenotype, either spontaneously or in response to treatment of the cells with phorbol ester tumor promoters (4 -7). Thus, PKC is implicated as a key enzyme in the pathological process of tumor promotion (8). PKC may also contribute to multiple drug resistance in cancer (9,10). Selective PKC inhibitors are therefore thought to be of potential value as cancer therapeutics (11,12).
Active-site affinity labels of the cAMP-dependent protein kinase that resemble its nucleotide (13,14) and peptide substrates (15)(16)(17) have been used to identify active-site residues located in or near the substrate binding domains. These reagents potently inactivate cAMP-dependent protein kinase by exploiting its substrate selectivity (13)(14)(15)(16)(17). Although numerous reversible inhibitors of PKC have been described (11,12,18,19), not a single active-site affinity label has been reported for the enzyme (20). Active-site affinity labels of PKC could be valuable tools in the characterization of active-site residues and topology and in the elucidation of the catalytic mechanism (21). Active-site affinity labels of PKC that resemble the peptide substrate may also offer an approach to potent and selective inactivation of PKC isozymes (19).
The PKC family consists of Ca 2ϩ -dependent, phorbol esteractivated isozymes (␣, ␤ 1 , ␤ 2 , ␥), Ca 2ϩ -independent, phorbol ester-activated isozymes (␦, ⑀, , ⍜, ), and Ca 2ϩ -and phorbol ester-independent isozymes (, ) (3). PS dependence is universal among PKC isozymes (3). We previously reported that the peptide substrate analog Arg-Lys-Arg-Cys-Leu-Arg-Arg-Leu (RKRCLRRL) irreversibly inactivates the Ca 2ϩ -and PS-dependent PKC isozymes ␣, ␤, and ␥ (20). The inactivated complex of PKC and RKRCLRRL was stable to dilution but not to exposure to the reducing thiol DTT (20). Protection against inactivation by substrates and substrate analogs provided evidence that the inactivator peptide was active site-directed (20). We hypothesized that the inactivation mechanism entailed covalent modification of the PKC isozymes at the active site (20). In this report, we demonstrate that N-biotinylated analogs of the inactivator peptide form covalent complexes with several PKC isozymes. We show that covalent modification of Ca 2ϩ -and PS-dependent PKC isozymes by the biotinylated peptides maps to the catalytic domain and is intimately associated with isozyme inactivation. The results provide evidence that N-biotinyl-RRRCLRRL and related inactivator peptides are active-site affinity labels of PKC isozymes.

MATERIALS AND METHODS
The synthetic peptides employed in this study were synthesized using the Vega coupler 250 peptide synthesizer and purified to Ͼ98% purity by reverse-phase high performance liquid chromatography using a Vydac C4 column and acetonitrile gradient elution at the M. D. Anderson Cancer Center Synthetic Antigen Facility. Rat brain PKC was purified to near homogeneity according to silver-stained polyacrylamide gels by a previously described method (22,23). The histone kinase activity of the purified PKC preparation was stimulated 10 -15fold by 0.2 mM Ca 2ϩ and 30 g/ml PS but was unaffected by either Ca 2ϩ or PS alone. Immunoblot analysis of the purified PKC isozyme mixture was done using an enhanced chemiluminescence (ECL) detection system as described previously (24) with the isozyme-specific monoclonal antibodies anti-PKC-␣, anti-PKC-␤, anti-PKC-␥, anti-PKC-␦, anti-PKC-⑀, anti-PKC-, anti-PKC-⍜, anti-PKC-, and anti-PKC-(Transduction Laboratories, Lexington, KY) and polyclonal anti-PKC-(Oxford Biochemicals, Oxford, MI). A catalytic domain fragment of PKC was generated from the purified PKC preparation by limited trypsinolysis with a yield of about 50%, as described previously (19,22). The catalytic domain fragment of the Ca 2ϩ -dependent PKC subfamily was detected by immunoblot analysis with the polyclonal antibody anti-pan-PKC, which recognizes the isozymes ␣, ␤, and ␥ and their catalytic domain fragments (Upstate Biotechnology, Inc., Lake Placid, NY).
Protein Kinase Assay-The Ca 2ϩ -and PS-dependent histone kinase activity of PKC was measured as described previously (19,22). The histone kinase reaction mixture (120 l) contained 20 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 0.2 mM CaCl 2 , 30 g/ml PS (or none), 6 M [␥-32 P]ATP (5,000 -8,000 cpm/pmol), 0.67 mg/ml histone III-S, and 5 ng of purified PKC. A 10-min histone kinase reaction period at 30°C, which yields linear kinetics (19,22), was initiated by the addition of [␥-32 P]ATP. The reaction was terminated on phosphocellulose paper, and histone phosphorylation was quantitated as described previously (19,22). In assays of synthetic peptide substrate phosphorylation, histone was replaced with the peptide substrate at concentrations indicated under "Results." Using previously described methods, synthetic peptide substrate analogs were tested for their ability to function as DTT-sensitive PKC inactivators (20). In these experiments, ␤-mercaptoethanol was removed from the purified PKC preparation by gel filtration on a 2-ml G-25 Sephadex column equilibrated in 20 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM EGTA, 10 g/ml leupeptin, and 0.4 mM phenylmethylsulfonyl fluoride (20). To measure the antagonism of the histone kinase reaction of PKC by the synthetic peptide substrate analogs, ␤-mercaptoethanol-free PKC was preincubated with the indicated synthetic peptide substrate analog for 5 min at 30°C. The preincubation mixture was placed on ice, and then a 10-l aliquot was added to the histone kinase reaction mixture (total volume, 120 l). To ascertain the DTT sensitivity of the antagonism of the histone kinase reaction, control assays were done in which 2.0 mM DTT was included in the preincubation mixture. All histone kinase assays were performed in triplicate and expressed as the mean value Ϯ S.D.
Detection of Stable Complexes of PKC Isozymes and N-Biotinylated Peptides-To detect complexes of PKC isozymes and N-biotinylated inactivator peptides that were stable under denaturing conditions in the absence of reducing thiol, a preincubation mixture of the purified ␤-mercaptoethanol-free PKC isozyme mixture (25 ng) and the N-biotinylated inactivator peptide under investigation was prepared as described under "Protein Kinase Assay" and boiled for 90 s in ␤-mercaptoethanol-free SDS-PAGE sample buffer (60 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol). The samples were subjected to the procedures and conditions employed in immunoblot analysis of PKC isozymes (24), except that membranes were probed with avidin instead of antibody. Briefly, samples were subjected to SDS-PAGE on 10% gels, and the protein was electrophoretically transferred to nitrocellulose mem-branes. Nonspecific binding sites were blocked by incubating membranes in 3% bovine serum albumin in Tris-buffered saline (TBS; 20 mM Tris HCl, 0.5 M NaCl, pH 7.5) at 4°C for 15-18 h. The membrane was probed with avidin conjugated to horseradish peroxidase (Bio-Rad) at the recommended working dilution of the commercial stock (1:3000) in the presence of 1% bovine serum albumin in TBS containing 0.05% Tween-20 (TTBS) for 2 h at room temperature. The blot was washed several times with TTBS, and avidin-reactive bands were detected using an ECL kit according to the manufacturer's instructions (Amersham Corp.) and quantitated by computerized densitometry. To quantitate the 82-kDa PKC subspecies in each lane, blots were stripped by a 30-min incubation at room temperature in 62.5 mM Tris-HCl, pH 6.7, 100 mM ␤-mercaptoethanol, 2% SDS. Stripped blots were probed with polyclonal anti-pan-PKC (1 g/ml), which recognizes PKC-␣, PKC-␤, and PKC-␥, using peroxidase-linked donkey anti-rabbit Ig (1:300) as the secondary antibody and previously described methods (24). The 82-kDa PKC subspecies was detected by ECL and quantitated with a computerized densitometer.
Immunoprecipitation of Complexes of PKC Isozymes and N-Biotinylated Peptides-The ␤-mercaptoethanol-free purified PKC preparation (25 ng) was incubated with N-biotinyl-RRRCLRRL (1 M) in the absence of DTT, as described above. Complexes of the peptide and PKC-␣, PKC-␤, and PKC-␥ were immunoprecipitated from the isozyme mixture with isozyme-specific monoclonal antibodies by previously described methods (9,10). Briefly, samples were incubated with 5 g of monoclonal anti-PKC-␣ (Transduction Laboratories), anti-PKC-␤, or anti-PKC-␥ (Seikagaku, Inc., Ijamsville, MD) by rotation for 2 h at 4°C. This was followed by further incubation in the presence of 5 g of rabbit anti-mouse IgG by rotation for 30 min at 4°C. Next, 50 l of protein A-Sepharose was added to the incubation mixture followed by rotation for 30 min at 4°C. The suspension was subjected three times successively to a 5-min spin in a microcentrifuge followed by washing of the Sepharose pellet. The wash buffer contained 10 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 0.5% Nonidet P-40, 0.2 mM sodium vanadate, and 0.2 mM PMSF. Immune complexes were recovered from the beads by suspension of the pellet in 100 l of 2 ϫ ␤-mercaptoethanol-free SDS-PAGE sample buffer and boiling for 5 min. Supernatants were subjected to SDS-PAGE, and avidin-reactive bands were detected by ECL, as described above.

Design of N-Biotinylated Peptide Substrate Analogs That
Inactivate the Calcium-and PS-dependent PKC Subfamily-We previously demonstrated that the synthetic peptide substrate analog RKRCLRRL is a DTT-sensitive irreversible inactivator of Ca 2ϩ -and PS-dependent PKC isozymes (20). The Ca 2ϩ -and PS-dependent PKC subfamily members ␣, ␤, and ␥ were similarly susceptible to inactivation by RKRCLRRL (20). Based on our results, we hypothesized that the inactivation mechanism most likely involved formation of a disulfide linkage between the inactivator peptide and an active-site Cys residue of PKC (20). The goal of the present study was to evaluate the sequence RKRCLRRL as a basis for the design of N-biotinylated active-site affinity labels of PKC isozymes. The first objective was to optimize the sequence for DTT-sensitive inactivation of the Ca 2ϩ -and PS-dependent PKC subfamily. We previously reported that RKRCLRRL analogs that lacked a single basic residue were weak or inactive against Ca 2ϩ -and PS-dependent PKC activity, indicating the importance of the basic residues in the inactivation potency (20). With this in mind, we synthesized RKRCLRRL analogs in which substitutions of Arg 3 Lys or Lys 3 Arg were made, or the positions of basic residues were changed. The RKRCLRRL analogs investigated for their ability to function as DTT-sensitive inactivators of Ca 2ϩ -and PS-dependent PKC were RRRCLRRL, KKKCLKKL, RKRCRRLL, and RRKCRRLL.
The purified PKC preparation that we employed in these studies is a mixture of isozymes. Fig. 1 shows the results of immunoblot analysis of the PKC preparation with isozymespecific monoclonal antibodies. The PKC preparation contained PKC-␣, PKC-␤, and PKC-␥, which co-migrated at 82 kDa, and PKC-⑀ and PKC-, which migrated at 90 and 73 kDa, respec-tively. The preparation did not contain detectable levels of PKC-␦, PKC-⍜, PKC-, PKC-or PKC-(data not shown). Thus, the preparation contained at least one representative of each of the three PKC subfamilies. The Ca 2ϩ -dependent PKC subfamily members ␣, ␤, and ␥ have equivalent K m (app) values for histone III-S (25). The histone kinase activity of the PKC preparation was stimulated by Ca 2ϩ and PS by 10 -15-fold and served as a sensitive assay of catalysis by the Ca 2ϩ -dependent PKC subfamily. Fig. 2A shows the concentration dependence of inactivation of the Ca 2ϩ and PS-dependent histone kinase activity of PKC by the RKRCLRRL analogs. Peptides were preincubated with PKC in the absence of reducing thiol at the peptide concentration shown for 5 min at 30°C followed by 12-fold dilution of the preincubation mixtures into histone kinase assay mixtures, as described under "Materials and Methods." None of the peptides inhibited the histone kinase activity when 2.0 mM DTT was included in the preincubation mixture (data not shown). RKRCLRRL (q), which is the Cys peptide analog that corresponds to the natural PKC substrate recognition sequence in the EGF receptor (-RKRTLRRL-) (26), was the most potent DTT-sensitive PKC inactivator of the peptides examined, but substantial inactivation of Ca 2ϩ -and PS-dependent PKC (30 -50%) was also achieved by each of the other Arg-containing peptides at concentrations of Յ10 M. In contrast, KKKCLKKL (å) was inactive across the concentration range tested ( Fig.  2A).
Ideally, an active-site affinity label should resemble the substrate as closely as possible to optimally exploit the substrate selectivity of the enzyme. Because the objective of this study was to identify potential sequences for active-site affinity labels of PKC isozymes, it was important to determine whether inactivator peptide analogs containing Cys 3 Thr substitutions could serve as substrates. Fig. 2B shows that, of the five Cys peptides under investigation, only two corresponded to Thr analogs that served as substrates of the PKC preparation. RKRTLRRL (q), which is a known synthetic peptide substrate of PKC (K m (app) ϭ 20 M) (19) and RRRTLRRL (f) were phosphorylated by the PKC preparation in a Ca 2ϩ -and PS-dependent manner. Stimulation of the phosphorylation of the peptides by Ca 2ϩ and PS was 3.0 -4.0-fold. Based on these results, we focused our investigation on the inactivator peptide sequences RKRCLRRL and RRRCLRRL, and we included the sequence KKKCLKKL in our study as a negative control.
As an approach to allow detection of covalent complex formation between PKC isozymes and inactivating peptide substrate analogs, N-biotinylated peptides were prepared. We first  Immunoblot analysis of PKC isozymes in the purified rat brain PKC preparation. The purified rat brain PKC preparation was subjected to Western analysis (50 ng of PKC/lane) using previously described conditions and procedures (24) with isozyme-specific monoclonal antibodies employed at dilutions of 1:5000 (anti-PKC-␣), 1:2500 (anti PKC-␤), 1:5000 (anti-PKC-␥), 1:500 (anti-PKC-⑀), and 1:250 (anti-PKC-). The PKC preparation was incubated with the primary antibodies for 2 h at 37°C, and peroxidase-linked sheep anti-mouse Ig (1:300) was employed as the secondary antibody. SDS-PAGE was performed with 10% gels. Immunoreactive bands were detected by ECL. For further details, see "Materials and Methods."

FIG. 2. PKC inactivation by Cys-containing peptide substrate analogs.
A, a ␤-mercaptoethanol-free PKC preparation was preincubated with the indicated peptide substrate analog at the concentration shown in 20 mM Tris-HCl, pH 7.5, for 5 min at 30°C. To ascertain the DTT sensitivity of PKC inactivation, control preincubation mixtures containing 2.0 mM DTT were included in the analysis. The preincubated mixtures were kept on ice prior to dilution (12-fold) into histone kinase reaction mixtures. The Ca 2ϩ -and PS-dependent histone kinase activity of PKC was assayed as described in "Materials and Methods." % INAC-TIVATION represents the DTT-sensitive antagonism of the Ca 2ϩ -and PS-dependent histone kinase activity of PKC by the peptide at the concentration shown. q, RKRCLRRL; f, RRRCLRRL; Ç, RRKCRRLL; E, RKRCRRLL; å, KKKCLKKL. 100% activity ϭ 276 Ϯ 24 nmol of 32 P transferred per min/mg. Each experimental value is an average of triplicate determinations; the experimental results shown were reproducible in a duplicate experiment. B, Cys 3 Thr analogs of the peptides in A were tested for their ability to serve as PKC substrates. PKCcatalyzed peptide phosphorylation was assayed as described under "Materials and Methods." PMOL/MIN represents the picomoles of 32 P transferred from [␥-32 P]ATP to the synthetic peptide per min, in a Ca 2ϩand PS-dependent manner, at the peptide concentration shown. Each assay mixture contained 5 ng of PKC. q, RKRTLRRL; f, RRRTLRRL; Ç, RRKTRRLL; E, RKRTRRLL; å, KKKTLKKL. Each point represents the mean value from two independent experiments done in triplicate. Error margins fell within the symbol size employed.
Having ascertained that the N-biotinylated peptides were recognized as substrates by the Ca 2ϩ -and PS-dependent PKC subfamily, we next prepared N-biotinylated analogs of the peptides RKRCLRRL, RRRCLRRL, and KKKCLKKL and determined their inactivation potencies against the histone kinase activity of PKC. Fig. 3 compares the concentration dependence of the N-biotinylated peptides with RKRCLRRL (å) in the DTT-sensitive inactivation of Ca 2ϩ -and PS-dependent PKC. None of the N-biotinylated peptide substrate analogs inhibited the histone kinase activity of PKC when PKC-peptide preincubation mixtures contained 2.0 mM DTT (data not shown). In the absence of DTT, N-biotinyl-RRRCLRRL (ࡗ) was the most potent PKC inactivator of the peptides examined, achieving 54 Ϯ 3% inactivation at a concentration of 0.5 M (Fig. 3). Both Arg-containing N-biotinylated peptides under investigation, Nbiotinyl-RRRCLRRL (ࡗ) and N-biotinyl-RKRCLRRL (Ç), effected potent inactivation (Ͼ75%) at concentrations of Յ 5 M (Fig. 3). In contrast, N-biotinyl-KKKCLKKL (f), like its nonbiotinylated counterpart, was by comparison a very weak PKC inactivator (Fig. 3). This underscores the importance of Arg residues in the potency of the RKRCLRRL inactivator peptide series against Ca 2ϩ -and PS-dependent PKC. These results show that the sequences N-biotinyl-RKRXLRRL and N-biotinyl-RRRXLRRL are comparably effective in producing excellent substrates (X representing T) and DTT-sensitive inactivators (X representing C) of the Ca 2ϩ -and PS-dependent PKC subfamily and that the substrate and inactivator peptides corresponding to the sequence N-biotinyl-KKKXLKKL are relatively weak.

Demonstration of Covalent Complex Formation between PKC Isozymes and Inactivating N-Biotinylated Peptide Substrate
Analogs-To determine whether the N-biotinylated inactivator peptides formed covalent complexes with PKC isozymes, the PKC preparation was preincubated with the N-biotinylated peptides under the conditions employed in the inactivation experiment (Fig. 3), and an aliquot of the preincubation mixture was boiled in ␤-mercaptoethanol-free SDS-PAGE sample buffer. The sample was subjected to SDS-PAGE on 10% gels, protein was transferred to nitrocellulose filters, and avidin was employed for the detection of biotinylated species (27,28) ) (lane 7). The 90-and 73-kDa bands are adjacent to the 82-kDa band, which is indicated by an arrow. Because the sample employed was a highly purified PKC isozyme mixture, the two relatively weak, lower molecular mass bands evident in lanes 5 and 6 at 62 and 66 kDa most likely represent copurifying proteolytic fragments of PKC isozymes. No bands were detected in PKC samples containing nonbiotinylated peptides (lanes 2-4) or PKC alone (lane 1). To determine whether there were differences in the amount of PKC per lane as a result of loading errors, we stripped the filter corresponding to the upper panel of Fig. 4A and subjected it to immunoblot analysis with anti-pan-PKC, which recognizes PKC-␣, PKC-␤, and PKC-␥ (29), i.e. the 82-kDa PKC subspecies. We detected similar amounts of the PKC subspecies, which migrated at 82 kDa, in each lane (Fig. 4A, lower panel). When blots were stripped and probed with anti-PKC-⑀ and anti-PKCsubsequent to the detection of avidin-reactive species at 90, 82, and 73 kDa, 90-and 73-kDa bands were detected by anti-PKC-⑀ and anti-PKC-, respectively (data not shown). The Ͼ90% inactivation of the Ca 2ϩ -and PS-dependent histone kinase activity of the PKC preparation achieved by both N-biotinyl-RRRCLRRL and N-biotinyl-RKRCLRRL (Fig. 3) suggested that the inactivator peptides complex with each of the Ca 2ϩ -and PS-dependent PKC isozymes in the preparation (␣, ␤, ␥). To determine whether this was the case, we complexed the PKC preparation with 1 M N-biotinyl-RRRCLRRL as described above and used isozyme-specific monoclonal antibodies to immunoprecipitate PKC-␣, PKC-␤, and PKC-␥ from the isozyme mixture (see "Materials and Methods"). Each immune complex was subjected to SDS-PAGE, electrophoretically transferred to a polyvinylidene difluoride membrane, and probed for avidin-reactive species. Fig. 4B shows that immunoprecipitated PKC-␣, PKC-␤, and PKC-␥ each produced an 82-kDa avidin-reactive species, indicating that each of the Ca 2ϩ -and PS-dependent PKC isozymes forms a complex with N-biotinyl-RRRCLRRL that is stable to denaturation.
The results shown in Fig. 4 demonstrate that the N-biotinylated inactivator peptides form complexes with PKC isozymes (␣, ␤, ␥, ⑀, ) that are stable to denaturing conditions. To test whether complex formation required specific interactions between the inactivator peptides and native PKC, we compared the ability of the N-biotinylated inactivator peptides to stably complex with native versus denatured PKC. In these experiments, N-biotinylated inactivator peptides were incubated with the PKC preparation either before (native state) or after (denatured state) the enzyme had been boiled in ␤-mercaptoethanol-free SDS-PAGE sample buffer. Results shown in the upper panel of  2 and 4) was readily detected. When the filter was stripped and probed with antipan-PKC, similar amounts of PKC were detected in each lane (Fig. 5, lower panel). The inability of the inactivator peptide to complex with the denatured PKC subspecies at 73 kDa (), 82 kDa (␣, ␤, ␥), and 90 kDa (⑀) indicates that the stability of each PKC isozyme-inactivator peptide complex to denaturation is due to a covalent intermolecular linkage.

Covalent Modification of the Calcium-and PS-dependent PKC Subspecies by N-Biotinylated Peptide Substrate Analogs Is
Associated with DTT-sensitive Inactivation-Having established that the N-biotinylated peptides covalently modify several PKC isozymes, we next focused on the relationship between the inactivation and covalent modification of the Ca 2ϩand PS-dependent (82-kDa) PKC subspecies by the N-biotinylated inactivator peptides. DTT sensitivity was an important feature of the inactivation of the Ca 2ϩ -and PS-dependent histone kinase activity of PKC by RKRCLRRL (20) and the Nbiotinylated inactivator peptides (Fig. 3). We next investigated whether the covalent complexes formed between the 82-kDa PKC subspecies and the N-biotinylated inactivator peptides were also DTT-sensitive. The PKC preparation was preincubated with 1 M N-biotinyl-RKRCLRRL and N-biotinyl-RRRCLRRL in the presence or absence of DTT. The upper panel of Fig. 6 shows that no complex was detected by avidin if 2 mM DTT was present when the PKC preparation was prein-  Fig. 6 were not due to loading errors (Fig. 6, lower panel). These results demonstrate the DTT sensitivity of the covalent linkage between the 82-kDa PKC subspecies and the N-biotinylated inactivator peptides.
Next, we compared the concentration dependence of N-biotinyl-RRRCLRRL in the inactivation (Fig. 3) and covalent mod-  2 and 3) for 5 min at 30°C either as described in the legend to Fig. 2A (lanes 1, 2, and 4) N-biotinyl-RKRCLRRL (lanes 2, 4, and 5), 1  M N-biotinyl-RRRCLRRL (lanes 3, 6, and 7), or alone (lane 1). The samples were subjected to a modified Western analysis, in which the blot was probed for biotinylated species with avidin conjugated to horseradish peroxidase (upper panel). The blot was then stripped and probed for the 82-kDa PKC subspecies with anti-pan-PKC (lower panel). For further details, see "Materials and Methods" and the legend to Fig. 4A. ification of the Ca 2ϩ -and PS-dependent 82-kDa PKC subspecies (Fig. 7, A and B). In each case, a maximal or near maximal effect was observed at 5 M N-biotinyl-RRRCLRRL (Figs. 3 and  7). 50% of the maximal effect corresponded to approximately 0.5 M N-biotinyl-RRRCLRRL in the inactivation of Ca 2ϩ -and PS-dependent PKC (Fig. 3) and 1-2 M N-biotinyl-RRRCLRRL in the covalent modification of the 82-kDa subspecies (Fig. 7B). Overall, there was generally a good correlation between the potencies of N-biotinyl-RRRCLRRL in the inactivation and covalent modification of the 82-kDa PKC subspecies across the peptide concentration range of 0.1-50 M, with the extent of covalent modification plateauing at peptide concentrations that achieved maximal inactivation (5-50 M) (Fig. 3 and Fig. 7, A and B).
Previously, we reported that substrates and substrate analogs protect the Ca 2ϩ -and PS-dependent PKC subfamily against inactivation by RKRCLRRL (20). The observed protection supported an active site-directed inactivation mechanism (20). We next compared the ability of 6 M MgATP and a pseudosubstrate peptide (200 M Phe-Ala-Arg-Lys-Gly-Ala-Leu-Arg-Gln) (30) to protect the Ca 2ϩ -and PS-dependent PKC subspecies against covalent modification and inactivation by 1 M N-biotinyl-RRRCLRRL. PKC and 1 M N-biotinyl-RRRCLRRL were preincubated under control conditions or in the presence of MgATP with or without pseudosubstrate peptide, and the inactivation and labeling of Ca 2ϩ -and PS-dependent PKC by the biotinylated peptide was measured. In Fig. 8 Fig. 2A. Covalent modification of the 82-kDa PKC subspecies by the peptide was analyzed as described under "Materials and Methods" and the legend to Fig. 4A. In the upper panel, avidin conjugated to horseradish peroxidase was employed as a probe of biotinylated species, and in the lower panel anti-pan-PKC (29) was used as a probe of the 82-kDa PKC subspecies. In the upper panel, the bands appearing immediately above and below 82 kDa have the electrophoretic mobilities of PKC-⑀ (90 kDa) and PKC-(73 kDa), respectively. B, the line graph shows the results of densitometric analysis of the avidin-reactive 82 kDa band in the xerogram shown in A. The inset shows the results of densitometric analysis at 82 kDa of a xerogram that corresponds to an analogous experiment that was done in parallel. For further details, see "Materials and Methods" and the legend to Fig. 4A. against the inactivation and covalent modification of the 82-kDa PKC subspecies by N-biotinyl-RRRCLRRL. To test the importance of the location of the Cys residue at the phosphoacceptor position in the inactivator peptide N-biotinyl-RRRCLRRL, we examined the labeling of PKC by analogs of the inactivator peptide that contained the Cys residue at alternative positions. At 1 M, N-biotinyl-CRRRLRRL, N-biotinyl-RRRLRRLC, and N-biotinyl-RRRLRRCL were all weaker than the parent peptide in labeling PKC, with respective efficacies in the labeling of the 82-kDa PKC species of 10%, 28%, and 7% of the labeling achieved by N-biotinyl-RRRCLRRL, according to densitometric analysis. We also investigated the effects of the stimulatory cofactors Ca 2ϩ and PS on the labeling of PKC by N-biotinyl-RRRCLRRL. Preincubation with 0.2 mM Ca 2ϩ plus 10 -30 g/ml PS failed to enhance the labeling of the PKC isozymes by the N-biotinylated peptide substrate analog (data not shown) (see "Discussion").
As a further test of whether the inactivator peptides operated by an active site-directed mechanism, we subjected Nbiotinyl-RRRCLRRL-modified PKC to limited trypsinolysis in an attempt to map the label to the catalytic or regulatory domain of the enzyme. Fig. 9 shows the avidin reactivity of modified PKC before (lane 1) and after (lane 2) limited trypsinolysis. A major broad avidin-reactive band is evident in lane 2 at 50 kDa, which corresponds to the molecular mass of the catalytic domain fragment of Ca 2ϩ -dependent PKC subspecies (29). Lanes 1 and 2 were stripped and probed with anti-pan-PKC, which recognizes the PKC isozymes ␣, ␤, and ␥ and their catalytic domain fragments (29). Results are shown in lanes 3 and 4. The immunoreactivity of the 50-kDa fragment with anti-pan-PKC (lane 4) confirmed its identity as the catalyticdomain fragment. Thus, the avidin-reactive label maps to the catalytic domain of the Ca 2ϩ -dependent PKC subfamily. These results, together with the protection against modification afforded by substrates, provide evidence that the peptide substrate analog N-biotinyl-RRRCLRRL covalently modifies and inactivates Ca 2ϩ -and PS-dependent PKC by binding at the active site. DISCUSSION We previously reported that the peptide substrate analog RKRCLRRL is an active site-directed irreversible inactivator of Ca 2ϩ -and PS-dependent PKC isozymes (20). We hypothesized that the inactivation mechanism entailed covalent modification of the PKC isozymes at the active site (20). To test this hypothesis, in this report we investigated the ability of N-biotinylated analogs of RKRCLRRL to inactivate and covalently modify PKC isozymes. N-biotinyl-RRRCLRRL covalently modified na-tive but not denatured PKC-␣, -␤, -␥, -⑀, and -. The Ca 2ϩ -and PS-dependent PKC isozymes ␣, ␤, and ␥ comigrated at 82 kDa in SDS-PAGE analysis. Inactivation of Ca 2ϩ -and PS-dependent PKC by the N-biotinylated inactivator peptides was associated with covalent modification of the 82-kDa PKC subspecies. Inactivation of Ca 2ϩ -and PS-dependent PKC and covalent modification of the 82-kDa PKC subspecies were both DTTsensitive. The relative potencies of N-biotinyl-RRRCLRRL, Nbiotinyl-RKRCLRRL, and N-biotinyl-KKKCLKKL in the inactivation of Ca 2ϩ -and PS-dependent PKC corresponded to their relative abilities to label the 82-kDa PKC subspecies. In addition, the concentration dependence of N-biotinyl-RRRCLRRL was similar in the inactivation of Ca 2ϩ -and PS-dependent PKC and the covalent labeling of the 82-kDa PKC subspecies. (The discrepancy between the concentration dependence curves for inactivation and covalent modification noted under "Results" may be due to a slow dissolution of the complex by residual ␤-mercaptoethanol during the comparatively lengthy procedure employed to detect the covalent complex). Because the RKRCLRRL peptide series inactivates PKC at an extremely rapid rate, achieving complete inactivation at 4°C within several seconds (20), the inactivation of PKC by the peptides was not amenable to kinetic analysis. Substantial protection against both inactivation and covalent modification of the 82-kDa PKC subspecies was achieved by a nucleotide substrate and a peptide substrate analog. These results provide strong evidence that the mechanism of PKC inactivation by the RKRCLRRL peptide series (20) entails covalent modification of the enzyme.
The protection afforded by substrates/pseudosubstrates against PKC inactivation and modification also provides evidence that the N-biotinylated inactivator peptides label PKC isozymes at the active site (15). Consistent with this, limited trypsinolysis mapped the covalent label to the catalytic domain. Additional support for active-site labeling is provided by the observation that Cys 3 Thr analogs of the inactivator peptides served as PKC substrates, with the potent inactivator peptides (N-biotinyl-RRRCLRRL and N-biotinyl-RKRCLRRL) yielding potent peptide substrates and the weak inactivator peptide (N-biotinyl-KKKCLKKL) yielding a weak peptide substrate. Taken together, our results provide evidence that the inactivating N-biotinylated peptide substrate analogs are active-site affinity labels of PKC isozymes. It should be noted that active-site labeling of PKC-␣, -␤, -␥, -⑀, and -by the inactivating peptide substrate analogs would be consistent with the known relatedness of the substrate selectivities of these isozymes (31)(32)(33).
Whereas labeling of PKC isozymes by the inactivator peptides at non-active-site residues cannot be ruled out, the specificity of the labeling is indicated by the observations that the peptides do not modify denatured PKC isozymes (Fig. 5) and that high concentrations of N-biotinyl-RRRCLRRL (Ն10 M) do not label additional sites of the fully inactivated 82-kDa PKCpeptide complex (Figs. 3 and 7). An active-site Cys residue is conserved in several mammalian protein kinases including the PKC isozyme family and cAMP-dependent protein kinase (34). In cAMP-dependent protein kinase, the active-site Cys exhibits high reactivity (15,17,35,36). Although the crystal structure of the catalytic domain of PKC is not yet known, structural modeling has shown that the catalytic core of PKC retains the secondary and tertiary elements of the crystal structure of cAMP-dependent protein kinase (34). The active-site Cys residue of cAMP-dependent protein kinase (Cys-199) corresponds to Cys-499 in PKC-␣, which is present in the peptide substrate binding lobe (C-4 region). Our hypothesis that peptide substrate analogs with Cys substitutions at the phosphoacceptor FIG. 9. Mapping of the covalent modification of PKC isozymes by an N-biotinylated inactivator peptide to the catalytic domain. PKC was preincubated with 1 M N-biotinyl-RRRCLRRL as described under "Materials and Methods," and the complex formed was subjected to limited trypsinolysis under previously described conditions (19,22). Covalent complex formation was detected by avidin reactivity as described under "Materials and Methods" and the legend to Fig. 4A. Lanes 1 and 2 show the avidin-reactive species detected in the peptidemodified PKC sample before and after trypsinolysis, respectively. The immunoreactive bands detected when Lanes 1 and 2 were stripped and reprobed with anti-pan-PKC, which recognizes the 82-kDa PKC subspecies and its catalytic domain fragment (50 kDa) (29) (see legend to residue S-thiolate Cys-499 (PKC-␣) is consistent with the placement of the active-site Cys residue in the crystal structure of cAMP-dependent protein kinase within the peptide substrate binding domain and proximal to the ␥-phosphate of ATP (35,36), and it is supported by our observation that moving the Cys residue of N-biotinyl-RRRCLRRL from the phosphoacceptor site to alternative positions reduced the potency of the inactivator peptide. Our hypothesis is also consistent with the covalent modification of the active-site Cys of cAMP-dependent protein kinase concomitant with enzyme inactivation by a peptide substrate analog bearing a disulfide at the phosphoacceptor position (15). The observations that RKRCLRRL is a better inactivator than RRRCLRRL and that the latter Thr analog is a much better substrate suggest that subtle differences in the positioning of the Cys inactivator and Thr substrate peptides in the active site of PKC may be required, respectively, for optimal alignment of the thiol of the inactivator peptide with the side chain of Cys-499 (PKC-␣) and for optimal alignment of the phosphoacceptor hydroxyl of the peptide substrate with the ␥-phosphate of active site-bound ATP. Studies are under way to determine whether the N-biotinylated inactivator peptides are active-site affinity labels of PKC that function by selective modification of the active-site Cys. In these studies, the modified site(s) in the catalytic domains of inactivated PKC isozymes will be mapped and sequenced.
Although the Cys 3 Thr analogs of the N-biotinylated inactivator peptides were phosphorylated in a Ca 2ϩ -and PS-dependent manner, the degree of dependence (3-4-fold) was much less than that observed with histone (10 -15-fold). The amount of N-biotinylated Thr peptide phosphorylation observed in the absence of Ca 2ϩ and PS was substantial, e.g. 1.6 -2.6 nmol of 32 P/min/g of PKC at 10 M of the Arg-containing N-biotinylated Thr peptides (data not shown). This indicates that the N-biotinylated Thr peptides bind to the active site of PKC in the absence of stimulating cofactors, albeit suboptimally. In their relatively modest dependence on activating cofactors, the N-biotinylated Thr peptide substrates bear some resemblance to the cofactor-independent substrate protamine sulfate, which has been shown to have in common with stimulating cofactors the ability to induce a structural change in PKC that displaces the pseudosubstrate domain from the active site (3,37). Our results suggest that the N-biotinylated Thr and Cys peptides may produce a modest induction of the activated conformation of PKC. This would account for the otherwise paradoxical observation that N-biotinylated Thr peptides displace the pseudosubstrate domain from the active site, which is clearly a prerequisite for the phosphorylation of the Thr peptides that was observed in the absence of Ca 2ϩ and PS, and a synthetic peptide corresponding to the pseudosubstrate domain protects the active site of PKC against labeling by an N-biotinylated Cys peptide in the absence of PKC-stimulatory cofactors (Fig. 8).
Furthermore, this is the most plausible model to reconcile the Ca 2ϩ and PS independence of the modification of PKC isozymes by the N-biotinylated Cys peptides with the evidence that the N-biotinylated Cys peptides modify the isozymes at the active site. The Ca 2ϩ and PS independence of the modification of PKC isozymes by N-biotinylated Cys peptides whose Thr analogs exhibit modest dependence on the stimulatory cofactors may be a kinetic effect, i.e. a result of the very rapid rate of the S-thiolation reaction, which is complete within several seconds, relative to the rate of peptide substrate phosphorylation, which is linear across a 20-min time course.
The mechanism of inactivation of Ca 2ϩ -and PS-dependent PKC most consistent with our results is that the inactivator peptide S-thiolates PKC at the active site, forming a disulfide linkage between the inactivator peptide and the active-site Cys residue of PKC. The conservation of the active-site Cys residue in only a subset of protein kinases and studies of its modification in cAMP-dependent protein kinase indicate that it is very unlikely that the active-site Cys residue participates in protein kinase catalysis (34 -36). It remains to be determined whether the active-site Cys is conserved in PKC for a regulatory function. S-Thiolation is a reversible modification of select proteins in cells undergoing mild oxidative stress that offers protection against permanent oxidative damage and in some cases regulates enzyme activity (37,38). Cellular PKC is subject to DTTreversible oxidative activation and inactivation under conditions of mild oxidative stress (39); the mechanisms underlying oxidative regulation of PKC have not yet been elucidated. The results presented here suggest that PKC may be subject to endogenous regulatory mechanisms in vivo that involve Sthiolation of the active-site Cys residue in response to oxidative stress.