Differential and Common Recognition of the Catalytic Sites of the cGMP-dependent and CAMP-dependent Protein Kinases by Inhibitory Peptides Derived from the Heat-stable Inhibitor Protein*

Synthetic peptides corresponding to the active do- main ,of the heat-stable inhibitor protein of CAMP-dependent protein kinase (Cheng, HA., Kemp, B. E., Pearson, R. B., Smith, A. J., Misconi, L., Van Patten, S. M., and Walsh, D. A. (1986) J. Biol. Chem. 261, 989-992) were tested as inhibitors of cGMP-depend-ent protein kinase. The peptides themselves were not substrates. cGMP-dependent protein kinase activity was assayed using histone H2B and two synthetic pep- tide substrates. Consistent with previous observations of other peptide inhibitors of this enzyme (Glass, D. B. Biochem. J. 213,159-164), the inhibitorypep-tides had no effect on the phosphorylation of histone H2B, but they competitively inhibited cGMP-depend-ent phosphorylation of the two peptide substrates. The parent inhibitor peptide, PKI(5-24)amide, and a series of analogs had Ki (or ICa,) values for cGMP-dependent protein kinase in the range of 15-190 PM. In contrast to their effects on the CAMP-dependent protein kinase, the inhibitory peptides were substantially less potent with cGMP-dependent

ent protein kinase. The peptides themselves were not substrates. cGMP-dependent protein kinase activity was assayed using histone H2B and two synthetic peptide substrates. Consistent with previous observations of other peptide inhibitors of this enzyme (Glass, D. B. (1983) Biochem. J. 213,[159][160][161][162][163][164], the inhibitorypeptides had no effect on the phosphorylation of histone H2B, but they competitively inhibited cGMP-dependent phosphorylation of the two peptide substrates. The parent inhibitor peptide, PKI (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)amide, and a series of analogs had Ki (or ICa,) values for cGMP-dependent protein kinase in the range of  In contrast to their effects on the CAMP-dependent protein kinase, the inhibitory peptides were substantially less potent with cGMP-dependent protein kinase, and potency was reduced by the presence of the NH,-terminal residues (residues 5-13). We conclude that the two protein kinases share a recognition of the basic amino acid cluster within the pseudosubstrate region of the peptide, but that the cGMP-dependent protein kinase does not recognize additional NHz-terminaf determinants that make the inhibitor protein extremely potent toward the CAMP-dependent enzyme. Even-when tested at high concentrations and with peptide substrates, the native inhibitor protein did not inhibit cGMP-dependent protein kinase under assay conditions in which the peptides derived from it were inhibitory. Thus, the native inhibitor protein appears to have structural features which block interaction with the cGMP-dependent enzyme and enhance its selectivity for CAMP-dependent protein kinase.
The cGMP-dependent and CAMP-dependent protein kinases are homologous proteins (1) which have similar but not identical substrate specificities (2)(3)(4). The subunit structures and mechanisms of activation are different, however, with the CAMP-dependent enzyme being dissociated into subunits by GM28144 (to D. B. G.) and AM21019 (to D. A. W) and by the * This work was supported by National Institutes of Health Grants National Health and Medical Council of Australia (to B. E. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. CAMP (5), while activation of the cGMP-dependent protein kinase by cGMP does not involve subunit dissociation (6). One of the other striking differences between the two enzymes is the selective inhibition of the CAMP-dependent protein kinase by the heat-stable inhibitor protein of that enzyme (7). Several reports (8)(9)(10)(11)(12)(13) have indicated that this inhibitor protein does not affect the cGMP-dependent enzyme, at least under the assay conditions used and with the concentration of native inhibitor protein available.

EXPERIMENTAL PROCEDURES
Synthetic Peptides-The 20-amino-acid peptide from the inhibitor protein of the CAMP-dependent protein kinase and ita analogs were synthesized as COOH-terminal amides by solid-phase synthesis and purified as described by Cheng et al. (25). The purity of these peptides has been previously confvmed by high performance liquid chromatography, amino acid analysis, and sequencing (25,26). The sequences of the parent peptide, PKI(5-24)amide, and the analogs used in this study are given in Table I.
Enzyme, Inhibitor Protein, and Histone Purification-Cyclic GMPdependent protein kinase was purified to homogeneity from bovine lung as described by Glass and Krebs (3). The enzyme had a specific activity of 3.5 pmol/min/mg when assayed with histone H2B as substrate at a concentration of 36 p~ under the conditions described previously (3). Cyclic GMP-dependent protein kinase was quantitated by the protein assay of Lowry et al. (31) using bovine serum albumin as standard.
Inhibitor Protein was purified by our recently described modifications (23). Purified histone H2B was prepared from calf thymus by the methods of Johns (32) and Oliver et al. (33). The concentrations of the stock solutions of both were determined by amino acid analysis.

3, show PKI
The kinetic mechanisms of the cyclic nucleotide-dependent protein kinases are probably ordered Bi-Bi reactions in which MgATP2-binds first (9, 40, 42, 43). It has also been shown

Protein kinase activity was assayed with
[T-~~PIATP as variable substrate as described under "Experimental Procedures." B, mixedtype noncompetitive inhibition versus MgATPz-by inhibitor peptide PKI (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)amide. Inhibitor peptide concentrations ( p~) were 0 (0), 53 (O), 106 (O), and 212 (m). Protein kinase activity was assayed with [y3'P]ATP as variable substrate. For both A and B, (Ala3')H2B (29)(30)(31)(32)(33)(34)(35) was at a concentration of 30 p~. The insets in both show the secondary replots versus l/V-from which were calculated the K,, (uncompetitive) inhibition constants in accord with Cleland (46). The that the interaction of the heat-stable inhibitor protein with the catalytic subunit of CAMP-dependent protein kinase is enhanced by preincubation with ATP (21,22). Because of this, we examined whether the order of addition of reagents might affect the potency of inhibition of the cGMP-dependent protein kinase by the inhibitory peptides. The enzyme was preincubated at 30 "C with PKI(7-22)amide, MgATP2-, and all the standard reaction components except phosphorylaccepting substrate, and the reaction was then initiated by the addition of peptide substrate. Under these conditions, exactly the same inhibitory potency was obtained as when the reaction was initiated by the addition of cGMP-dependent protein kinase (data not shown). In addition to these experiments, PKI(5-22)amide, PKI(10-24)amide, and PKI(l4-24)amide were directly tested as possible substrates of the cGMP-dependent protein kinase. Reaction conditions were as described under "Experimental Procedures." None of the inhibitory peptides was phosphorylated after 120 min of incubation at substrate concentrations of 50 p~ and a final concentration of cGMP-dependent protein kinase that was 50-fold greater than that used in the inhibition experiments. Under these same conditions, (Ala34)H2B (29)(30)(31)(32)(33)(34)(35) and Kemptide were stoichiometrically phosphorylated in less than 15 min (data not shown). Substrate-dependent Inhibition of the cGMP-dependent Protein Kinase-With previous studies, it has been shown that inhibition of the cGMP-dependent protein kinase by inhibitory peptides binding at the catalytic site was apparently peptide/protein substrate-dependent (9). Thus, (Ala3')H2B- (29)(30)(31)(32)(33)(34)(35) and (A1a)Kemptide were found to be effective inhibitors when a range of peptides were used as the phosphorylacceptor substrate, but not when one of several histones were substrate. Similar differences were not observed for inhibition of the CAMP-dependent protein kinase by either of these two peptides (9). Because of these observations, the protein/peptide substrate dependency of inhibition of cGMP-dependent protein kinase by the PKI(5-24) derivatives was tested. For this study (Table 11), three different substrates of the enzyme were used at or near their K, concentrations. Of these, (A1a3')H2B (29)(30)(31)(32)(33)(34)(35), the substrate used in the experiments of Figs. 1-3, and histone H2B are kinetically excellent substrates for the cGMP-dependent protein kinase (K, values = 28 pM and 1.5 p~, respectively; Refs. 28 and 3). Kemptide, which has been used extensively as a model peptide substrate for the CAMP-dependent protein kinase, is readily phosphorylated by the cGMP-dependent enzyme, although its K,,, value is much higher for the latter than for the former enzyme (K, values = 231 ptM and 5 pM, respectively; Refs. 40 and 42). In Table I1 are presented the results of testing the inhibition of cGMP-dependent protein kinase by (Ala3')H2B (29)(30)(31)(32)(33)(34)(35) and four of the inhibitor protein peptides using these three different protein/peptide substrates.

Inhibitor
Concen- did not block the phosphorylation of histone H2B. The same result was observed for the four inhibitor protein peptides tested. Each of these peptides inhibited the phosphorylation of (Ala34)H2B (29)(30)(31)(32)(33)(34)(35) and Kemptide with similar potencies but none, over the concentrations tested, caused significant inhibition of histone H2B phosphorylation. The same results were also obtained independent of the order of addition of reaction components (not shown). The reasons for this apparent substrate-dependent action of the inhibitor peptides are not yet understood. It is not a consequence of the differences in K,,, values for the peptides and histone, because each was examined at or near its K,,, concentration. For (A1a3')H2B (29)(30)(31)(32)(33)(34)(35), the inability to inhibit substrate phosphorylation appears unique to histones since, whereas it does not inhibit the phosphorylation of various histones, it does inhibit cGMP-dependent protein kinase autophosphorylation (40) and the phosphorylation by the cGMP-dependent protein kinase of other proteins such as phosphorylase kinase, troponin, and the Type I regulatory subunit of the CAMPdependent protein kinase? Possibly, these effects are related to the "poly(L-arginine) binding site" of the cGMP-dependent protein kinase that has been reported by Walton and Gill (44,45); this anionic site binds histones, but is distinct from the catalytic site and may be in the enzyme's regulatory domain. The use of histones as substrate for the cGMP-dependent protein kinase, however, explains why, when previously reported (23,24), no inhibition of the cGMP-dependent protein kinase by the peptides derived from the inhibitor protein was observed.
Previously, it has been stated (8)(9)(10)(11)(12)(13)) that the inhibitor protein of the CAMP-dependent protein kinase does not inhibit the cGMP-dependent protein kinase and this conclusion has come from experiments that have used both histone and the peptides, (Ala34)H2B (29)(30)(31)(32)(33)(34)(35) and Kemptide, as substrates (9). Since high concentrations of PKI(5-24)amide peptides inhibited the cGMP-dependent protein kinase, we have reevaluated whether very high concentrations of inhibitor protein might be inhibitory. The data on this are presented in Fig. 4; only one peptide, (Ala3")H2B (29)(30)(31)(32)(33)(34)(35), was tested as substrate because of a shortage of available inhibitor protein at the concentrations needed. Even at 100 p~ (i.e. -lo5 x Ki for the CAMP-dependent protein kinase), the inhibitor protein did not inhibit the cGMP-dependent protein kinase; this is a concentration where the peptides derived from it were markedly inhibitory.

DISCUSSION
The data presented in this report provide a comparison between the protein/peptide binding domains in the catalytic sites of the cGMP-dependent and CAMP-dependent protein kinases. As we have recently reported (25,26), the PKI(5-24) peptide derived from the inhibitor protein, and presumably the inhibitor protein itself, contain at least two critical regions of amino acids that are essential for the high affinity interaction at the catalytic site of the CAMP-dependent protein kinase. These two domains are the arginine cluster of Arg"-Arglg, and some of the first 7 NHz-terminal residues (residues [5][6][7][8][9][10][11]. For the CAMP-dependent protein kinase, deletion or substitution of residues in either of these two domains markedly modifies inhibitory peptide binding. In contrast to this, it is clear from the data provided in this report that the cGMP-dependent protein kinase contains in common with the CAMP-dependent protein kinase the recognition of the arginine cluster, but does not recognize the NH,-terminal D. B. Glass, unpublished observations. domain amino acid sequence. It is well documented that basic amino acid residues serve as determinants of peptide substrate specificity for the cGMP-dependent protein kinase (2-4, 28, 34). Thus, an ionic interaction with the guanidine groups of the arginine cluster of the inhibitory peptides is probably occurring in a similar pseudosubstrate manner. This recognition of the arginine cluster domain is not unexpected since it reflects the similarity (but not identity) of protein substrate specificity of the cGMP-dependent and CAMP-dependent protein kinases (2-4) and is presumably based on the high degree of homology between them (1). Nevertheless, there are also clear distinctions between the binding of proteins to the two enzymes, as further evidenced by the results presented here. As noted by the data of Fig. 1 and Table I, when the NH2-terminal region of PKI (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24) is also present in the inhibitory peptides, binding of the arginine cluster to the cGMP-dependent protein kinase is partially diminished. Given the difference in the activation mechanisms of the two cyclic nucleotide-dependent protein kinases and, in consequence, the continued proximity in the cGMP-dependent protein kinase of the regulatory and catalytic regions, then possibly the regulatory region by steric hindrance may diminish binding when the NHz-terminal domain of the inhibitory peptide is present. Presumably, this difference in recognition of the NHz-terminal domain is one reason why the inhibitor protein is selective for the CAMP-dependent protein kinase. Other features of the native inhibitor protein, however, must also contribute to its high specificity since the native inhibitor protein does not bind to the cGMP-dependent protein kinase with an affinity similar to that of the inhibitory peptides derived from it (Fig. 4), and may not bind at all. It would appear that the inhibitor protein has evolved not only specific recognition sequences for the substrate binding site of the CAMP-dependent protein kinase, but also additional features that have diminished its interaction with other protein kinases such as the cGMP-dependent enzyme.