Poly-L-proline type II peptide mimics as probes of the active site occupancy requirements of cGMP-dependent protein kinase.

Based on the X-ray crystal structure of cAMP-dependent protein kinase (PKA) with the endogenous inhibitor PKI and the X-ray crystal structure of cyclin-dependent kinase 2 (CDK2) with a substrate peptide, a proposal is put forth that some protein kinases bind peptide substrates in their active sites in the poly-L-proline type II (PPII) conformation. In this work, PPII peptide mimics are evaluated as pseudosubstrate inhibitors of cGMP-dependent protein kinase (PKG) to explore if PKG also binds peptide substrates in the PPII conformation. Inhibition data of our PPII mimetics provide evidence that the P-1, P-2, and P-3 residues of substrate peptides bind in the PPII conformation (phi approximately -75 degrees, psi approximately 145 degrees). In addition, the inhibition data also suggest that the P-1, P-2, and P-3 residues in substrate peptides bind with a gauche(-) chi1 angle.

Spatiotemporal dynamics of guanosine 3',5'-cyclic monophosphate revealed by a genetically encoded, fluorescent indicator.

To investigate the dynamics of guanosine 3',5'-cyclic monophosphate (cGMP) in single living cells, we constructed genetically encoded, fluorescent cGMP indicators by bracketing cGMP-dependent protein kinase (cGPK), minus residues 1-77, between cyan and yellow mutants of green fluorescent protein. cGMP decreased fluorescence resonance energy transfer (FRET) and increased the ratio of cyan to yellow emissions by up to 1.5-fold with apparent dissociation constants of approximately 2 microM and >100:1 selectivity for cGMP over cAMP. To eliminate constitutive kinase activity, Thr(516) of cGPK was mutated to Ala. Emission ratio imaging of the indicators transfected into rat fetal lung fibroblast (RFL)-6 showed cGMP transients resulting from activation of soluble and particulate guanylyl cyclase, respectively, by nitric oxide (NO) and C-type natriuretic peptide (CNP). Whereas all naive cells tested responded to CNP, only 68% responded to NO. Both sets of signals showed large and variable (0.5-4 min) latencies. The phosphodiesterase (PDE) inhibitor 3-isobutyl-1-methylxanthine (IBMX) did not elevate cGMP on its own but consistently amplified responses to NO or CNP, suggesting that basal activity of guanylate cyclase is very low and emphasizing the importance of PDEs in cGMP recycling. A fraction of RFL cells showed slowly propagating tides of cGMP spreading across the cell in response to delocalized application of NO. Biolistically transfected Purkinje neurons showed cGMP responses to parallel fiber activity and NO donors, confirming that single-cell increases in cGMP occur under conditions appropriate to cause synaptic plasticity.

Highly specific, membrane-permeant peptide blockers of cGMP-dependent protein kinase Ialpha inhibit NO-induced cerebral dilation.

Arrays of octameric peptide libraries on cellulose paper were screened by using (32)P-autophosphorylated cGMP-dependent protein kinase Ialpha (cGPK) to identify peptide sequences with high binding affinity for cGPK. Iterative deconvolution of every amino acid position in the peptides identified the sequence LRK(5)H (W45) as having the highest binding affinity. Binding of W45 to cGPK resulted in selective inhibition of the kinase with K(i) values of 0.8 microM and 560 microM for cGPK and cAMP-dependent protein kinase (cAPK), respectively. Fusion of W45 to membrane translocation signals from HIV-1 tat protein (YGRKKRRQRRRPP-LRK(5)H, DT-2) or Drosophila Antennapedia homeo-domain (RQIKIWFQNRRMKWKK-LRK(5)H, DT-3) proved to be an efficient method for intracellular delivery of these highly charged peptides. Rapid translocation of the peptides into intact cerebral arteries was demonstrated by using fluorescein-labeled DT-2 and DT-3. The inhibitory potency of the fusion peptides was even greater than that for W45, with K(i) values of 12.5 nM and 25 nM for DT-2 and DT-3, respectively. Both peptides were still poor inhibitors of cAPK. Selective inhibition of cGPK by DT-2 or DT-3 in the presence of cAPK was demonstrated in vitro. In pressurized cerebral arteries, DT-2 and DT-3 substantially decreased NO-induced dilation. This study provides functional characterization of a class of selective cGPK inhibitor peptides in vascular smooth muscle and reveals a central role for cGPK in the modulation of vascular contractility.

Delineation of selective cyclic GMP-dependent protein kinase Ialpha substrate and inhibitor peptides based on combinatorial peptide libraries on paper.

Peptide libraries on cellulose paper have proven to be valuable tools for the a priori determination of substrate specificities of cyclic AMP- and cyclic GMP-dependent protein kinases (cAMP-kinase and cGMP-kinase) on the basis of octa-peptide sequences. Here, we report the extension of our peptide library screens to 12-mer and 14-mer peptide sequences, resulting in highly cGMP-kinase Ialpha selective peptides. The sequences TQAKRKKSLAMA-amide and TQAKRKKSLAMFLR-amide, with Km values for cGMP-kinase Ialpha of 0.7 and 0.26 microM and Vmax values of 11.5 and 10.9 micromol/min/mg, respectively, display a high specificity for this enzyme. Furthermore, replacing the phosphate acceptor residue serine with alanine in TQAKRKKSLAMA-amide resulted in the highly cGMP-kinase Ialpha selective inhibitor peptide TQAKRKKALAMA-amide, with inhibitor constants for cGMP-kinase Ialpha and cAMP-kinase of 7.5 microM and 750 microM, respectively. Selective cGMP-kinase inhibitors have the potential to play an important role in the elucidation of the distinct cellular functions of cGMP-kinase separate from those activated by cAMP-kinases, and, therefore, may play an important role as pharmaceutical targets. Molecular docking experiments of the most cGMP-kinase selective sequences on a molecular model of the catalytic domain of cGMP-kinase Ialpha suggest that they adopt unique conformations, which differ significantly from those observed for the cAMP-kinase-specific inhibitor PKI(5-24). Our results suggest that despite their structural similarities, cAMP-kinase and cGMP-kinase use distinct peptide substrate and inhibitor conformations, which could account for their unique substrate specificities. These findings are further supported by cAMP- and cGMP-kinase-selective inhibitor analogs with (D)-Ala residues at the inhibitory positions.

Identification of the amino acid sequences responsible for high affinity activation of cGMP kinase Ialpha.

The cGMP-dependent protein kinases (cGK) Ialpha and Ibeta have identical cGMP binding sites and catalytic domains. However, differences in their first 100 amino acids result in 15-fold different activation constants for cGMP. We constructed chimeras to identify those amino acid sequences that contribute to the high affinity cGK Ialpha and low affinity cGK Ibeta phenotype. The cGK Ialpha/Ibeta chimeras contained permutations of six amino-terminal regions (S1-S6) including the leucine zipper (S2), the autoinhibitory domain (S4), and the hinge domain (S5, S6). The exchange of S2 along with S4 switched the phenotype from cGK Ialpha to cGK Ibeta and vice versa, suggesting that the domains with the highest homology between the two isozymes determine their affinity for cGMP. The high affinity cGK Ialpha phenotype was also obtained by a specific substitution within the hinge domain. Chimeras with the sequence of cGK Ialpha in S5 and cGK Ibeta in S6 were activated at up to 6-fold lower cGMP concentrations than cGK Ialpha. Based on the activation constants of all chimeras constructed, empirical weighting factors have been calculated that quantitatively describe the contribution of the individual amino-terminal domains S1-S6 to the high affinity cGK Ialpha phenotype.

The catalytic domain of the cGMP-dependent protein kinase Ialpha modulates the cGMP-binding characteristics of its regulatory domain.

The cGMP-dependent protein kinase Ialpha (PKG Ialpha) possesses two functional moieties, the regulatory and catalytic domains, which reside on a single polypeptide chain. Here we report on the influence of the catalytic domain on the binding of cGMP to the regulatory domain. A deletion mutant, delta352-670 of PKG Ialpha, lacking the catalytic domain, was constructed and expressed in E. coli. The purified 38 kDa mutant protein showed strong reactivity toward tryptic proteolysis at residue Arg77. Thus, a double deletion fragment delta1-77/352-670 PKG Ialpha, lacking the N-terminus, was also purified. Both proteins had functional cGMP binding, but differed kinetically from the wild-type protein. First the affinity constants for cGMP were modulated, second the constructs showed no signs of cooperative cGMP binding and third dimerization of the delta352-670 mutant was abolished. Our results provide evidence that the catalytic domain forms an intimate interaction with the regulatory domain and modulates the kinetics of cGMP binding.

Isoleucine 368 is involved in low-affinity binding of N6-modified cAMP analogues to site B of the regulatory subunit of cAMP-dependent protein kinase I.

The regulatory (R) subunit of cAMP-dependent protein kinase has a well-defined domain structure including the two in-tandem cAMP-binding sites that constitute the C-terminus of the protein. The N-terminal binding site (A) has a considerably higher affinity for analogues of cAMP that are substituted with bulky and hydrophobic substituents at the 6-amino group of the adenine ring compared to the affinity observed at the second site (B). On the basis of the crystal structure of the catabolite gene activator protein from Escherichia coli, molecular modelling of the binding domains suggested that a tyrosine (Y244) in site A could be involved in a high-affinity hydrophobic interaction, whereas a corresponding isoleucine (I368) in domain B could lead to steric hindrance in the binding of bulky N6-substituted analogues. Site-directed mutagenesis was used to construct mutations in Y244 and I368. Binding displacement experiments showed that replacing the tyrosine in site A with isoleucine (Y244I) did not affect the interaction of either N6-substituted or otherwise modified analogues with this site. However, replacing I368 with tyrosine (I368Y) led to a 3-4-fold increase in affinity for those N6-modified analogues that had a hydrophobic group attached directly or close to the 6-amino molecule. We conclude that I368 is involved in the molecular interaction between binding domain B and the 6-amino group of the adenine moiety of cAMP and that this residue is partly responsible for the reduced affinity of N6-substituted cAMP analogues for this site.

Active site mutations define the pathway for the cooperative activation of cAMP-dependent protein kinase.

cAMP-dependent protein kinase (cAPK) is a heterotetramer containing two regulatory (R) and two catalytic (C) subunits. Each R-subunit contains two tandem cAMP-binding domains, and activation of cAPK is mediated by the cooperative, high affinity binding of cAMP to these two domains. Mutant R-subunits containing one intact high affinity cAMP-binding site and one defective site were used to define the pathway for activation and to delineate the unique roles that each cAMP-binding domain plays. Two mutations were introduced by replacing the essential Arg in each cAMP-binding site with Lys (R209K in Site A and R333K in Site B). Also, the double mutant (R209/333K) was constructed. Analysis of cAMP binding and dissociation and the apparent constants for holoenzyme activation and R- and C-subunit interaction, measured by analytical gel filtration and surface plasmon resonance, established the following: (1) For rR(R209K), occupancy of Site B is not sufficient to activate the holoenzyme; the low affinity Site A must also be occupied. In rR(R333K), Site A retains its high affinity for cAMP, but Site A cannot bind until the low affinity Site B is occupied. Thus, both mutants, for different reasons, have similar Ka's for activation that are approximately 20-fold higher than that of the wild-type holoenzyme. The double mutant with two defective sites is no worse than either single mutant. (2) Kinetic analysis of cAMP binding showed that the mutation in Site A or B abolishes high affinity cAMP binding to that site and slightly weakens the affinity of the adjacent site for cAMP. (3) In the presence of MgATP, both mutants rapidly form a stable holoenzyme even in the presence of cAMP in contrast to the wild-type R where holoenzyme forms slowly in vitro and requires dialysis. Regarding the mechanism of activation based on these and other mutants and from kinetic data, the following conclusions are reached: Site A provides the major contact site with the C-subunit; Site B is not essential for holoenzyme formation. Occupancy of Site A by cAMP mediates dissociation of the C-subunit. Site A is inaccessible to cAMP in the full length holoenzyme, while Site B is fully accessible. Access of cAMP to Site A is mediated by Site B. Thus Site B not only helps to shield Site A, it also provides the specific signal that "opens up" Site A. Finally, a nonfunctional Site A in the holoenzyme prevents stable binding of cAMP to Site B in the absence of subunit dissociation.

(RP)-cAMPS inhibits the cAMP-dependent protein kinase by blocking the cAMP-induced conformational transition.

(RP)-cAMPS is known to inhibit competitively the cAMP-induced activation of cAMP-dependent protein kinase (PKA). The molecular nature of this inhibition, however, is unknown. By monitoring the intrinsic tryptophan fluorescence of recombinant type I regulatory subunit of PKA under unfolding conditions, a free energy value (delta GDH2O) of 8.23 +/- 0.22 kcal/mol was calculated. The cAMP-free form of the regulatory subunit was less stable with delta GDH2O = 6.04 +/- 0.05 kcal/mol. Native stability was recovered by treatment of the cAMP-free protein with either cAMP or (SP)-cAMPS but not with (RP)-cAMPS. Thus, (RP)-cAMPS binding to the regulatory subunit keeps the protein in a locked conformation, unable to release the catalytic subunit. This finding was further supported by demonstrating that holoenzyme formation was greatly accelerated only when bound cAMP was replaced with (RP)-cAMPS but not with cAMP or (SP)-cAMPS.

Expression of a chimeric, cGMP-sensitive regulatory subunit of the cAMP-dependent protein kinase type I alpha.

To study the fluctuations of cGMP in living cells through changes of energy transfer of dissociable fluorescence labeled subunits, we constructed a cGMP-sensitive probe by combining the N-terminus of the type I regulatory subunit of cAMP-dependent protein kinase (PKA) with the cGMP binding sites of cGMP-dependent protein kinase I alpha (PKG). This chimeric regulatory subunit retained PKA-like dimerization and PKG-compatible cGMP binding constants (Kd = 53 nM) for both binding sites. High affinity interaction with the PKA catalytic subunit was verified by Surface Plasmon Resonance (Kd = 3.15 nM). Additionally, the chimera inhibits the formation of wild-type holoenzyme with an apparent Ki of 1.05 nM. Furthermore, cGMP dissociated the mutant holoenzyme with an apparent activation constant of 146 nM. Thus, our construct provides all the requirements needed to investigate changes in intracellular cGMP concentrations.

Regulatory subunit of protein kinase A: structure of deletion mutant with cAMP binding domains.

In the molecular scheme of living organisms, adenosine 3',5'-monophosphate (cyclic AMP or cAMP) has been a universal second messenger. In eukaryotic cells, the primary receptors for cAMP are the regulatory subunits of cAMP-dependent protein kinase. The crystal structure of a 1-91 deletion mutant of the type I alpha regulatory subunit was refined to 2.8 A resolution. Each of the two tandem cAMP binding domains provides an extensive network of hydrogen bonds that buries the cyclic phosphate and the ribose between two beta strands that are linked by a short alpha helix. Each adenine base stacks against an aromatic ring that lies outside the beta barrel. This structure provides a molecular basis for understanding how cAMP binds cooperatively to its receptor protein, thus mediating activation of the kinase.

Determination of cyclic nucleotide-dependent protein kinase substrate specificity by the use of peptide libraries on cellulose paper.

An iterative approach to the a priori determination of the substrate specificity of cAMP- and cGMP-dependent protein kinases (PKA and PKG) by the use of peptide libraries on cellulose paper is described. The starting point of the investigation was an octamer library with the general structure Ac-XXX12XXX, where X represents mixtures of all 20 natural amino acids and 1 and 2 represent individual amino acid residues. The library thus contained all possible 2.56 x 10(10) octamers, divided into 400 sublibraries with defined amino acids 1 and 2 each consisting of 6.4 x 10(7) sequences. After phosphorylation with the kinases in the presence of [gamma-32P]ATP, the sublibrarys Ac-XXXRRXXX and Ac-XXXRKXXX were identified as the best substrates for PKA and PKG, respectively. The second-generation libraries had the structures Ac-XXXRR12X and Ac-XXXRK12X for PKA and PKG and resulted in the most active sequence pools Ac-XXXRRASX and Ac-XXXRKKSX. After delineation of every position in the octameric sequence and extension of the investigation to decameric peptides, the best sequences, Ac-KRAERKASIY and Ac-TQKARKKSNA, were obtained for PKA and PKG, respectively. Promising octameric and decameric peptides were assembled 5 or 10 times each and assayed in order to determine the experimental scatter inherent in the approach. The kinetic data of several octameric and decameric sequences were determined in solution and compared to data for known substrates. The recognition motif of PKA was confirmed by this approach, and a novel substrate sequence for PKG was identified.(ABSTRACT TRUNCATED AT 250 WORDS)

Ala335 is essential for high-affinity cAMP-binding of both sites A and B of cAMP-dependent protein kinase type I.

A single amino acid substitution (Ala335Asp) in cAMP binding site B of the regulatory subunit of cAMP-dependent protein kinase type I was sufficient to abolish high affinity cAMP binding for both cAMP binding sites A and B. Furthermore, the Ala335Asp mutation increased the activation constant for cAMP of the mutant holoenzyme 30-fold and also enhanced the rate of holoenzyme formation. Thus, the substitution was responsible for the dominant negative phenotype of the enzyme. Activation of mutant holoenzyme with site-selective cAMP analogs indicated that the enzyme dissociated through binding to site A only. Our results provide evidence that Ala335 is an essential residue for high affinity cAMP binding of both sites as well as for the functional integrity of the enzyme.

Crosstalk between domains in the regulatory subunit of cAMP-dependent protein kinase: influence of amino terminus on cAMP binding and holoenzyme formation.

The regulatory (R) subunit of cAMP-dependent protein kinase os an asymmetric multidomain protein with a dimerization domain at the N-terminus, an autoinhibitors site, and two cAMP binding domains at the C-terminus. Activation of the tetrameric holoenzyme is mediated by the cooperative binding of cAMP to the two cAMP binding sites. To better understand how the various domains influence each other, the N-terminus (delta 1-91) up to the autoinhibitor site was deleted. Not only did this monomeric deletion mutant, purified from Escherichia coli, still bind cAMP and the catalytic (C) subunit with high affinity, holoenzyme formation was actually accelerated by at least 50-fold. MgATP also was not required for rapid reassociation of (delta 1-91)R(cAMP)2 and C. The Kd(cAMP) and the Ka(cAMP) were similar to those for holoenzyme formed with full-length R; however, cooperatively was lost. Thus the N-terminus, either by inter- or intraprotomer contacts, not only impedes holoenzyme formation but also influences the cooperative binding of cAMP. The 1-91 deletion also renders the remaining fragment resistant to proteolytic degradation. Finally, unlike full-length R, the mutant protein can migrate freely into the nucleus. Surface plasmon resonance studies for the first time enabled direct measurements of the association and dissociation rate constants both for the intact R and for (delta 1-91)R. Both displayed very fast on-rates (1 x 10(-5) M-1 s-1 and 1.1 x 10(-5) M-1 s-1, respectively) and extremely slow off-rates (2.3 x 10(5) M-1 and 4.3 x 10(5) M-1, respectively). Thus, unlike the heat-stable protein kinase inhibitor, the region preceding the autoinhibitor site in R does not contribute in a quantitatively significant way to the high-affinity binding of C.

Crystallization of a deletion mutant of the R-subunit of cAMP dependent protein kinase.

A single deletion (delta 1 to 91) mutant of the regulatory subunit of the cAMP dependent protein kinase was crystallized. The crystals are hexagonal P6(1)22 (P6(5)22) with a = b = 88.7 A and c = 179.9 A. The crystals diffract to 3 A resolution. There is one molecule per asymmetric unit.

Identifying the molecular switches that determine whether (Rp)-cAMPS functions as an antagonist or an agonist in the activation of cAMP-dependent protein kinase I.

Previous investigations revealed that under physiological conditions in the presence of MgATP the phosphorothioate analogue of cAMP, (Rp)-cAMPS, is a competitive inhibitor and antagonist for cAMP for cAMP-dependent protein kinases I and II [DeWit et al., (1984) Eur. J. Biochem. 142, 255-260]. For the type I holoenzyme, the antagonist properties of (Rp)-cAMPS are shown here to be absolutely dependent on MgATP. In the absence of MgATP, (Rp)-cAMPS serves as a weak agonist with a Ka of 7.9 microM. The high-affinity binding of MgATP imposes a barrier on cAMP-induced activation of the homoenzyme--a barrier that both cAMP and (Sp)-cAMPS, but not (Rp)-cAMPS, can overcome. In the absence of MgATP, this barrier no longer exists, and (Rp)-cAMPS functions as an agonist. The holoenzyme also was formed with mutant regulatory subunits. Replacing the essential arginine, predicted to bind the exocyclic oxygens of cAMP, in site A with lysine abolishes high-affinity binding of cAMP to site A. The holoenzyme formed with this mutant R-subunit is activated by (Rp)-cAMPS in both the presence and absence of MgATP. These results suggest that the stereospecific requirements for holoenzyme activation involve this guanidinium side chain. Mutations that eliminate the high-affinity binding of MgATP, such as the introduction of an autophosphorylation site in the autoinhibitory domain, also generate a holoenzyme that can be activated by (Rp)-cAMPS. In the case of the type II holoenzyme, (Rp)-cAMPS is an antagonist in both the presence and absence of MgATP, emphasizing distinct roles for MgATP in these two forms of cAMP-dependent protein kinase.

Unfolding of the regulatory subunit of cAMP-dependent protein kinase I.

The unfolding of the recombinant regulatory subunit of cAMP-dependent protein kinase I was followed by monitoring the intrinsic protein fluorescence. Unfolding proceeds in at least two stages. First, the quenching of fluorescence due to cAMP binding is abolished at relatively low levels of urea (less than 2 M) and is observed as an increase in intensity at 340 nm. The high-affinity binding of cAMP is retained in 3 M urea even though the quenching is lost. The second stage of unfolding, presumably representing unfolding of the polypeptide chain, is seen as a shift in lambda max from 340 to 353 nm. The midpoint concentration, Cm, for this process is 5.0 M. Cyclic AMP binding activity is lost at a half-maximal urea concentration of 3.5 M and precedes the shift in lambda max. Unfolding of the protein in the presence of urea was fully reversible; furthermore, the presence of excess levels of cAMP stabilized the regulatory subunit. A free energy value (delta GDH2O) of 7.1 +/- 0.2 kcal/mol was calculated for the native form of the protein when denaturation was induced with either urea or guanidine hydrochloride. Iodide quenching of tryptophan fluorescence was used to elucidate the number of tryptophan residues accessible during various stages of the unfolding process. In the native cAMP-bound form of the regulatory subunit, only one of the three tryptophans in the regulatory subunit is quenched by iodide while more than two tryptophans can be quenched with iodide in the presence of 3 M urea.

Role of MgATP in the activation and reassociation of cAMP-dependent protein kinase I: consequences of replacing the essential arginine in cAMP binding site A.

The type I form of cAMP-dependent protein kinase binds MgATP with a high affinity, and binding of MgATP decreases the affinity of the holoenzyme for cAMP [Hofmann et al. (1975) J. Biol. Chem. 250, 7795]. Holoenzyme was formed here with a mutant form of the bovine recombinant type I regulatory subunit where the essential arginine in site A, Arg-209, was replaced with Lys. Although this mutation does not significantly change the high-affinity binding of MgATP to the holoenzyme, it does abolish high-affinity binding of cAMP to site A. In the absence of MgATP, binding of cAMP to site B is sufficient to promote dissociation of the holoenzyme complex and activation of the catalytic subunit [Bubis et al. (1988) J. Biol. Chem. 263, 9668]. In the presence of MgATP however, holoenzyme formed with this mutant regulatory subunit is very resistant to cAMP. The Kd(cAMP) was greater than 1 microM, and the Ka(cAMP) increased 60-fold from 130 nM to 6.5 microM in the presence of MgATP. Thus, MgATP serves as a lock that selectively stabilizes the holoenzyme and inhibits activation. Both site A and site B are shielded from cAMP in the presence of MgATP. These results suggest that Arg-209 may play a role in stabilizing the MgATP.holoenzyme complex in addition to its role in binding the exocyclic oxygens of cAMP when cAMP is bound to the regulatory subunit. The catalytic subunit also reassociates rapidly with this mutant regulatory subunit, and in contrast to the wild-type regulatory subunit, holoenzyme formation does not require MgATP.

Probing the cyclic nucleotide binding sites of cAMP-dependent protein kinases I and II with analogs of adenosine 3',5'-cyclic phosphorothioates.

A set of cAMP analogs were synthesized that combined exocyclic sulfur substitutions in the equatorial (Rp) or the axial (Sp) position of the cyclophosphate ring with modifications in the adenine base of cAMP. The potency of these compounds to inhibit the binding of [3H]cAMP to sites A and B from type I (rabbit skeletal muscle) and type II (bovine myocardium) cAMP-dependent protein kinase was determined quantitatively. On the average, the Sp isomers had a 5-fold lower affinity for site A and a 30-fold lower affinity for site B of isozyme I than their cyclophosphate homolog. The mean reduction in affinities for the equivalent sites of isozyme II were 20- and 4-fold, respectively. The Rp isomers showed a decrease in affinity of approximately 400-fold and 200-fold for site A and B, respectively, of isozyme I, against 200-fold and 45-fold for site A and B of isozyme II. The Sp substitutions therefore increased the relative preference for site A of isozyme I and site B of isozyme II. The Rp substitution, on the other hand, increased the relative preference for site B of both isozymes. These data show that the Rp and Sp substitutions are tolerated differently by the two intrachain sites of isozymes I and II. They also support the hypothesis that it is the axial, and not the previously proposed equatorial oxygen that contributes the negative charge for the ionic interaction with an invariant arginine in all four binding sites. In addition, they demonstrate that combined modifications in the adenine ring and the cyclic phosphate ring of cAMP can enhance the ability to discriminate between site A and B of one isozyme as well as to discriminate between isozyme I and II. Since Rp analogs of cAMP are known to inhibit activation of cAMP-dependent protein kinases, the findings of the present study have implications for the synthesis of analogs having a very high selectivity for isozyme I or II.