Mutating protein kinase cAMP-binding sites into cGMP-binding sites. Mechanism of cGMP selectivity.

The cAMP-dependent protein kinase contains two different cAMP-binding sites referred to as the slow and fast sites. Mutation of Ala-334 to a threonine in the slow site of the bovine type I regulatory subunit created a site with marked increase in cGMP affinity without changing cAMP affinity (Shabb, J. B., Ng. L., Corbin, J. D. (1990) J. Biol. Chem. 265, 16031-16034). The corresponding fast site residue (Ala-210) was changed to a threonine by oligonucleotide-directed mutagenesis, and a double mutant containing a threonine in each site was also made. Holoenzymes were formed from native catalytic subunit and each recombinant regulatory subunit. The fast site mutant holoenzyme exhibited an improved cGMP activation constant and an impaired cAMP activation constant. The double mutant cGMP/cAMP selectivity was 200-fold greater than that of wild-type holoenzyme, making it as responsive to cGMP as native cGMP-dependent protein kinase. The increased intrinsic binding energies of mutated sites for cGMP were 2.7-3.0 kcal mol-1, consistent with the presence of an extra hydrogen bond. Cyclic nucleotide analog studies implied that this hydrogen bond was between the threonine hydroxyl and the 2-amino of cGMP. Comparisons of amino acid sequences and cyclic nucleotide specificities suggested that the Ala/Thr difference may also impart cAMP/cGMP binding selectivity to related proteins such as cyclic nucleotide-gated ion channels.     

One amino acid change produces a high affinity cGMP-binding site in cAMP-dependent protein kinase

Discrimination between cAMP and cGMP is a critical feature of cAMP- and cGMP-dependent protein kinases. An alanine/threonine difference in the cyclic nucleotide-binding sites has been proposed to provide a structural basis for this functional distinction. Site-directed mutagenesis of this alanine to a threonine in a cAMP-binding site of cAMP kinase produced a mutant with markedly increased cGMP affinity as determined by cGMP binding and protein kinase activation assays. Studies of other mutants at this position support the role of the threonine hydroxyl group as the component that enhances cGMP binding affinity.     

Studies of cGMP analog specificity and function of the two intrasubunit binding sites of cGMP-dependent protein kinase

The specificity of the two intrasubunit cGMP binding sites of cGMP-dependent protein kinase was determined by measuring the ability of 46 cGMP analogs to compete with [3H]cGMP. Both sites of the enzyme exhibited high specificity for the ribose cyclic phosphate moiety, and lower specificity for the guanine moiety. Effects of modifications in the ribose cyclic phosphate moiety suggested that cGMP is bound at both sites by three hydrogen bonds at 2'-OH, 3'-O, and 5'-O. A negative charge in the cyclic phosphate is apparently required. Modifications of the pyrimidine part of guanine, particularly at C-1, generally caused selectivity for the rapidly exchanging site while modifications of the imidazole part of guanine at C-7 and C-8 caused selectivity for the slowly exchanging site. These increases in selectivity for a site were mainly due to losses in affinity of the other site. There was an apparent requirement of the intact amino group at C-2, particularly for the slowly exchanging site. Comparison of the molecular interactions of cAMP and cGMP with their specific protein kinases showed that both nucleotides are bound by similar forces in the 2', 3' and 5' region, both bases may be bound in syn conformation, but that each base moiety is bound by different molecular interaction, thus leading to the selectivity of the two enzymes. cGMP analogs which possessed strong selectivity for the rapidly exchanging site, but not those selective for the slowly exchanging site, stimulated the binding of [3H]cGMP. Only a few cGMP analogs were more potent than cGMP in stimulating protein kinase activity. The potency of cGMP analogs as stimulators of kinase activity correlated better with the mean binding affinity for both binding sites than with the affinity for either site alone. Two analogs added in combination were synergistic in kinase activation, particularly if one analog was selective for the slowly exchanging site and the other for the rapidly exchanging site. These observations are suggestive that cGMP binding at the rapidly exchanging site stimulates cGMP binding at the slowly exchanging site and that both sites are involved in the activation process.     

The amino-terminal cyclic nucleotide binding site of the type II cGMP-dependent protein kinase is essential for full cyclic nucleotide-dependent activation

For the type I cGMP-dependent protein kinases (cGKIalpha and cGKIbeta), a high affinity interaction exists between the C2 amino group of cGMP and the hydroxyl side chain of a threonine conserved in most cGMP binding sites. To examine the effect of this interaction on ligand binding and kinase activation in the type II isozyme of cGMP-dependent protein kinase (cGKII), alanine was substituted for the conserved threonine or serine. cGKII was found to require the C2 amino group of cGMP and its cognate serine or threonine hydroxyl for efficient cGMP activation. Of the two binding sites, disruption of cGMP-specific binding in the NH(2)-terminal binding site had the greatest effect on cGMP-dependent kinase activation, like cGKI. However, ligand dissociation studies showed that the location of the rapid and slow dissociation sites of cGKII was reversed relative to cGKI. Another set of mutations that prevented cyclic nucleotide binding demonstrated the necessity of the NH(2)-terminal, rapid dissociation binding site for cyclic nucleotide-dependent activation of cGKII. These findings suggest distinct mechanisms of activation for cGKII and cGKI isoforms. Because cGKII mediates the effects of heat-stable enterotoxins via the cystic fibrosis transmembrane regulator Cl(-) channel, these findings define a structural target for drug design.     

Studies of the molecular mechanism of discrimination between cGMP and cAMP in the allosteric sites of the cGMP-binding cGMP-specific phosphodiesterase (PDE5).

The regulatory domain of the cGMP-binding cGMP-specific 3':5'-cyclic nucleotide phosphodiesterase (PDE5) contains two homologous segments of amino acid sequence that encode allosteric cyclic nucleotide-binding sites, referred to as site a and site b, which are highly selective for cGMP over cAMP. The possibility that the state of protonation in these sites contributes to cyclic nucleotide selectivity was investigated. The binding of cGMP or cAMP was determined using saturation and competition kinetics at pH values between 5.2 and 9.5. The total cGMP binding by PDE5 was unchanged by variation in pH, but the relative affinity for cGMP versus cAMP progressively decreased as the pH was lowered. Using site-directed mutagenesis, a conserved residue, Asp-289, in site a of PDE5 has been identified as being important for cyclic nucleotide discrimination in this site. It is proposed that deprotonation of Asp-289 enhances the number and strength of bonds formed with cGMP, while concomitantly decreasing the interactions with cAMP.     

Mutagenesis of the cyclic AMP receptor protein of Escherichia coli: targeting positions 83, 127 and 128 of the cyclic nucleotide binding pocket.

The cyclic 3', 5' adenosine monophosphate (cAMP) binding pocket of the cAMP receptor protein (CRP) of Escherichia coli was mutagenized to substitute cysteine or glycine for serine 83; cysteine, glycine, isoleucine, or serine for threonine 127; and threonine or alanine for serine 128. Cells that expressed the binding pocket residue-substituted forms of CRP were characterized by measurements of beta-galactosidase activity. Purified wild-type and mutant CRP preparations were characterized by measurement of cAMP binding activity and by their capacity to support lacP activation in vitro. CRP structure was assessed by measurement of sensitivity to protease and DTNB-mediated subunit crosslinking. The results of this study show that cAMP interactions with serine 83, threonine 127 and serine 128 contribute to CRP activation and have little effect on cAMP binding. Amino acid substitutions that introduce hydrophobic amino acid side chain constituents at either position 127 or 128 decrease CRP discrimination of cAMP and cGMP. Finally, cAMP-induced CRP structural change(s) that occur in or near the CRP hinge region result from cAMP interaction with threonine 127; substitution of threonine 127 by cysteine, glycine, isoleucine, or serine produced forms of CRP that contained, independently of cAMP binding, structural changes similar to those of the wild-type CRP:cAMP complex.     

Interaction among the twelve-residue segment of antifreeze protein type I,or its mutants,water and a hexagonal ice crystal

We have carried out a molecular dynamics analysis on a mixture of supercooled water, a hexagonal ice crystal and segments of winter flounder antifreeze protein. The segment consists of nine alanine residues, two threonine residues and one asparagine residue. Mutant segments, in which the threonine residues are replaced with valine residues, or serine residues, are also used. It is found that the threonine residue near the asparagine residue of the original segment is located in the vicinity of the prism face of the ice crystal. This is due to the hydrogen bond between the hydrophilic sites of these residues and water molecules, and the hydrogen bond between these water molecules and the water molecules on the ice surface. The valine and serine residues in the mutant segments do not approach the prism face of the ice crystal compared with the threonine residue near the asparagine residue. The motion of five segments, closely located side by side, is not remarkable. This is because of the gathering of water molecules caused by hydrophobic hydration, not only around alanine residues but also around the methyl sites of threonine residues.     

Interaction between a twelve-residue segment of antifreeze protein type I,or its mutants,and water molecules

A molecular dynamics simulation has been carried out for water molecules with a rigid segment of antifreeze protein type I. The segment consists of nine alanine residues, two threonine residues and one asparagine residue. Mutant segments, in which the threonine residues are replaced with valine residues, or serine residues, are also used. It is predicted that the hydrogen site of asparagine residue, and that of threonine residue, play an important role in the hydrogen bond of water molecules in these sites. This hydrogen bond is not noticeable between water molecules and the valine residue, or serine residue. The existence of four hydrophilic sites enhances the mobility of water molecules close to the serine residue of the mutant segment. The difference in the zenith-angle fluctuations of the original segment and the valine-mutant segment is less noticeable in the case of 230 K. This is because the gathering of water molecules due to the hydrophobic hydration is predominant near the alanine residues of the segments at this temperature.     

Specific cGMP binding by the cGMP binding domains of cGMP-binding cGMP specific phosphodiesterase.

The structure of cyclic GMP (cGMP)-binding (cGB), cGMP specific phosphodiesterase (PDE5) comprises several domains. We have used RT-PCR methods to clone the noncatalytic cGB domains of PDE5 from human colon cancer cell RNA and constructed glutathione-S-transferase (GST) fusion proteins to express and study the domains. One fragment showed 94% identity to bovine PDE5 and coded for the high affinity cGB domain of PDE5 (Val(156)-Asp(394), cGB-I). Another cloned fragment showed 92% identity to bovine PDE5 and coded for the phosphorylation site plus both high and low affinity cGB domains of PDE5 (Val(36)-Glu(529), cGB-II). Both fragments expressed as GST-cGB fusion proteins bound cGMP specifically, as determined by competitive [3H]-cGMP ligand binding. We found that cGB-I showed high affinity cGMP binding with K(d)=0.33 microM. cGB-II showed two cGMP binding sites with similar affinities and specificity to the native enzyme. cGB-II was phosphorylated by cGMP-dependent protein kinase (PKG) as reported for bovine PDE5. These data show that recombinant regulatory regions of PDE5 form cGB sites similar to native enzyme sites and confirm proposed domain functions. These results establish that recombinant fusion proteins of PDE5 domains may be used to further characterize the structure of PDE5.     

Probing the interactions between cAMP and cGMP in cyclic nucleotide-gated channels using covalently tethered ligands

Cyclic nucleotide-gated channels contain four ligand-binding subunits, and they are directly activated by the binding of cGMP or cAMP. Channels with different combinations of subunits are known to have different sensitivities to the two nucleotides. However, the consequences of mixed occupancy by cGMP and cAMP are not well understood, and may have important implications for understanding the functions of these channels in different cell types. We studied the activation of homomeric and heteromeric retinal rod cyclic nucleotide-gated channels with the four ligand-binding sites occupied by different combinations of cGMP (a strong agonist) and cAMP (a weak agonist). Control of occupancy was obtained by covalently tethering different numbers of cGMP moieties using the photoaffinity analogue 8-p-azidophenacylthio-cGMP; the remaining sites were then saturated with cAMP, or cGMP, for comparison. The fractional current activated by cAMP increased dramatically as the number of tethered cGMP moieties increased. In homomeric channels comprised of the alpha subunit, cAMP became an effective agonist only after three of the four sites were occupied by tethered cGMP moieties. In contrast, in heteromeric channels comprised of two alpha and two beta subunits, cAMP caused significant activation after two sites were occupied by tethered cGMP moieties. In agreement with earlier work, a single residue on the beta subunit, N1201, accounted for much of the increased efficacy of cAMP on heteromeric channels. The results are consistent with significant interactions between subunits, including the two types of subunits in heteromeric channels.     

Cross-activation: overriding cAMP/cGMP selectivities of protein kinases in tissues.

cAMP- and cGMP-dependent protein kinases are homologous proteins and are predicted to exhibit very similar three-dimensional structures. Their cyclic nucleotide binding domains share a high degree of amino acid sequence identity. cAMP- and cGMP-dependent protein kinases are activated relatively specifically by cAMP and cGMP, respectively; and a single alanine-threonine difference between cAMP- and cGMP-binding domains partially accounts for this specificity. Thus, it would be expected that cAMP and cGMP mediate separate physiological effects. However, owing in part to the lack of absolute specificity of either enzyme and to the relatively high level of cAMP or cGMP in certain tissues, it is also possible that either cyclic nucleotide could cross-activate the other kinase. Increases in either cAMP or cGMP cause pig coronary artery relaxation. However, only cGMP-dependent protein kinase specific cyclic nucleotide analogues are very effective in causing relaxation, and cAMP elevation in arteries treated with isoproterenol or forskolin activates cGMP-dependent protein kinase, in addition to cAMP-dependent protein kinase. Conversely, increases in either cAMP or cGMP cause Cl- secretion in T-84 colon carcinoma cells, and the cGMP level in T-84 cells can be elevated sufficiently by bacterial enterotoxin to activate cAMP-dependent protein kinase. These results imply specific regulation of cAMP- and cGMP-dependent protein kinases by the respective cyclic nucleotides, but either cyclic nucleotide is able to cross-activate the other kinase in certain tissues.     

Solution structure of the cGMP binding GAF domain from phosphodiesterase 5: insights into nucleotide specificity, dimerization, and cGMP-dependent conformational change

Phosphodiesterase 5 (PDE5) controls intracellular levels of cGMP through its regulation of cGMP hydrolysis. Hydrolytic activity of the C-terminal catalytic domain is increased by cGMP binding to the N-terminal GAF A domain. We present the NMR solution structure of the cGMP-bound PDE5A GAF A domain. The cGMP orientation in the buried binding pocket was defined through 37 intermolecular nuclear Overhauser effects. Comparison with GAF domains from PDE2A and adenylyl cyclase cyaB2 reveals a conserved overall domain fold of a six-stranded beta-sheet and four alpha-helices that form a well defined cGMP binding pocket. However, the nucleotide coordination is distinct with a series of altered binding contacts. The structure suggests that nucleotide binding specificity is provided by Asp-196, which is positioned to form two hydrogen bonds to the guanine ring of cGMP. An alanine mutation of Asp-196 disrupts cGMP binding and increases cAMP affinity in constructs containing only GAF A causing an altered cAMP-bound structural conformation. NMR studies on the tandem GAF domains reveal a flexible GAF A domain in the absence of cGMP, and indicate a large conformational change upon ligand binding. Furthermore, we identify a region of approximately 20 residues directly N-terminal of GAF A as critical for tight dimerization of the tandem GAF domains. The features of the PDE5 regulatory domain revealed here provide an initial structural basis for future investigations of the regulatory mechanism of PDE5 and the design of GAF-specific regulators of PDE5 function.     

An equilibrium study of the cooperative binding of adenosine cyclic 3',5'-monophosphate and guanosine cyclic 3',5'-monophosphate to the adenosine cyclic 3',5'-monophosphate receptor protein from Escherichia coli

The binding of adenosine cyclic 3',5'-monophosphate (cAMP) and guanosine cyclic 3',5'-monophosphate (cGMP) to the adenosine cyclic 3',5'-monophosphate receptor protein (CRP) from Escherichia coli was investigated by equilibrium dialysis at pH 8.0 and 20 degrees C at different ionic strengths (0.05--0.60 M). Both cAMP and cGMP bind to CRP with a negative cooperativity that is progressively changed to positive as the ionic strength is increased. The binding data were analyzed with an interactive model for two identical sites and site/site interactions with the interaction free energy--RT ln alpha, and the intrinsic binding constant K and cooperativity parameter alpha were computed. Double-label experiments showed that cGMP is strictly competitive with cAMP, and its binding parameters K and alpha are not very different from that for cAMP. Since two binding sites exist for each of the cyclic nucleotides in dimeric CRP and no change in the quaternary structure of the protein is observed on binding the ligands, it is proposed that the cooperativity originates in ligand/ligand interactions. When bound to double-stranded deoxyribonucleic acid (dsDNA), CRP binds cAMP more efficiently, and the cooperativity is positive even in conditions of low ionic strength where it is negative for the free protein. By contrast, cGMP binding properties remained unperturbed in dsDNA-bound CRP. Neither the intrinsic binding constant K nor the cooperativity parameter alpha was found to be very sensitive to changes of pH between 6.0 and 8.0 at 0.2 M ionic strength and 20 degrees C. For these conditions, the intrinsic free energy and entropy of binding of cAMP are delta H degree = -1.7 kcal . mol-1 and delta S degree = 15.6 eu, respectively.     

Allosteric activation of cGMP-specific,cGMP-binding phosphodiesterase (PDE5) by cGMP

The effects of cGMP binding on the catalytic activity of cGMP-specific, cGMP-binding phosphodiesterase (PDE5) are unclear because cGMP interacts with both allosteric and catalytic sites specifically. We studied the effects of cGMP on the hydrolysis of a fluorescent substrate analogue, 2'-O-anthraniloyl cGMP, by PDE5 partially purified from rat cerebella. The preparation contained PDE5 as the major cGMP-PDE activity and was not contaminated with cAMP- or cGMP-dependent protein kinases. The Hill coefficients for hydrolysis of the analogue substrate were around 1.0 in the presence of cGMP at concentrations <0.3 microM, while they increased to 1.5 at cGMP concentrations >1 microM, suggesting allosteric activation by cGMP at concentrations close to the bulk binding constant of the enzyme. Consistent with an allosteric activation, increasing concentrations of cGMP enhanced the hydrolysis rate of fixed concentrations of 2'-O-anthraniloyl cGMP, which overcame competition between the two substrates. Such activation was not observed with cAMP, cyclic inosine 3',5'-monophosphate, or 2'-O-monobutyl cGMP, indicating specificity of cGMP. These results demonstrate that cGMP is a specific and allosteric activator of PDE5, and suggest that in cells containing PDE5, such as cerebellar Purkinje cells, intracellular cGMP concentrations may be regulated autonomously through effects of cGMP on PDE5.     

Communications between the high-affinity cyclic nucleotide binding sites in E. coli cyclic AMP receptor protein: effect of single site mutations

The binding of adenosine 3',5'-cyclic monophosphate (cAMP) and its nonfunctional analogue, guanosine 3',5'-cyclic monophosphate (cGMP), to the adenosine 3',5'-cyclic monophosphate receptor protein (CRP) from Escherichia coli was investigated by means of fluorescence and isothermal titration calorimetry (ITC) at pH 7.8 and 25 degrees C. A biphasic fluorescence titration curve was observed, confirming the previous observation reported by this laboratory (Heyduk and Lee (1989) Biochemistry 28, 6914-6924). However, the triphasic titration curve obtained from the ITC study suggests that the cAMP binding to CRP is more complicated than the previous conclusion that CRP binds sequentially two molecules of cAMP with negative cooperativity. The binding data can best be represented by a model for two identical interactive high-affinity sites and one low-affinity binding site. Unlike cAMP, the binding of cGMP to CRP exhibits no cooperativity between the high-affinity sites. The effects of mutations on the bindings of cAMP and cGMP to CRP were also investigated. The eight CRP mutants studied were K52N, D53H, S62F, T127L, G141Q, L148R, H159L, and K52N/H159L. These sites are located neither in the ligand binding site nor at the subunit interface. The binding was monitored by fluorescence. Although these mutations are at a variety of locations in CRP, the basic mechanism of CRP binding to cyclic nucleotides has not been affected. Two cyclic nucleotide molecules bind to the high-affinity sites of CRP. The cooperativity of cAMP binding is affected by mutation. It ranges from negative to positive cooperativity. The binding of cGMP shows none. With the exception of the T127L mutant, the free energy change for DNA-CRP complex formation increases linearly with increasing free energy change associated with the cooperativity of cAMP binding. This linear relationship implies that the protein molecule modulates the signal in the binding of cAMP, even though the mutation is not directly involved in cAMP or DNA binding. In addition, the significant differences in the amplitude of fluorescent signal indicate that the mutations also affect the surface characteristics of CRP. These results imply that these mutations are not perturbing specific pathways of signal transmission. Instead, these results are more consistent with the concept that CRP exists as an ensemble of native states, the distribution of which can be altered by these mutations.     

Pathway and endpoint free energy calculations for cyclic nucleotide binding to HCN channels

cAMP and cGMP differentially bind to and regulate a variety of proteins, including cyclic nucleotide-gated (CNG) channels and hyperpolarization-activated cyclic nucleotide-regulated (HCN) channels. Previous site-directed mutagenesis studies have isolated two conserved residues that are critical for enabling certain channels to selectively bind cGMP relative to cAMP. However, no definitive mechanism has been identified that explains the preferential activation of other channels by cAMP. Here we apply computational binding free energy methods, including thermodynamic integration, linear interaction energy, and continuum electrostatic calculations, to gain insights into the mechanisms of cyclic nucleotide selectivity. Consistent with experimental observations, computational results for the cAMP-selective HCN channels show that the binding free energy of cAMP is lower (more favorable) than that of cGMP. Surprisingly, cAMP selectivity is not due to its preferential contacts with protein, but rather reflects the greater hydration energy of cGMP relative to cAMP, resulting in a greater energetic cost for cGMP binding.     

Analysis of the cGMP/cAMP interactome using a chemical proteomics approach in mammalian heart tissue validates sphingosine kinase type 1-interacting protein as a genuine and highly abundant AKAP

The cyclic nucleotide monophosphates cAMP and cGMP play an essential role in many signaling pathways. To analyze which proteins do interact with these second messenger molecules, we developed a chemical proteomics approach using cAMP and cGMP immobilized onto agarose beads, via flexible linkers in the 2- and 8-position of the nucleotide. Optimization of the affinity pull-down procedures in lysates of HEK293 cells revealed that a large variety of proteins could be pulled down specifically. Identification of these proteins by mass spectrometry showed that many of these proteins were indeed genuine cAMP or cGMP binding proteins. However, additionally many of the pulled-down proteins were more abundant AMP/ADP/ATP, GMP/GDP/GTP, or general DNA/RNA binding proteins. Therefore, a sequential elution protocol was developed, eluting proteins from the beads using solutions containing ADP, GDP, cGMP, and/or cAMP, respectively. Using this protocol, we were able to sequentially and selectively elute ADP, GDP, and DNA binding proteins. The fraction left on the beads was further enriched, for cAMP/cGMP binding proteins. Transferring this protocol to the analysis of the cGMP/cAMP "interactome" in rat heart ventricular tissue enabled the specific pull-down of known cAMP/cGMP binding proteins such as cAMP and cGMP dependent protein kinases PKA and PKG, several phosphodiesterases and 6 AKAPs, that interact with PKA. Among the latter class of proteins was the highly abundant sphingosine kinase type1-interating protein (SKIP), recently proposed to be a potential AKAP. Further bioinformatics analysis endorses that SKIP is indeed a genuine PKA interacting protein, which is highly abundant in heart ventricular tissue.     

Cyclic nucleotide specificity of the activator and catalytic sites of a cGMP-stimulated cGMP phosphodiesterase from Dictyostelium discoideum

The cellular slime mold Dictyostelium discoideum has an intracellular phosphodiesterase which specifically hydrolyzes cGMP. The enzyme is activated by low cGMP concentrations, and is involved in the reduction of chemoattractant-mediated elevations of cGMP levels. The interaction of 20 cGMP derivatives with the activator site and with the catalytic site of the enzyme has been investigated. Binding of cGMP to the activator site is strongly reduced (more than 80-fold) if cGMP is no longer able to form a hydrogen bond at N2H2 or O2'H. Modifications at N7, C8, O3' and O5' induce only a small reduction of binding affinity. A cyclic phosphate structure, as well as a negatively charged oxygen atom at phosphorus, are essential to obtain activation of the enzyme. Substitution of the axial exocyclic oxygen atom by sulphur is tolerated; modification of the equatorial oxygen atom reduces the binding activity of cGMP to the activator site by 90-fold. Binding of cGMP to the catalytic site is strongly reduced if cGMP is modified at N1H, C6O, C8 and O3', while modifications at N2H2, N3, N7, O2'H, and O5' have minor effects. Both exocyclic oxygen atoms are important to obtain binding of cGMP to the catalytic site. The results indicate that activation of the enzyme by cGMP and hydrolysis of cGMP occur at different sites of the enzyme. cGMP is recognized at these sites by different types of molecular interaction between cGMP and the protein. cGMP derivatives at concentrations which saturate the activator site do not induce the same degree of activation of the enzyme (activation 2.3-6.6-fold). The binding affinities of the analogues for the activator site and their maximal activation are not correlated. Our results suggest that the enzyme is activated because cGMP bound to the activator site stabilizes a state of the enzyme which has a higher affinity for cGMP at the catalytic site.