Spectroscopic studies of the cAMP binding sites of the regulatory subunits of types I and II protein kinase

Endogenous tryptophan residues of cAPK regulatory subunit type IIbeta reveal local variations in environments and dynamics

The amino terminal dimerization/docking domain and the two-tandem, carboxy-terminal cAMP-binding domains (A and B) of cAMP-dependent protein kinase regulatory (R) subunits are connected by a variable linker region. In addition to providing a docking site for the catalytic subunit, the linker region is a major source of sequence diversity between the R-subunit isoforms. The RIIbeta isoform uniquely contains two endogenous tryptophan residues, one at position 58 in the linker region and the other at position 243 in cAMP-binding domain A, which can act as intrinsic reporter groups of their dynamics and microenvironment. Two single-point mutations, W58F and W243F, allowed the local environment of each Trp to be probed using steady-state and time-resolved fluorescence techniques. We report that: (a) the tryptophan fluorescence of the wild-type protein largely reflects Trp243 emission; (2) cAMP selectively quenches Trp243 and thus acts as a cAMP sensor; (3) Trp58 resides in a highly solvated, unstructured, and mobile region of the protein; and (4) Trp243 resides in a stable, folded domain and is relatively buried and rigid within the domain. The use of endogenous Trp residues presents a non-perturbing method for studying R-subunit subdomain characteristics in addition to providing the first biophysical data on the RIIbeta linker region.

Dynamic interaction of cAMP with the Rap guanine-nucleotide exchange factor Epac1

Epac1 is a Rap-specific guanine-nucleotide exchange factor (GEF) which is activated by the binding of cAMP to a cyclic nucleotide monophosphate (cNMP)-binding domain. We investigated the equilibrium and dynamics of the interaction of cAMP and Epac1 using a newly designed fluorescence analogue of cAMP, 8-MABA-cAMP. We observed that the interaction of cAMP, measured by competition with 8-MABA-cAMP, with an isolated cNMP binding domain of Epac1 has an overall equilibrium constant (Kd) of 4 microM and that the kinetics of the interaction are highly dynamic. The binding properties of cAMP are apparently not affected when the catalytic domain is present, despite the fact that binding of cAMP results in activation of Epac1. This indicates that for the activation process, no appreciable binding energy is required. However, when bound to Rap1b, the apparent Kd of Epac to cAMP was about fivefold lower, suggesting that substrate interaction stabilizes cAMP binding. Since the fluorescent analogues used here were either less able or unable to induce activation of Epac1, we concluded that the binding of nucleotide to Epac and the activation of GEF activity are uncoupled processes and that thus appropriate cAMP analogues can be used as inhibitors of the Epac1-mediated signal transduction pathway of Rap.

Probing the multidomain structure of the type I regulatory subunit of cAMP-dependent protein kinase using mutational analysis: role and environment of endogenous tryptophans

The regulatory R-subunit of cAMP-dependent protein kinase (cAPK) is a thermostable multidomain protein. It contains a dimerization domain at the N-terminus followed by an inhibitor site that binds the catalytic C-subunit and two tandem cAMP-binding domains (A and B). Two of the three tryptophans in the RIalpha subunit, Trp188 and Trp222, lie in cAMP-binding domain A while Trp260 lies at the junction between domains A and B. The unfolding of wild-type RIalpha (wt-RI), monitored by intrinsic fluorescence, was described previously [Leon, D. A., Dostmann, W. R. G., and Taylor, S. S. (1991) Biochemistry 30, 3035 (1)]. To determine the environment of each tryptophan and the role of the adjacent domain in folding and stabilization of domain A, three point mutations, W188Y, W222Y, and W260Y, were introduced. The secondary structure of wt-RI and the point mutants has been studied by far-UV circular dichroism spectropolarimetry (CD). The CD spectra of wt-RI and the three point mutants are practically identical, and the thermal unfolding behavior is very similar. Intrinsic fluorescence and iodide quenching in the presence of increasing urea established that: (a) Trp222 is the most buried, whereas Trp188 is the most exposed to solvent; (b) Trp260 accounts for the quenching of fluorescence when cAMP is bound; and (c) Trp222 contributes most to the intrinsic fluorescence of the wt-RI-subunit, while Trp188 contributes least. For wt-RI, rR(W188Y), and rR(W260Y), removal of cAMP causes a destabilization, while excess cAMP stabilizes these three proteins. In contrast, rR(W222Y) was not stabilized by excess cAMP.

Binding of 9-anthroylcholine monitors the interactions of adenosine cyclic 3',5'-phosphate-dependent protein kinase with MgATP, substrates, and regulatory subunits

The isolated catalytic subunit of cAMP-dependent protein kinase and smooth muscle myosin light chain kinase undergo interactions with the fluorescent dye 9-anthroylcholine (9AC) that are responsive to the two enzymes' associations with substrates and effectors. Additionally, the binding of 9AC is highly sensitive to subtle structural or functional differences among closely related protein kinases. Skeletal muscle myosin light chain kinase and the catalytically active chymotryptic fragment of the gamma-subunit of phosphorylase kinase do not associate with 9AC. The 1:1 fluorescent complex of the isolated catalytic subunit of cAMP-dependent protein kinase with 9AC exhibits a dissociation constant of 21 microM. The association of the catalytic subunit with either of the regulatory subunits, RI and RII, results in decreases in the observed 9AC fluorescence that are reversed upon the addition of cAMP. The effects of MgATP and of polypeptide substrates (Kemptide, troponin I, protamine) on the 9AC-catalytic subunit complex are consistent with a general noncompetitive model in which the interactions of 9AC and the other ligands with the enzyme are mutually antagonistic but not purely competitive.

Binding of two fluorescent cAMP analogues to type I and II regulatory subunits of cAMP-dependent protein kinases

Binding of two long wavelength fluorescent cAMP analogues, 8-thioacetamido-fluorescein-cAMP (SAF-cAMP) and 8-thioacetamido-rhodamine-cAMP (SAR-cAMP), to the RI (from bovine muscle) and RII (from bovine heart) regulatory subunits of cAMP dependent kinases has been studied. Displacement of [3H]cAMP from RI and RII and equilibrium dialysis measurements show that the fluorescent nucleotides are high affinity ligands for the cAMP binding sites. The binding is characterized by complex fluorescence spectral and fluorescence anisotropy changes, more evident for the fluorescein than for the rhodamine derivative. The fluorescence excitation spectrum of the bound SAF-cAMP is characterized by the appearance of a red shifted shoulder at 500-510 nm excitation wavelength region. Any change of the bound/free ratio in a solution equilibrium is accompanied by changes in fluorescence and anisotropy signals which are best detected at suitable wavelengths. It is proposed that fluorescence and anisotropy changes can distinguish between binding to type B (slow dissociating) and A (fast dissociating) cAMP binding sites of regulatory subunits. Applications of the fluorescent nucleotides to kinase localization and cAMP determination in living cells are discussed.

Comparison of the two classes of binding sites (A and B) of type I and type II cyclic-AMP-dependent protein kinases by using cyclic nucleotide analogs

cAMP analogs, all 96 of which were modified in the adenine moiety, were examined quantitatively for their ability to inhibit the binding of [3H]cAMP to each of the two classes (A and B) of cAMP-binding sites of type I (rabbit skeletal muscle) and type II (bovine heart) cAMP-dependent protein kinase. The study showed that analogs can be constructed that have a higher affinity than cAMP for a binding site. N6-phenyl-cAMP had 18-fold increased affinity for site A of RI (AI) and 40-fold increased affinity for site AII. 2-chloro-8-methylamino-cAMP had a 7-fold increased affinity for BI, and 8-(4-chlorophenylthio)-cAMP had 17-fold increased affinity for BII. Analogs could discriminate between the two classes of binding sites by more than two orders of magnitude in binding affinity: 2-chloro-8-methylamino-cAMP had 170-fold higher affinity for BI than for AI, and 2-n-butyl-8-thiobenzyl-cAMP had 700-fold higher affinity for BII than for AII. Analogs could also discriminate between the homologous binding sites of the isozymes: 2-n-butyl-8-bromo-cAMP had 260-fold higher affinity for AI than for AII (22-fold higher for BII than BI), and 8-piperidino-cAMP had 50-fold higher affinity for BII than for BI (and 50-fold higher for AI than for AII). The data suggest the following conclusions. (a) Stacking interactions are important for the binding of cAMP to all the binding sites. (b) Subtle differences exist between the sites as to the optimal electron distribution in the adenine ring since modifications that withdraw electrons at C2 and donate at C8 favour binding to BI, and disfavour binding to AI and AII. (c) There are no hydrogen bonds between the adenine ring of cAMP and any of the binding sites. (d) All sites bind cAMP in the syn conformation. (e) The subsites adjacent to the N6 and C8 positions may have nonpolar neighbouring regions since hydrophobic substituents at N6 could increase the affinity for AI and AII and similar substituents at C8 could increase the affinity for BII. Finally, (f) the sites differed in their ability to accomodate bulky substituents at C2 and C8. For all compounds tested, their potency as activators of protein kinases I and II was found to correlate, in a predictable fashion, to their mean affinity for the two classes of binding sites, rather than to the affinity for only one of the sites.

A comparison of the cyclic nucleotide-dependent protein kinases using chemical cleavage at tryptophan and cysteine

cGMP-dependent protein kinase (G-kinase) and the regulatory subunit of type I (RI) cAMP-dependent protein kinase (A-kinase) both contain a phosphorylation site located near the NH2 terminus of each enzyme. These sites can be utilized as convenient markers for the determination of the position of an amino acid residue susceptible to either chemical or enzymatic digestion. Using the tryptophan-specific reagent, N-chlorosuccinimide, the approximate location along the polypeptide chain of six reactive tryptophans in G-kinase and three reactive residues in RI were identified. Similarly, cleavage with cyanide was used to locate free and disulfide-bonded cysteines in both proteins. The approximate positions of nine cysteines in G-kinase were determined along with the location of the interchain disulfide bond and an intrachain disulfide bond. RI was found to contain three cyanide-reactive cysteines, two of which are involved in interchain disulfide bonding. A comparison of the positions of the cysteines and tryptophans determined by chemical cleavage in G-kinase and RI, with the positions of cysteine and tryptophan in the known sequence of the type II A-kinase, support the structural relationships between these enzymes. Comparison with subsequently reported primary sequences of all three enzymes indicates the limits of precision of this chemical cleavage procedure.

Interaction of cAMP derivatives with the 'stable' cAMP-binding site in the cAMP-dependent protein kinase type I

cAMP binding to the 'stable' cAMP-binding sites in the regulatory subunit of the cAMP-dependent protein kinase type I was investigated using a set of 18 selected derivatives. All the tested analogues were competitive with [3H]cAMP and inhibitor constants from 12 nM to 20 microM with the free regulatory subunit were determined. The cAMP molecule seemed to be bound by these specific hydrogen bonds to the 5' and 3' oxygen, the 2' hydroxyl, and an ion pair interaction between the negative charge in equatorial position and a positively charged amino acid side chain. The adenine base is rather unspecifically bound with no hydrogen bonds involved. This binding specificity of the 'stable' site is similar to the requirement for dissociation as determined by the activation of the kinase by a respective analogue. This indicates that occupation of the 'stable' sites leads to activation of the protein kinase. The presence of the catalytic subunit reduced the affinity of most analogues. The binding of one derivative with the negative charge fixed in the axial position is not influenced by the addition of the catalytic subunit and ATP. A plausible model for a conformational change during the activation process in the 'stable' site is discussed.

Phosphorylation of cAMP-dependent protein kinase subunits

The cAMP-dependent protein kinases comprise two enzyme forms designated as type I and type II. The type II enzyme can catalyze an autophosphorylation reaction whereby phosphate is transferred from ATP to one seryl residue on each regulatory subunit monomer. Since this reaction can occur in the absence of cAMP-induced enzyme dissociation, it has been used as a probe to identify one site of interaction between the catalytic subunit (C) and the type II regulatory subunit (R11). The type I cAMP-dependent protein kinase does not catalyze an analogous reaction; however, if cGMP-dependent protein kinase is substituted for C, the type I regulatory subunit (R1) becomes phosphorylated. The effects of cyclic nucleotides on this reaction, coupled with the ability of R1 to serve as an inhibitor of cGMP-dependent protein kinase suggest that this phosphorylation also occurs within an important functional domain on R1. A comparison of the autophosphorylation site on R11 with the cGMP-dependent protein kinase catalyzed phosphorylation site on R1 indicates that each modification takes place within a similar proteolytically sensitive region. On each subunit, this sensitive "hinge" region lies distal to the functional domain responsible for regulatory subunit dimerization and proximal to that responsible for cAMP binding. Phosphorylation of the "hinge" region decreases the affinity of each regulatory subunit for C, although the magnitude of this change appears greater for R1 than for R11. Phosphorylation of R1 also reduces the stoichiometry of cAMP binding from two to one mole of cAMP bound per mole of R1 monomer. These results suggest that the "hinge" regions of both R1 and R11 form part of the interaction site between the regulatory subunit and C; and, in the case of R1, it also forms a portion of one of two cAMP-binding sites. The amino acid sequence surrounding the phosphorylated serine of each regulatory subunit has been determined: R11: D-R-R-V-S(P)-V R1: R-R-R-R-G-A-I-S(P)-A It is thought that the number and position of the basic amino acid residues proximal to the modified serine may be responsible, in part, for determining the susceptibility of each site to phosphorylation by cAMP or cGMP-dependent protein kinase. Both R1 and R11 exist as phosphoproteins in vivo. Dephosphorylation of purified "native" phospho-R1 is without effect on the ability of R1 to interact with either C or cAMP. The site phosphorylated in vivo is therefore distinct from that modified in vitro by cGMP-dependent protein kinase. In addition to the autophosphorylation site, R11 possesses a second, less enzymatically reactive, phosphorylation site that is modified in vivo. Dephosphorylation of this site is also without apparent effect on the functional properties of R11. The kinases responsible for catalyzing the phosphorylation of R1 and the cryptic site on R11 and the role that these modifications play in modulating kinase activity are currently unknown but are under active investigation.

Preparation of a fluorescent-labeled derivative of calmodulin which retains its affinity for calmodulin binding proteins

Calmodulin was derivatized with 5-[[[(iodoacetyl)amino]ethyl]amino]-1-naphthalenesulfonic acid to fluorescently label the protein. This derivative (AEDANS.CaM) stimulated the Ca2+-sensitive cyclic nucleotide phosphodiesterase and formed Ca2+-dependent complexes with troponin I and the phosphodiesterase. Association between AEDANS.CaM and these proteins was directly monitored by changes in fluorescence anisotropy. The dissociation constants for the AEDANS.CaM--troponin I and AEDANS.CaM--phosphodiesterase complexes were 60 nM and 4 nM, respectively. This fluorescent derivative of calmodulin appears suitable for direct monitoring of the complexes between calmodulin and calmodulin binding proteins. Rotational diffusion of AEDANS.CaM was also measured with fluorescence anisotropy. These measurements indicated that the shape of calmodulin in solution is best approximated by a prolate ellipsoid.

Calcium-induced exposure of a hydrophobic surface on calmodulin

Interactions between calmodulin (CaM) and several hydrophobic fluorescent probes were characterized in order to determine if CaM expresses hydrophobic binding sites in the presence of Ca2+. Several classes of fluorescent probes capable of sensing exposure of hydrophobic binding sites on proteins were found to bind to CaM, and these interactions were greatly enhanced by Ca2+. In the presence of Ca2+, the fluorescence intensity of 9-anthroylcholine (9AC) was increased 24-fold by CaM, with a shift in the fluorescence emission maximum from 514 to 486 nm. The fluorescence intensity of 8-anilino-1-naphthalenesulfonate (Ans) was enhanced 27-fold with an emission maximum shift from 540 to 488 nm in the presence of CaM and Ca2+. Similar results were obtained with the uncharged fluorescent ligand, N-phyenyl-1-naphthylamine. With all three fluorescent dyes, the fluorescence changes caused by CaM in the absence of Ca2+ were minor compared to those observed with CaM and Ca2+. Direct binding studies using equilibrium dialysis demonstrated that CaM can bind four to six molecules of 9AC or two to three molecules of Ans in a calcium-dependent manner. The effects of various amphiphilic compounds on the Ca2+-dependent complex formation between CaM and the Ca2+-sensitive phosphodiesterase or troponin I were investigated. Trifluoperazine (TFP) and 9AC inhibited CaM stimulation of the Ca2+-sensitive phosphodiesterase. The Ca2+-dependent binding of the phosphodiesterase to CaM-Sepharose was also inhibited by TFP, 9AC, and Ans. Furthermore, binding of CaM to troponin I-Sepharose was inhibited by these ligands. Consistent with these data was the observation that troponin I antagonized binding of 9AC to CaM. These data indicate that binding of Ca2+ to CaM results in exposure of a domain with considerable hydrophobic character, and binding of hydrophobic ligands to this domain antagonizes CaM-protein interactions. It is proposed that this hydrophobic domain may serve as the interface for the Ca2+-dependent binding of CaM to the phosphodiesterase or troponin I.