Effects of CAMP-binding Site Mutations on Intradomain Cross-communication in the Regulatory Subunit of CAMP-dependent Protein Kinase I*

Each protomer of the regulatory subunit dimer of cAMP-dependent protein kinase contains two tandem and homologous cAMP-binding domains, A and B, and cooperative cAMP binding to these two sites promotes holoenzyme dissociation. Several amino acid residues in the type I regulatory subunit, predicted to lie in close proximity to each bound cyclic nucleotide based on affinity labeling and model building, were replaced using recombinant techniques. The mutations included replacement of 1) Glu-200, predicted to hydrogen bond to the 2'-OH of cAMP bound to site A, with Asp, 2) Tyr-371, the site of affinity labeling with 8-N3-cAMP in site B, with Trp, and 3) Phe-247, the position in site A that is homologous to Tyr-371 in site B, with Tyr. Each mutation caused an approximate 2-fold increase in both the Ka(cAMP) and Kd(cAMP); however, the off-rates for cAMP and the characteristic pattern of affinity labeling with 8-N3-cAMP differed markedly for each mutant protein. Furthermore, these mutations affect the cAMP binding properties not only of the site containing the mutation, but of the adjacent nonmutated site as well, thus confirming that extensive cross-communication occurs between the two cAMP-binding domains. Photoaffinity labeling of the native R-subunit results in the covalent modification of two residues, Trp-260 and Tyr-371, by 8-N3-cAMP bound to sites A and B, respectively, with a stoichiometry of 1 mol of 8-N3-cAMP incorporated per mol of R-monomer (Bubis, J., and Taylor, S. S. (1987) Biochemistry 26, 3478-3486). A stoichiometry of 1 mol of 8-N3-cAMP incorporated per R-monomer was observed for each mutant regulatory subunit as well, even when 2 mol of 8-N3-cAMP were bound per R-monomer; however, the major sites of covalent modification were altered as follows: R(Y371/W), Trp-371; R(E200/D), Tyr-371, and R(F247/Y), Tyr-371.

and &(cAMP); however, the offrates for CAMP and the characteristic pattern of affinity labeling with S-N3-CAMP differed markedly for each mutant protein.
Furthermore, these mutations affect the CAMP binding properties not only of the site containing the mutation, but of the adjacent nonmutated site as well, thus confirming that extensive crosscommunication occurs between the two CAMP-binding domains.
Photoaffinity labeling of the native R-subunit results in the covalent modification of two residues,  and Tyr-371, by 8-Ns-CAMP bound to sites A and B, respectively, with a stoichiometry of 1 mol of B-N3-CAMP incorporated per mol of R-monomer  Biochemistry 26, 3478-3486).
A stoichiometry of 1 mol of 8-N3-CAMP incorporated per R-monomer was observed for each mutant regulatory subunit as well, even when 2 mol of 8-N3-CAMP were bound per R-monomer; however, the major sites of covalent modification were altered as follows: R(Y371/W), Trp-371; R(EBOO/D), Tyr-371, and R(F247/Y), Tyr-371. CAMP-binding domains at the carboxyl terminus (1,2). These sites can be distinguished on the basis of preference for binding different analogs of CAMP as well as kinetically by measuring off-rates of bound CAMP (3,4). The first site in the linear sequence, site A, preferentially binds analogs of CAMP containing substitutions at the Ns position of the adenine ring, and CAMP also dissociates rapidly from site A.
In contrast, site B shows a preference for analogs having substitutions at the C-8 position of the adenine ring, and CAMP dissociates very slowly from site B in the absence of catalytic subunit (3,4). Binding of CAMP as well as activation of the catalytically inert holoenzyme complex by CAMP show positive cooperativity.
This positive cooperativity results primarily from interactions between the two tandem CAMPbinding domains within each protomer of the regulatory subunit (5, 6). The dimeric form of the regulatory subunit is not essential for this positive cooperativity (7), whereas deletion of CAMP-binding site B abolishes positive cooperativity (8). These two CAMP-binding domains found in all regulatory subunits share extensive sequence similarities not only with each other but also with the catabolite gene activator protein (CAP)' of Escherichia coli. On this basis, Weber et al. (9) constructed a model of each CAMP-binding domain by building the amino acid sequences of the CAMP-binding domains for the type I and type II regulatory subunits into the CAP crystal structure. This model of the CAMP-binding domains of the regulatory subunits is consistent with photoaffinity labeling using El-azido-cyclic adenosine 3':5'-monophosphate (8-N:<-CAMP) that identified specific amino acid residues in proximity to the C-8 position of the adenine ring of bound CAMP for each CAMP-binding site (10, 11). Site-specific mutations in the regulatory subunit also confirm the validity of this CAP-based model as a general framework for the folding of the polypeptide chain in each CAMP-binding domain (12,13).
Several of the amino acids that contribute directly to CAMP binding in CAP are highly conserved in all known regulatory subunit sequences. Each CAMP-binding domain contains, in particular, an invariant Arg and Glu. The conserved Arg in CAP is thought to ion pair with the exocyclic phosphate oxygens of bound CAMP while the conserved Glu is predicted to hydrogen bond with the 2'-OH of the ribose ring of bound CAMP (9). The corresponding residues in the type I regulatory subunit are   cAPKI: Point Mutations in the R-subunit CAMP-binding Sites bound CAMP (13). Whether or not Glu-200 is likewise essential has not been tested previously. Three new point mutations in the RI-subunit are described here. All three are localized in the CAMP-binding domains. Two of these mutations were designed as a result of photoaffinity labeling studies which revealed that Tyr-371 and Tyr-381 in the CAMP-binding site B of the R' and RI'-subunits, respectively, were close to the C-8 position of the adenine ring of bound 8-N3-CAMP (10, 11). One mutation replaced Tyr-371 in the B domain with a Trp. The other mutation replaced Phe-247 in site A with a Tyr. The position of Phe-247 in domain A of the RI-subunit is homologous to Tyr-371 in domain B. The third mutation replaced the conserved residue in site A, Glu-200, with an Asp in order to evaluate possible interactions between Glu-200 and the 2'-OH of bound CAMP. The effects of all three mutations on CAMP binding affinities, ATP binding properties, CAMP off-rates, and photoaffnity labeling were studied. All other reagents were analytical grade. Catalytic subunit was purified from porcine heart according to Nelson and Taylor (14). Mutagenesis-The bacterial expression vector for the type I regulatory subunit, pLST2, was described previously (15). The procedure used for mutagenesis was described by Durgerian and Taylor (16). The oligonucleotide probes used to introduce the point mutations are shown in Fig. 1

Construction of the Mutant
Plasmids-Three point mutations in the CAMP-binding domains of the RI-subunit were introduced using the synthetic oligonucleotides described in Fig. 1. The three mutant proteins produced were: 1) rR(Y371/ W), where the Tyr in the CAMP-binding domain B known to be in the vicinity of the C-8 position of the adenine ring of bound CAMP (10,ll) was converted to a Trp; 2) rR(F247/Y), where the residue in CAMP-binding domain A homologous to Tyr-371 (9) was converted to Tyr; and 3) rR(E200/D), where the Glu-200 residue in the A domain, thought to form a hydrogen bond with the 2'-OH of bound CAMP (9), was converted to Asp. The location of each of these sites of mutation with respect to the overall domain structure of the regulatory subunit is shown in Fig. 2 The mutant proteins expressed in these cells were then purified to homogeneity by affinity chromatography using CAMP-Sepharose. A typical yield was 15-20 mg per liter of culture.
General Structural Features-All the mutant rR-subunits retained many of the general properties of the native RIsubunit. The expressed proteins did not appear to be significantly more susceptible to proteolysis than the native protein, since no major degradation products were observed. A slight amount of proteolysis was observed on storage with both the mutant rR-subunits and the wild type rR-subunit. Sequence analysis revealed the proteolytic fragment to be a mixture of two fragments beginning at either General features of the domain structure include a dimer interaction site at the amino terminus, the "hinge" region essential for interaction with the C-subunit, and two tandem CAMP-binding sites, A and B, at the carboxyl terminus.
endogenous cleavage site lies in the hinge region known to be susceptible to proteolysis in the native protein (24-26).
Holoenzyme Formation-All three mutant rR-subunits reassociated with the catalytic subunit yielding an enzymatitally inactive holoenzyme complex in uitro. Based on staining after sodium dodecyl sulfate-polyacrylamide gel electrophoresis, the rR-subunit and the catalytic subunit were present in a molar ratio of approximately 1:l in each holoenzyme. Activation of each holoenzyme complex was achieved by the addition of CAMP.

CAMP
Binding-As seen in Fig. 3, each holoenzyme bound 2 mol of CAMP per regulatory subunit monomer. The apparent Kd for the wild type holoenzyme was 17 nM with a Hill coefficient of 1.3. The apparent Kd values for the rR(Y371/ W), rR(F247/Y), and rR(E200/D) holoenzyme complexes were each 34 nM, with Hill coefficients of 1.4. CAMP binding was measured only in the absence of MgATP.
CAMP Exchange Rates-The off-rate of bound [3H]cAMP was measured for the wild type and mutant rR-subunits under both low and high salt conditions, and the results are summarized in Table I and Fig. 4. For the wild type RI-subunit, the CAMP off-rates in low salt were 4.5 and 55 min for sites A and B, respectively. Under high salt conditions, the off-rate for CAMP from site A was increased to 1.8 min while the offrate for CAMP bound to site B was reduced to 140 min.
All three mutant proteins had faster CAMP exchange rates than the wild type protein.
The rR(F247/Y)-subunit most resembled the wild type rR-subunit.
The low salt off-rates were 1.5 and 12.6 min, respectively, for site A, while the high salt off-rates were 1.3 and 45 min, respectively.
The CAMP off-rates for the rR(YS'Il/W)-subunit were much faster. The off-rates in low salt were 1 and 4.2 min for sites A and B, respectively. The high salt off-rates were 23 s and 2.2 min for sites A and B, respectively.
Lastly, the rR(EBOO/D)-subunit  formed with all of the mutant rR-subunits retained the high affinity MgATP-binding site observed in the wild type holoenzyme. As seen in Fig.  5, the wild type and mutant holoenzyme complexes each bound 2 mol of MgATP per holoenzyme complex, had identical Kd values of 15-20 nM, and had identical Hill coefficients of 1.6. Thus, although these three mutations exhibited altered CAMP binding properties to varying degrees, each showed no alterations in the MgATP-binding site in contrast to two deletion mutants where slight differences were observed (8).
The effect of MgATP on the apparent Kd for CAMP and on the induced off-rate for CAMP in the presence of the catalytic subunit has not been determined yet for these mutants.
Photolabeling of the Mutant Proteins-The stoichiometry of 8-N3-[3H]~AMP incorporation revealed that the rR-subunit, rR(Y371/W), rR(Y247/Y), and rR(E200/D) all bound 2 mol of label per mol of subunit but covalently incorporated only 1 mol of label per mol of subunit (Table II) In order to analyze photoaffinity labeling of the mutant subunits more precisely, 2 mg of each protein were photolyzed in solution with 8-N,-[3H]cAMP, digested with TPCK-trypsin, and the resulting tryptic peptides resolved by HPLC as indicated in Fig. 6. Major fractions containing radioactivity were pooled and rechromatographed using the same column but a different gradient. Each of the radioactive peptides from this second gradient were lyophilized, resuspended in 0.1% trifluoroacetic acid, and sequenced. The radioactivity remained bound to the filter during sequencing, and hence it was not possible to directly correlate the radioactivity with the step missing in the sequence. Instead, at the position that is photolabeled, no residue was identified in contrast to a clearly resolved PTH-derivative at the corresponding position in the unmodified peptide. The missing residue at these steps can thus be identified unambiguously as the site of covalent modification.
The mutation in rR(Y371/W) eliminates most of the photolabeling of the A domain. In this case, the tryptophan in the B domain that replaced Tyr-371, normally labeled in the wild type RI-subunit, was the primary site of photolabeling.
The two conservative point mutations in rR(F247/Y) and rR(E200/D) also were sufficient to eliminate most photolabeling of the A domain. In both of these mutants, Tyr-371 in domain B was the dominant site of photolabeling. The reason why this peptide containing Tyr-371 elutes in various positions is not clear but may be due to different addition products or to rearrangement products once a covalent adduct was formed. In each case the sequence was identical and the site of labeling was the Tyr; hence, the variability does not appear to be due to incomplete digestion with trypsin. Since all of the minor peaks were not sequenced, we cannot rule out unambiguously that no labeling of Trp-260 occurs; however, in comparison with labeling of the wild type R'subunit (ll), labeling of Trp-260, if it occurs at all, is very low.

DISCUSSION
Three different point mutations, summarized in Fig. 2, were introduced into the RI-subunit of CAMP-dependent protein  for the CAMP-binding site mutant holoenzymes. High performance liquid chromatography was as described under "Experimental Procedures." A, radioactive fractions from the phosphate gradient separation for the photolabeled rR(E200/D) holoenzyme tryptic digest. The labeled fractions corresponding to Tyr-371 are indicated. The radioactive fractions for the rR(Y371/W) holoenzyme tryptic digest and the corresponding Trp-371 photolabeled residue are indicated in B. The radioactive fractions and the absorbance at 219 nm for the tryptic digest of the photolabeled rR(F247/Y) holoenzyme are shown in C and D, respectively. kinase in order to further probe the two tandem CAMPbinding sites. The K, (cAMP) and Kd(cAMP) were increased approximately 2-fold for holoenzyme formed with each of the mutant R-subunits.
In contrast, the off-rates were increased by as much as 60-fold and also varied significantly for each mutant. As a general rule, measurement of the CAMP offrates was more sensitive for detecting differences between the mutant and wild type proteins than the measurement of K,JcAMP) or KJcAMP). Each mutation not only perturbed the CAMP-binding properties of the site that was altered but also affected the CAMP-binding properties of the adjacent site. Furthermore, this cross-communication can occur in both directions.
The potential importance of Glu-200 is based on several independent criteria. The general importance of hydrogen bonding of the 2'-OH of CAMP to the protein was indicated initially by mapping the CAMP-binding sites with analogs of CAMP. In this case, replacing the 2'-OH with hydrogen  or methylating the 2'-OH were sufficient to increase the K, by 170-and 500-fold, respectively (27). The crystal structure of CAP identified Glu-70 as the most likely acceptor group (28). Glu-70 in CAP corresponds to  in site A of the RI-subunit (9). Finally, the invariance of this Glu in all of the R-subunits is consistent with its playing an important functional role (9, 28). The role of this Glu in the cyclic nucleotide-free structure is not known since the crystal structure has not been solved for CAP in the absence of CAMP. Since Glu and Asp differ by only a single methylene, this conservative replacement should maintain the overall charge environment while potentially disrupting hydrogen bond interactions.
Replacement of Glu-200 with Asp results in only a 2-fold increase in the KJcAMP).
Thus, Glu-200 cannot be considered as essential in spite of its conservation throughout evolution. This is in contrast to the other invariant residue, Arg-209, where the guanidinium group, not just a positive charge, does appear to be necessary for high affinity binding of CAMP (13). Although rR(E200D) still supports high affinity binding of CAMP, the off-rate for CAMP from site A is enhanced nearly l&fold.
This rapid off-rate may reflect weakened hydrogen bonding. This site A mutation also increased the observed CAMP exchange rate for site B by llfold, demonstrating significant intrasubunit communication between sites A and B. As indicated in Fig. 7, Glu-200 is thought to lie in a loop between p strands 6 .and 7. The potential flexibility of this loon might minimize the consequences of replacing Glu with Asp thus enabling the mutant R-subunit to retain high affinity binding for CAMP. The other two mutations probe the region surrounding the adenine ring of CAMP. Tyr-371 in site B, located near the C-8 position of bound 8N3- ["HIcAMP (11,12), lies on the inward facing surface of the long C-helix according to the CAP-based model (9,28). This mutation reduced the affinity for CAMP by 2-fold relative to the wild type rR-subunit, and increased the off-rate for CAMP from site B I3-fold. The offrate for CAMP from site A also was enhanced, once again emphasizing that these two sites communicate closely. However, unlike an earlier mutation replacing Tyr-371 with Phe (12), replacement with Trp had little effect on the cooperative interaction between the CAMP-binding sites. One feature that Tyr and Trp share, in contrast to Phe, is the capacity to provide r-donor interactions.
Based on kinetic arguments, CAMP binds first to site B which causes a conformational change that makes site A, otherwise shielded, more accessible to CAMP (30,31). A highly favorable stacking interaction between the aromatic side chain of residue 371 and the adenine ring of CAMP, coupled with movement of the long Chelix, may be required for inducing the conformational changes that lead to dissociation of the holoenzyme. Though Phe can provide hydrophobic interactions, Tyr and particu- The corresponding residues in site A are  and Arg-209. The region surrounding the bound CAMP is shown below. The model is based on the crystal structure of CAP (9).
larly Trp, are energetically favored for stacking. Another feature that may be important for residue 371 is the intrinsic dipole moment of the side chain. While Phe, Tyr, and Trp all will contribute to a hydrophobic environment, their dipole moments differ significantly.
Phe, in particular, has a very small dipole moment in comparison to Tyr and Trp. The potential importance of this dipole was suggested earlier based on mapping with CAMP analogs (32). Clearly additional mutations need to be introduced at this site in order to understand more clearly the interactions that occur between the adenine ring and the protein.
A second mutation replacing Phe-247 in site A with Tyr also was constructed based on the role of Tyr-371 and the long C-helix in site B. The model proposed by Weber et al. (9) suggests that the C-helix lining the CAMP-binding pocket is shorter in site A than in site B, and the position corresponding to  in site A is Phe-247. Based on sequence similarities, the two C-helices in sites A and B are not as highly conserved as other regions of the CAMP-binding domains ( Figs. 7 and 8). However, both RI-and R"-subunits have a conserved Phe at positions 247 and 251, respectively, in site A and a conserved tyrosine, 371 (discussed above), at the same position in site B (3,4). If one considers this to be an amphipathic helix (Fig. 8), the residues lining the surface that faces CAMP are hydrophobic in both sites and are highly conserved in R' and R" while those on the outer surface, presumably in contact with another domain or with solvent, tend to be hydrophilic and variable.

Replacing
Phe-247 with Tyr is a relatively conservative change that would not be expected to greatly disrupt the protein structure. Indeed only slight differences were observed in the binding of CAMP and in the off-rates for CAMP, and CAMP binding still showed cooperativity.
While the  fold increase in the off-rates for CAMP was seen for site A.
This site A mutation also increased the CAMP off-rate from site B consistent with previous studies showing that CAMP bound to site A alters the CAMP off-rate of site B for the wild type RI-subunit (5, 6, 33). Phe-247 in the A domain is located in the same position of the C-helix as Tyr-371 in the B domain (9). The possibility existed that the photoaffinity labeling of Trp-260 in site A of the wild type RI-subunit was due to the relatively nonnucleophilic nature of Phe-247. Therefore, changing this residue to a potentially more reactive tyrosine might result in labeling at position 247. In the two deletion mutants indicated in Table II, site B, as well as Trp-260, are missing. In this case Tyr-244 gets photolabeled stoichiometrically so clearly several aromatic rings are in relatively close proximity to the adenine ring. Tyr-247 did not get photolabeled to any significant extent; however, neither did Trp-260, the normal site of modification in the wild type protein (ll), or Tyr-244, the site of modification in the site B deletion mutant (8). Labeling of rR(F247Y) occurred predominantly at Tyr-371, in contrast to the native RI-subunit where labeling occurs at both Trp-260 (40%) and Tyr-371 (60%) (11). This conservative point mutation apparently altered somewhat the location of Trp-260 with respect to the C-8 position of the adenine ring of 8-Nz-CAMP but did not compete effectively for photoincorporation.
In general, photoaffinity labeling proved to be a sensitive method for detecting changes, perhaps subtle ones, in the two CAMP-binding sites. When all of the photoaffinity labeling results are considered (Table II), including labeling of the native type I and II R-subunits (10, ll), a proteolytic fragment of the R"-subunit (34), and various mutant forms of the RIsubunit (12,13,22), several general observations are apparent. First, and perhaps most striking, is that the stoichiometry for covalent modification typically does not exceed 1 mol of 8-NEj-cAMP incorporated per mol of R-monomer even though one or both sites can be labeled. This is true even though in most cases 8-N3-CAMP binds with a high affinity to both CAMP-binding sites. Furthermore, even after one site is covalently modified, the remaining site is still functional and can still bind 8-Ns-cAMP (ll), there is just no stable photoincorporation.
The second conclusion is that slight perturbations of CAMP-binding site A or B, either by proteolysis or mutagenesis, can alter the overall pattern of photoaffinity labeling even though the total stoichiometry remains relatively unchanged. It is as though covalent modification at one site precludes covalent modification of the adjacent site. Whichever site competes most efficiently will be the dominant site for covalent modification.
In the case of the mutant proteins, the mutation could either change the geometry of the CAMP-binding site or increase the off-rate sufficiently so that the relative opportunity for covalent labeling is reduced. We see here examples where a mutation at one site can improve photolabeling at the mutated site as in the case of R(Y371W).
On the other hand, both R(E200D) and R(F247Y) show reduced labeling of the mutated site in favor of the nonmutated site.
The results described here confirm earlier kinetic results (33, 35) indicating that considerable communication occurs between the two tandem CAMP-binding sites in each Rmonomer. A mutation at one site consistently perturbs the CAMP binding properties of the adjacent site. Measurement of CAMP off-rates and of sites of photoaffinity labeling are two very sensitive methods for detecting changes in each CAMP-binding site. Clearly further structural studies are required before the detailed molecular events that are associated with the cooperative communication between the two CAMPbinding sites can be elucidated. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. N&l.