Monoclonal Antibodies as Probes for Functional Domains in CAMP-dependent Protein Kinase 11”

The antigenic regions of the type I1 regulatory subunit of CAMP-dependent kinase from bovine heart have been correlated with the previously established domain structure of the molecule. Immunoblotting with both serum and monoclonal antibodies of fragments generated by limited proteolysis or chemical cleavage of the R-subunit established that the major antigenic sites were confined to the amino-terminal portion of the polypeptide chain (residues 1-146). Radioimmu- noassays using two different antisera suggested that one or more of the high affinity serum antibody rec- ognition sites were further restricted to residues 91-145. This amino-terminal portion of the R-subunit in- cludes the hinge region which is particularly sensitive to proteolysis, allowing the R-subunit to be cleaved readily into a COOH-terminal domain which retains the CAMP-binding sites and an NH,-terminal fragment which appears to be the major site for interaction of the R-subunits in the native dimer. Monoclonal anti- bodies that recognized determinants on both sides of this hinge region were characterized and their specific recognition sites localized. in contrast was compared. did slightly between

plexed with the C-subunit, indicating that the arginine residues not only contribute to the specificity of the phosphorylation site but also are an essential component for energetically stabilizing the holoenzyme complex.
The regulatory subunit of CAMP-dependent protein kinase has a well-defined domain structure that has been characterized by a variety of approaches. The domain structure was first apparent from distinct proteolytic fragments that were generated during purification (1). Limited proteolysis was utilized subsequently to more precisely define that domain structure. To summarize, these studies demonstrated that the regulatory subunit, normally present as a dimer, has a socalled "hinge" region which is susceptible to proteolytic cleavage by a variety of proteases (2). Following such limited proteolytic cleavage, each regulatory subunit is cleaved into two fragments, a COOH-terminal domain (31,000-37,000) which retains both of the high affinity CAMP-binding sites of the native protein but which no longer dimerizes and an NHZterminal domain (10,500-19,000) which appears to retain the sites that are important for interaction between the two promoters of the native dimeric protein. More extended proteolysis suggested that the domain structure in the COOHterminal CAMP-binding region of the molecule could be defined even further, since proteolytic fragments of molecular weights of 12,000-14,000 could be generated which retained the ability to bind cAMP (3).
The domain structure in the COOH-terminal region of the molecule was reinforced when the homology between the R'subunit and the CAMP-binding protein from Escherichia coli became apparent. This protein has two domains which correspond to a COOH-terminal DNA-binding domain and an NH2-terminal CAMP-binding domain (4). Sequence homology strongly suggested that the in-tandem gene duplicated regions in the COOH-terminal portion of bovine heart R (5) corresponded to two CAMP-binding domains which are both homologous to the CAMP-binding domain of the CAMP-binding protein of E. coli. (6).
The antigenic properties of the R-subunit also have been characterized using a variety of methods. Although all Rsubunits have the general domain structure described above, the major forms of CAMP-dependent protein kinase in cells are antigenically distinctive, and these differences are associated exclusively with the R-subunits (7, 8). In the present studies we have correlated the antigenic structure of the type The abbreviations used are: R, the regulatory subunit of CAMPdependent protein kinase; C, the catalytic subunit of CAMP-dependent protein kinase; R", type I1 R-subunit; MES, 2-(N-morpho-1ino)ethanesulfonic acid; RIA, radioimmunoassay; dansyl, 5-dimethylaminonaphthalene-1-sulfonyl; HPLC, high performance liquid chromatography; SDS, sodium dodecyl sulfate.

4203
I1 R-subunit with the overall domain structure of the molecule using serum antibodies as a more general probe and monoclonal antibodies (9) as probes for very specific regions of the molecule. Finally, the monoclonal antibodies have been used to define specific functional properties that are associated with isolated domains of the R-subunit.
The type I regulatory subunit from porcine skeletal muscle was purified according to Zick and Taylor (12) and cGMP-dependent protein kinase, purified from bovine lung (13), was provided by G. N.
Gill (Department of Medicine; University of California, San Diego).
Preparation of Serum Antibodies-Polyclonal antibodies against R" from porcine cardiac tissue were prepared by initial injection of 50 pg of antigen in 0.1 ml of phosphate-buffered saline and 0.1 ml of Freund's complete adjuvant into the hind leg lymph nodes of rabbits (16). Animals were boosted at 1-2-month intervals with 100 pg of R" in Freund's adjuvant by intramuscular injection in hind legs and subcutaneous injections along the spinal column. Blood (20 ml) was collected from ear vein puncture 1 week after the first booster and collection was repeated every 2-3 weeks. Clots were removed, and serum was separated from blood cells by centrifugation for 10 min at 13,000 rpm in a Sorvall RC-5 centrifuge. Serum was stored with an equal volume of saturated ammonium sulfate (pH 7.4), 1 mM EDTA. Production of antibodies was monitored by Ouchterlony diffusion and RIA. Preparation of Hybrid Cell Lines and Ascitic Fluids-Hybrid cells producing monoclonal antibodies against R" from bovine cardiac muscle were isolated from fusions between NS-I myeloma cells and ously described by Mumby and Beavo (9).
lymphocytes from BALB/c mice immunized with antigen as previ-SDS-Polyacrylamide Gel Electrophoresis and Zmmunoblotting-Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was carried out according to Laemmli (14) utilizing 15% acrylamide in the lower gels. Proteins were either stained with 0.025% Coomassie Blue R-250, 25% isopropyl alcohol, 10% acetic acid or transferred immediately to 0.1-pm nitrocellulose sheets. Transfers were achieved in 25 mM Tris base, 192 mM glycine, 20% methanol (pH 8.3) at room temperature for 12 h at 0.1 A in a Hoeffer Transphor apparatus.
Nitrocellulose sheets were blocked in 25 ml of 2.5% bovine serum albumin in buffer A (10 mM Tris-HC1 (pH 7.4), 0.9% NaCl) for 2 h. All blotting procedures were performed at room temperature with constant mixing. Serum or ascitic fluic was added to fresh blocking buffer and incubated 90 min with the blotted sheets. Blots were washed four times with 100 ml of buffer A, then incubated with 50 pl of biotinylated second antibody in 25 ml of blocking buffer for 30 min. The washing steps were repeated and an avidin-biotinylated horseradish peroxidase mixture (50 p1 of avidin and 50 pl of biotinylated horseradish peroxidase in 25 ml of buffer A mixed 15 min before use) was incubated for 45 min with the nitrocellulose sheets. The blots were washed as before and then developed with 25 ml of 50 mM Tris-HC1, 200 mM NaCl, plus 5 ml of 0.30% 4-chloro-1-naphthol in methanol and 10 pl of 30% hydrogen peroxide.
Autoradiographs were recorded on Kodak X-Omat film and developed manually with Kodak GBX-developer/replenisher and rapid fixer.
Radioimmunoassay of R"-All radioimmunoassays were performed as described previously (17) except that bovine .heart R" was used as the competing label.
High Performance Liquid Chromatography-HPLC was carried out analytically and preparatively using an Altex 3200 system, monitoring absorbance at 219 nm. Enzymatic digests were separated on a Waters CI8-pBondapak column (0.39 X 30 cm), using 0.1% trifluoroacetic acid (solvent A) and acetonitrile (solvent B). Conditions for the separations are detailed in the figure legends.
Amino Acid Analyses and Sequencing-Amino acid analyses and dansyl Edman degradation were performed as described previously

(17).
Preparation of Type ZZ Holoenzyme-Type I1 regulatory (or CAMPbinding fragments) and catalytic subunits, purified to homogeneity from porcine heart, were combined in equal weights and dialyzed for 24 h at 4 "C uersus 10 mM potassium phosphate (pH 6.5), 1 mM EDTA, 10 mM 2-mercaptoethanol. To remove excess C and verify holoenzyme formation, the complex was bound to a column of DEAE-52 (2 ml) equilibrated with 40 mM potassium phosphate (pH 6.51, washed with the same buffer, and eluted with a 200-ml linear gradient of 40-250 mM potassium phosphate (pH 6.5), containing 1 mM EDTA and 10 mM 2-mercaptoethanol. Fractions exhibiting both phosphotransferase and CAMP-binding activity were pooled and concentrated.
Assay for the Catalytic Subunit-ATPphosphotransferase activity was measured employing a modification of the filter disc method of Wastila et al. (18) using histone IIA as a substrate.
Zmmunoadsorptwn of Holoenzyme-Ascitic fluid (0.5 ml) containing monoclonal antibody or normal rabbit serum (0.5 ml) was reacted with protein A-Sepharose (0.5 ml) in phosphate-buffered saline, 0.1% Nonidet-P-40, for 4 h at 4 "C. The solid was washed extensively with phosphate-buffered saline to remove unbound protein. The R".C complex (0.4 mg) was incubated with the immobilized antibody for 2 h at 4 'C in the presence or absence of 0.3 mM CAMP in 50 mM potassium phosphate, 1 mM EDTA (pH 6.5). Following the elution of nonbinding protein, the residue was washed extensively with buffer. CAMP (0.3 mM) was added to release any bound catalytic activity. Column fractions were monitored by absorbance at 280 nm, phosphotransferase activity, and SDS-polyacrylamide gel electrophoresis.

RESULTS
In order to correlate the antigenic properties of R" with its domain structure, fragments were prepared following proteolysis with a variety of proteases. The proteases used were chymoptrypsin, thermolysin, and S. aureus V8 protease, and the specific sites of cleavage for these proteases are indicated polypeptide chain.  in Fig. 1. These sites of cleavage were verified by sequencing the NH2-terminal residues of the larger isolated domains. Visualization of these proteolytic fragments and correlation with the site of autophosphorylation are shown in Fig. 2. In contrast to digestion with chymotrypsin, thermolysin, and trypsin,, proteolysis with the S. aureus V8 protease results in the autophosphorylation site being retained with the smaller NH2-terminal fragment. Fragments also were generated by chemical cleavage with cyanogen bromide, and the results likewise are summarized in Fig. 1. Immunoblotting following gel electrophoresis was used as a tool for qualitatively identifying immunoreactive proteins and peptide fragments whereas RIAs were utilized for a more Serum H 5

B10
FIG. 3. Immunoblotting of cyanogen bromide and limited proteolysis fragments of heart R". R" (0.5 pg), its CNBr peptides (0.5 pg), and the fragments derived from limited proteolysis (4 pg) were run on 15% polyacrylamide gels and transferred to nitrocellulose by Western blotting. Immunoreactive bands were visualized with polyclonal antibodies prepared against porcine heart R" and the H5 and B10 monoclonal antibodies to bovine heart R". It should be noted that the phosphorylated form of heart R" migrates more slowly than dephosphorylated R" (ll), and this observation was true for all of the phosphorylated fragmerits as well (data not shown). quantitative measure of cross-reactivity. The serum antibodies were generated against porcine heart R" and were specific for R" versus other functionally related proteins. Immunoblotting with the serum antibodies showed no cross-reactivity with porcine R' or C or with bovine cGMP-dependent protein kinase (data not shown). These polyclonal antibodies were used initially to ascertain whether the R-subunit contains select regions that are preferentially antigenic. Antigenicity was monitored by immunoblotting following polyacrylamide gel electrophoresis in the presence of SDS as described under "Experimental Procedures.'' When a total cyanogen bromide digest was characterized, the results (Fig. 3, lane 1) indicated that all of the cross-reactivity was associated with one major CNBr peptide having a molecular weight of 26,000 on SDS gels, which corresponded to CNBr I. The only other crossreactive band that was visualized was a minor band at 37,000 which corresponded to an incomplete digestion product con- taining CNBr I since the autophosphorylation site was retained in this fragment. Immunoblotting of R" that had first been subjected to limited proteolysis with either chymotrypsin or S. aureus V8 protease showed that both the NHz-terminal fragment and the COOH-terminal domain were immunoreactive, indicating that antigenic determinants were localized on both sides of the hinge region (Fig. 3, lanes 3 and 4 ) . A more quantitative assessment of the relative antigenicity of the COOH-terminal domain is seen with the competitive displacement RIA shown in Fig. 4. Under these conditions the majority of the immunoreactivity was associated with the domain generated by chymotrypsin. The fact that a large portion of the immunoreactivity was lost with the thermolytic fragment suggests that the charged residues associated with the hinge region may be dominant determinants. Similar results were observed with antisera from two different rabbits. Antigenic sites on R" were more carefully defined using monoclonal antibodies that were originally generated to bovine heart R". Immunoblotting of the CNBr peptides and the fragments generated by limited proteolysis gave the patterns shown in Fig. 3 for the two monoclonal antibodies, B10 and H5.' Both antibodies cross reacted only with CNBr 1. However, immunoblotting of the proteolytic fragments showed that the recognition sites for these antibodies were located on opposite sides of the hinge region. Unlike H5 which consistently cross-reacted with the COOH-terminal domain generated by both proteases, B10 was found to cross-react with the smaller NHz-terminal fragment. A third antibody, B6, was more difficult to evaluate in that although cross-reactivity was also associated with CNBrI, immunoreactivity was lost following digestion with trypsin, chymotrypsin, or thermolysin. Antigenicity was retained only when R" was proteolyzed with S. aureus V8 protease, and this cross-reactivity was associated with the NH2-terminal fragment (data not shown).
The antigenic site recognized by H5 was previously characterized (17) and localized to a 20-residue peptide (residues 102-121) located just beyond the site of autophosphorylation, serine 95. This is consistent with the immunoblotting pattern seen in Fig. 3. Like H5, the second monoclonal antibody, B10, also recognized fragments of R" in immunoblotting experi-B6, B10, and H5 were originally referred to as 3PI-B6, 3PI-B10, and 3DI-H5 by Mumby and Beavo (9). ments. Based on these results, further localization of the B10 antigenic site was pursued. Peptides generated from an extended V8 protease digestion of R" autophosphorylated with [y3'P]ATP were separated by HPLC. Immunoreactivity of the resultant fractions coincided with the 32P-labeled autophosphorylation site as seen in Fig. 5. The major antigenic peptide (A) was cleaved with trypsin which removed the autophosphorylation site (A2). The largest tryptic peptide (Al) retained the immunoreactivity and its amino acid composition (Table I) indicated that the peptide was located in region 46-89 of the polypeptide chain in the native protein.
Although no catalytic activity is associated with the regulatory subunit, it has a number of functional properties. Specifically, the R-subunit serves as a receptor for CAMP, is a substrate for autophosphorylation, and functions as a reversible inhibitor of the catalytic subunit. Previous studies (17) have suggested that interaction of R with monoclonal antibodies H5 and B10 was not altered by autophosphorylation and also did not significantly affect the affinity or maximal levels of cAMP binding. The other important role of the R-subunit is its function as an inhibitor of the C-subunit. Since a number of indications implicate the hinge region in R-C interaction and since the antigenic sites described above lie close to this hinge region, the interaction of the monoclonal antibodies with holoenzyme was characterized to see if aggregation with the C-subunit masked any antigenic sites.
Holoenzyme competed as effectively as R2(cAMP)4 in competitive displacement RIAs with 32P-R11 for most of the monoclonal antibodies studied. However, intact holoenzyme was not as efficient as the free subunit or holoenzyme incubated with cAMP for monoclonal antibody H5. In this case, the addition of cAMP to the reconstituted holoenzyme increased the relative affinity of R" in the complex to that of Rz(cAMP)4 at concentrations of holoenzyme above 5 nM (Fig. 6). To determine whether binding of the antibody promoted dissociation of the holoenzyme, the H5 monoclonal antibody was immobilized on protein A-Sepharose and subsequently mixed with holoenzyme that had been reconstituted as described under "Experimental Procedures." The gel was collected on a small fritted glass filter, washed with buffer, and eluted with buffer containing 0.3 mM CAMP. Fractions were monitored for absorbance and for phosphotransferase activity. As seen in Fig. 7, when a control gel containing immunoglobulins from a nonimmunized rabbit bound to protein A-Sepharose was used, no protein was bound to the gel; both R-and C-subunits were found in the initial eluant. In contrast, when the holoenzyme was bound to H5-protein A-Sepharose no protein or phosphotransferase activity appeared in the eluant until the column was washed with CAMP. Only at that time did the Csubunit appear in the eluate. If the holoenzyme was incubated with cAMP prior to mixing with H5-protein A-Sepharose, the phosphotransferase activity did not bind to the column. Similar results were obtained with B10 and B6 and serum antibodies. Immunoprecipitation from crude extracts confirmed that both R-and C-subunits were precipitated by the antibodies when cAMP was absent (data not shown).
Since the above experiment indicated that binding of antibody to the holoenzyme did not promote dissociation, H5protein A-Sepharose was utilized to determine whether or not any of the fragments of R" generated by limited proteolysis retained the ability to form a stable complex with the Csubunit. Two proteolytic fragments were used, the COOHterminal domains generated by chymotrypsin and thermolysin. The individual fragments were initially recombined with C-subunit under conditions normally sufficient to regenerate holoenzyme. The mixture was then combined with H5-protein    A-Sepharose, collected on a fritted glass filter, and eluted first with buffer and then with buffer containing CAMP. The results shown in Fig. 8 indicated that phosphotransferase activity associated with the C-subunit was retained on the protein A-Sepharose column only with the chymotryptic fragment, in which case the C-subunit could be selectively eluted with CAMP. When the thermolytic fragment was utilized, no phosphotransferase activity was retained, indicating that no stable complex was formed between this fragment and the Csubunit.

DISCUSSION
The domain structure of the R-subunit of CAMP-dependent protein kinase can be defined experimentally by limited pro- teolysis, and that domain structure is consistent with the amino acid sequence (5) and with the homologies of R with the CAMP-binding protein from E. coli (6). The essential features of that structure have been summarized in Fig. 9 which shows the hinge region that is susceptible to proteolytic cleavage and also indicates the site of autophosphorylation, the CAMP-binding domain that contains two CAMP-binding sites, and the R-R interaction site. The antigenic features of R" also are summarized in Fig. 9 and are based on the characterization of both serum and monoclonal antibodies with proteolytic fragments and CNBr peptides of R". Immunoblotting with serum antibodies indicated that only the NH2terminal third of the polypeptide chain was antigenic, since immunoblotting demonstrated that all of the sites of antigenic cross-reactivity were localized in the large cyanogen bromide fragment that includes residues 1-145. Radioimmunoassays using antisera from two rabbits suggested that most of the high affinity sites are further localized between residues 91 and 145. The COOH-terminal two-thirds of the molecule which contains the two in-tandem CAMP-binding domains is remarkably nonantigenic.
Monoclonal antibodies confirm the results obtained with serum antibodies in that those sites which have been localized to specific regions are NH2-terminal to the two CAMP-binding sites. The hinge region which joins NHz-terminal domain with the COOH-terminal CAMP-binding domain is included within this antigenic segment. Most sites of proteolytic cleavage in this hinge region fall on either side of the site of autophosphorylation at serine 95. Since the antigenic site for one of the monoclonal antibodies, B10, lies NHP-terminal to the hinge region and another, H5, is COOH-terminal to the hinge region, these two antibodies in particular promise to be useful tools for further investigating the domain structure of

ANTIGENIC REGION
these regulatory subunits, especially with regard to potential interaction sites with membranes and other proteins.
Although the antigenic sites recognized by these monoclonal antibodies are in close proximity in terms of the sequence to several functional sites, binding of antibody does not appear to markedly affect a number of functional properties. For example, antibody binding does not significantly alter CAMP binding, and antibody interaction with R" is not affected by autophosphorylation (17). Most of the sites thought to be involved with interaction between the Rand C-subunits in the holoenzyme are also localized in this NH2-terminal region of the molecule (2). The autophosphorylation site is clearly a recognition site for the C-subunit, and since autophosphorylation can occur in the absence of dissociation of the holoenzyme, it has been assumed that this site masks a portion of the catalytic site of the C-subunit in the inactive holoenzyme complex. The importance of this autophosphorylation site for R-C recognition was indicated by Weber and Hilz (19) who demonstrated that R" was still capable of being autophosphorylated following limited proteolysis with chymotrypsin. They showed, furthermore, that the chymotryptic digest, but not the tryptic digest, was still capable of inhibiting the catalytic subunit (20). Nelson and Taylor (21) showed that Cys 97 in R" is also protected against alkylation in the holoenzyme but not in RZ, further supporting the importance of this region for interaction between the R-and C-subunits. Finally, Takio et al. (5) have predicted that the two anionic sites flanking the autophosphorylation site at serine 95 are important components of R-C interaction since basic proteins such as histones will promote dissociation of the holoenzyme.
In light of the apparent importance of the hinge region for providing an interaction site for the C-subunit, it was somewhat surprising to find that all of the sites recognized by the monoclonal antibodies were just as accessible in the holoenzyme as in the dissociated regulatory subunit. The affinity of H5 for R did appear to be reduced in the holoenzyme in the absence of CAMP; however, the extent of this difference is difficult to evaluate experimentally since the Kd for antigen interaction with the antibody is very close to the Kd (lo-' M) for R-C interaction. These results suggest that perhaps only a small region of R is masked by the C-subunit. In any case, it is apparent that even though some of the antigenic sites may be somewhat constrained in the holoenzyme, the major conformational integrity of those sites is retained and accessible. The results also demonstrate that binding by antibody did not dissociate the R-and C-subunits.
It was possible to evaluate the role of specific residues because of the known amino acid sequence in the hinge region and because functional proteolytic fragments differing by only a few amino acid residues can be purified. Immunoabsorption was used for these studies by immobilizing H5 to protein A-Sepharose. Limited proteolysis with chymotrypsin and thermolysin generated two large fragments which were compared for their respective ability to bind the C-subunit. These COOH-terminal fragments both retain the CAMP-binding site and serine 95. Both are also monomeric, in contrast to the native R-subunit which is dimeric. The fragments differ by only three residues, the Asp-Arg-Arg that precedes the site of autophosphorylation being present in the chymotryptic frag-ment but missing in the thermolytic fragment. The results show that the chymotryptic fragment retains the ability to form a stable complex with the catalytic subunit and that dissociation of the complex is promoted by CAMP. The inability of the thermolytic fragment to form a stable complex with C indicates that the two arginine residues that precede the site of autophosphorylation not only provide a site that is recognized by the active site of the C-subunit, but that these residues are essential for interaction between the Rand Csubunits and make a major contribution to the energy of binding between the subunits. Since the chymotryptic CAMPbinding domain is missing the first acidic region and H5 binds to at least part of the second, the formation of a stable complex between the chymotryptic CAMP-binding domain and the Csubunit indicates that the presence of the two arginine residues is sufficient for R-C interaction.