Structural Similarities between the Ca2+-dependent Regulatory Proteins of 3’:5’-Cyclic Nucleotide Phosphodiesterase and Actomyosin ATPase*

Results of studies of the CaZ+ -dependent protein modulator of 3’:5’-cyclic nucleotide phosphodiesterase isolated from bovine brain are presented which show its structural similarity to the Ca’+-binding subunit of muscle troponin. Both proteins have blocked NH, termini, similar and characteristic ultraviolet absorption spectra, similar Ca’+- binding properties, very similar amino acid compositions, and co-migrate on sodium dodecyl sulfate-polyacrylamide gels. The primary structures of selected tryptic peptides isolated from bovine brain modulator protein are similar or identical with regions of the primary sequences of rabbit skeletal muscle and bovine cardiac muscle troponin C. Bovine brain modulator protein contains an unidentified ninhydrin-positive basic compound not found in muscle troponin C. An improved procedure is presented which yields 40 to 70 mg of modulator protein per kg of bovine brain. The divalent cation, CaZ+, is involved in the regulation of numerous physiological and biochemical processes in


Results
of studies C. An improved procedure is presented which yields 40 to 70 mg of modulator protein per kg of bovine brain.
The divalent cation, CaZ+, is involved in the regulation of numerous physiological and biochemical processes in animal tissues including such diverse processes as muscle contraction (I), microtubule assembly (2,3), stimulus-secretion coupling (4), and the regulation of several enzyme activities.
Because of its apparent role as a generalized regulatory signal, Rasmussen (5)  The most conclusive of these studies to date has involved the activation of cyclic nucleotide phosphodiesterase (EC 3.4.1.17) by Ca". Kakiuchi et al. (6) and Cheung (7,8) independently demonstrated the presence of a factor in brain homogenates which, in the presence of Ca*', stimulated the activity of a brain cyclic nucleotide phosphodiesterase. This factor was subsequently shown to be a heat-stable phosphodiesterase activator protein that was present in high concentrations in numerous vertebrate neurosecretory tissues (9). This protein will be referred to as the modulator protein in this paper. ' Modulator protein activity has also been found in high concentrations in homogenates of a number of invertebrates has been shown to be the functional species (11,12) which, when bound to the phosphodiesterase, increases V,,. and lowers the K, for cyclic nucleotides (8).
Recently, Lin et al. (12) have purified modulator protein to apparent homogeneity from bovine brain and shown it to be an acidic calcium-binding protein (p1 4.3) which is monomeric in solution and possesses a molecular weight equivalent to 15,000 to 18,000 daltons. Teo and Wang (11) obtained similar results from studies of the modulator protein isolated from bovine heart.
Stevens and co-workers (13. 14) noted that this bovine heart protein is very similar in physicochemical properties to the calcium-binding subunit of rabbit skeletal muscle troponin, which provides Ca2+ sensitivity to actomyosin ATPase activity.
We have presented a preliminary report (15) of the isolation and characterization of TN-C-like2 proteins from terase and adenyl cyclase preparations. We have previously referred to this protein as a troponin C-like protein because of its structural similarity to skeletal muscle troponin C. The modulator protein will also function as a calcium-dependent stimulator of the ATPase activity of reconstituted skeletal muscle actomyosin (G. Amphlett, S. V. Perry, T. C. Vanaman, manuscript in preparation All other chemicals were reagent grade and utilized without further purification. Bovine brain S-106, prepared as described by Dannies and Levine (18), had an amino acid composition identical with that previously published (18)  Modulator protein was prepared using the procedure described below. All operations were performed at 4'.
Step I: Homogenization-Brains, obtained fresh from the slaughterhouse or from exsanguinated animals, were washed in physiological NaCl solution, stripped of membranes and collagenous material, and the spinal cord removed at the base of the brain. They were cut into sections and stored frozen at 20" in 500-g batches until used.
Frozen brains obtained from Pel-Freeze Biological, Inc. (Rogers, Ark.) were thawed, washed in physiological saline, and cut into sections after removal of spinal cord and cerebellum.
One kilogram of tissue, thawed at 4", was homogenized in 2-volumes of 0.1 M sodium acetate, 0.001 M 2-mercaptoethanol, and 0.001 M EDTA, pH 7.2 (Buffer A), at low speed in a Waring Blendor for 90 s. The homogenate was centrifuged at 10,000 rpm in a GSA rotor for 1 hour. The supernatant fluid was decanted and saved. The pellet was homogenized in an equal volume of buffer, centrifuged again, and the resulting supernatant fluid added to the original supernatant fraction. The pellet was discarded.

C-50 Batch
Filtration-The combined supernatant fractions were mixed with 600 ml of CM-Sephadex C-50 (packed volume) equilibrated with Buffer A. This mixture was allowed to stand for 30 min; then the resin was removed by filtration on a sintered-glass funnel. The packed resin bed was washed with 2 bed volumes of Buffer A and the washings added to the eluate.
Step 3: First DEAE-Sephadex Column Chromatography-The en- Fractions were pooled and treated exactly as described in Step 4.
Step 6: Gel I&&ion-The lyophilized protein from Step 5 was dissolved in 20 ml of 0.05 M Tris-HCl. 0.001 M EDTA. and 0.001 M 2-mercaptoethanol, pH 7.5, and applied to a column (5.4 x 133 cm) of Sephadex G-106 (fine) in the same buffer. The column was developed at 50 ml/hour with lo-ml fractions collected and monitored as described above. The appropriate fractions were pooled, desalted, and lyophilized.

Preparations
of Actiuator-Deficient Enzyme Partially purified phosphodiesterase was prepared from bovine brain using a modification of the procedure of Cheung (8). Bovine brain was homogenized and the homogenate treated exactly as described for Step 1 of the modulator protein purification scheme described above. The resultant supernatant was adjusted to 30% saturation by the addition of solid ammonium sulfate (176 g/liter). After standing for 1 hour at 4", the suspension was centrifuged at 10,000 x g for 30 min. The supernatant fluid was decanted and adjusted to 55% saturation by the addition of solid ammonium sulfate (162 g/liter Phosphodiesterase activity and the ability of modulator protein to stimulate the activity of bovine brain cyclic nucleotide phosphodiesterase (prepared as described in the previous section) were assaved using a modification of the procedure of Teo and Wang (11) then applied to the gels. Electrophoresis was performed at 150 volts for 5 hours. Gels were stained with 0.25% (w/v) Coomassie brilliant blue R250 in 50% (v/v) methanol/10% (v/v) acetic acid for 1 hour at 37", and destained for 2 hours in 7.5% (v/v) acetic acid and 5% (v/v) methanol at 37".
Gel Electrofocusing-Isoelectric focusing gels were made with 7.5% and 0.12% (w/v) ammonium persulfate. The upper reservoir (cathode vessel) buffer was 1% (v/v) ethylenediamine, and lower reservoir (anode vessel) buffer was 1.4% (v/v) orthophosphoric acid. Samples dissolved in 6 M urea/l% (v/v) 2-mercaptoethanol were layered on gels and focused at 200 volts until the amperage dropped to zero (4 to 5 hours). Gels were stained and destained as described by Awdeh (21) A partial specific volume of 0.72 cc/g was calculated for bovine brain modulator protein from its amino acid composition by the procedure of Cohn and Edsall (22).

Co'+-Binding
Sites-Equilibrium dialysis was performed at various concentrations of 'YZa*+ in 0.1 M imidazole-HCl, pH 7.0. All reagents were prepared and stored in polycarbonate vessels. Plastic laboratory ware and pipettes were used for all determinations. One-milliliter samples containing ligand and buffer with or without protein were placed on opposite sides of a dialysis membrane (acetylated 18/32 Visking cellulose casing prepared as previously described (23)) separating the chambers of a microdialysis cell. Dialysis cells were sealed and gently agitated for 8 hours in a New Brunswick shaker at 25'. Samples were withdrawn after incubation and the concentration of '5Ca2+ determined by scintillation counting. Protein concentration was determined by measuring absorbance at 276 nm using an E:&l,,,,,, 276nm = 0.18 (see "Results") and by amino acid analysis of acid-hydrolyzed aliquots.
As a further internal control, the specific activity of rsCaz+ used in each set of experiments was determined by atomic absorption spectrometry and scintillation counting. Data from these determinations were treated by the method of Scatchard (24).
Amino Acid Analysis-Amino acid analyses were performed as previously described (20). For composition analysis, time-course hydrolysates were analyzed to correct for destructive losses of serine and threonine and for the slow release of valine and isoleucine. Tryptophan was determined spectrophotometrically by the method of Bencze and Schmid (25 (29). PTH-derivatives were identified by gas chromatographic analysis (30) and by amino acid analysis subsequent to back hydrolysis as described by Smithies et al. (31). Amino acid sequences of tryptic peptides, prepared as described below, were determined by a modification of the procedure of Gray (32). Samples of peptide (10 to 20 nmol) were evaporated to dryness in tubes equal to the number of Edman cycles to be performed plus one for NH,-terminal determination only. Peptides were then degraded the required number of cycles by the Edman procedure as described by Peterson et al. (33). NH,-terminal amino acids were determined as the dansyl derivatives exactly as previously described (34).
Z'rypsin Digestion-Trypsin digestion was performed as previously described (35) both on bovine brain modulator protein which had been oxidized with performic acid by the method of Hirs (26) and on the untreated protein in the presence of 1 rn~ EGTA (11). The resulting tryptic peptides were purified by a combination of ion exchange chromatography and gel filtration using volatile buffer systems as described under "Results." Peptides were detected in column effluents using a modified Technicon Autoanalyzer system as previously described (35). The appropriate fractions were pooled, the solution was evaporated to dryness, and the residues were dissolved in 50% (v/v) acetic acid and stored at -20".
A trace identical with that shown in Fig. 3 was obtained when the S-100 pool (Fractions 120 to 140) from Step 5 was chromatographed on the same Sephadex G-100 column. The overall yield of bovine brain modulator protein obtained with this procedure (-40 mg/kg of brain) is 13% of that present in the original homogenate. The purity of this material was assessed by gel electrophoresis and electrofocusing. As shown in Fig. 4, one major band and some minor, slower migrating bands are seen by nondenaturing discontinuous electrophoresis (gel A), while only one major band is detected when the same amounts of protein are examined by discontinuous electrophoresis in the presence of 8 M urea (gel B), or isoelectric focusing in 6 M urea (gel C). A single band is seen when this sample is subjected to sodium dodecyl sulfate-electrophoresis (see Fig.  7). These results suggest that the apparent contaminants seen in Fig. 4A

FIG. 2. DEAE-Sephadex
A-50 chromatography (Step 5). Chromatography was performed as described in the text with lo-ml fractions collected.
Modulator protein (MP) and S-100 were detected in fractions by discontinuous polyacrylamide electrophoresis as shown by the gels positioned above the fractions examined.
A pin was inserted in each gel at the position of the bromophenol blue tracking dye after electrophoresis.
S-100 co-migrated with tracking dye in this system. Fractions 155 to 170 were pooled for modulator protein purification, while Fractions 129 to 140 were pooled for S-100 purification.
Fro. 8. Gel filtration on Sephadex G-100 (Step 6). Gel filtration was performed as described under "Methods." Ten-milliliter fractions were collected as indicated on the figure and treated as described in the text. when analyzed using standard Beckman single column methodology as described under "Methods." Under these conditions, histidine eluted at 210.8 min after sample injection followed closely by N'-methylhistidine (212.3 min), N"methyllysine (224.0 min), and ammonia (232.0 min). N'-diand trimethyllysines were not tested. As shown in Table II, 24-hour hydrolysates of bovine brain modulator protein contained 1.1 mol of Compound X/18,000 g of protein calculated using the ninhydrin color constant for histidine. No such ninhydrin-positive compound was detected in hydrolysates of rabbit skeletal muscle TN-C, in agreement with the amino acid sequence previously published (40).
Analyses of acid hydrolysates of l-to 2-mg quantities of purified bovine brain modulator protein for inorganic phosphorus show less than 1 mol of phosphorus/18,000 g of protein. Analyses for carbohydrate were negative.
The amino acid compositions of rabbit muscle TN-C, hake parvalbumin, and bovine brain S-100 are also shown in Table  II. The amino acid compositions of bovine brain modulator protein and rabbit muscle TN-C are strikingly similar, as has also been noted for the bovine heart protein (13,14). The only significant differences are (a) the presence of 12 residues of threonine in the modulator protein as compared to 7 in muscle TN-C; (b) the presence of a single residue of cysteine in rabbit muscle TN-C which is absent in the brain protein; and (c) the presence of an unknown ninhydrin-positive compound in the modulator protein and not in skeletal muscle TN-C. The amino acid composition of hake parvalbumin is quite distinct from either the brain or rabbit muscle protein, despite the fact that parvalbumin appears to share a common genetic ancestry with skeletal muscle TN-C (40). The amino acid composition of the S-100 protein fraction from bovine brain is clearly not related to brain modulator protein.
NH&erminal Determinations-No cu-dansyl amino acids could be detected when 10 nmol of bovine brain modulator protein were subjected to procedures previously described for determination of free NH,-terminal amino acids by the dansyl technique. When 250 nmol (4.52 mg) of this protein were subjected to automated Edman degradation in a Beckman model 890B Sequencer as described under "Methods," no PTH-derivatives were detected during 10 cycles of degradation. Sperm whale myoglobin applied to the cup immediately after the 10th cycle gave only the appropriate amino acids at each cycle for 10 cycles of degradation with an average repetitive yield of 92%.
These data suggest that, as with skeletal and cardiac muscle 42), bovine brain modulator protein has a blocked NH, terminus. The bovine heart protein also appears to have a blocked NH, terminus (13).
Spectral Properties- Fig.  8 shows the spectrum obtained with bovine brain modulator protein at neutral pH. This spectrum is striking due to the presence of the phenylalanine Similarities between Calcium-dependent Regulatory Proteins These data are very similar to those previously reported by Lin et al. (12) for the bovine brain protein and by Teo and Wang (11)   NH, terminus could be detected for this peptide by the dansyl method (see Table  III). The peptides produced by trypsin digestion of bovine brain modulator protein in the presence of 1 mM EGTA have also been purified exactly as described in the legend to Fig. 10. The elution profile (data not shown) obtained for these peptides was similar to that shown in Fig. 10 except that (a) peptide 4 was not resolved from the group of peaks labeled 5 in Fig. 10; and, (b) a new peptide (Tp-SA) was isolated eluting between peptides 8 and 9 in the profile shown in Fig. 10. The pooled fractions containing the unresolved peptide 4 from this separation were resolved into three pure peptides by gel filtration as shown in Fig. 11B Fig. 10. B, Pool 5 from a separation similar to that shown in Fig. 10 of a trypsin digest of bovine brain modulator protein in the presence of EGTA (see "Results"). e-Dansyllysine (1 rmol) was added to each sample as a marker. Tp-3 : Tp-4: Tp-5B: Tp-6 7'11-8: Tp-8A: Tp-9: Tp-10: Tp-11:  The results of purification and characterization of bovine brain modulator protein reported in this study differ in a number of ways from the results of Lin et al. (12). The purification procedure reported here appears to yield 10 times more material per kg of bovine brain than obtained by their procedure. The yield of pure modulator protein which they reported (4.2 mg/kg of brain) was measured by the method of Warburg and Christian (451, a spectrophotometric technique using the empirical relationship: Protein concentration (mg/ml) = 1.55 (A 280 .m) 0.67 (A,,, .m) The modulator protein does not obey this empirical relationship, as noted under "Results." For example, the absorbances at 280 and 260 nm of 1 mg/ml solutions of the bovine brain protein would give a value of 0.17 mg/ml if substituted into Equation 1. This would indicate that Lin et al. (12) actually obtained approximately 25 mg of pure modulator protein/kg of brain with their purification procedure. This also reduces the specific activity of the pure protein substantially.
Using a specific activity corrected by multiplying that reported by 0.17, the total amount of modulator protein in their brain homogenate would have been 348 mg/kg of tissue. This is very close to the value of 310 mg/kg which was obtained in our homogenate supernatant as judged by discontinuous electrophoresis (see Table I). Therefore, it is concluded that the assay of modulator protein by discontinuous electrophoresis is at least as accurate as that obtained by enzymic methods. In addition, it appears that the procedure described under "Methods" solubilizes bovine brain modulator protein efficiently. In contrast, brain homogenates prepared by the procedure of Lin et al. (12) in a buffer containing no chelating agents yielded only % of the total modulator protein in a soluble form, the remainder being lost in the insoluble fraction after centrifugation of the crude homogenate. The procedure described under "Methods" utilized buffers containing 1 mM EDTA in most steps including homogenization.
Only trace amounts of modulator protein were detected by discontinuous electrophoresis of samples of the resuspended pellet obtained from the crude homogenate. This more efficient extraction during homogenization probably accounts for the higher yields of the pure protein obtained using the procedure described under "Methods." As noted under "Results," one of the major problems encountered in the purification of modulator protein from brain is to separate it from the acidic S-100 protein fraction. In our procedure, this was accomplished by ion exchange chromatography on DEAE-Sephadex A-50 with a relatively shallow salt gradient (Step 5). Gel filtration on Sephadex G-100 is then sufficient to yield modulator protein which is more than 95% pure as judged by gel electrophoresis under a number of different conditions. prepared from porcine, rabbit, rat, and chicken brains.' The structure of Compound X is currently under investigation.
No NH,-terminal amino acid was detected in the current studies for the purified bovine brain modulator protein by either dansylation or automated Edman degradation. Similar observations have been made for the bovine heart protein (13). Using the dansyl method, Lin et al. (12) detected an NH,-terminal valine in their preparations of bovine brain modulator protein. The reason for these differences is unclear. However, primary structure analyses currently in progress in our laboratory should provide precise information concerning the NH, terminus of this protein.
The other physicochemical properties of bovine brain modulator protein presented here are in general agreement with those determined by Lin et al. (12) and are identical with those of the bovine heart protein (13).
A number of Ca'+-binding proteins have been isolated from brain and other neurosecretory tissue. The S-100 protein fraction from brain, originally isolated by Moore and McGregor (49) and subsequently shown to be a calcium-binding protein (50) has been studied extensively. However, no biological function has yet been determined for S-100. In addition, the heterogeneity of this protein fraction (51-53) and the fact that physiological concentrations of monovalent cations abolish high affinity Ca2+ binding indicate that S-100 is not a unique, specific Ca'+-binding protein.
The amino acid composition of bovine S-100 is clearly distinct from bovine brain modulator protein as can be seen in Table II. It is not possible to directly correlate the specific activity of the bovine brain modulator protein prepared here with the value published by Lin et al. (12) because of differences in assay procedures. However, it does appear to have a specific activity identical with the bovine heart protein. 3 The procedure described in this paper has been used to prepare modulator protein from a number of sources including porcine, rabbit, rat, and chicken brains.' In all cases, 40 to 70 mg of protein were isolated/kg of brain. The purity of these preparations was comparable to that of the bovine brain protein shown here. In large scale purification (2 to 6 kg of brain) a modification of the procedure described under "Methods" has been devised which yields 70 to 100 mg of modulator protein/kg of bovine brain.5 Bovine brain modulator protein isolated by the procedures reported here is identical in physicochemical properties with that isolated and characterized by Lin et al. (12) with two important exceptions. It is reported here that this protein contains approximately 1 mol of an unknown ninhydrin-positive compound, Compound X/mol of protein. This compound which is not well resolved from lysine with the standard two-column methodology of Moore and Stein (48) used for amino acid analysis in previous studies (11,12) has been localized in a single tryptic peptide isolated from bovine brain protein, peptide Tp-5A (see Table III). It also has been found in bovine heart modulator protein (13)

and in modulator proteins
In addition to S-100, a number of other small acidic Ca*+-binding proteins have been isolated from brain. Wolff and Siegel (54) have reported the isolation and characterization of a Ca'+-binding phosphoprotein from porcine brain. Brooks and Siegel (55) have reported the isolation of a similar protein from bovine brain and adrenal medulla. In both cases, the molecular masses of these proteins were determined by sedimentation to be approximately 12,000 daltons. However, these analyses were performed in dilute buffers (0.01 M phosphate, pH 7.0), conditions which give anomalously low molecular weights for bovine brain modulator protein as noted under "Results." The amino acid compositions of the bovine brain and adrenal medulla proteins calculated using a molecular mass of 18,000 daltons are very similar to that of bovine brain modulator protein reported here. However, unlike these phosphoproteins which contain 2 mol of phosphorus/12,000 g of protein, bovine brain modulator protein prepared by our procedure contains no detectable phosphate, in agreement with the results of Lin et al. (12) and Teo and Wang (11) for bovine brain and heart proteins, respectively. The amino acid composition reported for the porcine brain phosphoprotein (54) is quite distinct from either the bovine brain and adrenal medulla phosphoproteins or bovine modulator protein. Porcine brain modulator protein, prepared by our procedure, appears to be very similar in structure to bovine brain modulator protein as judged by amino acid composition, NH,-terminal analysis, and tryptic peptide mapping, and is devoid of phosphorus.' 'T. C. Vanaman, unpubl i shed observations.
Recently, Wolff and Brostrom (56) and Brostrom et al. (57) have shown that the phosphoprotein isolated from bovine brain is an activator of cyclic nucleotide phosphodiesterase (57) and also appears to activate adenylate cyclase (57). Cheung et al. (58) have also shown that bovine brain modulator protein can activate adenylate cyclase and appears to be bound to detergent-solubilized adenylate cyclase preparations. As is noted under "Results," the physicochemical properties of bovine brain modulator protein are very similar to those of rabbit skeletal muscle TN-C, as has also been noted for bovine heart activator (13,14).