The contribution of phosphorylation and loss of COOH-terminal arginine to the microheterogeneity of myelin basic protein.

Two guinea pig myelin basic protein preparations which differed markedly in their contents of high pH electrophoretic or chromatographic forms were studied in an attempt to elucidate the causes of their microheterogeneity. Both total preparations and components isolated therefrom were examined for their amino acid compositions, NH2-terminal and COOH-terminal residues, total phosphorus contents, amd contents of phosphamino acids. The results showed that the five components differed sequentially by a single charge and that the microgeterogeneity arose as a result of secondary modifications of a single secies (Component 1) Of basic protein. Two modifications were demonstrated; viz. phosphorylation of serine and threonine and loss of COOH-terminal arginine. These two modifications were insufficient to account completely for the observed microheterogeneity; an additional cause, deamidation, was postulated. From the relationship between the number of components present in the total basic protein, the phosphorus and phosphoamino acid contents of the components, and the changes in relative electrophoretic mobility of the components which accompanied their phosphorylation and dephosphorylation we conclude that in the native basic protein no more than two sites in any polypeptide chain are phosphorylated.

From the Research Division and Clinical Laboratory, Nakamiya Mental Hospital, Hirakata, Osaka, Japan Two guinea pig myelin basic protein preparations which differed markedly in their contents of high pH electrophoretic or chromatographic forms were studied in an attempt to elucidate the causes of their microheterogeneity.
Both total preparations and components isolated therefrom were examined for their amino acid compositions, NH,-terminal and COOH-terminal residues, total phosphorus contents, and contents of phosphoamino acids. The results showed that the five components differed sequentially by a single charge and that the microheterogeneity arose as a result of secondary modifications of a single species (Component 1) of basic protein. Two modifications were demonstrated; uiz. phosphorylation of serine and threonine and loss of COOH-terminal arginine. These two modifications were insufficient to account completely for the observed microheterogeneity; an additional cause, deamidation, was postulated. From the relationship between the number of components present in the total basic protein, the phosphorus and phosphoamino acid contents of the components, and the changes in relative electrophoretic mobility of the components which accompanied their phosphorylation and dephosphorylation we conclude that in the native basic protein no more than two sites in any polypeptide chain are phosphorylated.
Myelin basic protein isolated from central nervous system tissue of vertebrates exists in several forms which differ in charge, as judged by their behavior during alkaline-pH gel electrophoresis (1,2) or ion exchange chromatography (2)(3)(4). These different forms have been isolated and shown to have essentially the same amino acid compositions (2,4) and encephalitogenic activities (4). All forms contained unsubstituted arginine, NC-monomethylarginine, and N",N'"-dimethylarginine at position 106, a finding which eliminated differential methylation of this residue as a cause of the microheterogeneity (2,4). A possible cause of at least some of the observed microheterogeneity has been suggested by Martenson et al. (3), who pointed out that since the COOH terminus of the protein is -Ala-Arg-Arg, partial loss of these arginines would yield three forms of the protein differing by a single charge.
Recently the basic protein has been found to exist in a partially phosphorylated state. Direct chemical analyses of the *This work was supported in part by Grant 828-A-4 from the United States Multiple Sclerosis Society.
protein from bovine brain have demonstrated that the phosphorylation involves both serine and threonine residues (5, 6).
In viuo studies in the rat have shown the rapid uptake of radioactive orthophosphate into serine and threonine residues of the basic protein (5-7). Studies by Carnegie et al. (8,9) have shown that endogenous and exogenous protein kinases phosphorylate specific serine and threonine residues in the myelin basic protein in uitro. As an explanation of the microheterogeneity of the basic protein, Miyamoto and Kakiuchi (6) have suggested that the different forms of the protein result from different extents of phosphorylation.
In the present study we describe experiments designed to examine these two possibilities (loss of COOH-terminal arginine and phosphorylation) for their role in producing the protein's microheterogeneity. When maximal resolution of' components for the purpose 01' quantitation was desired, the amount of protein applied to the gels was limited to 50 pg. llnder these conditions repeated electrophoretic analvses of the same sample yielded relative area values agreeing to with"in 5% for well resolved components and 10'; for less well resolved ones.

Treatment of Basic
Protein with Carbox.ypeptidases A and H-It is known from studies with carboxypeptidases A and B (21, 22) that the COOH terminus of the intact guinea pig myelin basic protein is -Ala-Arg-Arg and that removal of the fourth residue, methionine, is inhibited (22). In the bovine protein this inhibition is due to the presence of an adjacent prolyl residue (231, and it can be assumed that the same is true for the guinea pig protein as well. In order to assess the importance of the intact COOH terminus in contributing to the over-all pattern of microheterogeneity, the basic protein preparations were treated with carboxypeptidases A and B (Fig. !, left). Loss of COOHterminal residues from preparation A resulted in a complete anodic shift such that the relative proportions of Components 3, 4, and 5 in the new electrophoretic pattern were essentially the same as those of Components 1. 2, and 3 in the pattern obtained prior to enzyme treatment. It appeared as if Components 1 to 5 differed sequentially by a single charge and that removal of the two COOH-terminal arginines had converted Components 1, 2, and 3 to Components 3, 4, and 5, respectively. Loss of COOH-terminal residues from preparation B, on the other hand, resulted in relatively little change in the pattern other than the loss of Components 1 and 2. From the behavior of the two preparations it was obvious that the and after (---) enzyme treatment. Corresponding components are labeled 1 to 5. Left, treatment with carboxypeptidases A and B for 30 min: 50 pg of preparation A and 100 peg of preparation B were applied to the gels. Right, treatment with alkaline phosphatase; 75 fig of each preparation were applied to the gels. Electrophoresis was carried out at high pH toward the cathode for Z hours at 3.75 ma/gel. The resolution of preparation A (right) was inferior to that of the same preparation (left) because a larger amount of protein was applied to the former gels, and the solutions contained sufficient (NH,),SO, to raise their conductivities and broaden the starting zones.
fraction of polypeptide chains having the intact COOH terminus was much greater in preparation A than in B. Treatment of Basic Protein with Alkaline Phosphatase-The role of phosphorylation in contributing to the microheterogeneity was examined in an analogous fashion ( Fig. 1,  right). Dephosphorylation of the basic protein resulted for both preparations in a partial shift to Components 1 and 2. The effect was shown most clearly with preparation A, where appreciable reduction in the relative amounts of Components 3 and 4, as well as 5, occurred. Chemical analyses of these preparations before and after enzyme treatment showed that the electrophoretic shift was accompanied by loss of 98 to 100% of the phosphorus originally present. These results indicated that Components 0, 4, and 5 existed in a partially phosphorylated form, whereas Components 1 and 2 were unphosphorylated.
Isolation of Components-Detailed studies on the relationship between microheterogeneity and modification of the basic protein were made possible by the isolation of the individual components by ion exchange chromatography. The partial characterization of components obtained from preparation B has been described previously (4). Those obtained from preparation A appeared to be essentially homogeneous by electrophoresis at alkaline pH (Fig. 2). Electrophoresis at acid pH, however, revealed that Components 3, 4, and 5 were still contaminated to the extent of approximately lo'%> with protein of slightly higher electrophoretic mobility. The corresponding components of preparations A and B had identical electrophoretie mobilities at acid and alkaline pH.
Amino Acid Compositions of Components-As shown in Table I there were no significant differences in amino acid FG. 2 (left). Electrophoresis in parallel of 75 pg of total preparation A (2') and 50 pg of each of its Components 1 to 5. Electrophoretic conditions were the same as in Fig. 1.
Flo. 3 (right). Electrophoresis in parallel of 50 pg of each component of preparation A after treatment with carboxypeptidases A and B for 10 min. Starting components are designated by the numbers below the gels; components arising as the result of enzyme treatment are designated by numbers at the sides of the gels. Electrophoretic conditions were the same as in Fig. 1 (Table II). No conversion of any components occurred upon incubation for 90 or 120 min in the absence of the enzymes. A further difference between the two preparations lay in the nature of the final products of Components 3 and 5. In preparation A Components 3 and 4, respectively, were ultimately converted to Components 5 and 7 only; in preparation B the final spectrum included Components 4 and 6 as well (Table II).
Since the conversion of any component to the one of next higher number involved the loss of one positive charge, the data in Table II (2) + 0.13(l) = 0.63 mol/mol.
Portions of the samples analyzed electrophoretically were analyzed for free arginine and alanine on the amino acid analyzer. The actual yields of these amino acids obtained after 90. or 120-min treatment with carboxypeptidases A and B are presented in Table III, together with the theoretical yields of arginine calculated on the basis of the data in Table II. All components released approximately 1 mol of alanine/mol of protein but no additional amino acids other than arginine and a trace of methionine. The actual yields of arginine determined directly were in reasonably good agreement with the theoretical contents of COOH-terminal arginine estimated by electrophoretic analysis.
As a further comparison of the direct and indirect methods for determining the COOH terminus of each component, the time courses of amino acid release and component conversion were followed. Each component released arginine in parallel with its conversion to the next higher homologue that differed by two charges. For example, it can be seen (Fig. 4, left) that conversion of Components 3 and 5 to Components 5 and 7, respectively, proceeded via Intermediates 4 and 6 and that the extent of conversion of Components 3 and 5 was much greater for preparation A than for B. As shown in Fig. 4 (right) arginine was released faster than alanine from the A components, whereas the reverse was true for the B components. The results obtained with Component 4 paralleled those shown for Components 3 and 5.

Relative
Electrophoretic Mobility versus Extent of Phosphorylation-Component 1, which is phosphorus-free (see below), was phosphorylated by brain protein kinase. In the course of 24 hours approximately 3.4 g atom of phosphorus/mol were incorporated.
As shown in Fig. 5 (top) a whole series of  III COOH-terminal residues of myelin basic protein and its components Preparation A and its components were treated with carboxypeptidases A and B for 120 min. Preparation B and its components were treated for 90 min. The amino acids released were quantitated on an amino acid analyzer. Values in parentheses are theoretical ones calculated from electrophoretic data in Table II (see text). to higher numbered components upon incubation with carboxypeptidases A and B. Right, corresponding release of COOHterminal amino acids from Components 3 (top) and 5 (bottom) upon incubation with the enzymes. -, preparation A; ---, preparation B.
components differing from one another by two charges was generated. The earliest products resulting from phosphorylation of Component 1 were Components 3 and 5. These must contain one and two phosphate groups in their respective polypeptide chains. The other components must have arisen from the successive incorporation of additional phosphate groups. It is obvious that Components 7, 9, and 11 are not detectable in the total isolated basic protein. The importance of carrying out electrophoresis above the pK, of phosphoric acid and, more significantly, close to the isoelectric point of the basic protein is shown in Fig. 5 (bottom). Electrophoresis at acid pH failed to reveal any microheterogeneity, and only a slight progressive decrease in electrophoretic mobility accompanied the increase in phosphorylation of Component 1. Phosphorus Analyses-The extent to which each component of the basic protein was phosphorylated was examined in a manner analogous to that used for COOH-terminal analyses. Incubation of Components 1 and 2 from either preparation with alkaline phosphatase for 24 hours had no effect on their electrophoretic mobilities. Components 3 and 4, however, underwent partial conversion to Components 1 and 2, respec-  Fig.  1. Bottom, electrophoresis in parallel in acid gels for 2 hours at 2.5 ma/gel. tively, while Component 5 underwent partial conversion to Components 1 and 3. These two-charge changes were precisely those expected from dephosphorylation.
No changes in electrophoretic mobility occurred when incubation was carried out in the absence of the enzyme. The capacities of Components 3,4, and 5 of preparation A to convert to components of higher electrophoretic mobility (lacking two or four negative charges) were much greater than those of preparation B. As an example, Fig. 6 shows that more than one-half of Component 3 of preparation A underwent conversion to Component 1, whereas the conversion of Component 3 of preparation B to Component FIG. 6. Electrophoresis in parallel of Component 3 (50 lg) before and after treatment with alkaline phosphatase. A, preparation A; B, preparation B. Control and enzyme-treated samples are designated C and E, respectively. Electrophoretic conditions were the same as in Fig. 1. 1 was barely significant. The results of treatment of Components 1 to 5 of both preparations with alkaline phosphatase are summarized in Table IV. They indicate the fraction of each initial component which contained either one or two phosphate groups; e.g. for preparation A Component 3 consisted of polypeptides 53% of which were monophosphorylated, while Component 5 consisted of polypeptides 44% of which were monophosphorylated and 29% of which were diphosphorylated. As described above for the electrophoretic data on the carboxypeptidases A and B treatment of components, the data in Table IV permitted the calculation of theoretical phosphorus contents: for preparation A Component 3 contained 0.53 g atom of phosphorus/mol, while Component 5 contained 0.29 (2) + 0.44(l) = 1.02 g atom of phosphorus/mol. Table V presents the experimental data from chemical analyses of phosphorus in the components, together with the theoretical values calculated from the electrophoretic data in Table IV. The agreement between the two sets of data was excellent. Summation of the products (g atom of phosphorus/ mol x fraction of total basic protein) for each component yielded a value of 0.16 or 0.18 g atom of phosphorus/mol of total preparation B, depending upon whether the actual or theoretical phosphorus values were used. Similarly, for total preparation A a single value of 0.23 g atom/m01 was calculated. These values agreed well with those (0.16 and 0.20 g atom/mol, respectively) determined directly.

Analyses for Phosphoamino
Acids-Examination of 8-hour hydrolysates of the total basic protein preparations and their Components 3, 4, and 5 on an amino acid analyzer revealed a single symmetrical peak of strongly acidic material eluting in the position of authentic phosphoserine and/or phosphothreonine.* High voltage electrophoresis of these hydrolysates demonstrated in each case the presence of both phosphoserine ?After shorter hydrolysis times (3 to 6 hours) phosphopeptides appeared to be present also. On the 20-cm DC-GA column a shoulder was present on the trailing edge of the phosphoamino acid peak. The material had a higher 440:570 absorbance ratio than the phosphoamino acids and could be completely separated from the latter on the 55-cm AA-15 column.  and phosphothreonine. These two amino acids together accounted for 67 to 85% of the phosphorus present in the samples prior to hydrolysis. On occasion, hydrolysates of phosphoruscontaining basic proteins also contained a highly acidic ninhydrin-positive substance which migrated ahead of phosphoserine. Its occurrence was not eliminated by trichloroacetic acid precipitation and ethanol-ether washing of the protein prior to hydrolysis. Acid hydrolysates (8 hour) of pepsin and ovalbumin contained the material as well. Since the phosphorylated residue in pepsin and ovalbumin is known to be 0-phosphoserine (24), it would appear that the unknown compound might be a phosphorus-containing degradation or rearrangement product of this amino acid which forms during acid hydrolysis of certain phosphoproteins. An electrophoretogram illustrating the composition of 8-hour hydrolysates is depicted in Fig. 7. DISCUSSION In the present study we have examined two preparations of myelin basic protein with regard to the COOH termini and phosphoamino acid contents of their constituent components. The results are summarized in Table VI. The two preparations differed markedly in the relative proportion of their components: in one preparation (A) it was the most basic form of the protein (Component 1) and in the other (B) the intermediate   (53) Mono- (55) Mono-(44); di- (29) None None Mono- (10) Mono- (11) Mono-(33); di- (6) a Values calculated from electrophoretic data in Tables II and IV. in preparation A 83% of the polypeptides in Component 2 were intact, yet they differed from Component 1 by a single charge. Possibly, this additional mechanism of component conversion might have been deamidation of a limited number of the 10 amidated dicarboxylic acid residues reported (23) to be present in the protein. This possibility is currently under investigation.
The preparations which have been described represent two extremes in a large number of basic protein preparations which have been studied in our laboratory. Most resemble preparation A in containing predominantly Component 1, with Component 3 next in quantitative significance.
Recently, we have prepared preparations of basic protein obtained from freezeblown guinea pig brain. In the freeze-blowing procedure the tissue is instantaneously removed from the conscious animal and frozen so as to virtually eliminate postmortem changes (25). The electrophoretic pattern of basic protein obtained from tissue obtained in this manner is identical with that of preparation A.3 It is obvious, therefore, that the basic protein of preparation B cannot be representative of the basic protein as it exists in uiuo, inasmuch as it has undergone extensive alterations at the COOH terminus. Recently Chou et al. (26) have described an analysis of the COOH termini of basic proteins corresponding to Components 1 and 3. These investigators treated the components with BrCN, isolated the fragment COOH-terminal to the methionine near the end of the polypeptide chain, and found the amino acid composition in each case to correspond to the sequence Ala-Arg-Arg.
These data are in complete agreement with our results with preparation A. Phosphorylated polypeptides in Component 5 consist of (a) chains containing both a phosphoserine and a phosphothreonine residue and (6) monophosphorylated chains with two additional negative charges resulting from arginine and/or amide loss.
It is important to note that in the native basic protein no more than two sites in any polypeptide chain are phosphorylated. Our studies have shown that brain phosphokinase converts Component 1 in vitro to Components 3 and 5, but not to Components 2 and 4, and that Components 1 to 5 differ sequentially by a single charge. Therefore, Component 5 must differ from Component 1 by no more than four charges; i.e. two phosphate groups. A phosphoamino acid present at three sites in the polypeptide chain would yield an electrophoretic species differing from Component 1 by six charges and would appear as Component 7. This species as well as Components 9 and 11 can, in fact, be generated from Component 1 by prolonged incubation with brain phosphokinase. However, no forms of basic protein corresponding to Component 7, 9, or 11 have been detected either in basic protein preparations or in crude pH 3.0 extracts obtained from quick-frozen or freeze-blown tissue.
'l'he phosphorus content of a basic protein preparation is bound to vary depending upon which components are removed in the course of its purification.
Since Components 1 and 2 contain no phosphorus, it is obvious that "highly purified" preparations consisting only of these two components would be phosphorus-free. Inclusion of Component 3 would yield a preparation containing significantly less than the 0.2 g atom of phosphorus/mol of protein found in our studies and in those of Miyamoto and Kakiuchi (6). This relationship between phosphorus content and component composition could explain why Eylar and Thompson (27) reported a phosphorus content of less than 0.1 g atom/m01 for their basic protein preparation.