Methylation at specific altered aspartyl and asparaginyl residues in glucagon by the erythrocyte protein carboxyl methyltransferase.

Protein carboxyl methyltransferases from erythrocytes and brain appear to catalyze the esterification of L-isoaspartyl and/or D-aspartyl residues but not of normal L-aspartyl residues. In order to identify the origin of these unusual residues which occur in subpopulations of a variety of cellular proteins, we studied the in vitro methylation by the erythrocyte enzyme of glucagon, a peptide hormone of 29 amino acids containing 3 aspartyl residues and a single asparagine residue. Methylated glucagon was digested with either trypsin, chymotrypsin, pepsin, or endoproteinase Arg C, and the labeled fragments were separated by high-performance liquid chromatography and identified. In separate experiments, methyl acceptor sites were determined by digesting glucagon first with proteases and then assaying purified glucagon fragments for methyl acceptor activity. Using both approaches, we found that the major site of methylation, accounting for about 62% of the total, was at the position of Asp-9. Chemical analysis of fragments containing this residue indicated that this site represents an L-isoaspartyl residue. A second site of methylation, representing about 23% of the total, was detected at the position of Asn-28 and was also shown to represent an L-isoaspartyl residue. Methyl acceptor sites were not detected at the positions of Asp-15 or Asp-21. Preincubation of glucagon under basic conditions (0.1 M NH4OH, 3 h, 37 degrees C) increased methylation at the Asn-28 site by 4-8-fold while methylation at the Asp-9 site remained unchanged. These results suggest that methylation sites can originate from both aspartyl and asparaginyl residues and that these sites may be distinguished by the effect of base treatment.

Previous studies have focused on defining the chemical nature of the methylatable aspartyl residue and its esterified product. For instance, aspartic acid /?-[3H]methyl ester isolated from proteolytic digests of 3H-methylated erythrocyte membrane and cytosolic proteins is entirely in the D- Configuration (McFadden and Clarke, 1982;O'Connor andClarke 1983, 1984). Since the maximal amount of D-aSpartyl 8methyl ester from these preparations accounts for only 15% of the 3H label, it is possible that the remaining methylated sites might represent other types of esters. One such candidate is an L-isoaspartyl a-methyl ester. L-Isoaspartyl residues have been shown in fact to be good methyl accepting substrates in a variety of synthetic peptides (Murray and Clarke, 1984;McFadden and Clarke, 1986;Johnson et al., 1987). Homologous synthetic peptides containing L-aspartyl residues are not substrates.
We were interested in this work in asking what specific residues in what protein sequences contribute to the formation of methylatable sites. It appears that at least one source of substrate sites is L-asparagine residues. Previous work has suggested that the 39-residue polypeptide hormone ACTH' could be methylated at an ~-isoaspartyl-25-glycyl-26 site derived from original L-Asn-Gly residues by mild base treatment (Aswad, 1984a). Asn-Gly sequences may be particularly prone to L-isoaspartyl formation via succinimide intermediates in proteins (Bornstein and Balian, 1977;Bernhard, 1983;Blodgett et al., 1985;Geiger and Clarke, 1987). Base treatment of bovine brain calmodulin greatly increases the methylatability of this substrate, and it has been suggested that this increase in activity may also be due to L-isoaspartyl formation from the two L-Asn-Gly sequences that are present in this protein (Johnson, et al. 1985). On the other hand, heating erythrocyte calmodulin increases its methylatability, and this increase is independent of pH in the range 5-9 (Brunauer and Clarke, 1986). Because L-isoaspartyl formation from asparagine residues might not be expected to be independent of pH, it is possible that these heat-induced sites originate from aspartyl residues. We were thus interested to know whether both aspartyl and asparaginyl residues could give rise to methylatable sites.
To answer this question and to more precisely define the origin of methylatable sites in proteins, we have chosen to study the well characterized 29-residue polypeptide hormone glucagon as a model for protein methyl acceptors. This peptide can be methylated to approximately the same extent as many red cell proteins, and because of its small size it is amenable to peptide mapping techniques (Fig. 1). In this study, we have found that the major methyl acceptor site in glucagon is derived from Asp-9 and that this site represents an L-isoas-party1 residue. A minor site of methylation occurs at an Lisoaspartyl site derived from Asn-28. Base treatment was found to enhance methylation at the Asn-28 site but cause little change in methylation at the Asp-9 site.

EXPERIMENTAL PROCEDURES'
Portions of this paper (including "Experimental Procedures," part of "Results," Figs. 6 and 7, and Tables I11 and IV) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. &quest Document No. 86M-2283, cite the authors, and include a check or money order for $8.00 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press. cagon was methylated with [3H]AdoMet and erythrocyte protein carboxyl methyltransferase to a maximum extent of 0.004 mol of methyl groups/mol of polypeptide, and the 3H-methylated'hormone was purified by HPLC (Fig. 2). To identify the sites of methylation, 3H-methylated glucagon was digested with trypsin, the fragments were separated by HPLC, and the major UV-absorbing peaks were identified by amino acid analysis (Fig. 3, Table I; see Fig. 1 for nomenclature). All of the five major UV-absorbing peaks were found to be composed of amino acids corresponding to sequences in glucagon (Fig.  1). These fragments resulted from tryptic cleavages after Lys-12, Arg-17, or Arg-18, except for the ArgT3' fragment which was produced by a cleavage after Trp-25. Cleavage at this position has been observed with trypsin treatment of glucagon by others (Bromer et al., 1957b;Maroux et al., 1966) and can be attributed to either residual chymotrypsin activity in the trypsin preparation or an activity inherent to trypsin (Maroux et al., 1966).  HPLC. The glucagon fragment, TI, HPLC purified from 200 pg of trypsin-digested glucagon, was resuspended in 150 pl of 0.2 M sodium citrate, pH 6.0, and 105 pl of this was incubated at 37 "C for 1 h with 42 pl of [3H]AdoMet (15 Ci/mmol, 1 mCi/ml in dilute H2SOI (pH 2.5-3.5):ethanol, 9:l (v:v)) and 63 pl of partially purified erythrocyte carboxyl methyltransferase isozymes (see "Experimental Procedures"). This mixture was quenched by the addition of 210 pl of 10% trifluoroacetic acid, and 200 pgof trypsin-digested glucagon was added to the methylated TI as markers. This material was chromatographed on HPLC by methods identical to those used for the purification of tryptic fragments of glucagon.
None of the 3H-methylated peptides comigrated exactly with any of the identified tryptic peptide-containing UVabsorbing peaks (Fig. 3A). For instance, the major peak of radioactivity migrating at 37-38 min and representing 66% of the total radioactivity eluted between TI and ArgT3'. There are at least two reasons why a methylated peptide might not comigrate with an unmethylated fragment. First, methylatable peptides, which are likely to contain L-isoaspartyl or Daspartyl peptides, could migrate slightly differently than normal L -A s~ or L-Asn-containing peptides. Furthermore, a peptide containing a methyl ester residue may be retained longer than an unmethylated peptide on reverse phase HPLC columns because of its increased hydrophobicity.
Since the addition of a methyl group was found to have a  Fig. 2) were incubated with chymotrypsin at 0.7 mg/ml glucagon, 0.05 mg/ml chymotrypsin, 1 mg/ml calcium chloride, and 0.15 M sodium citrate at pH 6.0 for 15 min at 37 "C. The incubation was quenched by addition of 140 p1 of 10% trifluoroacetic acid, and the fragments were separated by HPLC as described under "Experimental Procedures." One-ml fractions were collected, and 0.5 ml of each fraction was counted in 10 ml of scintillation fluid. E, methylation of chymotryptic fragments of glucagon. Chymotryptic fragments of glucagon were prepared from 156 pg of glucagon by incubating glucagon at a final concentration of 0.62 mg/ml, with chymotrypsin at 0.05 mg/ml and calcium chloride at 1 mg/ml in 0.15 M sodium citrate buffer as described above and under rated on HPLC (as above), and 0.8 ml of each I-ml fraction was "Experimental Procedures." The chymotryptic fragments were sepalyophilized and resuspended in 60 pl of 0.2 M sodium citrate, pH 6.0. A portion of these solutions (22.5 pl) was incubated with 13.5 p1 of erythrocyte methyltransferase isozymes and 9 pl of [3H]AdoMet (200 cpm/pmol) at a final concentration of 10 p~. This mixture was incubated at 37 "C for 1 h, quenched with an equal volume of 0.6 M sodium borate, 1% sodium dodecyl sulfate, pH 10.2, and 60 pl was assayed for methyl esters as described under "Experimental Procedures."  FIG. 8. Tryptic fragments of methylated base-treated and untreated glucagons purified by HPLC. Top, glucagon was base treated at a final concentration of 1.25 mg/ml in 0.1 M NHIOH at 37 "C for 3 h (final pH is 10.1). This material was lyophilized, resuspended in 0.2 M sodium citrate, pH 6.0, and methylated with partially purified erythrocyte methyltransferases and [3H]AdoMet at final concentrations of approximately 1.2 mg/ml glucagon, 13.3 pM [3H]AdoMet (9990 cpm/pmol), and 0.16 M sodium citrate, pH 6.0, for 1 h at 37 'C. This material was chromatographed by HPLC to separate [3H]AdoMet from 3H-methylated glucagon. The glucagoncontaining fractions were pooled, lyophilized, resuspended in 0.2 M sodium citrate buffer, and digested with trypsin as described under "Experimental Procedures." The fragments were separated by HPLC as described under "Experimental Procedures," and 100 p1 of each 1ml fraction was counted directly in 10 ml of scintillation fluid. A radioactive peak migrating after T2, an Asp-15-containing peptide, increased by 2-fold; however, we have not been able to identify the composition of this radioactive material. In chymotrypsin digests, peptides containing Asp-15 did not appear to be methyl acceptors with base treatment (data not shown). Bottom, untreated glucagon was methylated, and the 3H-methylated glucagon was purified by HPLC as above. The radiolabeled glucagon-containing fractions were pooled, lyophilized, resuspended in 0.2 M sodium citrate, pH 6.0, and digested with trypsin as described above. The fragments were separated as above, and 100 p1 of each l ml fraction was counted in 10 ml of scintillation fluid. greater impact on peptide migration position than the isomerization or racemization of an aspartyl residue in previous studies (Murray and Clarke, 1984;, we hypothesized that the radioactive peak eluting between T1 and ArgT3' was methylated TI peptide (glucagon 1-12). To test this hypothesis, we purified TI by HPLC from trypsin digests of unmethylated glucagon and methylated this fragment with [3H] AdoMet and erythrocyte protein carboxyl methyltransferase (cf Fig. 3B). As shown in Fig. 4, when the methylated TI fraction is rechromatographed, a single peak of radioactivity (other than [3H]AdoMet) elutes at the same position as the large peak of radioactivity in trypsin digests of 3H-methylated glucagon (cf. Fig. 3A). As a further test, the major 3H-methylated peptide from digested glucagon was found to coelute with the product of enzymatic TI methylation when these preparations were chromatographed together (data not shown).
Based on these results, as well as the specificity of the erythrocyte protein carboxyl methyltransferase for altered aspartyl residues (Clarke, 1985) and the base lability of the methyl groups on TI (96% of the radioactivity purified with 3H-methylated T1 was base labile and volatile), we concluded that the Asp-9 residue in TI was the best candidate for this major methylation site in glucagon. To determine whether this site represented a racemized residue, purified TI methyl ester was analyzed for its maspartate content. This preparation was found to contain 1.0 mol of aspartyl residues/mol of Wrnethyl groups, and 97% of the aspartate was in the Lconfiguration. Because the small amount of D-aspartic acid found could be accounted for by spontaneous racemization of aspartic acid during acid hydrolysis of TI (Manning and Moore, 1968), these data indicated that methylatable TI contains aspartate in the L-configuration. Taken together with the ability of the erythrocyte enzyme to catalyze methylation of L-isoaspartyl sites in a variety of peptides (and its failure to methylate L-aspartate residues) (Clarke, 1985;McFadden and Clarke, 1986) as well as the fact that the bulk of the glucagon preparation, which contains an L -A s~ residue at position 9, cannot be methylated (see "Experimental Procedures''), these results are consistent with the hypothesis that the methylatable site in TI is an L-isoaspartate residue in the Asp-9 position.
To confirm that Asp-9 is the major methyl acceptor site in glucagon, other proteases were used to digest 3H-methylated glucagon in experiments similar to those described above. Chymotrypsin digestion was performed as described in Fig. 5. Amino acid analyses of these peptides were consistent with glucagon chymotryptic peptides (Bromer et al., 1957a), except for two peptides (Car and C3") which could have been produced from glucagon by residual trypsin activity (Figs. 1 and 5, and Table 11). The major peak of radioactivity eluting at 67 min on HPLC, representing about 60% of the total radioactivity, migrated several minutes after Cz (glucagon 7-12) (Fig. 5A).
Since C2 was the likely methylated species, C2 purified by HPLC from digests of unmethylated glucagon was enzymatically methylated with [3H]AdoMet (cf. Fig. 5B). The resulting radiolabeled fragment purified by HPLC was found to migrate at the same position as the large radioactive peak from 3Hmethylated glucagon digests (data not shown). Because C, contains Asp-9 and no other possible methylation sites, this confirms the localization of the major site of methylation to Similar experiments with endoproteinase Arg C and pepsin digests of 3H-methylated glucagon produced radioactive fragments consistent with the results of the trypsin and chymotrypsin digestions (see the Miniprint supplement).
A Second Site of Methylation Originates from the Asn-28 Site-Although it appears that most of the methylation of glucagon occurs at a site derived from Asp-9, a significant amount of 3H-methyl groups, accounting for an average of 23% of the total radioactivity in digests of 3H-methylated glucagon and about 21% of the total methyl acceptor activity in glucagon fragments, has been found in all of the experiments described above. From the evidence described below, this site appears to originate from Asn-28. In tryptic digests, a peak representing about 20% of the radioactivity migrates ahead of the ArgT3 peptide (Fig. 3A). This radioactive peak appears to represent the methyl ester of ArgTa (see below). In chymotrypsin digests of 3H-methylated glucagon, a peak representing 21.5% of the radioactivity eluted in the region between the UV-absorbing peaks of Cl (glucagon 1-6) and C2 (glucagon 7-13) (Fig. 5A). Although from the labeled peak's migration position it might be considered C1 methyl ester, this seemed unlikely because C1 contains no aspartate or asparagine. A likely candidate for the parent of this methyl ester was the tetrapeptide, C4, which could be prone to Lisoaspartyl formation at Asn-28 (Johnson et al., 1985). Since the tryptic peptide A r c 3 contains this site as well, it appeared that in this peptide Asn-28 was the site of methylation and not Asp-21.
In order to confirm that the radioactive peak from trypsin digestion designated as ArgT3 methyl ester did indeed contain a methylation site at the Asn-28 position, the radioactive peak eluting at 52-53 min (Fig. 3A) was isolated and digested further with chymotrypsin. This treatment should result in the formation of C, methyl ester. Upon digestion of the putative ArgT3 methyl ester peptide and HPLC we found a radiolabeled peak migrating between C1 and CP in the position of C4 methyl ester (data not shown). No radiolabel was found in a region expected of C3' methyl ester suggesting that no methylation occurred at Asp-21.
To test methylation at the Asn-28 site more directly, we decided to isolate a peptide containing Asn-28 and show that such a peptide could be a methyl acceptor. Unfortunately, none of the UV-absorbing peaks containing Asn-28 peptides were methylatable. Instead, the methylatable fractions which contained about 20% of the methyl acceptor activity appeared to migrate just ahead of the Asn-28-containing peptide peaks. In methylation of tryptic fragments, a small amount of methyl Asp-9.
acceptor activity migrates just ahead of ArgT3 (Fig. 3B). In methylation of chymotryptic fragments, methyl acceptor activity elutes 10 min ahead of C4 (Fig. 5B). It appears that these fractions, probably containing L-isoaspartic acid or Daspartic acid, migrate ahead of their counterparts containing asparagine. The methylatable chymotryptic fragment eluting ahead of C4 was found to elute in the position expected for C, methyl ester when it was enzymatically 3H methylated and rechromatographed (see above).
To demonstrate that the methylatable C, peptide could be derived from Asn-28, we employed the base treatment technique of Johnson et al. (1985) which appears to produce methylatable sites from asparagines, probably by promoting L-isoaspartyl formation. Thus, we isolated the unmethylatable peptide C4 (glucagon 26-29) from a chymotrypsin digest of glucagon, treated it with 0.1 M ammonium hydroxide at 37 "C, and tested it for methyl acceptor activity. The base-treated C4 was found to be a methyl acceptor whereas untreated C4 was not. Furthermore, when the methylated base-treated C, was rechromatographed with chymotryptic peptide markers, it was found that the methylated species migrated in the same position as the C4 methyl ester previously described.

Identification of the Methylatable Residue Originating from
Asp-9 and Asn-28 as t-Isoaspartate-The results described above are consistent with the assignment of the Asp-9 and Asn-28 sites of methylation with L-isoaspartyl residues. To directly test these assignments, we chemically synthesized authentic standards of peptides corresponding to the chymotryptic peptides C, and C, but with the substitution of an Lisoaspartyl residue for the normal L-aspartyl and L-asparaginyl residues, respectively. These peptides were found to be stoichiometrically methylated (1.2 and 0.79 mol of methyl groups/mol of peptide) by the erythrocyte methyltransferase under the same conditions that the C, and C, peptides derived from glucagon as well as intact glucagon were substoichiometrically methylated (less than 0.01 mol of methyl groups/ mol of glucagon or peptide) (see "Experimental Procedures"). If the methylation sites in 3H-methylated glucagon were at Lisoaspartyl residues, we would expect that the methylated synthetic and natural peptides would be identical. We tested this possibility by HPLC. When the synthetic isoC, peptide was enzymatically methylated and chromatographed with chymotryptic peptides of glucagon the synthetic peptide ester was found to elute at the same time as the C , ester from chymotrypsin-digested 3H-methylated glucagon (data not shown; cf. Fig. 5A). When the synthetic isoC2 ester and C, ester from 3H-methylated glucagon were isolated by HPLC and co-chromatographed on the same HPLC system a single peak of radioactivity was found. Similarly, when the synthetic isoC4 peptide was methylated and chromatographed on HPLC, the peptide ester eluted in the same position as C4 ester from chymotrypsin-digested 3H-methylated glucagon (data not shown; cf. Fig. 5A). Each esterified C4 peptide, either isolated from the methylation of the synthetic isoC4 peptide or digestion of 3H-methylated glucagon, was found to spontaneously produce a radiolabeled degradation product migrating 6 min earlier than the parent ester peak on HPLC. Both the parent ester and the new radiolabeled peak from both the synthetic peptide and 3H-methylated glucagon were found to co-migrate on HPLC.
Site of Methylation in Synthetic Glucagon-All of the studies above were performed with glucagon isolated from bovine and porcine pancreas. To determine whether the isoaspartyl groups at positions 9 and 28 were generated during the biosynthesis, cellular processing, or purification of glucagon, we enzymatically methylated a sample of glucagon, containing the same amino acid sequence, but which had been chemically synthesized (see "Experimental Procedures"). We found that the degree of methylation of this material was similar to that of the natural sample. Tryptic digestion of 3H-methylated synthetic glucagon under the conditions used in Fig. 3 revealed that the bulk of the radioactivity (87%) was associated with the Asp-9-containing fragment; no radioactivity was detected in the fragment containing the Asn-28 site. These results indicate that the formation of an isoaspartyl residue at Asp-9 does not require cellular processes and may reflect an inherent chemical instability of this aspartyl residue. On the other hand, these data suggest that this is not the case for the asparagine 28 residue.

Base Treatment of Glucagon Increases Methylation at Asn-28 but Does Not Affect Methylation at
Asp-9"It has been suggested that base treatment of proteins enhances their methylatability by catalyzing L-isoaspartyl formation from asparagines (Johnson et al., 1985). As shown above, base treatment of C., (L-Leu-L-Met-L-Asn-L-Thr) leads to the production of methylatable peptide. We were interested to know whether this treatment could lead to the production of methylatable sites from aspartyl residues such as Asp-9, the major methyl acceptor in native glucagon. Therefore, glucagon, which was base treated under the conditions described by Johnson et al. (1985) was methylated with [3H]AdoMet and erythrocyte methyltransferases, digested with trypsin, and chromatographed by HPLC. There was no difference between the UV profiles of the base-treated and untreated glucagons, while there was a significant difference between the radiolabel profiles of these two samples (Fig. 8). The major radioactive peak in the base-treated glucagon digest was ArgT3-methyl ester, an Asn-28-containing peptide. The amount of ArgT3 methyl ester per mol of glucagon increased 6-fold over that amount in methylated untreated glucagon. Similarly, an increase in the amount of radiolabel with T3 methyl ester, another Asn-28-containing peptide, is observed. However, no change in the amount of radiolabel a t TI methyl ester (an Asp-9-containing peptide) is seen. Similarly, in chymotrypsin digests of methylated base-treated glucagon a large increase in the amount of Asn-28-containing peptide, C, methyl ester, is observed while no change in the amount of C2 methyl ester (an Asp-9-containing peptide) is seen (data not shown). These data indicate that in whole glucagon, as well as in the C, peptide, base treatment produces methylatable substrates by altering the Asn-28 site. However, these conditions did not enhance methylation at the Asp-9 site.  or Asp-21"Methylation of sites derived from Asp-9 and Asn-28 together accounts for about 85% of the total methylation of glucagon. The remainder of the methylation does not appear to be derived from the other two possible origins of methylatable sites, Asp-15 or Asp-21, since peptides containing these sites do not appear to be methyl acceptors (cf. Figs. 3B and 5B-7B) and peptides containing these sites produced by digestions of enzymatically 3H-methylated glucagon do not appear to be associated with radiolabel (cf. Figs. 3A and 5A-7A). In any case, if methylation occurred at either of these sites they could account for a maximum of 15% of the total methylation of glucagon.

DISCUSSION
Although it is now clear that widely distributed type 11 protein carboxyl methyltransferases catalyze the methyl esterification of unusual D-aspartyl and L-isoaspartyl residues, the origin of these residues in cellular proteins has not been established (Clarke, 1985). Previous work has indicated that base treatment of the peptide hormone ACTH can generate a methylatable L-isoaspartyl residue from the asparagine residue in position 25 (Aswad, 1984a), but the situation for other proteins and peptides, especially those studied under physiological conditions, is not known. The mechanisms of isoas-party1 formation from an asparagine residue have been proposed to be the hydrolysis of a succinimide intermediate formed in an intramolecular attack of the peptide nitrogen atom on the side chain carbonyl carbon. The rate of this reaction is dependent upon the bulk of the side chain of the adjoining residue on the carboxyl side. Maximal succinimide formation is found when there is little or no steric hindrance from this group, such as in an Asn-Gly sequence (Bornstein and Balian, 1977;Aswad, 1984a;Murray and Clarke, 1984;Blodgett et al., 1985;Geiger and Clarke, 1987).
An analogous reaction has been shown to occur with aspartyl residues in model peptides, but at a much slower rate. For example, the rate of isoaspartyl formation from Val-Tyr-Pro-Asp-Gly-Ala is about 38 times slower than that for the corresponding Asn-containing peptide at pH 7.4, 37 "C (Geiger and Clarke, 1987). From these considerations, one might predict that the order of the rates of isoaspartyl formation in glucagon might be Asn-28-Thr-29, followed by Asp-15-Ser-16, and then by Asp-9-Tyr-10 and Asp-21-Phe-22. Although in this work we did find that a small fraction of the methylatable sites was derived from the Asn-28 residue, the major site was unexpectedly found to be at the Asp-9 residue. It does not appear that isoaspartyl residues did form at other sites but were not recognized by the methyltransferase because all peptides tested to date containing L-isoaspartyl residues have been shown to be substrates (Murray and Clarke, 1984;McFadden and Clarke, 1986;Johnson et al., 1987; cf. "Results"), and we do not observe any methylation of glucagon fragments containing Asp-15 or Asp-21 residues (cf. Figs. 3B and 5B). Our results suggest that isoaspartyl residue formation in proteins may not be strictly limited to asparagine-glycine sequences or even to asparagine residues.

Intrinsic Structural Features of Glucagon Appear to Account for L-Isoaspartyl Formation at
Asp-9"Why should isoaspartyl residues occur preferentially at Asp-9 in glucagon? We have considered the possibility that this abnormal residue is generated by an error in the biosynthesis of the hormone precursor (cf. Clarke, 1985) or by chemical modification at this site during the cellular processing of glucagon precursor. However, our observation that Asp-9 is preferentially isomerized in chemically synthesized glucagon as in glucagon isolated from pancreatic tissue indicates that there is an instability inherent in the chemical structure of glucagon that may enhance isoaspartyl formation at Asp-9. Assuming that a succinimide is involved, this formation could be explained by intramolecular interactions that result in an increase in the nucleophilicity of the Tyr-10 peptide nitrogen or an increase in the susceptibility of the Asp-9 carbonyl carbon to attack or an increase in the leaving group potential of the Asp-9 y-hydroxyl group. It is of interest to note that the tyrosine residue at position 10 has an unusually low pK. value and that this region of the polypeptide appears to have multiple interactions in the binding of glucagon to its receptor (Krstenansky et al., 1986).
Our results suggest that the identification of methylation sites may be more complex than we originally anticipated and that other factors such as the interaction of aspartic acid or asparagine residues with neighboring groups brought into proximity with these sites by the three-dimensional structure of the peptide or protein may be crucial in determining which sites wiil ultimately be degraded to isoaspartyl residues and then recognized by the cellular methyltransferases.

Base Treatment Enhances Methylation at Sites That Originate from Asparagine Residues but Does Not Affect Methylation at Aspartate-derived Sites-
In this study we have used glucagon as a model peptide to determine the effects of mild base treatment on the methylatability of sites derived from aspartyl and asparaginyl residues. We have found that the methylatability of the Asn-28 site was increased 6-fold upon base treatment. However, methylation at the Asp-9 site, the major methyl acceptor site in untreated glucagon, was unchanged by base. Asp-15 and Asp-21 did not appear to be affected by base because we found no evidence that either became a methyl acceptor upon base treatment.
These data indicate that base does not appear to catalyze the formation of methylatable residues from aspartyl residues. If isoaspartates are produced from aspartates via succinimide formation, the increase in the nucleophilicity of the peptide bound N-H may be balanced by the deprotonation of the pcarboxyl group of aspartate. Basic conditions may also reduce the capacity of other protonated groups to stabilize the pcarboxyl group of aspartate during succmimide formation. In the base-enhanced formation of methylatable residues from asparagine, the increased nucleophilicity of the peptide bond N-H appears to outweigh any reduction in the leaving group potential of the asparagine -NH2. Therefore, base treatment could be a useful experimental method to distinguish between methylation sites arising from aspartyl and asparaginyl residues. The usefulness of this method has been confirmed in preliminary studies using synthetic asparagine-and aspartic acid-containing peptides.