Human Hypoxanthine-Guanine Phosphoribosyltransferase TRYPTIC PEPTIDES AND POST-TRANSLATIONAL MODIFICATION OF THE ERYTHROCYTE ENZYME*

We describe the isolation and characterization of tryptic peptides of human erythrocyte hypoxanthine-guanine phosphoribosyltransferase. The digest was separated by reverse-phase high pressure liquid chromatography into 30 peaks, 26 of which contained puri- fied peptides. The four complex peaks were resolved by high pressure liquid chromatography with a different reverse-phase column. Each peptide predicted from the recently described amino acid sequence of human erythrocyte hypoxanthine-guanine phosphoribosyltransferase was isolated from this peptide map. Se- quence analysis of the purified peptides identified two peptides, 14 and 14d, that differed by an Asn/Asp het- erogeneity at the position corresponding to residue 106 of the intact protein. Additional studies indicate that this heterogeneity is due to deamidation of the protein in vivo. This post-translational modification appears to be responsible for some of the electrophoretic heterogeneity observed in the normal erythrocyte enzyme.

We describe the isolation and characterization of tryptic peptides of human erythrocyte hypoxanthineguanine phosphoribosyltransferase. The digest was separated by reverse-phase high pressure liquid chromatography into 30 peaks, 26 of which contained purified peptides. The four complex peaks were resolved by high pressure liquid chromatography with a different reverse-phase column. Each peptide predicted from the recently described amino acid sequence of human erythrocyte hypoxanthine-guanine phosphoribosyltransferase was isolated from this peptide map. Sequence analysis of the purified peptides identified two peptides, 14 and 14d, that differed by an Asn/Asp heterogeneity at the position corresponding to residue 106 of the intact protein. Additional studies indicate that this heterogeneity is due to deamidation of the protein in vivo. This post-translational modification appears to be responsible for some of the electrophoretic heterogeneity observed i n the normal erythrocyte enzyme.
Hypoxanthine-guanine phosphoribosyltransferase is a purine salvage enzyme that plays a key role in the regulation of purine metabolism in man. An inherited deficiency of this Xlinked enzyme is associated with two distinct clinical syndromes in man. A virtually complete deficiency of hypoxanthine-guanine phosphoribosyltransferase activity has been described in patients with the Lesch-Nyhan syndrome (l), while a partial deficiency of enzyme activity is found in some male patients who present with hyperuricemia and a severe form of gout (2). In an attempt to understand the molecular basis of these deficiency states, we have focused our recent studies on the structural and functional properties of hypoxanthine-guanine phosphoribosyltransferase from normal and enzyme-deficient patients.
Human hypoxanthine-guanine phosphoribosyltransferase exists, in its native state, as a tetramer of enzyme subunits coded for by a single genetic locus (3). We and others (4-7) have described multiple isoelectric forms of the enzyme subunit in erythrocytes. This heterogeneity appears to be due to a series of undefined post-translational modifications (4,5,7,8). We have recently defined the entire amino acid sequence of hypoxanthine-guanine phosphoribosyltransferase purified from normal human erythrocytes (9). The enzyme has 217 residues with M, = 24,470. The precise structural heterogene-* This work was supported by Grant R01-AM19045 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ities responsible for the isoelectric complexity of the erythrocyte enzyme were not identified in that study.
Insight into the mechanisms responsible for a deficiency of hypoxanthine-guanine phosphoribosyltransferase activity in man was recently provided by the isolation and characterization of structural variants of the human enzyme. Mutant forms of hypoxanthine-guanine phosphoribosyltransferase were purified from lymphoblasts (10) and erythrocytes (6) of five unrelated patients with a deficiency of enzyme activity. Each enzyme variant demonstrated a unique set of structural and functional abnormalities.
As a prelude to future studies of 1) the primary structure of mutant forms of human hypoxanthine-guanine phosphoribosyltransferase and 2) functionally important residues of the normal enzyme (ie. essential active site amino acids), we have developed a method for mapping and isolating tryptic peptides of normal human hypoxanthine-guanine phosphoribosyltransferase. We have used this approach to study the structural basis of the electrophoretic heterogeneity observed in the erythrocyte enzyme.

DISCUSSION
We describe, in this report, the isolation and characterization of the tryptic peptides of human hypoxanthine-guanine phosphoribosyltransferase. Peptides were separated by reverse-phase high pressure liquid chromatography using the trifluoroacetic acid solvent system described by Mahoney and Hermodson (11). This method of peptide mapping was superior in virtually every aspect to the more conventional methods (ie. 2-dimensional chromatography and electrophoresis). Several notable advantages of the reverse-phase high pressure liquid chromatography method include its simplicity, reproducibility, sensitivity (100-200 pmol/map), high recovery of large hydrophobic peptides, and easy sample recovery.
The tryptic peptides of human hypoxanthine-guanine phosphoribosyltransferase were initially separated on a large pore reverse-phase column to assure high recovery and good resolution of the larger more hydrophobic peptides (12, 13). This peptide mapping procedure afforded impressive resolution of the complex trypsin digestion mixture; 26 of the 30 peaks contained purified peptides. The four complex peaks were easily resolved by high pressure liquid chromatography with a different reverse-phase column. A thorough analysis of the peaks recovered from the peptide map identified peptides that span the entire sequence of human hypoxanthine-guanine phosphoribosyltransferase. These results, therefore, corroborate the previously described primary structure of hypoxanthine-guanine phosphoribosyltransferase (9) and validate this approach as a method for mapping the entire hypoxanthineguanine phosphoribosyltransferase molecule. We have recently used this peptide mapping method to identify a single amino acid substitution in a structural variant isolated from a patient with gout.' Several investigators have described tryptic peptide maps of rodent (14,15) and human (16) hypoxanthine-guanine phosphoribosyltransferase that had been labeled in culture with radioactive amino acids. The COOH-terminal tryptic peptide was identified, in each study, as a peptide that incorporated [35S]methionine but not ['Hllysine or [?H]arginine. This is in disagreement with our studies of human erythrocyte hypoxanthine-guanine phosphoribosyltransferase which identitied the COOH-terminal tryptic peptide as Tyr 215-Lys 216-Ala 217. This COOH-terminal amino acid sequence is identical to that predicted from the nucleotide sequence of human hypoxanthine-guanine phosphoribosyltransferase cDNA.~ The discrepancy between our studies and the previous studies of the radiolabeled enzymes therefore cannot be explained by selective proteolysis of the COOH terminus of hypoxanthineguanine phosphoribosyltransferase in erythrocytes. The COOH-terminal peptide identified in the previous studies may in fact be an internal peptide formed by the cleavage of a nonlysine/arginine bond. We describe two anomalous cleavage sites (Leu 64-Cys 65 and Met 94-Thr 95) in human erythrocyte hypoxanthine-guanine phosphoribosyltransferase that result in the formation of peptide fragments that contain methionine but not lysine or arginine residues. Similar aberrant cleavages are possible in the previous studies of the radiolabeled enzyme (14-16) since these proteins were digested with large quantities of tosylphenylalanyl chloromethyl ketone-treated trypsin (25% trypsin by weight) for long periods of time (14-20 h).
Hypoxanthine-guanine phosphoribosyltransferase in erythrocytes exhibits substantial electrophoretic heterogeneity which is caused, in part, by post-translational modification of its primary structure (4,5,7,8). Three major isoelectric forms of the enzyme subunit have been described suggesting at least two post-translational modifications. The most basic erythrocyte isozyme is indistinguishable from the unprocessed enzyme subunit found in cultured cells (4, 5, 7). Johnson et al. (7) have recently localized one post-translational modification in the erythrocyte enzyme to a cysteine-containing tryptic peptide. However, their study was limited to the autoradiographic detection of peptides labeled with [14C}iodoacetamide and, therefore, compared only 4 of 27 peptides. Furthermore, they were unable to identify the structural basis of the charge heterogeneity in this peptide.
In this study, we have analyzed peptides that span the entire sequence of human hypoxanthine-guanine phosphori- bosyltransferase and have detected heterogeneity in only one peptide, a 12-residue cysteine-containing peptide corresponding to peptide 14 in Fig. 3. Amino acid sequence analysis indicated that this heterogeneity is due to partial deamidation of asparagine 106.
Several independent findings support the hypothesis that this Asn/Asp heterogeneity is caused by deamidation of the enzyme in vivo. The hydrolysis of an amide to an acid is consistent with the shift in the isoelectric points of the erythrocyte subunit isozymes to more acidic values. In addition, the partial deamidation of asparagine 106 described in this study is similar in nature and location to the in vivo post-translational modification described by Johnson et al. (7). Both studies localized the structural heterogeneity to a cysteinecontaining tryptic peptide. Furthermore, the modification described by Johnson et al. (7) increased the net negative charge of the peptide at pH 3.5, a finding which is consistent with the conversion of an asparagine to an aspartic acid. Finally, we were unable to detect any artifactual deamidation of asparagine 106 during the enzyme purification, S-pyridylethylation, or trypsin digestion.
In summary, we have described an efficient and sensitive high pressure liquid chromatography method for mapping the tryptic peptides of human hypoxanthine-guanine phosphoribosyltransferase. A thorough analysis of the tryptic peptides isolated by this peptide mapping procedure revealed a site of post-translational modification in the erythrocyte enzyme, partial deamidation of asparagine 106.  Hudng. 1-1.. Rublnflcn. E.. and 1oshidl. A. (19001 J. 010l. Chen. 255. 6438  Thr n e t

3.3
Gl" i y r n e t V~I 6.