Isoaspartyl Protein Methyltransferase from Human Erythrocytes

Vol. 264,No. 33,Issue of’ November 25, pp , 20131-20139,1989 Printed in U.S.A.

THEJOURNALOF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc.

Sequence of the D-Aspartyl/r,-Isoaspartyl Protein Methyltransferase from HumanErythrocytes COMMON SEQUENCE MOTIFS FOR PROTEIN, DNA, RNA, AND SMALL MOLECULE S-ADENOSYLMETHIONINE-DEPENDENT METHYLTRANSFERASES* (Received for publication, June 26, 1989)

Diego Ingrosso$Q**, Audree V. FowlerQll, Janice BleibaumQll, and StevenClarkeSQII From the $Departmentof Chemistry and Biochemistry, the 7lDepartmentof Biological Chemistry, and the §Molecular Biology Institute, University of California, Los Angeles, California 90024-1569

A widely distributed protein methyltransferase cat- found in both procaryotes and eucaryotes, methylates abnoralyzes the transfer of a methyl group from S-adenosyl-mal D-aSpartyl and L-isoaspartyl residues present in proteins methionine to the free carboxyl groups of D-aspartyl that have undergone spontaneous degradation reactions and/or L-isoaspartyl derivatives of L-aspartyl and L- (Clarke, 1985, 1988 Aswad et al., 1988; Galletti et al., 1988b). asparaginyl residues. This enzyme has been postulated Human erythrocytes are one of the best characterized systo function in the repair or the catabolism of age- tems for understanding the role of this class of methyltransdamaged proteins.We present here the complete amino ferases in intact cells (Lowenson and Clarke, 1988). TWO acid sequence of the more basic isozyme I of this en- isozymes of the human erythrocyte D-aspartyl/L-isoaspartyl zyme from human erythrocytes. The sequence was de-methyltransferase have been purified to homogeneity (Gilbert termined by Edmandegradationandmassspectral et al., 1988). These isozymes are monomeric polypeptides of analysis of overlapping trypsin, Staphylococcus auabout 25,000 Da and have similar catalytic properties, but r e m V8 protease, Pseudomonas fragi endoproteinase differ by about 1pH unit inisoelectric point (Ota et al., 1988). Asp-N, cyanogen bromide, and hydroxylamine-gener- Several membrane (Freitag and Clarke, 1981) and cytosolic is modified by acetated fragments. The NHz-terminus 226 amino acids fora (O’Connor and Clarke, 1984) proteins have been found to be ylation and the protein contains calculated molecular weight of 24,575. This value is the methyl-accepting substrates for this enzyme inintact in good agreement with the molecular weight deter- cells, and a significant increase in theirmethylation has been mined for the purified protein by polyacrylamidegel shown to occur during cell aging (Barber and Clarke, 1983; electrophoresis in the presenceof dodecyl sulfate and Galletti et al., 1983). D-Aspartic acid [3H]methyl ester has by gelfiltration chromatography under nondenaturing been identified as a product of this enzyme in proteolytic conditions. The identificationof 2 different amino acid digests of both membrane and cytosolic proteins from intact labeled with [ methyl-3H]methionine residues at both positions 22 and 119 may indicate the humanerythrocytes presence of allelic variants or of two or more closely (McFadden and Clarke, 1982; O’Connor and Clarke, 1984; related structural genes. Finally, comparison of this Lou and Clarke, 1987). From these results, it was proposed sequence with those of methyltransferases for RNA, that the methyltransferase functions as part of a mechanism DNA, and small molecules, as well as other S-adeno- capable of repairing altered aspartyl residues (McFadden and sylmethionine-utilizing enzymes, shows that many of Clarke, 1982).In vitro evidence to support a repairmechanism these proteins share elements of three regions of se- has been obtained recently (McFadden and Clarke, 1987; quence similarity and may be structurally or evolu- Johnson et al., 1987a, 1987b; Galletti et al., 1988a). tionarily related. An analysis of tryptic fragments accounting for about 50% of the sequence of the two isozymes of the human erythrocyte methyltransferase and one isozyme of the bovine erythrocyte Protein carboxyl methyltransferases catalyze the transfer enzyme revealed that these enzymes are very similar to each of a methyl group from S-adenosylmethionine (AdoMet)’ to other (Gilbert et al., 1988). Wenow present the complete an acceptor protein or peptide. Three classes of such enzymes sequence of the 226 amino acid residues of the human erythhave been found to date. The most widely distributed class, rocyte isozyme I. This sequence is similar in three regions to those of methyltransferases that can act on widely different * This work was supported by Grant GM-26020 from the National substrates, and thissuggests possible common structural features and/or a common evolutionary origin for these S-adeInstitute of Health and by Grant DMB-8602102 from the National Science Foundation. The costs of publication of this article were nosylmethionine-dependent enzymes. 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. The aminoacid sequence(s) reported in this paperhas been submitted to the GenBankTM/EMBL Data Bank with accession number(s) 505115. ** During the course of this study on leave of absence from the Department of Biochemistry of Macromolecules, 1st Medical School, IJniversity of Naples, Naples, Italy. 11 To whom correspondence should be addressed. The abbreviations used are: AdoMet, S-adenosylmethionine; AdoHcy, S-adenosylhomocysteine; HPLC, high performance liquid chromatography.

EXPERIMENTAL PROCEDURES*

RESULTS

General Strategy forthe Sequence Determination-Isozyme I of the D-aspartyl/L-isoaspartyl methyltransferase was puriPortions of this paper(including“ExperimentalProcedures,” Tables I-VIII, and Figs. 2,4, and 5) 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 included in the microfilm edition of the Journal thatis available from Waverly Press.

20131

D-Aspartyllt-Isoaspartyl Protein Methyltransferase Structure

20132

fied from human erythrocytesand subjected to fragmentation byStaphylococcus aureus V8 protease, Pseudomonas fragi endoproteinaseAsp-N, cyanogen bromide, and hydroxylamine as described under “Experimental Procedures.” Reduced and 14C-alkylatedenzyme, prepared as described under “Experimental Procedures,” was used to identify cysteinecontaining peptides. Fragments were purified by HPLC and characterized by a combination of amino acid analysis, automated Edman sequencing, and fast atom bombardment mass spectral analysis as described under “Experimental Procedures.” The final sequence was obtained by overlapping peptide segments and is summarized in Fig. 1. Alignment of Overlapping Peptides-The starting point of our studywas the analysis by Edman degradation of the eight tryptic peptides previously described (Gilbert et al., 1988; Table I), followedby that of nineendoproteinase Asp-N peptides (Figs. 2 and 3, Tables I1 and 111), six V8 protease peptides (Fig. 4,Tables IV and V), and three cyanogen bromide peptides (Tables VI and VII, Fig. 1).The sequences of these peptides could be overlapped straightforwardly to give a sequence corresponding to residues 30-226 in the final sequence. At position 119, we obtained evidence for two different residues in three different preparationsof the meth10

20

yltransferase. For example, peptide T12 gave valine at this position (Table I; Gilbert et al., 1988), while isoleucine was found at thisposition in peptide D8 (Table 111) and in peptide CB5 (Table VII) (see “Discussion”). Determination of the NHz-terminal Structure-Both isozymes of the intact D-aspartyl/L-isoaspartylmethyltransferase were not susceptible to Edman degradation, suggesting the presence of a blocked NHa-terminal residue (Gilbert et al., 1988). Amino acid analysis of a blocked tryptic peptide (T3) found inour previous study showed stoichiometric amounts of alanine and lysine; the presence of a tryptophan was surmised from its UV absorbance (Gilbert et al., 1988). We utilized the specific absorption of tryptophan at 288 nm to obtain alarger, overlapping fragment of this putative NHzterminal peptide. Fig. 2 shows the chromatogram from the HPLC separation of the endoproteinase Asp-N fragments. Two major peptides were found that absorbed at 288 nm, Dl’ and D15. Peptide Dl’ was blocked to Edman degradation and, in addition to theAla, Lys, and Trpresidues seen in T3, contained another Ala, 2 Gly, 3 Ser, and1 His residue (Table 11). This peptide was analyzed by fast atom bombardment mass spectrometry and revealed a parent ion with mass (M+H)’ of 1029.48, corresponding to an acetylated peptide 30

I

50

40

130 110 v120 150 Q T G K V I G I D H I K E L V D D S I N N V R K D D P T L L S S G R V Q L V V G D G R M G Y A E E A T13 v5 ’ I T4 I D8 I T5 I D10”-

0

-.-

I

D9 D6.7 CB5 170

220

I v7-

D10’

I

I

CB6

160 200 180 190 P Y D A I H V G A A A P V V P Q A L I D Q L K P G G R L I L P V G P A G G N Q M L E Q Y D K L Q D G .-I T13 . I T10 D10”-I v7 . . I V8 CB6 I Dl 4210 S I K M K P L M G V I Y V P L T D K E K Q W S R W K T8 I V8I Dl 5 I Dl4 . . I v9 I ” -

FIG. 1. Sequence of the D-aspartyl/L-isoaspartylmethyltransferase isozyme I from human erythrocytes. The alignments of the peptide fragments used to determine the sequence are shown. Peptides from trypsin ( T ) ,S. aureus V8 protease (VI, P. frugi endoproteinase Asp-N (D), and cyanogen bromide ( C B ) fragmentation are listed in order of their relative sequence position exceptfor the trypticpeptides which are given the designations assigned in Gilbert et al., 1988. C-NHZOH is the COOH-terminal fragment from the hydroxylamine cleavage. Peptides were sequenced by automated Edmandegradation, except Dl’, whose sequence was determined by tandem mass spectrometry. Both isoleucine and leucine have been detected at position 22 and both isoleucine and valine were detected at position 119. The presence of a dot in a sequence indicates that anunambiguous identification of an amino acid residue could not be made in this cycle. Vertical bars indicate the last residue experimentally identified in a given peptide. Ac, N-acetyl group.

D-Aspartyl/L-Isoaspartyl Methyltransferase Protein Structure .

"

(M+H)+

20133

(Ile/Leu)-Lys-Thr-Asp-Lys-Val-Phe-Glu-Val-Met-Leu-Ala

(Fig. 1).The amino-terminal sequence of this fragment, representing residues 20-226, can now be used to link peptides V2, V3, and CB2.3 and establish thecomplete structure. The NHz-terminal sequence of the C-NH,OH fragment was further confirmed by direct Edman sequence analysis of this large fragment blotted ontopolyvinylidene difluoride after gel electrophoresis by the procedure of Matsudaira (1987). Both sequence analyses revealed the presence of both isoleucine and leucine in the third position, corresponding to position 22 in the final structure. Sequence Comparison of the D-Aspartyl/Isoaspartyl Methyltransferme with Other Proteins-The sequence of the methm/z yltransferaseshownin Fig. 1 was comparedwiththeseFIG. 3. Collision-induced decomposition mass spectrum of quences of 6843 proteins in Release 15 (28 December 1987) the (M+H)+ ion of the D l ' NHz-terminal peptide. The D l ' peptide from the endoproteinase Asp/N cleavage of D-Asp/L-isoAsp of theNational Biomedical ResearchFoundation/Protein Identification Resource bank (NBRF/PIR, Washington, D. methyltransferase was obtained as detailedunder"Experimental Procedures." Tandem mass spectrometry of the m/z (1029.48) parent C.; George et al., 1986) using the FASTP programof Lipman ion was performed by Dr. Stephen Martin at the Mass Spectrometry and Pearson (1985). Thebestmatch(optimized score 84; Facility at the Massachusetts Institute of Technology. Peaks corre- mean score 26.5 f 7.0 standard deviation) was obtained with sponding to fragment ions are designated according to Biemann and the sequence of scallop myosin EDTA light chain, representScoble (1987) and Johnsonet al. (1988).The sequence can be deduced from the difference between two consecutive fragment ions of the ing 19.7% sequence identity over 127 residues. We also comD-aspartyl/L-isoaspartyl methyltransferasesesame series: z, t; y,c----; x,