THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1988 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 263, No. 9, Issue of March 25, pp. 4430-4433,1988 Printed in U.S.A.

Functional Domains and Methyl Acceptor Sites of the Escherichia coli Ada Protein*

(Received for publication, July 9, 1987)

Barbara SedgwickS, Peter Robins$, Nick TottyQll, and Tomas Lindahl$ From the imperial Cancer Research Fund. iChre Hall Laboratories, South Mimms, Herts EN6 3LD, and the ILincolnS Inn Fields Laboratories, London WC2, United’Kingdom

The ada gene of Escherichia coli encodes a 39-kDa protein which serves both as a transcriptional activa- tor of the adaptive response to alkylating agents and as a DNA repair enzyme demethylating O’-methyl- guanine and phosphotriester residues. Here, the iso- lated Ada protein was found to be readily cleaved into two fragments of similar size by treatment with tryp- sin, chymotrypsin, subtilisin, or VS protease. The frag- ments retained their respective methyltransferase ac- tivities. The Ada protein is, therefore, comprised of two stable active domains united by a central hinge region of about 10 amino acids. Post-translational modification of the Ada protein by methylation of a specific cysteine residue in the NHz-terminal domain is known to convert it to an efficient transcriptional activator. This residue has now been identified as Cys- 69.

Escherichia coli respond to alkylation damage by induction of a DNA repair pathway which increases the cellular resist- ance to the mutagenic and toxic effects of alkylating agents (1). The inducible Ada protein acts both as a DNA repair enzyme and as a transcriptional activator of expression of several genes involved in this adaptive response to alkylating agents, including the ada gene itself as well as alkB, alkA, and aidB (2, 3). The Ada protein repairs one of the two diaster- eoisomers of methylphosphotriesters in DNA by transferring the methyl group on to one of its own cysteine residues (4). This self-methylation converts the Ada protein from a weak to a strong activator of transcription (2). The scavenging cysteine residue within the 354 amino acids of the 39-kDa Ada protein has not been previously identified. It is believed, however, to be located within the NH2-terminal 90 amino acids bzcause such a fragment of the protein remains active in methylphosphotriester repair (5). The Ada protein also accepts a methyl group from the highly mutagenic lesion 06- methylguanine onto a different cysteine known to be residue 321 in the COOH-terminal region of the protein (6).

On cell lysis the Ada protein is readily cleaved at two sites by an endogenous proteolytic activity (7, 8). The 19-kDa COOH-terminal fragment has been purified as an active methyltransferase which repairs @-methylguanine (9). The two cleavage sites have considerable sequence homology (10). A very similar amino acid sequence also occurs in the UvrB protein and is cleaved in cell extracts (10, 11). These obser-

* 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.

n Present address: Ludwig Institute for Cancer Research, London W1, U.K.

vations have indicated that the limited fragmentation of the Ada and UvrB proteins by the endogenous protease is due to the restrictive substrate specificity of this activity, in partic- ular because most E. coli enzymes seem resistant to cleavage.

In this work, we have partially cleaved the Ada protein with several different reagent proteolytic enzymes of broad speci- ficity to investigate whether certain regions of the protein are particularly susceptible. We have also identified the cysteine residue in the NH,-terminal domain which accepts a methyl group from a DNA phosphotriester.

MATERIALS AND METHODS

Enzymes and Substrates-Trypsin (L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated) and chymotrypsin were obtained from Millipore, subtilisin from Sigma, and Staphylococcus aureus V8 pro- tease from Miles Laboratories.

Micrococcus luteus DNA (Sigma) and poly(dT) (Pharmacia LKB Biotechnology Inc.) were methylated by treatment with [3H]N- methyl-N-nitrosourea (29.1 Ci mmol”, obtained from Amersham Corp.). The M. luteus DNA was freed of N-alkylated purines by heat treatment (12) leaving 80% of the radioactivity as 06-methylguanine and 15% as methylphosphotriesters.

The 39-kDa Ada protein and its 19-kDa COOH-terminal fragment, which was generated by cleavage of the intact protein in cell extracts, were purified as described previously (4, 9). A mixture of 20-kDa NH,-terminal and 19-kDa COOH-terminal endogenous fragments of the Ada protein was obtained as an active side fraction during purification of the overexpressed 39-kDa protein by gel filtration (4). The intact Ada protein (3 pg) was labeled at its active sites by incubation for 20 min at 37 “C with methylated poly(dT) (5.4 lo4 cpm) annealed to poly(dA) and then with methylated M. luteus DNA (1.2 lo‘ cpm) for 15 min at 37 ‘C. The activities of the protein in repairing 06-methylguanine and methylphosphotriesters were as- sayed as previously described (4).

Proteolytic Digests-Aliquots of the Ada protein (3 pg each) in 10 pl70 mM Hepesl-KOH (pH 7.8), 1 mM EDTA, 10 mM dithiothreitol, and 5% glycerol were digested with varying amounts of proteolytic enzymes for 20 min at 37 “C. The reactions were stopped by addition of 0.5 mM phenylmethylsulfonyl fluoride and aliquots removed for enzymatic assays. The remaining samples were supplemented with 50 mM Tris-HC1 (pH 6.8), 2% SDS, 2 mM EDTA, 10% glycerol, 5% B mercaptoethanol, and bromphenol blue, heated at 95 “C for 10 min, and resolved by 15% polyacrylamide-SDS gel electrophoresis. The peptides were visualized by silver staining (13). 3H-Labeled peptides were visualized by autoradiography after enhancing with “Amplify” (Amersham Corp.).

IsoEatian of NH,-terminul Active Site Peptide-The purified Ada protein (125 pg) was methylated by incubation with alkylated M. luteus DNA (25 pg, 5 lo5 cpm) in 1 ml of 70 mM Hepes-KOH (pH 7.6), 0.2 mM EDTA, and 10 mM dithiothreitol at 37 “C for 30 min. To remove the substrate DNA, the mixture was subsequently supple- mented with 6 mM MgCl,, 1 mM CaC12, and 4 pg ml” pancreatic deoxyribonuclease I and incubated at 20 “C for 1 h. The digest was dialyzed extensively against 100 mM ammonium bicarbonate (pH

The abbreviations used are: Hepes, 4-(2-hydroxyethyl)-l-pipera- zineethanesulfonic acid; SDS, sodium dodecyl sulfate; HPLC, high pressure liquid chromatography.

4430

Ada Domains 4431

8.0). Ten mM CaC12 and 10 pg of trypsin were then added to the methylated Ada protein (1.8 lo4 cpm), followed by incubation at 20 "C for 24 h. The resulting peptides were fractionated by HPLC on a C8 Aquapore column (220 x 2.1 mm) using a linear gradient of 1-90% acetonitrile in 0.1% trifluoroacetic acid at a flow rate of 0.2 ml min". The effluent was monitored for absorbance at 214 and 280 nm. Fractions (100 pl) were collected and 2 pl removed to determine radioactivity. The major labeled peptide was subjected to total amino acid analysis and to microscale sequencing by automated Edman degradation (14). An aliquot (10%) of this peptide was digested to free amino acids with proteinase K and leucine aminopeptidase and analyzed for radioactive S-methylcysteine (15).

RESULTS

Limited Trypsin Digestion of Ada Protein-Complete diges- tion of the Ada protein with trypsin should yield 46 fragments after cleavage at all susceptible arginine and lysine residues (6). Limited digestion of the protein, however, revealed several particularly sensitive sites (Fig. 1). At the lowest trypsin concentration used, a fragment of approximately 33 kDa was produced (Fig. 1, lune 1 ). On increasing the amount of trypsin added, additional polypeptides of 25, 20, and 19 kDa were observed (Fig. 1, lunes 2 4 ) . The latter two products are similar in size to those generated when the Ada protein is cleaved by the E. coli endogenous protease (Fig. 1, lanes B and C). The tryptic 19-kDa fragment was found to correspond to the COOH-terminal Ada fragment by immunoblotting with a rabbit antiserum raised against the latter peptide (Ref. 7, and data not shown), This tryptic fragment appeared as a double band on silver staining, probably due to cleavage at Ada residue Arg-181 as well as Lys-178. At higher trypsin concentrations the NHp-terminal half of the Ada protein was more readily cleaved into smaller fragments than the COOH- terminal domain (Fig. 1, lanes 5-8). A possible Ada cleavage pattern to yield the polypeptides observed is shown in Fig. 4. Trypsin digestion of Ada protein labeled with 3H at the cysteines of the two active sites yielded labeled fragments of approximately 20, 19, 14, and 8 kDa (Fig. 2). This is in agreement with the proposed digestion pattern. A labeled fragment of 33 kDa was not observed, however, indicating

A 1 2 3 4 5 6 7 8 9 B C M 13

Ada - - 29

N- -20.1 C-

-1 2.4

- 6.5

FIG. 1. Tryptic digests of Ada protein. Ada protein (3 pg aliquots) was digested with varying amounts of trypsin: lane 1, 0.3 ng; lane 2,O.g ng; lane 3, 2 ng; lane 4, 4 ng; lane 5,s ng; lane 6, 20 ng; lane 7,40 ng; lane 8,80 ng; lane 9,80 ng trypsin alone. Lane A, Ada protein alone. Lane B, 20 kDa NH,- and 19 kDa COOH-terminal fragments of endogenously cleaved Ada protein. Lane C, 19-kDa COOH-terminal (C) fragment of endogenously cleaved Ada protein separated from the 20-kDa NH2-terminal ( N ) fragment but contam- inated with two larger polypeptides. Lane M, ovalbumin 45 kDa, carbonic anhydrase 29 kDa, trypsin inhibitor 20.1 kDa, cytochrome c 12.4 kDa, and aprotinin 6.5 kDa. The peptides were resolved on a 15% polyacrylamide-SDS gel and visualized by silver staining.

A 1 2 3 B -"" - 39

- N -C

I - FIG. 2. Tryptic digest of self-methylated Ada protein. Self-

methylated Ada protein (3 pg aliquots) was digested with varying amounts of trypsin. Lane 1 , 3 ng; lane 2,15 ng; lane 3,60 ng. Lane A, methylated Ada protein alone. Lane B, NH,-(N) and COOH-terminal (C) fragments of endogenously cleaved Ada protein. The peptides were fractionated on a 15% polyacrylamide-SDS gel and visualized by autoradiography.

chymotrypsin subtilisin V8 protease " -

A 1 2 3 4 5 A 6 789BA10111213BM

Ada-

N - C-

-45

-29

-20.1

-1 2.4

-6.5

FIG. 3. Digests of Ada protein by chymotrypsin, subtilisin, and V8 protease. Ada protein (3 pg aliquots) was digested with varying amounts of several proteases. Chymotrypsin: lane 1 , 1.2 ng; lane 2,4 ng; lane 3, 12 ng; lane 4,40 ng; lane 5,40 ng chymotrypsin only. Subtilisin: lane 6, 0.8 ng; lane 7, 2 ng; lane 8, 4 ng; lane 9, 4 ng of subtilisin only. V8 protease: lane 10, 0.8 ng; lane 1 1 , 2 ng; lane 12, 4 ng; lane 13, 4 ng of V8 protease alone. Lane A , Ada protein alone. Lane B, NH2- ( N ) and COOH-terminal (C) domains of the endoge- nously cleaved Ada protein. Lane M, markers as in Fig. 1. The peptides were resolved by SDS-polyacrylamide gel electrophoresis and silver stained.

that the methylated protein is less sensitive than the non- methylated protein to proteolytic cleavage in the NH2-termi- nal region in the vicinity of the methyl acceptor site.

The tryptic digests of Ada shown in Fig. 1, lanes 1-5, were fully active in repairing 06-methylguanine and methylphos- photriester residues in alkylated DNA (data not shown). This agrees with the previously observed activity of the 19-kDa

4432 Ada Domains

14kO.+i I 25k0.

Z O k O . - : F ISLO. ~-4 4

,: .I ' C L I 4 k D . : : t L 8kDa-j "_ - - I W l O S CIW".O. mea

T f adogaou. CI.*".P. sile.

FIG. 4. Diagram of the Ada protein showing the active sites, the hinge region, and the proteolytic cleavage sites. The cys- teine residues and the amino acid residues at endogenous proteolytic cleavage sites are labeled and numbered (6, 10). The sizes of tryptic peptides were obtained from Fig. 1 and the approximate cleavage sites indicated by dashed arrows. The most sensitive cleavage sites are indicated by double-headed arrows. The stars indicate labeled active sites in tryptic fragments of the self-methylated protein.

5000

4000

3000

2000

1 000

0

0.100

0 075

A 214

0 050

0.025

3 0 20 40 60 80

Fraction No.

FIG. 5. HPLC elution profile of tryptic peptides of ['HI methyl-labeled Ada protein. After incubation with [3H]alkylated DNA, the methylated Ada protein (125 fig, 1.8 10' cpm) was digested with trypsin and the insoluble material removed by centrifugation. HPLC separation of the soluble peptides was performed as described under "Materials and Methods." The solid l i n e represents radioactive material, and the dashed l ine Azl,.

COOH-terminal fragment isolated after endogenous cleavage of Ada (9). The NHz-terminal fragment generated in cell extracts was also found to be active (Fig. 2, lane B ) . On further tryptic digestion (Fig. 1, lanes 6-8) both methyltrans- ferase activities were reduced. The digests in lane 8 were approximately 20% active in both standard assays. The dis- appearance of the COOH-terminal 19-kDa fragment corre- sponded well with the loss of 06-methylguanine repair activ- ity. The residual ability to repair methylphosphotriesters after further digestion of the 20-kDa NH,-terminal half of Ada indicated that a smaller fragment, possibly the 8-kDa poly- peptide, was active in this reaction. An NH2-terminal frag- ment of 90 amino acids (10 kDa), encoded by a truncated udu gene, has been shown previously to have low but significant activity in methylphosphotriester repair (5).

Sensitivity of Ada Protein to Other Proteases-Limited digestion of the Ada protein with either chymotrypsin or V8 protease yielded two major products between 19 and 20 kDa in size (Fig. 3). Subtilisin treatment produced an intense band of 19 kDa (Fig. 3) which was probably comprised of the two halves of the Ada protein, generated by more than one subtil- isin cleavage in the middle of the Ada molecule. A major cleavage site of all four reagent proteolytic enzymes is, there- fore, close to the center of the Ada protein near the site of endogenous cleavage. The apparent cleavage sites of trypsin, and of V8 protease which cleaves preferentially at glutamic

acid residues (16), suggest an exposed central hinge region of about 10 amino acids (Fig. 4). The Ada protein, therefore, is comprised of two major active domains. The resistance of the COOH-terminal domain to further cleavage by these enzymes indicates that it is stably folded. The NH,-terminal domain is more sensitive to further degradation and can be cleaved into smaller active fragments which may represent subdo- mains.

Methyl Group Acceptor Sites-The Ada protein acceptor site for a methyl group derived from an 06-methylguanine residue in DNA was shown previously to be Cys-321, the most COOH-terminal of the 12 cysteine residues in the protein (6, see Fig. 4). In order to define the acceptor residue for the more recently discovered methylphosphotriester transferase activity, the Ada protein was labeled by incubation with [3H] methylnitrosourea-treated DNA. Sequencing of the first 20 residues of the labeled Ada protein confirmed that its NH,- terminal sequence is Met-Lys-Lys-Ala-Thr-Cys-Leu-etc. (8). No radioactive material was released in the sixth sequencing cycle, which ruled out residue Cys-6 as the acceptor. The labeled Ada protein was extensively digested with trypsin. This proteolysis generates a hydrophobic, insoluble tryptic peptide containing the acceptor site for 06-methylguanine (6). A soluble radioactive peptide would be expected to contain the acceptor site for methyl groups derived from phosphotries- ters. HPLC chromatography identified a single major peak of radioactive material (Fig. 5). The amino acid composition of the peak fraction strongly indicated that it represented the tryptic peptide corresponding to Ada residues 50-70. The radioactive residue was chromatographically identified as S- methylcysteine as described previously (15). Edman degra- dation of the peptide yielded a single major sequence corre- sponding to Ada residues 50-70. Since there is only one cysteine residue in this peptide (6), the methyl group acceptor site for phosphotriesters in DNA can be identified as residue Cys-69. This tryptic peptide, one of the four largest from the Ada protein, contains an -Arg-Pro- sequence, which is not cleaved efficiently by trypsin.

DISCUSSION

Several different proteolytic enzymes cleave the Ada pro- tein preferentially within a short region of about 10 amino acids at the center of its sequence. As a result, two fragments of similar size are produced, each of which retains a separate methyltransferase activity for alkylated DNA. We conclude that the Ada protein is comprised of two separate functional domains, united by a protease-sensitive hinge region. Exam- ples of other enzymes and regulatory proteins which contain separate domains of defined functions are aminoacyl-tRNA synthetases (17, 18) the cyclic AMP receptor protein (19), and X phage repressor (20). Inactivation of the X repressor in vivo occurs by cleavage within its "hinge region." Cleavage of the Ada protein by a highly sequence-specific endogenous activity also occurs within the central protease-sensitive re- gion and could be of physiological importance in down-regu- lating the adaptive response to alkylating agents (21).

The active methyl group acceptor site in the COOH-ter- minal domain of the Ada protein was identified in our earlier work as Cys-321 (6). The second acceptor site in the NH2- terminal domain, which is of critical importance for the positive regulatory function of the protein (Z), is now shown to be Cys-69. The amino acid sequences surrounding these two methyl scavenging sites show considerable homology. The sequences 68-71 and 320-323 of the Ada protein are -Pro- Cys-Lys-Arg- and -Pro-Cys-His-Arg-, respectively, in both cases preceded by a hydrophobic sequence. A mechanistic

Ada Domains 4433

model proposed for the 06-methylguanine acceptor site (6) is therefore similarly applicable to the second Ada acceptor site for methylphosphotriester residues.

Proteins comprised of domains may have evolved either by gene duplication or by fusion of two different DNA sequences. In the case of the Ada protein, we favor the second possibility because no significant sequence homology seems to occur between the two domains outside the acceptor regions. Mam- malian cells contain a constitutively expressed 06-methyl- guanine-DNA methyltransferase of 24 kDa with a cysteine residue at its active site but no methyl phosphotriester repair capacity (22, 23). The mammalian enzyme may therefore be similar to the COOH-terminal domain of the Ada protein.

The Ada protein binds specifically to the DNA sequence 5’-AAAGCGCA-3’, which occurs in the promoter region of genes under ada gene control (2,21). Methylation of the Ada protein at Cys-69 is required for efficient binding. The 20- kDa methylated NH2-terminal fragment of the protein also binds specifically to this sequence, as judged from DNase1 “footprinting” experiments (24). However, the entire protein (or a large NH,-terminal fragment containing 88% of the total sequence) is required for transcriptional activation (5, 25, 26). Demple (27) investigated the properties of several mutant Ada proteins and suggested that the COOH-terminal domain might be required for direct interaction with RNA polymerase.

A more detailed understanding of the mechanisms of methyl group transfer and transcriptional activation by the Ada protein will now require physicochemical analyses, including x-ray crystallography and NMR. Isolation and crystallization of the functional domains of the Ada protein as identified in the present work should facilitate such further studies.

Acknowledgments-We thank Dr. T. McCarthy, Dr. P. Karran, Dr. M. Waterfield, and F. Fitzjohn for their help in the early part of this work and R. Philp for assistance with protein sequencing.

REFERENCES 1. Samson, L., and Cairns, J. (1977) Nature 267, 281-283 2. Teo, I., Sedgwick, B., Kilpatrick, M. W., McCarthy, T. V., and

3. Nakabeppu, Y., and Sekiguchi, M. (1986) Proc. Natl. Acad. Sci. Lindahl, T. (1986) Cell 45, 315-324

U. S. A. 83,6297-6301

4. McCarthy, T. V., and Lindahl, T. (1985) Nucleic Acids Res. 13,

5. Kataoka, H., Hall, J., and Karran, P. (1986) EMBO J. 5, 3195- 3200

6. Demple, B., Sedgwick, B., Robins, P., Totty, N., Waterfield, M. D., and Lindahl, T. (1985) Proc. Natl. Acad. Sci. U. S. A. 82,

7. Teo, I., Sedgwick, B., Demple, B., Li, B., and Lindahl, T. (1984)

8. Nakabeppu, Y., Kondo, H., Kawabata, S., Iwanga, S., and Seki-

9. Demple, B., Jacobson, A., Olsson, M., Robins, P., and Lindahl,

2683-2698

2688-2692

EMBO J. 3,2151-2157

guchi, M. (1985) J. Biol. Chem. 260 , 7281-7288

T. (1982) J. Bwl. Chem. 257,13776-13780 10. Teo, I. A. (1987) Mutat. Res. 183, 123-127 11. Arikan, E., Kulkani, M. S., Thomas, D. C., and Sancar, A. (1986)

Nucleic Acids Res. 14,2637-2650 12. Karran, P., Lindahl, T., and Griffin, B. (1979) Nature 280 , 76-

77 13. Nielsen, B. L., and Brown, L. R. (1984) Anal. Biochem. 141,

311-315 14. Downward, J., Yarden, Y. , Mayes, E., Scrace, G., Totty, N.,

Stockwell, P., Ullrich, A., Schlessinger, J., and Waterfield, M. D. (1984) Nature 307 , 521-527

15. Olsson, M., and Lindahl, T. (1980) J. Biol. Chem. 255, 10569- 10571

16. Houmard, J., and Drapeau, G. R. (1972) Proc. Natl. Acad. Sci. U.

17. Waye, M. Y., Winter, G., Wilkinson, A. J., and Fersht, A. R.

18. Kohda, D., Yokoyama, S., and Miyazawa, T. (1987) J. Biol. Chem.

19. Angulo, J., and Krakow, J. S. (1986) J. Biol. Chem. 261 , 11315-

20. Pabo, C. O., Sauer, R. T., Sturtevant, J. M., and Ptashne, M.

21. Lindahl, T., Sedgwick, B., Sekiguchi, M., and Nakabeppu, Y.

22. Harris, A. L., Karran, P., and Lindahl, T. (1983) Cancer Res. 43,

23. Myrnes, B., Nilsen, I. W., Haugen, A., and Krokan, H. (1986) in Repair of DNA Lesions Introduced by N-Nitroso Compounds (Myrnes, B., and Krokan, H., eds) pp. 112-134, Norwegian University Press, Oslo

24. Sedgwick, B., and Hughes, S. (1988) in Mechanisms and Conse- quences of DNA Damage Processing (Friedberg, E., and Hana- Walt, P., eds) Alan R. Liss, New York, in press

S. A. 69,3506-3509

(1983) EMBO J. 2 , 1827-1829

262,558-563

11319

(1979) Proc. Natl. Acad. Sci. U. S. A. 76, 1608-1612

(1988) Annu. Reu. Biochem. 57,133-157

3247-3252

25. Sedgwick, B. (1983) Mol. Gen. Genet. 191,466-472 26. LeMotte, P. K., and G. C. Walker (1985) J. Bacteriol. 161,888-

27. Demple, B. (1986) Nucleic Acids Res. 14,5575-5589 895