- Email: [email protected]
Functional sites of the Ada regulatory protein of Escherichia coli
J. Mol. Hiol. (198X) 201. 661-271
Functional
Sites of the Ada Regulatory of Escherichia coli
Protein
Analysis by Amino Acid Substitutions Kyoko
Takano,
Yusaku
Nakabeppu
and Mutsuo
Sekiguchi
Department of Biochemistry Faculty of Medicine Kyushu University 60 Fukuoka 812, Japan (Received 10 September 1987) Specific cysteine residues at possible methyl acceptor sites of the Ada protein of Escherichia coli were converted to other amino acids by site-directed mutagenesis of the cloned ada gene of E. coli. Ada protein with the cysteine residue at 321 replaced by alanine was capable of accepting the methyl group from the methylphosphotriester but not from 06methylguanine or 04-methylthymine of alkylated DNA, whereas the protein with alanine at position 69 accepted the methyl group from the methylated bases but not from the methylphosphotriester. These two mutants were used to elucidate the biological significance of repair of the two types of alkylation lesions. Introduction of the ada gene with the Ala69 mutation into an ada- cell rendered the cell more resistant to alkylating agents with respect to both killing and induction of mutations, but the gene with the Ala321 mutation exhibited no such activity. Replacement’ of the cysteine residue at position 69, but not at position 321, abolished the ability of Ada protein to promote transcription of both ada and alkA genes in vitro. These results are compatible with the idea that methylation of the cysteine residue at position 69 renders Ada protein active as a transcriptional regulator, whilst the cysteine residue at position 321 is responsible for repair of pre-mutagenic and lethal lesions in DNA. The actions of mutant Ada proteins on the ada and alkA promoters in viva were investigated using an artificially composed gene expression system. When the ada gene wit,h the Ala69 mutation was introduced into the cell, there was little induction of expression of with an alkylating agent, in either the ada or the alkA genes. even after treatment agreement. with the data obtained from st,udies in vitro. With the Ala321 mutation, however, a considerable degree of ada gene expression occurred without adaptive treatment. The latter finding suggests that the cysteine residue at position 321, which is located near the C terminus of the Ada protein, IS involved in regulating activity, as the transcriptional activator.
The Ada protein, with a molecular weight of 39,400, possesses two methyltransferase activities. It’ transfers a methyl group from the 06-methylguanine residue of alkylated DNA to a specific cysteine residue close to the C terminus of its own molecule (Demple et al., 1985). It also cat,alyzes the transfer of a methyl group from one of two stereoisomers of methylphosphotriesters of alkylated DNA to a cysteine residue in the N-terminal half of the protein (McCarthy & Lindahl, 1985; Margison et al., 1985). The Ada protein is cleaved by cellular protease to two fragments, each carrying one of the methyltransferase activities
1. Introduction Enzymes that repair alkylated DNA are formed in Escherichia co& cells on exposure to low concentrations of alkylating agents (Karran et al., 1982; Evensen & Seeberg, 1982), and this induced synthesis is controlled by the ada gene (Jeggo, 1979; Sedgwick, 1983; Teo et al., 1984; Lemotte & Walker, 1985; Nakabeppu et al., 1985a,b). Recent studies have revealed that the Ada protein, coded by the ada gene, functions as a DNA repair enzyme as well as a transcriptional regulator (for a review, see Sekiguchi & Nakabeppu, 1987). 261
K. Takano
(Nakabeppu et al., 1985a: Teo, 1987; Yoshikai et al., unpublished results). Expression of the ada and the related genes, such as alkA, belonging to the ada regulon is positively controlled by the Ada protein. Large amounts of ada and alkrl-specific RNAs are formed in cells treated with a methylating agent. whereas lit’tle such RiVA is produced in untreated cells. The in&vo transcription initiation sites for the two genes were determined by primer-extension cDNA synthesis and S, nuclease mapping (Nakabeppu et al.. 19&a). In in-vitro reconstituted systems, bot,h ada and alkA transcripts were formed in an Ada protein-dependent manner (Teo et al., 1986; Nakabeppu & Sekiguchi. 1986). It was shown that the methylated form of Ada protein is more effective in promoting ada t,ranscription than is the unmethylated form, whilst the effects of both forms are much the same with regard t’o aZkA transcription. The DNA sequencesrequired for the regulat,ed expression of the ada and alkA genes have been determined by introducing various mutations in the promoter regions (Xakamura et 01.. 1988). A model for the molecular mechanism of a,daptive response was designed on the basis of these findings. We constructed altered forms of Ada protein. whose specific amino acid residues at the possible methyl acceptor sites were replaced by other amino acids. by the site-directed mutagenesis of the cloned ada gene. This facilitated an examination of the effects of specific methylation on the activities of Ada protein, and revealed the biological consequencesof the two types of methylation of Ada protein.
2. Materials (a)
Chemicala
and Methods and
et al.
E. coli st’rain YN;18O j A(ada-alkB) : krrtr] was (aon strutted as follows. A 1.3 kb Km’ fragment derived front
pUC4K (Vieira & Messing, 1982) was 2.5 kb .V/uI
Jll~rT fragment
cxjntaining
substituted for a. 1)ar’ts of tht, c/d/c
and a&B genesin pYpu‘3028(Xakabeppu rt ccl.. I985r’i). Strain YY22 (poL4’“) cells (Yamamoto Pu Hekiguchi. 1979) were transformed Cm’ transformants DNA is replaced
with the resulting plasmid and Km’ were selected at, 30°C‘. If the c+mecl wit,h t)hr corresponding region of thr
chromosomeby double crossing-over,Km’ selected at the restrictive vclc*tor plasmid ])A(‘I’(‘184
CInP
cells
cat,
temperature. 42’C’. since the ((‘hang R (‘oh~n. l!fi8) is
be
unableto replicate without I)NA polymrrasr I. Such culls were infected with 1’1 phage and the A(a,dn-alkR) region was transduced into thr pol=l ’ strain.
: ku,c YN3
(Nakabeppuet al.. 19856).The resultingstrain wasnatnptl YN180. Strain KTIO (dam . F + 1 was c~onstruc*ted by in1 1.0 ducing F plasmid, pWH373. derived from WB373 (Barnes et al.. 1983) into GM48 (da& ( F-) (Marinus. 1973).
Strain KTIO. WE373 (Barnes
cl
al.,
1983)
and wtarf’
vwtor
used
phagr
for
~WWWL’
sit,?-directed
mutagenesisand for sequencing. As the parental plasmid t,o be mutated. pYS:W:i!) carrying the ada gene with intact rrdcr promotel (Nakabeppu ut al., 1985a) was used. Since introduc*t,ion of the Ala321 mutant gene into t’his plasmid resulW in ;I poor growth of cells. probably because of ovrrI)roduc,tiolI of the Ada protein. we deleted a part of t,he ada promotjer (a 57 bp Hl;ndIII-HstETT fragment) to eliminate the effect, of mutant Ada protein on its own promot,rr. The resulting plasmid was named pYN3089. in which the ndo gene wa,s transcribed only from the lac promoter existing in the v&or plasmid, pUC9 (Vieira & Messing. 198%). To investigate the effect of Ada protein on the rxd~~ promot#er in PEWS. plastnid pKT410 was c~onstructrtl, A 176bp HindTTI-~k=coRlfragment containing the trrlrl promoter. derived from pYN3059. was inserted into the C’lnI site upstrearn from the chloramphenic~ol acrtyltrxnsferase gene in the plasmid pCM7 (Close & Rodriguez. 1982).
enzymes
Restriction enzymes, bacteriophageT4 polynucleotide kinase. DNase I, bacterial alkaline phosphatase, DNA sequencing kits. and DNA ligation kit,s were obtained f’rom Takara Shuzo (lo.. Lttl (Kyot)o). Toyobo (‘0.. Ltd (Osaka). and Nippon (Gene (‘0.. Ltd (Toyama). E. co/i RNA polymerase holoenzyme, phosphodiesterasr 1 and II, bovine serum albumin. poly[d(A.T)]. poly(dA), poly(dT) and calf thymus DNA were purchased from Pharmacia. Micrococcus Zutec~s DN-A. heparin. rifampicin. chloramphenicol, acetyl coenzyme A and 5.5’-dithiobis-2nitrobenzoic acid were obtained from Sigma. Ribosomal RNA was from Hoehringer-Mannheim Riochem. /cI-~~PJ~TTP(>ciOOC”i/mmol). [Y-~~I’JATP (> 7000(‘ii mmol). [a-32PIdCTP ( > 400 (‘i/mmol), and [n~othyZ-3H 1 MSl’t (0.5 to 2 (‘i/mmol) were purchased from N-e\+ England Nuclear and Amrrsham. MNNG and MMS were purcahasrd from Nakarai (‘hrmicals (‘0.. Ltd (Tokyo).
t Abbreviations used: LXNTJ, S-methylLLV-nitrosourea: MNNG. N-methyl-N’-nitro-N-nitrosoguanidine; MMS, methyl methanesulfonate; kb, lo3 bases; Km. kanamycin; Cm, chloramphenicol: bp. base-pair(s): ssDNA, single-stranded DNA.
Oligon~rcleotides for site-direct,ed mutagenesis ~‘rre chemically synthesized by an Applied Hiosyst)ems I)NA synthesizer 381A. The strategy of site-specific. mutsgenesis was according t,o Mark PI al. ( 1984). Th~x mutations were verified to he introduced at the designed sites by srquencc~ analysis (Sanger rt ol 1!)77) of the rnutant grnrs.
Methylated DNA was prepared as tltascribe
E. coli Ada Regulatory iMutant Ada proteins were purified from YN180 cells harboring plasmids derived from pYN3089, as described (Nakabeppu et al., 198%) with some modifications. Newly transformed cells were cultured in 1 to 4 1 of LB broth containing (50pg/ml each) ampicillin and kanamycin at 30”(” to ‘466o = 1-O. Crlls were collected and sonirated in buffer A (20 mM-Tris. HCl (pH 8.5), 1 mM-EDTA, 1 mm2-mercaptoethanol. 5% (w/v) glycerol) containing NaCl at 0.3 M. After treatment with 0.57& Polymin P to remove the nucleic acids, the sample was subjected to precipitation with ammonium sulfate. applied to a phosphocellulose P-11 column. and then eluted with a linear gradient from 0.15 to 0.5 M-NaCl in buffer A. Peak fractions were concentrated and applied to a Bio Gel P-60 column (1.5 cm x 100 cm). The eluted fractions were collected and dialyzed against’ a 1 : 1 (v/v) mixture of glycerol and buffer A containing 0.2 M-ammonium sulfate overnight at 4°C. (e) Nurvival und mutation frequency
Protein
263
graph equipped with a Waters ~BOSDAPAK column. General methods for DNA manipulation according to Maniatis et al. (1982).
Cle were
3. Results (a) Replacement of cysteine residues 321 of Ada protein by alar&P
6.9 and
Direct sequenceanalysis of the [J-terminal half of the methylated Ada protein revealed that cysteine 321, which is located between a proline and a histidine residue, is the acceptor site for a methyl group from 06-methylguanine of alkylated DNA (Demple et al., 1985). Based on the amino acid sequence similarity, it was proposed that cysteine 69, flanked by proline and lysine, may be the acceptor site for a methyl group from the methylphosphotriester of the DNA (Teo et al., 1986). We constructed mutant forms of Ada protein whose
(‘ells harboring various plasmids were grown at 30°C to .Abh,, =(bfi to 0.X. Samples (5 ml) of the rult,ures were centrifuged. and the cells were resuspended in 5 ml of 1M9 salt medium (pH 6.0). A portion (1 ml) of the suspension was added to 1 ml of M9 salt medium containing MNNG. After incubation at 37°C for 10 min, 2 ml of LB broth was added and the samples were plated on LB plates after appropriate dilution. The number of surviving cells was counted after incubation at 30°C overnight. To determine the mutat’ion frequency. MNNG-treated cells wpre placed on an LB plate containing rifampicin at lOO~g/ml. After incubation at 30°C for 2 days, the number of rifampicin-resistant cells was scored.
cysteine
residues
at
the
plausible
methyl
acceptor sites are replaced by other amino acids. To introduce the mutations into the cloned ada gene, we applied site-directed mutagenesis using synthetic oligonucleotides (Fig. 1). Ml3 ssDNA carrying a part of the ada gene was annealed with the synthetic primer, and the complementary &and was synthesized by the DNA polymerase I reaction. After transfection of E. coli WB373, the mutant phages were selected by hybridization, using the same oligonucleotides as probes. The cloned DNAs carrying the mutant’ ada gene were introduced into plasmid pYN3089, in which the gene, without its promoter region, can be transcribed under the control of the lac promoter. The plasmids thus formed would produce mutant Ada proteins whose cysteine residues at positions 69 and 321 are replaced by alanine. These were t,ermed pKT101 and pYN3087, respectively. These plasmids were introduced into E. coli strain YS180. which deletes most of the a,da gene.
(f) Other meth,ods .Methyltransferasr and 3-methyladenine-DNA glycosylase II activities were determined as described (Nakabeppu et al.. 1984, 1985a). Chloramphenicol acetyltransferase was assaved by the spectrophotometric method described by khan (1975). High-pressure liquid chromatographic analysis of alkylated product’s was performed using an J,KR BROMMA liquid chromato-
Ala69 -e--m
5' 3'
69 T CGC CCC TGC AAA CGT Arg Pro Cys Lys Arg
3' T
w--m-
Template
DNA
5'
A GCG GGG CGG TTT GCA A
Synthetic
primer
Ala
Ala321 5' v---m
321 ATA
ATA
CCC TGT
Ile
Ile
Pro
3' CAT
Cys His
CGG
---mm
Template
DNA
Arg 5'
3' TAT TAT GGG CGA GTA GCC
Synthetic
primer
Ala Figure 1. Synthetic primers used for site-directed mutagenesis. The primers are given from 3’ to 5’; the complementary strands of template DNA are from 5’ to 3’. The expected amino acid sequences are shown under the DNA sequences.
K. Takarw
et al.
-___
-__--..-.
samples. From the elect,rophoreticA mohilities of’ t htl minor bands, their molecular weights were &imated t’o he approximately 24,000 (for AlaGS) and 15,000 (for Ala321). which may represent cleavage products originating from t,he (!-terminal and the N-terminal half of the protein, respeAively. Sincat> the wild-t,ype prot,ein wa,s not significantly degraded under t,he conditions used. it is likely that, t,he mutant proteins are more susceptible to proteoIytjic. cleavage, probably reflecting subtAr alterations tiuc, to the amino acid substitutions.
(lalf’ thymus I)NA and poly[d(A T)] rverc’ trrat!ed wit,h [H3]MNV and used as subxtjratmes for methyltransferase. The former c*ont,ains 06-met h~l~uarlil~t~ methylphosphotriester. whilst the latkr
(0)
Figure 2. Methyl acceptor activities of normal and mutant ,4da proteins. E. coli YN180 cells harboring plasmid pYN3089 and its derivatives were grown at 30°C’ overnight. The crude extracts were prepared in buffer A and incubated with [3H]MNu’C-treated calf thymns Dh’A (54.000 disints/min) at 37”r for I.? min in 40 PI of brrfipr M (70 mu-Hrpes. KOH (pH 7.X). I mu-dithiothrc+tol. 5 mlw-EDTA). The reac%ion was trrminated by heat,ing at 90°C for 3 min aft,er adding 20 ~1 of Smrrcap 2.50 mM-Tris. HCI (pH 6.X). 2Oy:, (v/v) toethanol, 9.206 (w/v) SDS. -IO”;, (wiv) gl?/cerol. Thr samplrs were applied to a loo’, to PO”,, SDS/polyacrylamidr gradient gel. rlfter electrophoresis. the gel wax fixed in lOoi (v/v) met,hanol. IO”,, (v.‘v) acetic acid for 1 h and [3H]meth$-arrepted proteins were detected hy fuorography. Lane I 1 1.5 pg of protein from thr carudr rxtract derived from cells carrying pYrV3089 (wild-type): lane 2. 3 pg of protein from the extract from cells c*arr+ng pYN3087 (Ala321): lane :S. I2 pg of protein from the extract from cells carrying pKTIO1 (.4ia69). An arro\\ indicates the intact Ada protein.
When the cells were cultivat>ed in t,he prrsenc~t~ of ampicillin, the plasmids were stably maintained. These cells produced a large amount of protein indistinguishable from normal Ada protein in molecular size. When the rxtract,s were incubated with [ 3H]MNU-treated calf thymus I>SA and then subjected to SDS/polyacrylamide gel elertrophoresis followed by fluorographg. t’hr result shown In addition to the main in Figure 2 was obtained. band corresponding to the intact rIda protein (39,000 lVr): thin bands were detected in the mutant
0
(b)
)--I A
Oh IO
20 Ada
.40 protein
A ^^ tw ( pmoi i
Figure 3. ~lrthyltransferasr a(atlvltit+ of mutjant Adi1 wr’w protrins. I’arious amounts of puritird Ads protf~in inctubatt~d with [3HjMNII-trratrd (*aIf thymus I)SA ((a) .59,000 disintsjmin par assay) or j 3HjMNl:-trc~att~ti poly[d(A ‘T)] ((h) 7400 disintsimin per assay) at X7-1’ f’or 15 min in lOOp1 of buffer M, .\ftt:r t)hca rraction. thp mixtuns was heated at 90°C” for 15 min with IO0 pl c)t 1 mg bovinr serum albumin/ml and 100 pl I)t’ 0.8 M-tricShloroacaetic acid and the prot,tlin prrcipitattLtl. The pcllrt was ri nsrd wit,t1 500/d of AO<> (a-,!.) trichloroac.eticx acid and dissolvt~d in 200 pi Uf 0. I ~-.kiaOH. Thp radioactivity was detrrmined in ii liquid sc-intillation counter. (0) Wild-type Ada protrirl, (0) AlaX .-Ida prot4n: (A) Ala69 Ada protein.
E. coli Ada Regulatory contains methylphosphotriester alone, as the major methyl donors for the methyltransferase reaction. Although hot,h contain 06-methylthymine, the content is too scanty to be accounted for the assay. Figure 3(a) shows transfer of the methyl group from meth?;lated DNA to purified preparations of r\da proteins derived from t’hree types of cells. The radioactivit>- retained in the protein reached sat,uration at caertain levels, thereby reflecting the numlter~ of methyl acceptor sites of the protein. 12oth the mutant proteins were able to accept thr m&h-l group at a level just half of that accepted I)!. t.he wild-type prot,ein. 12’ith mrthylatrd poI~rld(A. T)] as a methyl donor. thcb result shown in Figure 3(b) was ohtainetl. In this ca~sr,Ala321 protein accepted the tncth?-t group to the same extent as seen with wildt!~l~~ protein. The Ala60 prot’ein could not accept the methyl group. \l’e obtained similar results with crude extracts of cells (data not shown), These results are compatible with the idea that the Ala69 and A4ta321 mut,ant proteins accept the methyl group from 06-methylguanine and methyt-
Elui~on
(0)
(b)
Protein
265
phosphotriester residues of the methylated respectively. (c) Repair
DNA,
of 04-methylthym,ine
It has been shown that 04-methylthymine, which is less abundant in methylated DNA, is repaired by the ada gene product (McCarthy et al., 1984). To observe the cysteine residues that accept the methyl group from 04-methylthymine, we performed the following experiment. Poly(dT) treated with [3H]MNU was annealed with poly(dA), and the polymer was incubat,ed with purified Ada proteins derived from the three types of extracts. The DNA samples were digested with a mixture of DNase I, phosphodiesterase T and II, and bacterial alkaline phosphatase. The hydrolyzates were then analyzed by high-pressure liquid chromatography. Figure 4 shows the distribution of the radioactivity of each sample fractionated. There were three peaks, two of which comigrated with the authentic markers for nucleosides. The first peak corresponds to N3-methylthymidine and the second
time
(rnln) (C)
id)
Figure 4. High-pressure liquid chromatographic analysis of 04-methylthymine and other methytated products. (3H]MNU-trrated pofT(dT) was annealed with poly(dA), and 20 nmol of the polymer (90,000 disints/min) was treated \vith PO0 pmot of purtf~rd Ada prot’ein in 100 ~1 of buffer M at 37°C’ for 15 min. The polynucleotides were then incubated in 200~1 of a mixture containing 300 units of DNase I, 0.2 unit each of phosphodiesterase I and II, 1.5 units of bacterial alkaline phosphntasr, 20 mM-MgCi, and 50 mm-Tris. HCI (pH 7%) at 37°C for 20 h. After removing the proteins by precipitation with ethanol. the supernatants were concentrated to about 250~1. The sample (45~1 each) was applied, together with 5 ~1 (IO nmol) of 04-methylthymidine, to the column. Elution was performed at a flow-rate of 0.5 ml/min with a linear gradient from OojO t,o 100° ,. methanol in 50 mM-ammonium format’r (pH 4.6). Fractions were c*ollectrd ~rery 0.5 min, and the radioactivity was determined in a liquid scintillation counter. Circles indicate the ratioactivity of racsh fraction. and broken lines the absorbance of 04-methylthymidine at 260 nm. (a) Ko protein: (b) wild-type Ada protein: (c) Ala69 Ada prot’rin; (d) Ala321 Ada protein.
266
K.
Takano
et al
20
I 2
IO ”
‘I
L
MNNG
0
(Fg/ml)
Figure 5. Survival of cells aft’er t’reatment with MSXG. E. coli YN180 [A(ada-aMI)] cells harboring various f)lasniids were grown at SO”(’ to .&60=O% treated with various concentrations of MNNG
to 0.8. and
at 37 “V for 10 min in M9 salt medium (pH 6.0). After appropriate dilution, portions of the cultures were plated on LH plates. and the plates were incubated at 30°C overnight. (0) pUC9 (vector); (a) pYN3089 (wild-type): (A) pYN3087 (Ala321); (A) pKTlOl (Ala69).
to 04-methylthymidine (broken line in Fig. 4). The second radioactivity peak (for 04-methylthymidine) disappeared when the DNA was pretreated with the wild-type and Ala69 Ada proteins, but not with Ala321 protein. This result is compatible with the idea that the methyl group of 04-methylthymine can be transferred to the cysteine residue at position 321, but not that at position 69, of Ada protein. The third peak, which may represent’ the radioactivity derived from methylphosphotriesters, also gave characteristic patterns for three types of Ada protein. The peaks for the wild-type and Ala321 samples were almost half those for the untreated and Ala69 samples. This is consistent with the notion that cysteine 69, but not cysteine 321, accepts a methyl group from one of two stereoisomers of methylphosphotriesters in the DNA. (d) Biological speci$c
The availability a single methyl ample opportunity
consequences lesions of
of repair
of
DNA
of mutant Ada protein with only acceptor site would provide an to see which types of lesions
4 MNNG
6
(pg/mlI
Figure 6. Mutation frrquency of cells after treatment with MNNG. E. roli YNl80 [A(ada-nlkH)] cells harboring various plasmids were incubated in I119 salt medium containing various concenkations of MNN(: at 37°C’ for 10 min. The cells were collected by caentrifugation. resuspended in LB broth and then incubated at 30°C for I h to fix the mutations arisen. Portions of t,hr ~11 suspension were plat’ed on LB plates containing 100 jig of rifampic.in,‘tnl. and the plates were inc~rrl)atc~ti at 30 (’ for, 2 t1aj.s. (0) plTC’9 (vrc.tot,): (0) pYN3089 (wihl-tyl~~~): (a) pYS:H)xi (;\Ia:El): (A) p[iTIoI (:Iiafi!)).
produced in DNA by alkylating agents art‘ responsible for killing a,nd for induction of mutations. For this. bacteria, wit,h t’he mutant or the wild-type ada gene were exposed to various concentrations of MNNG at 37°C for t,en minutes. Figure 5 shows the survival of bacteria after such treatment. Cells carrying pKTlO1 (AlaSS) showed much t’he same high level of survival as thosr carrying pYN3089 (wild-type), whereas (sells carrying pYN3087 (Ala321 ) were as sensitive as those carrying only the vector plasmids When the occurrence of mutations in cells afit-’ treatment with MNNG was examined, pKTIO1 (Ala69) was as effective as pYN3089 (wild-type) for the suppressing induction of mut8at’ions (Fig. 6). 1 n cells carrying pYN3087 (Ala321), a considerable number of mutations were induced by MNNG as in cells without ada+ plasmids. These observa,tionx indicate that 06-methylguanine and 04-methylthymine, rather than the methylphosphotriester~ produced in the DNA are responsible both for killing of the cells and for the induct)ion of mutations.
E. coli Ada Regulatory
Protein
267
K. Takano
26X
(e) Effects
ada mutations on the of the ada and alkA genes in vitro
of the
transcription
Since both the ada and alkA gene products are known to be inducible as part, of an adaptive response (Demple et al., 1982; Evensen &, Seeberg, 1982), an in-vitro transcription system was used to examine the ability of mutant Ada proteins to promote transcription of the ada and aEkA genes. Various amounts of wild-type and the mutant forms of Ada protein were incubated with normal or methylated DNA, and then subject,ed to an in-vitro transcription system, which contain the E. coli RT\;A polymerase holoenzyme, DNA templates containing the ada, alkA and lacrJV5 promoters, and labeled ribonucleoside triphosphates. The transcription products were resolved by polyacrylamide gel electrophoresis followed by autoradiography (Fig. 7). As already noted (Nakabeppu & Sekiguchi, 1986), two types of RXAs, 90 and 98 nucleotides long, were formed from the ada, promoter. Formation of the 98-nucleotide transcript was greatly enha,nced by the methylated form of the wild-type Ada protein, and a lesser degree of stimulation was observed wit,h the nonmethylated form. The 90-nucleotide t’ranscript. whose formation was suppressed by the methylated form of Ada protein, appears to be an in-vitro artifa,ct (Nakamura uf al.. 1988). Format,ion of the 23 nuclrotide alkA transcript, was stimulated by both normal and met,hylat’ed forms of Ada protein. The result obtained with Ala321 Ada protein (Fig. 7(b)) was essentially similar to that with the wild-type Ada protein, except that a greater extent of stimulation of alkA transcription occurred with the methylated form. With the Ala69 protein, a different result was obt)ained (Fig. 7(c)). Only a basal level of formation of the 90 nucleotide ada and the 23 nucleotide alkA products occurred in the of the unmet’hylated form of Ala69 presence protein. Even this basal level of transcription was not induced by t,he methylated form of Ala69 protein. Since normal transcription occurred from the cont,rol Zac promot,er. this inhibition was specific: for both the ada and alkA promoters. Tt is evident t,hat the cysteine residue at, position 69 is important for accepting the methyl group and for maintaining the conformation of Ada protein as a transcriptional regulator.
(f) Effects of the
of
et al.
ada smutations ada and alkA
on the expression genes in vivo
To investigate the effects of mutant Ada proteins on the expression of the ada and alkA genes in viao. we devised a system that allows for a quantitative estimation of activity of Ada protein on the two types of promoters, Pada and PalkA (Fig. 8). In this system, Ada protein would be constitutivel) produced, since the lac promot,er (Plac) is placed upstream from the ada gene lacking its own
PlasmId
Pado -m
cat
PalkA
alkA
Pfac
ada
Figure 8. The system fhr measuring transcription promoting act,ivit’y of mutant Ada Fxotein in GCV. PvuJ~ PvuII fragments (1.6 kb) containing t,he 1~ promoter (Plac) attached to the promoter-deleted normal and mutant forms of the adn gene were t)urified from pYN3089 and its derivatives, and inserted into the EcoRV site of pKT410 plasmid. which carries the a&r promoter (Pada) attached to the chloramphenicol acsrt>lltransferase (cat) gene. The resulting plasmids carrying th(l 2 genes with different promoters. Pada-ml and Plctc-uh. were introduced into E. coli YXl80. which possesses the intact alkA gene with its own promoter hut deletes the n*ln-alkH region of the cshromosomr. ---. _---.--__-.---.---~~_ ._..._.~.. promoter. Ada protein thus formed c~mltl lw methylated when cells were exposed to alkylating agents. Normal or methylat,ed forms of’ Ada protein would then act’ on the a& promoter (Pa&). to which the chloramphenicol acetylt,ransferasr gpnc’ (cat) was attached, and t,he aZkA promoter (I’alkrl ) with its own gene, alkrf. Thus, by measuring chloramphenicol acrtyltransferasc and S-met 1-1~1~ adenine-DNA glycosylastl IT activitichs, u v (::III estimat)r the transc:ription-I)romotitl~ act’ivity of Ada protein on the two types of promoters. I’sing this sy.stem, we measured transcriptiolipromoting activities of three types of Ada protein: the results are shown in Table 1. In culls wit,h pKT41 I carrying the wild-type adn gtLnt>, consider able amounts of thr t,wo enzymes wt’rtt producac4. even without adaptive tseat)ment. N’frc~n treated with 0.005 94, AIMS. the cells produced more thatr tenfold acetyltransf&ase. On t lrc~ other hand. ;I relatively small increase in glyc~osylaw~ activity was observed, probably reflecting the finding that t IIP normal and methylated forms of Ada protein ar(’ equally effective for promoting t’ranscription of t’hr alkA gent. In cells harboring plasrnid pKT4I 2 (Ala32 I ). both enzymes were produced t’o levels cLotnparabl(l with or even higher than those in ~~11s c*arrying t h(l wild-t,ype ada gene. Of interest is t,hc> observat,iorr that a very high level of cat gene expression wit.< induced by the Ala321 protein even when t ht. (*-. In cells harboring pKT413 (Alafit)). the Irvtals of both enzyme activities were low under non-induc~t~tl c:onditJions and did 1101 incrt~asc signifi(-ant 1). ;~f’t’(br
E. coli Ada Regulatory Protein
269
Table 1 Transcription-promoting activity of Ada protein in vivo
Ada protein
MMS
Chloramphenicol acetyltransferase (nmol min-’ mg-‘)
+ + + +
\Vild-typr Ala32l .4la69
4.90 64K 24.7 391 329 706 5.78 8-10
:l-.\leth~latlenilleDNA glycos~lase II (pm01 min mp-‘) 0694 0.736 8.86 12.8 4.02 13.5 1.31 1.15
YN180 cells harboring pKT410 (without a&), pKT4ll(wild-type u&z). pKT412 (Ala321 ada), or pIiT (,\hB9 cdu) were grown at 37 Y to A,,, = 0.2 to 0.3. After shaking for 2 h with or without 0~005°~,1 MMS, the crude extracts were prepared for the assays of the 2 enzymes. For the acetyltransferase assay, 10 to 8Opg of protein from the crude extract were added to a mixture wntaining 0.1 iwTris HCI (pH 7,8), 0.1 M-aCt%yl coenzgme A, and 0.4 mg of 5,5’-dithiobis-2nitrobenzoic acid/ml. After adding chloramphenicol to a final concentration of 0.2 mM, changes in the ;~bsorl)tion at -112 nm were followed at 37°C’ for 5 min. For the gl?;cosylasr assay, 5 to 20 pg of protein ti,om the crude extract were incubat,ed with [3H JMSI’-treated calf thymus DNA in 60 ~1 of a mistuw c,ontaining 70 miwHrpes~ KOH (pH 7.8). 1 rnM-tiithiothrritol. 5 mwEI)TA. and 3 mw3-methyladeninr at 37°C for 15 min. After removing protein by precipitation with ethanol, the radioactivity of the supernatant was determined. c
.
the inducing treatment. This result is consistent with findings in the in-vitro experiments, presented in the preceding section. It seems that the cysteine residue at position 69 is important for accepting the methyl group and to maintain the conformation of Ada prot’ein so as to exhibit the basic transcriptionpromoting activity.
4. Discussion The DNA sequence analysis of the cloned ada gene revealed that the Ada protein is composed of 354 amino acid residues and has a molecular weight of 39,400 (Demple rt al., 1985; Nakabeppu et al.. 1985a). The protein is readily cleaved by cellular proteinase(s) to generate smaller polypeptides possessing the methyltransferase activities (Teo al., 1984: Nakabeppu et al., 1985a,; Yoshikai et al., unpublished result). The 19,000 Jf, polypeptide. which was initially characterized as V-methylguanine-I)XA methyltransferase (Demple et al.. 1982), appears to be one of the cleavage products. Direct amino acid analysis of 3H-labeled, 19,000 M, protein (Demplr et al., 1985) revealed that the methyl acceptor site is a cvsteine residue, which corresponds to the 321st amino acid residue of the int,act Ada protein. It, was later found that the intact Ada protein possesses anot,her methyltransferase activity, which (Batalyzes transfer of a methyl group from one of the two stereoisomers of methylphosphotriesters in alkplated DNA to a cysteine residue in the N-terminal half of the protein (Margison et al., 1985: Mc(:arthy & LindahI, 1985). On the basis of the amino acid sequence homology, it was suggested that the cysteine residue at position 69 may be the mththyl n,ccept,or site (Teo rt al., 1986). If such is indeed the case. t.hen the Ada protein whose et
cysteine residue at position 69 is converted to another amino acid could not accept t,he methyl group from methylphosphotriester, but would retain the ability to accept the methyl group from 06-methylguanine. In this work, we constructed such a mutant form of Ada protein by site-directed mutagenesis of the cloned ada, gene, and found that this is actually the case. It has been found that the cysteine 69 residue accepts the methyl group from methylphosphotriesters, as determined by peptide mapping (Sedgwick et aE., 1987). In a parallel experiment, we constructed Ada protein whose cysteine residue at position 321 is converted to alanine, thereby supporting the notion that this cysteine residue is the acceptor sit)e for t,he methyl group from 06-methylguanine. Using the mutant, we demonstrated that the methyl group of 04-methylthymine is also transferred to the cysteine at position 321. The availability of mutant Ada proteins carrying only one of the methyl acceptor sites would provide a unique opportunity to investigate the biological significance of repair of specific lesions. By introducing the mutant ada gene into ada- cells, we demonstrated that the cysteine residue at position 321, but not that at position 69, is required for suppression of alkylating agent-induced mutagenesis. This is in accord with the notion t,hat 06-methytguanine would be the primary mutagenic lesion produced by alkylating agents (Loveless. 1969). We found that cells producing Ada protein with alanine at position 69 are as resistant to alkylating agents as ada+ cells producing normal Ada protein. Since Ada protein with alanine at position 321 had no such effect on ada- cells, it is likely that 06-methylguanine and 04-methylthymine, the repair of which is handled by the cysteine residue at position 321, are prematagenic
270
K. Takano
and lethal t’o cells. Evidence that) 06-methylguanine is a major lethal as well as mutagenic lesion has been obtained with mammalian cells (Brennand 8r Margison, 1986; Ishizaki et al., 1987). To elucidate the role of specific methylation of Ada protein in regulation of expression of the inducible genes, we performed in-vitro transcription experiments. Using a reconstituted syst’em with purified components, we demonstrated that met,hylation of the cysteine residue at position 69 is responsible for activation of the Ada protein, as a transcriptional regulator. The Ada protein with the alanine residue at, position 69 was inactive for promoting in-vitro transcription of the ada gene. whereas the Ada protein with alanine at position 321 retains the full activity for ada transcription. This is consistent with the proposal of Teo ef al. (1986), that transfer of the methyl group from methylphosphotriester of DNA to Ada protein is a trigger for induction of the adaptive response. \\‘e developed a system to measure the activitv of Ada protein to induce the ada and alkA expression in viva (see Fig. 8), and obtained results t,hat conform in part with those of experiments Zn vitro. The Ala69 protein is unable to induct expression of the ada gene in vivo. However. in the case of the Ala321 protein we obtained an unexpected result. Cells carrying the Ala321 protein exhibited high levels of expression of the ada gene, even though they had not been exposed to methylating agents. The apparent discrepancy from the results of experiment,s in vitro may be explained by the amplification in vivo of relatively small changes that cannot’ be detected in vitro or by the presence of unknown factor(s) that may be involved in t’he transcriptional regulation in ZGO. The result obtained in vivo suggests that conversion of cysteine at position 321 to alanine would ca,use some subtle conformational alteration of Ada protein so as to make for more ready accessto the promoter regions. This may relate to t~heobservation t’hat a deletion that truncated the 3’ end of t’he ada gene resulted in the production of Ada derivatives constitutively act’ivated for transcription (Lemotte & Walker, 1985; Nakabeppu et al.. 19856; Shevell & Walker, personal communication). The involvement of the C-terminal domain in gene activation was also suggested by analysis of the ada mutants (Temple, 1986). Based on the results obtained in t,his and previous studies (Nakabeppu & Sekiguchi, 1986: Teo et al., 1986), a model for the activation of Ada protein is depicted (Fig. 9). The cysteine residue at position 321 would accept the met’hyf group from both 06-methylguanine and 04-methylthymine of alkylated DNA, whilst. the cyst’eine residue at position 69 would accept the methyl group from methylphosphotriester of the DNA. The t*wo reactions occur independently and lead to different) biological consequences; the former is responsible for removing premutagenic and lethal lesions from the DNA, and the latter for converting Ada protein to an active form as a transcriptional regulator.
et al.
1 Proteolytic cleovoge
Tronscrlptlonol octlvotlon of the ado reqlon
Repolr of lethal and premutagenic lesions
Figure 9. A model for the actions of Ada prot,ein. I’“. G” and T” st,and for methylphosphotriestcr, 06-n~rthylguanine and 04-methylthymine. respectively.
Ada protein is cleaved by a cellular proteinase(s) to produce smaller polypeptides consisting of the N-terminal and the (‘-terminal half of the protein. \Z’r recently obtained evidence that the cleaved products carry no act,ivity to promote transcription of crdn but do ret,ain activity to promote a&4 t,ranscription in citm (Voshikai et al.. unpublished results). So. it is likely that both the N-terminal half and the (I-terminal moiety of the protein are required for transcription, ;tnd this ma! relate to our observation that modification of the cysteine residue at position 321 affect’s the tran scription
patterns
in
viva.
We thank $1. Ohara for comments on the manuscript. This work has been supported by grants from the Ministry of Education, Science and (‘ult~urr (61065007) and from t)he Jnstitute of Protein Engineering.
References Barnes. LV. Il., Hevan, M. & Son. I’. H. (IYX3). M~th,ods Enzymol. 101. 98-122. Brennand. J. h Margison. G. I’. (1986). (‘ctrcinc~tenesis. 7. 18.5188. Cathcart. R. & (ioldthwait, I). A. (1981). Hiochpmistry 20. 273 -286. (Ihang, A. C. Y. & (‘ohen. S. N. (1978). .I Hrxc~eriol. 134. 1141-I 156. (‘lose, T. ,J. & Rodriguez, R. 1,. (198%). (&r~c. 20. X&316. Demple, H. (1986). Nu,rl. ilcids Res. 14, 557.55589. Temple, B., ~Jacobsson. A.. Olsson. M., Robins. 1’. & Lindahl. T. (1982). J. Riol. Uum. 257. 13776~~13780. Temple, R.. Sedgwick. ES., R,obins. P.. Totty. S.. Waterfield. M. I). B Lindahl. T. (1985). I’roc. .\:tsl. Acad. Sri., IY3.A. 82. 2688-269~. Evensen. G. & Seeberg. E. (1982). Rjlt~re flondon). 296. 773-775. Ishizaki. K.. Tsujimura. ‘I‘.. F’ujio. (’ Yangpf~l. %.
E. coli Ada Regulatory Yawata, H Nakabeppu, Y., Sekiguchi, M. & Ikenaga, M. (1987). Mutat. Res. 184, 121-128. ,Jeggo. I?. (1979). J. Bacterial. 139, 783-791. Karran, I’., Hjelmgren, T. & Lindahl, T. (1982). Nature (London), 296, 77&773. Lemotte. I’. K. & Walker, G. C. (1985). J. Bactehol. 161, 888- 895. Loveless. A. (1969). Nature (London), 223, 206-207. Maniatis. T., Frimch, E. F. & Sambrook, J. (1982). In Molecular (Toning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor. NY. Margison. G. P.. Cooper, D. P. & Brennand.
Protein
271
Nakabeppu, Y., Kondo, H., Kawabata, S.. Iwanaga, S. & Sekiguchi, M. (1985a). J. Biol. Chem. 260, 7281-7288. Nakabeppu, Y.. Mine, Y. & Sekiguchi, M. (19856). Mutat. Res. 146, 155-167. Sakamura. T.. Tokumoto. Y.. Sakumi. K.. Koikr. (i.. Kakabeppu, Y. & Srkiguchi. M. (198X). .1. llfol. Hiol. in the press. Sanger, F.. Nicklen, S. & Coulson, A. R. (I 977). Proc. Xat. Acad. Sci.. U.S.A. 74. 5463-5467. Sedgwick. B. (1983). Mol. Gen. Genet. 191. 46fGi72. Sedgwick, B., Robins, P., Totty, K’. & Lindahl. T. (1988). J. Biol. Chem. in the press. Sekiguchi. M. & Nakabeppu, Y. (1987). Trends Genet. 3. 51-54. Shaw. W. V. (1975). Methods Enzymol. 43. 737~-755. Teo, I. (1987). Mutat. Res. 183, 123-127. Teo, I.. Sedgwick, B.. Demple, B.. Li. B. & Lindahl, T. (1984). EMBOJ. 3. 2151-2157. Teo, I., Sedgwick, B. Kilpatrick. M. W., McCarthy, T. V. & Lindahl, T. (1986). Cell, 45. 315-324. Vieira, J. & Messing. J. (1982). Gene, 19, 259-268. Yamamoto. Y. & Sekiguchi, M. (1979). Mol. Ben. Genet. 171. 251-256.
by K. Matsubara
Functional
Sites of the Ada Regulatory of Escherichia coli
Protein
Analysis by Amino Acid Substitutions Kyoko
Takano,
Yusaku
Nakabeppu
and Mutsuo
Sekiguchi
Department of Biochemistry Faculty of Medicine Kyushu University 60 Fukuoka 812, Japan (Received 10 September 1987) Specific cysteine residues at possible methyl acceptor sites of the Ada protein of Escherichia coli were converted to other amino acids by site-directed mutagenesis of the cloned ada gene of E. coli. Ada protein with the cysteine residue at 321 replaced by alanine was capable of accepting the methyl group from the methylphosphotriester but not from 06methylguanine or 04-methylthymine of alkylated DNA, whereas the protein with alanine at position 69 accepted the methyl group from the methylated bases but not from the methylphosphotriester. These two mutants were used to elucidate the biological significance of repair of the two types of alkylation lesions. Introduction of the ada gene with the Ala69 mutation into an ada- cell rendered the cell more resistant to alkylating agents with respect to both killing and induction of mutations, but the gene with the Ala321 mutation exhibited no such activity. Replacement’ of the cysteine residue at position 69, but not at position 321, abolished the ability of Ada protein to promote transcription of both ada and alkA genes in vitro. These results are compatible with the idea that methylation of the cysteine residue at position 69 renders Ada protein active as a transcriptional regulator, whilst the cysteine residue at position 321 is responsible for repair of pre-mutagenic and lethal lesions in DNA. The actions of mutant Ada proteins on the ada and alkA promoters in viva were investigated using an artificially composed gene expression system. When the ada gene wit,h the Ala69 mutation was introduced into the cell, there was little induction of expression of with an alkylating agent, in either the ada or the alkA genes. even after treatment agreement. with the data obtained from st,udies in vitro. With the Ala321 mutation, however, a considerable degree of ada gene expression occurred without adaptive treatment. The latter finding suggests that the cysteine residue at position 321, which is located near the C terminus of the Ada protein, IS involved in regulating activity, as the transcriptional activator.
The Ada protein, with a molecular weight of 39,400, possesses two methyltransferase activities. It’ transfers a methyl group from the 06-methylguanine residue of alkylated DNA to a specific cysteine residue close to the C terminus of its own molecule (Demple et al., 1985). It also cat,alyzes the transfer of a methyl group from one of two stereoisomers of methylphosphotriesters of alkylated DNA to a cysteine residue in the N-terminal half of the protein (McCarthy & Lindahl, 1985; Margison et al., 1985). The Ada protein is cleaved by cellular protease to two fragments, each carrying one of the methyltransferase activities
1. Introduction Enzymes that repair alkylated DNA are formed in Escherichia co& cells on exposure to low concentrations of alkylating agents (Karran et al., 1982; Evensen & Seeberg, 1982), and this induced synthesis is controlled by the ada gene (Jeggo, 1979; Sedgwick, 1983; Teo et al., 1984; Lemotte & Walker, 1985; Nakabeppu et al., 1985a,b). Recent studies have revealed that the Ada protein, coded by the ada gene, functions as a DNA repair enzyme as well as a transcriptional regulator (for a review, see Sekiguchi & Nakabeppu, 1987). 261
K. Takano
(Nakabeppu et al., 1985a: Teo, 1987; Yoshikai et al., unpublished results). Expression of the ada and the related genes, such as alkA, belonging to the ada regulon is positively controlled by the Ada protein. Large amounts of ada and alkrl-specific RNAs are formed in cells treated with a methylating agent. whereas lit’tle such RiVA is produced in untreated cells. The in&vo transcription initiation sites for the two genes were determined by primer-extension cDNA synthesis and S, nuclease mapping (Nakabeppu et al.. 19&a). In in-vitro reconstituted systems, bot,h ada and alkA transcripts were formed in an Ada protein-dependent manner (Teo et al., 1986; Nakabeppu & Sekiguchi. 1986). It was shown that the methylated form of Ada protein is more effective in promoting ada t,ranscription than is the unmethylated form, whilst the effects of both forms are much the same with regard t’o aZkA transcription. The DNA sequencesrequired for the regulat,ed expression of the ada and alkA genes have been determined by introducing various mutations in the promoter regions (Xakamura et 01.. 1988). A model for the molecular mechanism of a,daptive response was designed on the basis of these findings. We constructed altered forms of Ada protein. whose specific amino acid residues at the possible methyl acceptor sites were replaced by other amino acids. by the site-directed mutagenesis of the cloned ada gene. This facilitated an examination of the effects of specific methylation on the activities of Ada protein, and revealed the biological consequencesof the two types of methylation of Ada protein.
2. Materials (a)
Chemicala
and Methods and
et al.
E. coli st’rain YN;18O j A(ada-alkB) : krrtr] was (aon strutted as follows. A 1.3 kb Km’ fragment derived front
pUC4K (Vieira & Messing, 1982) was 2.5 kb .V/uI
Jll~rT fragment
cxjntaining
substituted for a. 1)ar’ts of tht, c/d/c
and a&B genesin pYpu‘3028(Xakabeppu rt ccl.. I985r’i). Strain YY22 (poL4’“) cells (Yamamoto Pu Hekiguchi. 1979) were transformed Cm’ transformants DNA is replaced
with the resulting plasmid and Km’ were selected at, 30°C‘. If the c+mecl wit,h t)hr corresponding region of thr
chromosomeby double crossing-over,Km’ selected at the restrictive vclc*tor plasmid ])A(‘I’(‘184
CInP
cells
cat,
temperature. 42’C’. since the ((‘hang R (‘oh~n. l!fi8) is
be
unableto replicate without I)NA polymrrasr I. Such culls were infected with 1’1 phage and the A(a,dn-alkR) region was transduced into thr pol=l ’ strain.
: ku,c YN3
(Nakabeppuet al.. 19856).The resultingstrain wasnatnptl YN180. Strain KTIO (dam . F + 1 was c~onstruc*ted by in1 1.0 ducing F plasmid, pWH373. derived from WB373 (Barnes et al.. 1983) into GM48 (da& ( F-) (Marinus. 1973).
Strain KTIO. WE373 (Barnes
cl
al.,
1983)
and wtarf’
vwtor
used
phagr
for
~WWWL’
sit,?-directed
mutagenesisand for sequencing. As the parental plasmid t,o be mutated. pYS:W:i!) carrying the ada gene with intact rrdcr promotel (Nakabeppu ut al., 1985a) was used. Since introduc*t,ion of the Ala321 mutant gene into t’his plasmid resulW in ;I poor growth of cells. probably because of ovrrI)roduc,tiolI of the Ada protein. we deleted a part of t,he ada promotjer (a 57 bp Hl;ndIII-HstETT fragment) to eliminate the effect, of mutant Ada protein on its own promot,rr. The resulting plasmid was named pYN3089. in which the ndo gene wa,s transcribed only from the lac promoter existing in the v&or plasmid, pUC9 (Vieira & Messing. 198%). To investigate the effect of Ada protein on the rxd~~ promot#er in PEWS. plastnid pKT410 was c~onstructrtl, A 176bp HindTTI-~k=coRlfragment containing the trrlrl promoter. derived from pYN3059. was inserted into the C’lnI site upstrearn from the chloramphenic~ol acrtyltrxnsferase gene in the plasmid pCM7 (Close & Rodriguez. 1982).
enzymes
Restriction enzymes, bacteriophageT4 polynucleotide kinase. DNase I, bacterial alkaline phosphatase, DNA sequencing kits. and DNA ligation kit,s were obtained f’rom Takara Shuzo (lo.. Lttl (Kyot)o). Toyobo (‘0.. Ltd (Osaka). and Nippon (Gene (‘0.. Ltd (Toyama). E. co/i RNA polymerase holoenzyme, phosphodiesterasr 1 and II, bovine serum albumin. poly[d(A.T)]. poly(dA), poly(dT) and calf thymus DNA were purchased from Pharmacia. Micrococcus Zutec~s DN-A. heparin. rifampicin. chloramphenicol, acetyl coenzyme A and 5.5’-dithiobis-2nitrobenzoic acid were obtained from Sigma. Ribosomal RNA was from Hoehringer-Mannheim Riochem. /cI-~~PJ~TTP(>ciOOC”i/mmol). [Y-~~I’JATP (> 7000(‘ii mmol). [a-32PIdCTP ( > 400 (‘i/mmol), and [n~othyZ-3H 1 MSl’t (0.5 to 2 (‘i/mmol) were purchased from N-e\+ England Nuclear and Amrrsham. MNNG and MMS were purcahasrd from Nakarai (‘hrmicals (‘0.. Ltd (Tokyo).
t Abbreviations used: LXNTJ, S-methylLLV-nitrosourea: MNNG. N-methyl-N’-nitro-N-nitrosoguanidine; MMS, methyl methanesulfonate; kb, lo3 bases; Km. kanamycin; Cm, chloramphenicol: bp. base-pair(s): ssDNA, single-stranded DNA.
Oligon~rcleotides for site-direct,ed mutagenesis ~‘rre chemically synthesized by an Applied Hiosyst)ems I)NA synthesizer 381A. The strategy of site-specific. mutsgenesis was according t,o Mark PI al. ( 1984). Th~x mutations were verified to he introduced at the designed sites by srquencc~ analysis (Sanger rt ol 1!)77) of the rnutant grnrs.
Methylated DNA was prepared as tltascribe
E. coli Ada Regulatory iMutant Ada proteins were purified from YN180 cells harboring plasmids derived from pYN3089, as described (Nakabeppu et al., 198%) with some modifications. Newly transformed cells were cultured in 1 to 4 1 of LB broth containing (50pg/ml each) ampicillin and kanamycin at 30”(” to ‘466o = 1-O. Crlls were collected and sonirated in buffer A (20 mM-Tris. HCl (pH 8.5), 1 mM-EDTA, 1 mm2-mercaptoethanol. 5% (w/v) glycerol) containing NaCl at 0.3 M. After treatment with 0.57& Polymin P to remove the nucleic acids, the sample was subjected to precipitation with ammonium sulfate. applied to a phosphocellulose P-11 column. and then eluted with a linear gradient from 0.15 to 0.5 M-NaCl in buffer A. Peak fractions were concentrated and applied to a Bio Gel P-60 column (1.5 cm x 100 cm). The eluted fractions were collected and dialyzed against’ a 1 : 1 (v/v) mixture of glycerol and buffer A containing 0.2 M-ammonium sulfate overnight at 4°C. (e) Nurvival und mutation frequency
Protein
263
graph equipped with a Waters ~BOSDAPAK column. General methods for DNA manipulation according to Maniatis et al. (1982).
Cle were
3. Results (a) Replacement of cysteine residues 321 of Ada protein by alar&P
6.9 and
Direct sequenceanalysis of the [J-terminal half of the methylated Ada protein revealed that cysteine 321, which is located between a proline and a histidine residue, is the acceptor site for a methyl group from 06-methylguanine of alkylated DNA (Demple et al., 1985). Based on the amino acid sequence similarity, it was proposed that cysteine 69, flanked by proline and lysine, may be the acceptor site for a methyl group from the methylphosphotriester of the DNA (Teo et al., 1986). We constructed mutant forms of Ada protein whose
(‘ells harboring various plasmids were grown at 30°C to .Abh,, =(bfi to 0.X. Samples (5 ml) of the rult,ures were centrifuged. and the cells were resuspended in 5 ml of 1M9 salt medium (pH 6.0). A portion (1 ml) of the suspension was added to 1 ml of M9 salt medium containing MNNG. After incubation at 37°C for 10 min, 2 ml of LB broth was added and the samples were plated on LB plates after appropriate dilution. The number of surviving cells was counted after incubation at 30°C overnight. To determine the mutat’ion frequency. MNNG-treated cells wpre placed on an LB plate containing rifampicin at lOO~g/ml. After incubation at 30°C for 2 days, the number of rifampicin-resistant cells was scored.
cysteine
residues
at
the
plausible
methyl
acceptor sites are replaced by other amino acids. To introduce the mutations into the cloned ada gene, we applied site-directed mutagenesis using synthetic oligonucleotides (Fig. 1). Ml3 ssDNA carrying a part of the ada gene was annealed with the synthetic primer, and the complementary &and was synthesized by the DNA polymerase I reaction. After transfection of E. coli WB373, the mutant phages were selected by hybridization, using the same oligonucleotides as probes. The cloned DNAs carrying the mutant’ ada gene were introduced into plasmid pYN3089, in which the gene, without its promoter region, can be transcribed under the control of the lac promoter. The plasmids thus formed would produce mutant Ada proteins whose cysteine residues at positions 69 and 321 are replaced by alanine. These were t,ermed pKT101 and pYN3087, respectively. These plasmids were introduced into E. coli strain YS180. which deletes most of the a,da gene.
(f) Other meth,ods .Methyltransferasr and 3-methyladenine-DNA glycosylase II activities were determined as described (Nakabeppu et al.. 1984, 1985a). Chloramphenicol acetyltransferase was assaved by the spectrophotometric method described by khan (1975). High-pressure liquid chromatographic analysis of alkylated product’s was performed using an J,KR BROMMA liquid chromato-
Ala69 -e--m
5' 3'
69 T CGC CCC TGC AAA CGT Arg Pro Cys Lys Arg
3' T
w--m-
Template
DNA
5'
A GCG GGG CGG TTT GCA A
Synthetic
primer
Ala
Ala321 5' v---m
321 ATA
ATA
CCC TGT
Ile
Ile
Pro
3' CAT
Cys His
CGG
---mm
Template
DNA
Arg 5'
3' TAT TAT GGG CGA GTA GCC
Synthetic
primer
Ala Figure 1. Synthetic primers used for site-directed mutagenesis. The primers are given from 3’ to 5’; the complementary strands of template DNA are from 5’ to 3’. The expected amino acid sequences are shown under the DNA sequences.
K. Takarw
et al.
-___
-__--..-.
samples. From the elect,rophoreticA mohilities of’ t htl minor bands, their molecular weights were &imated t’o he approximately 24,000 (for AlaGS) and 15,000 (for Ala321). which may represent cleavage products originating from t,he (!-terminal and the N-terminal half of the protein, respeAively. Sincat> the wild-t,ype prot,ein wa,s not significantly degraded under t,he conditions used. it is likely that, t,he mutant proteins are more susceptible to proteoIytjic. cleavage, probably reflecting subtAr alterations tiuc, to the amino acid substitutions.
(lalf’ thymus I)NA and poly[d(A T)] rverc’ trrat!ed wit,h [H3]MNV and used as subxtjratmes for methyltransferase. The former c*ont,ains 06-met h~l~uarlil~t~ methylphosphotriester. whilst the latkr
(0)
Figure 2. Methyl acceptor activities of normal and mutant ,4da proteins. E. coli YN180 cells harboring plasmid pYN3089 and its derivatives were grown at 30°C’ overnight. The crude extracts were prepared in buffer A and incubated with [3H]MNu’C-treated calf thymns Dh’A (54.000 disints/min) at 37”r for I.? min in 40 PI of brrfipr M (70 mu-Hrpes. KOH (pH 7.X). I mu-dithiothrc+tol. 5 mlw-EDTA). The reac%ion was trrminated by heat,ing at 90°C for 3 min aft,er adding 20 ~1 of Smrrcap 2.50 mM-Tris. HCI (pH 6.X). 2Oy:, (v/v) toethanol, 9.206 (w/v) SDS. -IO”;, (wiv) gl?/cerol. Thr samplrs were applied to a loo’, to PO”,, SDS/polyacrylamidr gradient gel. rlfter electrophoresis. the gel wax fixed in lOoi (v/v) met,hanol. IO”,, (v.‘v) acetic acid for 1 h and [3H]meth$-arrepted proteins were detected hy fuorography. Lane I 1 1.5 pg of protein from thr carudr rxtract derived from cells carrying pYrV3089 (wild-type): lane 2. 3 pg of protein from the extract from cells c*arr+ng pYN3087 (Ala321): lane :S. I2 pg of protein from the extract from cells carrying pKTIO1 (.4ia69). An arro\\ indicates the intact Ada protein.
When the cells were cultivat>ed in t,he prrsenc~t~ of ampicillin, the plasmids were stably maintained. These cells produced a large amount of protein indistinguishable from normal Ada protein in molecular size. When the rxtract,s were incubated with [ 3H]MNU-treated calf thymus I>SA and then subjected to SDS/polyacrylamide gel elertrophoresis followed by fluorographg. t’hr result shown In addition to the main in Figure 2 was obtained. band corresponding to the intact rIda protein (39,000 lVr): thin bands were detected in the mutant
0
(b)
)--I A
Oh IO
20 Ada
.40 protein
A ^^ tw ( pmoi i
Figure 3. ~lrthyltransferasr a(atlvltit+ of mutjant Adi1 wr’w protrins. I’arious amounts of puritird Ads protf~in inctubatt~d with [3HjMNII-trratrd (*aIf thymus I)SA ((a) .59,000 disintsjmin par assay) or j 3HjMNl:-trc~att~ti poly[d(A ‘T)] ((h) 7400 disintsimin per assay) at X7-1’ f’or 15 min in lOOp1 of buffer M, .\ftt:r t)hca rraction. thp mixtuns was heated at 90°C” for 15 min with IO0 pl c)t 1 mg bovinr serum albumin/ml and 100 pl I)t’ 0.8 M-tricShloroacaetic acid and the prot,tlin prrcipitattLtl. The pcllrt was ri nsrd wit,t1 500/d of AO<> (a-,!.) trichloroac.eticx acid and dissolvt~d in 200 pi Uf 0. I ~-.kiaOH. Thp radioactivity was detrrmined in ii liquid sc-intillation counter. (0) Wild-type Ada protrirl, (0) AlaX .-Ida prot4n: (A) Ala69 Ada protein.
E. coli Ada Regulatory contains methylphosphotriester alone, as the major methyl donors for the methyltransferase reaction. Although hot,h contain 06-methylthymine, the content is too scanty to be accounted for the assay. Figure 3(a) shows transfer of the methyl group from meth?;lated DNA to purified preparations of r\da proteins derived from t’hree types of cells. The radioactivit>- retained in the protein reached sat,uration at caertain levels, thereby reflecting the numlter~ of methyl acceptor sites of the protein. 12oth the mutant proteins were able to accept thr m&h-l group at a level just half of that accepted I)!. t.he wild-type prot,ein. 12’ith mrthylatrd poI~rld(A. T)] as a methyl donor. thcb result shown in Figure 3(b) was ohtainetl. In this ca~sr,Ala321 protein accepted the tncth?-t group to the same extent as seen with wildt!~l~~ protein. The Ala60 prot’ein could not accept the methyl group. \l’e obtained similar results with crude extracts of cells (data not shown), These results are compatible with the idea that the Ala69 and A4ta321 mut,ant proteins accept the methyl group from 06-methylguanine and methyt-
Elui~on
(0)
(b)
Protein
265
phosphotriester residues of the methylated respectively. (c) Repair
DNA,
of 04-methylthym,ine
It has been shown that 04-methylthymine, which is less abundant in methylated DNA, is repaired by the ada gene product (McCarthy et al., 1984). To observe the cysteine residues that accept the methyl group from 04-methylthymine, we performed the following experiment. Poly(dT) treated with [3H]MNU was annealed with poly(dA), and the polymer was incubat,ed with purified Ada proteins derived from the three types of extracts. The DNA samples were digested with a mixture of DNase I, phosphodiesterase T and II, and bacterial alkaline phosphatase. The hydrolyzates were then analyzed by high-pressure liquid chromatography. Figure 4 shows the distribution of the radioactivity of each sample fractionated. There were three peaks, two of which comigrated with the authentic markers for nucleosides. The first peak corresponds to N3-methylthymidine and the second
time
(rnln) (C)
id)
Figure 4. High-pressure liquid chromatographic analysis of 04-methylthymine and other methytated products. (3H]MNU-trrated pofT(dT) was annealed with poly(dA), and 20 nmol of the polymer (90,000 disints/min) was treated \vith PO0 pmot of purtf~rd Ada prot’ein in 100 ~1 of buffer M at 37°C’ for 15 min. The polynucleotides were then incubated in 200~1 of a mixture containing 300 units of DNase I, 0.2 unit each of phosphodiesterase I and II, 1.5 units of bacterial alkaline phosphntasr, 20 mM-MgCi, and 50 mm-Tris. HCI (pH 7%) at 37°C for 20 h. After removing the proteins by precipitation with ethanol. the supernatants were concentrated to about 250~1. The sample (45~1 each) was applied, together with 5 ~1 (IO nmol) of 04-methylthymidine, to the column. Elution was performed at a flow-rate of 0.5 ml/min with a linear gradient from OojO t,o 100° ,. methanol in 50 mM-ammonium format’r (pH 4.6). Fractions were c*ollectrd ~rery 0.5 min, and the radioactivity was determined in a liquid scintillation counter. Circles indicate the ratioactivity of racsh fraction. and broken lines the absorbance of 04-methylthymidine at 260 nm. (a) Ko protein: (b) wild-type Ada protein: (c) Ala69 Ada prot’rin; (d) Ala321 Ada protein.
266
K.
Takano
et al
20
I 2
IO ”
‘I
L
MNNG
0
(Fg/ml)
Figure 5. Survival of cells aft’er t’reatment with MSXG. E. coli YN180 [A(ada-aMI)] cells harboring various f)lasniids were grown at SO”(’ to .&60=O% treated with various concentrations of MNNG
to 0.8. and
at 37 “V for 10 min in M9 salt medium (pH 6.0). After appropriate dilution, portions of the cultures were plated on LH plates. and the plates were incubated at 30°C overnight. (0) pUC9 (vector); (a) pYN3089 (wild-type): (A) pYN3087 (Ala321); (A) pKTlOl (Ala69).
to 04-methylthymidine (broken line in Fig. 4). The second radioactivity peak (for 04-methylthymidine) disappeared when the DNA was pretreated with the wild-type and Ala69 Ada proteins, but not with Ala321 protein. This result is compatible with the idea that the methyl group of 04-methylthymine can be transferred to the cysteine residue at position 321, but not that at position 69, of Ada protein. The third peak, which may represent’ the radioactivity derived from methylphosphotriesters, also gave characteristic patterns for three types of Ada protein. The peaks for the wild-type and Ala321 samples were almost half those for the untreated and Ala69 samples. This is consistent with the notion that cysteine 69, but not cysteine 321, accepts a methyl group from one of two stereoisomers of methylphosphotriesters in the DNA. (d) Biological speci$c
The availability a single methyl ample opportunity
consequences lesions of
of repair
of
DNA
of mutant Ada protein with only acceptor site would provide an to see which types of lesions
4 MNNG
6
(pg/mlI
Figure 6. Mutation frrquency of cells after treatment with MNNG. E. roli YNl80 [A(ada-nlkH)] cells harboring various plasmids were incubated in I119 salt medium containing various concenkations of MNN(: at 37°C’ for 10 min. The cells were collected by caentrifugation. resuspended in LB broth and then incubated at 30°C for I h to fix the mutations arisen. Portions of t,hr ~11 suspension were plat’ed on LB plates containing 100 jig of rifampic.in,‘tnl. and the plates were inc~rrl)atc~ti at 30 (’ for, 2 t1aj.s. (0) plTC’9 (vrc.tot,): (0) pYN3089 (wihl-tyl~~~): (a) pYS:H)xi (;\Ia:El): (A) p[iTIoI (:Iiafi!)).
produced in DNA by alkylating agents art‘ responsible for killing a,nd for induction of mutations. For this. bacteria, wit,h t’he mutant or the wild-type ada gene were exposed to various concentrations of MNNG at 37°C for t,en minutes. Figure 5 shows the survival of bacteria after such treatment. Cells carrying pKTlO1 (AlaSS) showed much t’he same high level of survival as thosr carrying pYN3089 (wild-type), whereas (sells carrying pYN3087 (Ala321 ) were as sensitive as those carrying only the vector plasmids When the occurrence of mutations in cells afit-’ treatment with MNNG was examined, pKTIO1 (Ala69) was as effective as pYN3089 (wild-type) for the suppressing induction of mut8at’ions (Fig. 6). 1 n cells carrying pYN3087 (Ala321), a considerable number of mutations were induced by MNNG as in cells without ada+ plasmids. These observa,tionx indicate that 06-methylguanine and 04-methylthymine, rather than the methylphosphotriester~ produced in the DNA are responsible both for killing of the cells and for the induct)ion of mutations.
E. coli Ada Regulatory
Protein
267
K. Takano
26X
(e) Effects
ada mutations on the of the ada and alkA genes in vitro
of the
transcription
Since both the ada and alkA gene products are known to be inducible as part, of an adaptive response (Demple et al., 1982; Evensen &, Seeberg, 1982), an in-vitro transcription system was used to examine the ability of mutant Ada proteins to promote transcription of the ada and aEkA genes. Various amounts of wild-type and the mutant forms of Ada protein were incubated with normal or methylated DNA, and then subject,ed to an in-vitro transcription system, which contain the E. coli RT\;A polymerase holoenzyme, DNA templates containing the ada, alkA and lacrJV5 promoters, and labeled ribonucleoside triphosphates. The transcription products were resolved by polyacrylamide gel electrophoresis followed by autoradiography (Fig. 7). As already noted (Nakabeppu & Sekiguchi, 1986), two types of RXAs, 90 and 98 nucleotides long, were formed from the ada, promoter. Formation of the 98-nucleotide transcript was greatly enha,nced by the methylated form of the wild-type Ada protein, and a lesser degree of stimulation was observed wit,h the nonmethylated form. The 90-nucleotide t’ranscript. whose formation was suppressed by the methylated form of Ada protein, appears to be an in-vitro artifa,ct (Nakamura uf al.. 1988). Format,ion of the 23 nuclrotide alkA transcript, was stimulated by both normal and met,hylat’ed forms of Ada protein. The result obtained with Ala321 Ada protein (Fig. 7(b)) was essentially similar to that with the wild-type Ada protein, except that a greater extent of stimulation of alkA transcription occurred with the methylated form. With the Ala69 protein, a different result was obt)ained (Fig. 7(c)). Only a basal level of formation of the 90 nucleotide ada and the 23 nucleotide alkA products occurred in the of the unmet’hylated form of Ala69 presence protein. Even this basal level of transcription was not induced by t,he methylated form of Ala69 protein. Since normal transcription occurred from the cont,rol Zac promot,er. this inhibition was specific: for both the ada and alkA promoters. Tt is evident t,hat the cysteine residue at, position 69 is important for accepting the methyl group and for maintaining the conformation of Ada protein as a transcriptional regulator.
(f) Effects of the
of
et al.
ada smutations ada and alkA
on the expression genes in vivo
To investigate the effects of mutant Ada proteins on the expression of the ada and alkA genes in viao. we devised a system that allows for a quantitative estimation of activity of Ada protein on the two types of promoters, Pada and PalkA (Fig. 8). In this system, Ada protein would be constitutivel) produced, since the lac promot,er (Plac) is placed upstream from the ada gene lacking its own
PlasmId
Pado -m
cat
PalkA
alkA
Pfac
ada
Figure 8. The system fhr measuring transcription promoting act,ivit’y of mutant Ada Fxotein in GCV. PvuJ~ PvuII fragments (1.6 kb) containing t,he 1~ promoter (Plac) attached to the promoter-deleted normal and mutant forms of the adn gene were t)urified from pYN3089 and its derivatives, and inserted into the EcoRV site of pKT410 plasmid. which carries the a&r promoter (Pada) attached to the chloramphenicol acsrt>lltransferase (cat) gene. The resulting plasmids carrying th(l 2 genes with different promoters. Pada-ml and Plctc-uh. were introduced into E. coli YXl80. which possesses the intact alkA gene with its own promoter hut deletes the n*ln-alkH region of the cshromosomr. ---. _---.--__-.---.---~~_ ._..._.~.. promoter. Ada protein thus formed c~mltl lw methylated when cells were exposed to alkylating agents. Normal or methylat,ed forms of’ Ada protein would then act’ on the a& promoter (Pa&). to which the chloramphenicol acetylt,ransferasr gpnc’ (cat) was attached, and t,he aZkA promoter (I’alkrl ) with its own gene, alkrf. Thus, by measuring chloramphenicol acrtyltransferasc and S-met 1-1~1~ adenine-DNA glycosylastl IT activitichs, u v (::III estimat)r the transc:ription-I)romotitl~ act’ivity of Ada protein on the two types of promoters. I’sing this sy.stem, we measured transcriptiolipromoting activities of three types of Ada protein: the results are shown in Table 1. In culls wit,h pKT41 I carrying the wild-type adn gtLnt>, consider able amounts of thr t,wo enzymes wt’rtt producac4. even without adaptive tseat)ment. N’frc~n treated with 0.005 94, AIMS. the cells produced more thatr tenfold acetyltransf&ase. On t lrc~ other hand. ;I relatively small increase in glyc~osylaw~ activity was observed, probably reflecting the finding that t IIP normal and methylated forms of Ada protein ar(’ equally effective for promoting t’ranscription of t’hr alkA gent. In cells harboring plasrnid pKT4I 2 (Ala32 I ). both enzymes were produced t’o levels cLotnparabl(l with or even higher than those in ~~11s c*arrying t h(l wild-t,ype ada gene. Of interest is t,hc> observat,iorr that a very high level of cat gene expression wit.< induced by the Ala321 protein even when t ht. (*
E. coli Ada Regulatory Protein
269
Table 1 Transcription-promoting activity of Ada protein in vivo
Ada protein
MMS
Chloramphenicol acetyltransferase (nmol min-’ mg-‘)
+ + + +
\Vild-typr Ala32l .4la69
4.90 64K 24.7 391 329 706 5.78 8-10
:l-.\leth~latlenilleDNA glycos~lase II (pm01 min mp-‘) 0694 0.736 8.86 12.8 4.02 13.5 1.31 1.15
YN180 cells harboring pKT410 (without a&), pKT4ll(wild-type u&z). pKT412 (Ala321 ada), or pIiT (,\hB9 cdu) were grown at 37 Y to A,,, = 0.2 to 0.3. After shaking for 2 h with or without 0~005°~,1 MMS, the crude extracts were prepared for the assays of the 2 enzymes. For the acetyltransferase assay, 10 to 8Opg of protein from the crude extract were added to a mixture wntaining 0.1 iwTris HCI (pH 7,8), 0.1 M-aCt%yl coenzgme A, and 0.4 mg of 5,5’-dithiobis-2nitrobenzoic acid/ml. After adding chloramphenicol to a final concentration of 0.2 mM, changes in the ;~bsorl)tion at -112 nm were followed at 37°C’ for 5 min. For the gl?;cosylasr assay, 5 to 20 pg of protein ti,om the crude extract were incubat,ed with [3H JMSI’-treated calf thymus DNA in 60 ~1 of a mistuw c,ontaining 70 miwHrpes~ KOH (pH 7.8). 1 rnM-tiithiothrritol. 5 mwEI)TA. and 3 mw3-methyladeninr at 37°C for 15 min. After removing protein by precipitation with ethanol, the radioactivity of the supernatant was determined. c
.
the inducing treatment. This result is consistent with findings in the in-vitro experiments, presented in the preceding section. It seems that the cysteine residue at position 69 is important for accepting the methyl group and to maintain the conformation of Ada prot’ein so as to exhibit the basic transcriptionpromoting activity.
4. Discussion The DNA sequence analysis of the cloned ada gene revealed that the Ada protein is composed of 354 amino acid residues and has a molecular weight of 39,400 (Demple rt al., 1985; Nakabeppu et al.. 1985a). The protein is readily cleaved by cellular proteinase(s) to generate smaller polypeptides possessing the methyltransferase activities (Teo al., 1984: Nakabeppu et al., 1985a,; Yoshikai et al., unpublished result). The 19,000 Jf, polypeptide. which was initially characterized as V-methylguanine-I)XA methyltransferase (Demple et al.. 1982), appears to be one of the cleavage products. Direct amino acid analysis of 3H-labeled, 19,000 M, protein (Demplr et al., 1985) revealed that the methyl acceptor site is a cvsteine residue, which corresponds to the 321st amino acid residue of the int,act Ada protein. It, was later found that the intact Ada protein possesses anot,her methyltransferase activity, which (Batalyzes transfer of a methyl group from one of the two stereoisomers of methylphosphotriesters in alkplated DNA to a cysteine residue in the N-terminal half of the protein (Margison et al., 1985: Mc(:arthy & LindahI, 1985). On the basis of the amino acid sequence homology, it was suggested that the cysteine residue at position 69 may be the mththyl n,ccept,or site (Teo rt al., 1986). If such is indeed the case. t.hen the Ada protein whose et
cysteine residue at position 69 is converted to another amino acid could not accept t,he methyl group from methylphosphotriester, but would retain the ability to accept the methyl group from 06-methylguanine. In this work, we constructed such a mutant form of Ada protein by site-directed mutagenesis of the cloned ada, gene, and found that this is actually the case. It has been found that the cysteine 69 residue accepts the methyl group from methylphosphotriesters, as determined by peptide mapping (Sedgwick et aE., 1987). In a parallel experiment, we constructed Ada protein whose cysteine residue at position 321 is converted to alanine, thereby supporting the notion that this cysteine residue is the acceptor sit)e for t,he methyl group from 06-methylguanine. Using the mutant, we demonstrated that the methyl group of 04-methylthymine is also transferred to the cysteine at position 321. The availability of mutant Ada proteins carrying only one of the methyl acceptor sites would provide a unique opportunity to investigate the biological significance of repair of specific lesions. By introducing the mutant ada gene into ada- cells, we demonstrated that the cysteine residue at position 321, but not that at position 69, is required for suppression of alkylating agent-induced mutagenesis. This is in accord with the notion t,hat 06-methytguanine would be the primary mutagenic lesion produced by alkylating agents (Loveless. 1969). We found that cells producing Ada protein with alanine at position 69 are as resistant to alkylating agents as ada+ cells producing normal Ada protein. Since Ada protein with alanine at position 321 had no such effect on ada- cells, it is likely that 06-methylguanine and 04-methylthymine, the repair of which is handled by the cysteine residue at position 321, are prematagenic
270
K. Takano
and lethal t’o cells. Evidence that) 06-methylguanine is a major lethal as well as mutagenic lesion has been obtained with mammalian cells (Brennand 8r Margison, 1986; Ishizaki et al., 1987). To elucidate the role of specific methylation of Ada protein in regulation of expression of the inducible genes, we performed in-vitro transcription experiments. Using a reconstituted syst’em with purified components, we demonstrated that met,hylation of the cysteine residue at position 69 is responsible for activation of the Ada protein, as a transcriptional regulator. The Ada protein with the alanine residue at, position 69 was inactive for promoting in-vitro transcription of the ada gene. whereas the Ada protein with alanine at position 321 retains the full activity for ada transcription. This is consistent with the proposal of Teo ef al. (1986), that transfer of the methyl group from methylphosphotriester of DNA to Ada protein is a trigger for induction of the adaptive response. \\‘e developed a system to measure the activitv of Ada protein to induce the ada and alkA expression in viva (see Fig. 8), and obtained results t,hat conform in part with those of experiments Zn vitro. The Ala69 protein is unable to induct expression of the ada gene in vivo. However. in the case of the Ala321 protein we obtained an unexpected result. Cells carrying the Ala321 protein exhibited high levels of expression of the ada gene, even though they had not been exposed to methylating agents. The apparent discrepancy from the results of experiment,s in vitro may be explained by the amplification in vivo of relatively small changes that cannot’ be detected in vitro or by the presence of unknown factor(s) that may be involved in t’he transcriptional regulation in ZGO. The result obtained in vivo suggests that conversion of cysteine at position 321 to alanine would ca,use some subtle conformational alteration of Ada protein so as to make for more ready accessto the promoter regions. This may relate to t~heobservation t’hat a deletion that truncated the 3’ end of t’he ada gene resulted in the production of Ada derivatives constitutively act’ivated for transcription (Lemotte & Walker, 1985; Nakabeppu et al.. 19856; Shevell & Walker, personal communication). The involvement of the C-terminal domain in gene activation was also suggested by analysis of the ada mutants (Temple, 1986). Based on the results obtained in t,his and previous studies (Nakabeppu & Sekiguchi, 1986: Teo et al., 1986), a model for the activation of Ada protein is depicted (Fig. 9). The cysteine residue at position 321 would accept the met’hyf group from both 06-methylguanine and 04-methylthymine of alkylated DNA, whilst. the cyst’eine residue at position 69 would accept the methyl group from methylphosphotriester of the DNA. The t*wo reactions occur independently and lead to different) biological consequences; the former is responsible for removing premutagenic and lethal lesions from the DNA, and the latter for converting Ada protein to an active form as a transcriptional regulator.
et al.
1 Proteolytic cleovoge
Tronscrlptlonol octlvotlon of the ado reqlon
Repolr of lethal and premutagenic lesions
Figure 9. A model for the actions of Ada prot,ein. I’“. G” and T” st,and for methylphosphotriestcr, 06-n~rthylguanine and 04-methylthymine. respectively.
Ada protein is cleaved by a cellular proteinase(s) to produce smaller polypeptides consisting of the N-terminal and the (‘-terminal half of the protein. \Z’r recently obtained evidence that the cleaved products carry no act,ivity to promote transcription of crdn but do ret,ain activity to promote a&4 t,ranscription in citm (Voshikai et al.. unpublished results). So. it is likely that both the N-terminal half and the (I-terminal moiety of the protein are required for transcription, ;tnd this ma! relate to our observation that modification of the cysteine residue at position 321 affect’s the tran scription
patterns
in
viva.
We thank $1. Ohara for comments on the manuscript. This work has been supported by grants from the Ministry of Education, Science and (‘ult~urr (61065007) and from t)he Jnstitute of Protein Engineering.
References Barnes. LV. Il., Hevan, M. & Son. I’. H. (IYX3). M~th,ods Enzymol. 101. 98-122. Brennand. J. h Margison. G. I’. (1986). (‘ctrcinc~tenesis. 7. 18.5188. Cathcart. R. & (ioldthwait, I). A. (1981). Hiochpmistry 20. 273 -286. (Ihang, A. C. Y. & (‘ohen. S. N. (1978). .I Hrxc~eriol. 134. 1141-I 156. (‘lose, T. ,J. & Rodriguez, R. 1,. (198%). (&r~c. 20. X&316. Demple, H. (1986). Nu,rl. ilcids Res. 14, 557.55589. Temple, B., ~Jacobsson. A.. Olsson. M., Robins. 1’. & Lindahl. T. (1982). J. Riol. Uum. 257. 13776~~13780. Temple, R.. Sedgwick. ES., R,obins. P.. Totty. S.. Waterfield. M. I). B Lindahl. T. (1985). I’roc. .\:tsl. Acad. Sri., IY3.A. 82. 2688-269~. Evensen. G. & Seeberg. E. (1982). Rjlt~re flondon). 296. 773-775. Ishizaki. K.. Tsujimura. ‘I‘.. F’ujio. (’ Yangpf~l. %.
E. coli Ada Regulatory Yawata, H Nakabeppu, Y., Sekiguchi, M. & Ikenaga, M. (1987). Mutat. Res. 184, 121-128. ,Jeggo. I?. (1979). J. Bacterial. 139, 783-791. Karran, I’., Hjelmgren, T. & Lindahl, T. (1982). Nature (London), 296, 77&773. Lemotte. I’. K. & Walker, G. C. (1985). J. Bactehol. 161, 888- 895. Loveless. A. (1969). Nature (London), 223, 206-207. Maniatis. T., Frimch, E. F. & Sambrook, J. (1982). In Molecular (Toning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor. NY. Margison. G. P.. Cooper, D. P. & Brennand.
Protein
271
Nakabeppu, Y., Kondo, H., Kawabata, S.. Iwanaga, S. & Sekiguchi, M. (1985a). J. Biol. Chem. 260, 7281-7288. Nakabeppu, Y.. Mine, Y. & Sekiguchi, M. (19856). Mutat. Res. 146, 155-167. Sakamura. T.. Tokumoto. Y.. Sakumi. K.. Koikr. (i.. Kakabeppu, Y. & Srkiguchi. M. (198X). .1. llfol. Hiol. in the press. Sanger, F.. Nicklen, S. & Coulson, A. R. (I 977). Proc. Xat. Acad. Sci.. U.S.A. 74. 5463-5467. Sedgwick. B. (1983). Mol. Gen. Genet. 191. 46fGi72. Sedgwick, B., Robins, P., Totty, K’. & Lindahl. T. (1988). J. Biol. Chem. in the press. Sekiguchi. M. & Nakabeppu, Y. (1987). Trends Genet. 3. 51-54. Shaw. W. V. (1975). Methods Enzymol. 43. 737~-755. Teo, I. (1987). Mutat. Res. 183, 123-127. Teo, I.. Sedgwick, B.. Demple, B.. Li. B. & Lindahl, T. (1984). EMBOJ. 3. 2151-2157. Teo, I., Sedgwick, B. Kilpatrick. M. W., McCarthy, T. V. & Lindahl, T. (1986). Cell, 45. 315-324. Vieira, J. & Messing. J. (1982). Gene, 19, 259-268. Yamamoto. Y. & Sekiguchi, M. (1979). Mol. Ben. Genet. 171. 251-256.
by K. Matsubara