Increased spontaneous mutation and alkylation sensitivity of ...

3 downloads 0 Views 2MB Size Report
6. Increased Spontaneous Mutation and Alkylation Sensitivity of. Escherichia coli Strains Lacking the ogt 06-Methylguanine. DNA Repair Methyltransferase.
JOURNAL

OF

BACTERIOLOGY, Mar. 1991,

p.

2068-2076

Vol. 173, No. 6

0021-9193/91/062068-09$02.00/0 Copyright © 1991, American Society for Microbiology

Increased Spontaneous Mutation and Alkylation Sensitivity of Escherichia coli Strains Lacking the ogt 06-Methylguanine DNA Repair Methyltransferase G. WILLIAM REBECK AND LEONA SAMSON* Laboratory of Toxicology, Harvard School of Public Health, 665 Huntington Avenue, Boston, Massachusetts 02115 Received 30 October 1990/Accepted 18 January 1991

Escherichia coli expresses two DNA repair methyltransferases (MTases) that repair the mutagenic 06-methylguanine (06MeG) and 04-methylthymine (04MeT) DNA lesions; one is the product of the inducible ada gene, and here we confirm that the other is the product of the constitutive ogt gene. We have generated various ogt disruption mutants. Double mutants (ada ogt) do not express any 06MeG/04MeT DNA MTases, indicating that Ada and Ogt are probably the only two 06MeG/04MeT DNA MTases in E. coli. ogt mutants were more sensitive to alkylation-induced mutation, and mutants arose linearly with dose, unlike ogt+ cells, which had a threshold dose below which no mutants accumulated; this ogt+-dependent threshold was seen in both ada' and ada strains. ogt mutants were also more sensitive to alkylation-induced killing (in an ada background), and overexpression of the Ogt MTase from a plasmid provided ada, but not ada&, cells with increased resistance to killing by alkylating agents. The induction of the adaptive response was normal in ogt mutants. We infer from these results that the Ogt MTase prevents mutagenesis by low levels of alkylating agents and that, in ada cells, the Ogt MTase also protects cells from killing by alkylating agents. We also found that ada ogt E. coli had a higher rate of spontaneous mutation than wild-type, ada, and ogt cells and that this increased mutation occurred in nondividing cells. We infer that there is an endogenous source of 06MeG or 04MeT DNA damage in E. coli that is prevalent in nondividing cells. The ogt gene was originally cloned by using a functional assay for MTase activity and was shown to encode a 19-kDa O6MeG/O4MeT DNA MTase which bears regions of homology to the Ada MTase (30, 56). However, there has been no direct evidence that DNA MTase II, characterized in cell extracts, and the Ogt MTase are the same protein. Here we describe the generation of E. coli ogt mutants which, as expected, do not express DNA MTase II, and we characterize these mutant cells to determine some of the roles that this DNA repair MTase plays in protecting E. coli against the effects of alkylating agents.

06-Methylguanine (O6MeG) and 04-methylthymine (O4MeT) are mutagenic DNA lesions because they can base mispair during DNA replication (18, 31). DNA methyltransferases (MTases) that repair O6MeG and O4MeT lesions irreversibly transfer methyl groups from the methylated base to specific cysteine residues in the MTase (27). MTases that repair O6MeG have been found in many organisms, including bacteria (14, 25, 40), yeasts (38), insects (11), fish (26), and mammals (3), suggesting that 06MeG DNA damage is commonly encountered. There are two known 06MeG/ 04MeT DNA repair MTases in Escherichia coli, the Ada MTase (21, 40, 50) and DNA MTase II (32, 44). The 39-kDa Ada MTase is inducible by DNA alkylation damage (40) and repairs methylphosphotriester lesions in addition to 06MeG and 04MeT lesions (22). Methyl groups from methylphosphotriester lesions are transferred to Cys-69 of the Ada MTase (43), and methyl groups from 06MeG and 04MeT lesions are transferred to Cys-321 (8). Upon methylation at Cys-69, Ada becomes a transcriptional activator of the genes of the adaptive response, which consist of the adaalkB operon, the alkA gene, and the aidB gene (17, 51, 55). alkA encodes a DNA glycosylase which removes o2_ methylpyrimidines and 3-methylpurines (10, 13), lethal lesions that block DNA replication and thus cause cell death. The functions of the alkB and aidB gene products remain unknown (17). The induction of the adaptive response after exposure to alkylating agents protects E. coli from both the mutagenic and the lethal effects of alkylation damage (37). The 19-kDa DNA MTase II is constitutively expressed in E. coli and is thought to be the product of the ogt gene (30).

*

MATERIALS AND METHODS E. coli strains. All strains used are listed in Table 1. P1 transduction (23) was used for transfer of the ada-JO::TnlO (16) and ogt-J::Kanr alleles into various strains. All strains were grown in LB or in M9 salts supplemented with glucose and the required amino acids (23). Isolation and disruption of the ogt gene. To clone the ogt gene, we synthesized a 20-nucleotide oligomer (5'-GTAATC TTAGCATTATTGAT-3') that is present in the published sequence of ogt but not present in ada (30). This oligonucleotide was 32P end labeled and used to probe an E. coli genomic DNA library in XSE6 (9) (American Type Culture Collection, Rockville, Md.) by standard techniques (20). One positive X clone was isolated, and from this clone a 3.1-kb EcoRI DNA fragment carrying the ogt gene was subcloned into a modified pUC vector lacking a HindIII site to create pUCogt (Fig. 1A). A gene for kanamycin resistance was isolated on a HindIll fragment (Pharmacia, Piscataway, N.J.) and inserted into a unique HindIII site in ogt to generate the ogt-J::Kanr allele (see Results). The 3.1-kb

Corresponding author. 2068

VOL. 173, 1991

IN VIVO ROLE OF Ogt DNA METHYLTRANSFERASE

TABLE 1. Strains Strain

manipulations (DNA restriction, ligation, purification, and Southern blots) were conducted as described previously (20). Chromosomal mapping of ogt. The ogt-J::Kanr allele was mapped by using a collection of strains containing genetically linked Tetr markers (45). This collection consisted of seven Tetr Hfr strains for initial mapping to a 10-min region of the chromosome and 95 P1 mapping strains with Tetr elements at approximately 1-min intervals. DNA was transferred from the different Hfr strains to ogt-J :Kanr E. coli;

Reference

Genotype

or source

V355

recF143 his-4 met rpsL31 argA recD1014 GWR105 As V355 plus ogt-]::Kanr AB1157 K-12, argE3 hisG4 leuB6 proA2 thr-J ara-14 galK2 lacYlI mtl-l xyl-J thi-J rpsL31 supE44 tsx-33 GW7101 As AB1157 plus Aada-25 GWR111 As AB1157 plus Aada-25 ogt-J:: Kanr GW5352 As AB1157 plus ada-JO::TnlO GWR107 As AB1157 plus ogt-J::Kanr MV1902 As AB1157 plus rfa-550 alkA105: :ApSG1 GWR161 As MV1902 plus ada-JO::TnJO GWR162 As MV1902 plus ogt-J::Kanr GWR163 As MV1902 plus ada-lO::TnJO ogt-l:: Kanr F26 B/r, his thy GWR110 As F26 plus ogt-J::Kanr BS23 As F26 plus Aada-alkB GWR109 As BS23 plus ogt-J::Kanr Mapping strains Tetr

2069

44

This work

44 This work

these cells were scored for kanamycin sensitivity to determine whether the ogt-J:: Kanr allele was replaced during Hfr transfer. P1 transduction was then conducted from ogt-1::Kanr cells to 10 Tetr P1 mapping strains, which were selected for Kanr and scored for tetracycline sensitivity, thus localizing the ogt-J::Kanr to less than 1 min of the E. coli chromosome. O6MeG DNA MTase assay. E. coli strains were grown in LB to log phase, harvested by centrifugation, and disrupted by sonication in 50 mM N-2-hydroxyethylpiperazine-N'-2ethanesulfonate (pH 7.8)-10 mM dithiothreitol-1 mM EDTA-5% glycerol. Cell extract proteins (50 ,ug or 1 mg) were incubated with Micrococcus Iysodeikticus DNA (ICN, Costa Mesa, Calif.) alkylated with [3H]methylnitrosourea (20 Ci/mmol; Amersham, Arlington Heights, Ill.) as described previously (32). Proteins were separated on a sodium dodecyl sulfate-12% polyacrylamide gel (20), which was then cut into 2-mm slices, incubated in Scintilene (Fisher, Pittsburgh, Pa.) with 5% protosol (NEN, Wilmington, Del.), and scintillation counted.

16 This work 55

This work This work This work This work 42 This work 45

EcoRI DNA fragment carrying ogt was also subcloned into the EcoRI site of pBR322 to generate pBRogt. Plasmid pDS400, a pBR322 derivative containing the ada gene (43a), was the gift of D. Shevell and G. Walker. Further genetic

A

H

B

p

\-

J

H

2

1

ogt pUCogt E~~~~~~~~~~~~~~~~

3

4

5

6

7

8

4.8 kb _

H

3.2 kb

-k,anr pUCogt::kanr E

C 100/

E

H

H kan

E r

75

25

tag\~

~K7 \

tranisformn.colti recD

cells

/

A

acda-alkB

c)~~~~~gt

alkA

FIG. 1. Generation of ogt E. coli. (A) Strategy for the generation of an ogt::Kanr allele in vitro and introduction into recD E. coli. (B) Southern blot analysis of EcoRI-digested DNA from wild-type and Kanr E. coli, probed with 32P-labeled ogt DNA. Lanes: 1, AB1157; 2, GWR107 (ogt); 3, GW7101 (Aada); 4, GWR111 (Aada ogt); 5, F26; 6, GWR110 (ogt); 7, BS23 (Aada-alkB); 8, GWR109 (Aada-alkB ogt). (C) Map position of ogt as determined by Hfr crosses and P1 transductions using ogt::Kanr E. coli and Tetr mapping strains.

2070

REBECK AND SAMSON

Induction of the adaptive response. Induction of the ada gene was measured by immunoblot analysis of the Ada protein, and induction of the alkA gene was measured by monitoring ,B-galactosidase production in strains containing a chromosomal alkA-lacZ fusion (55). (i) Western immunoblot analysis. Log-phase E. coli in minimal medium were treated with 1 pxg of N-methyl-N'nitro-N-nitrosoguanidine (MNNG) per ml, and aliquots of cells were removed at various times. Cells were harvested by centrifugation and resuspended in MTase buffer (see above), and cell extracts were prepared as described previously (32). Samples (200 ,ug) of cell extract proteins were separated by 12% polyacrylamide gel electrophoresis, transferred to nitrocellulose, and probed with Ada polyclonal antiserum (0.5% [vol/vol] serum). Antibody binding was detected by horseradish peroxidase enzyme activity linked to goat anti-rabbit immunoglobulin G (Bio-Rad, Richmond, Calif.). (ii) B-Galactosidase measurements. E. coli alkA4-lacZ fusion strains were exposed to 1 ,ug of MNNG per ml, samples were removed at the indicated times, and P-galactosidase was measured, correcting for cell optical density at 600 nm as described previously (23). Alkylation-induced mutation and cell killing. To measure mutation, log-phase his E. coli growing in LB were exposed to MNNG for 15 min. The cells were then concentrated 20-fold into M9 salts (23), diluted, and plated on either minimal plates or minimal plates lacking histidine to estimate the number of surviving cells and the number of his' revertants, respectively. Mutation frequencies are expressed as the number of induced his+ revertants per 108 surviving cells. To measure cell killing, log-phase cells growing in LB were exposed to MNNG or methyl methanesulfonate (MMS) for the indicated times, after which samnples were removed, diluted in M9 salts, and spread on LB plates to estimate cell survival. Alternatively, relative sensitivities were estimated by using MNNG gradient plates. Gradient plates were prepared by pouring 30 ml of complete M9 agar containing 0 to 10 ,ug of MNNG per ml into a square petri dish at an angle of about 60 from the horizontal; after the agar wedge solidified, 30 ml of complete M9 agar was added to the horizontal dish, and the plate was used within 2 h of solidifying. Overnight cultures were spread across the gradient in M9 top agar, and the plates were incubated for 2 days at 37°C. Survival was estimated as percent growth across the gradient. Spontaneous mutation. We used two types of mutation assay, one in which cells were grown on plates (the plate assay) and one in which cells were grown in liquid cultures and then transferred to plates (the liquid assay). (i) Plate assay. Approximately 107 his E. coli were spread on minimal plates containing 10 ,uM histidine to allow limited growth of the his E. coli. Plates were incubated at 37°C, and his+ revertant colonies were scored every day after plating. (ii) Liquid assay. Thirty 3-ml cultures, inoculated with approximately 105 his E. coli, were grown to 2 x 108 cells per ml at 37°C with aeration. In one case, 10 cultures were inoculated (see Table 2). Small aliquots were removed from at least three cultures, diluted, and plated on minimal plates containing histidine or on LB plates to estimate the final cell density. The cultures were then harvested by centrifugation, resuspended in 0.5 ml of M9 salts, and plated on minimal plates lacking histidine (to measure reversion to his+) or on LB plates with 100 jig of rifampin per ml (to measure forward mutation to Rif). The mean number of mutations per culture (m) was calculated from the median number (r) of

J. BACTERIOL.

mutants per distribution, according to the equation (rim) ln(m) = 1.24 (15). Mutation rates per generation were recorded as m/N, with N representing the average number of cells per culture.

RESULTS Generation of an ogt disruption mutant. To study the in vivo role of DNA MTase II in E. coli, we generated a mutant strain deficient in this DNA repair protein. Since it seemed likely that the ogt gene identified by Potter et al. (30) encodes DNA MTase II, we generated an ogt disruption mutant. The ogt gene (cloned as described in Materials and Methods), present on a 3.1-kb EcoRI DNA fragment, was disrupted in vitro by insertion of a 1.6-kb DNA cassette encoding kanamycin resistance (Fig. 1A). A 4.7-kb EcoRI DNA fragment containing the disrupted ogt gene was introduced into E. coli recD cells, and Kanr derivatives were selected. The Kanr allele was then transferred from the recD strain into wildtype and ada E. coli B/r and K-12 strains by P1 transduction. Southern blot analysis of EcoRI-digested genomic DNA confirmed that the ogt gene was disrupted in the Kanr mutants (Fig. 1B; Table 1); as expected, the fragment that hybridized to the ogt probe increased from 3.2 to 4.8 kb in kanamycin-resistant strains bearing the 1.6 kb Kanr cassette insertion. The Kanr marker in the ogt gene was used to map the location of ogt in the E. coli chromosome. Using a set of Tetr E. coli consisting of seven Hfr strains and 95 P1 mapping strains (45; see Materials and Methods), we mapped the ogt gene between 29.1 and 29.3 min (Fig. 1C). This position is in agreement with the recently reported position of ogt as 200 bp 5' of the fnr gene, which maps at 29.4 min (28). MTase activity in ogt mutants. DNA repair MTases are irreversibly methylated by the transfer of methyl groups from damaged DNA, and so they can be identified and quantified by incubating cell extracts with DNA containing 06-[3H]MeG lesions and then separating the proteins by polyacrylamide gel electrophoresis. Most of the DNA MTase activity in wild-type E. coli was due to expression of the 19-kDa DNA MTase II; only very low levels of 39-kDa Ada MTase were present (33; Fig. 2A). Extracts of ogt-J: :Kanr E. coli did not contain the 19-kDa DNA MTase in either the ada or ada+ background (Fig. 2B and C), suggesting that ogt does indeed encode DNA MTase II. Thus, DNA MTase II will henceforth be referred to as the Ogt MTase. Since ada ogt E. coli cell extracts showed no DNA MTase activity (Fig. 2B), Ogt appears to be the only constitutively expressed O6MeG DNA MTase in E. coli. The level of 39-kDa Ada is not dramatically different in ogt E. coli than in wild-type cells (Fig. 2C and A), suggesting that the Ogt MTase may not influence the level of active Ada MTase. Sensitivities of E. coli ogt cells to alkylating agents. As expected, both ogt and ada ogt E. coli strains were more sensitive to MNNG-induced mutation than were wild-type cells (Fig. 3). ada ogt+ cells showed a threshold in their resistance to MNNG; at less than 0.3 ,ug of MNNG per ml no mutations were induced, but beyond this dose there was dramatic mutation induction (Fig. 3A). In contrast, mutation induction in ada ogt E. coli was linear with dose, to even very low doses of MNNG (Fig. 3A). Similarly, ada+ ogt+ cells exhibited a threshold in their resistance to MNNG; again, at less than 0.3 ,ug of MNNG per ml no mutations were induced, but beyond this dose some mutations were induced (Fig. 3B). ada+ ogt E. coli were slightly more

VOL. 173, 1991

IN VIVO ROLE OF Ogt DNA METHYLTRANSFERASE 39

19

2071

2000

L 0

1600

L

1200

O

800

N-

aL)

u

U, L

Q)

C 0)

E U6

0

4

","

400 0 0

0.3

0.7

1.0

pg/ml MNNG 0

10 20 30 Slice Number FIG. 2. O6MeG DNA MTase activity. Cell extract proteins were incubated with 06-[3H]MeG DNA substrate and separated by 12% polyacrylamide gel electrophoresis. MTase activity was determined by measuring tritium in 2-mm slices of the gel. Slice 1 is the top of the gel; arrows mark the positions of 39- and 19-kDa proteins. (A) 50 ,ug of AB1157 (wild type); (B) 1 mg of GWR111 (Aada ogt); (C) 1 mg

of GWR107 (ogt).

sensitive to mutagenesis than were ada+ ogt+ cells, and mutation induction was linear with increasing MNNG dose (Fig. 3B). Similar results were observed after treatment of E. coli K strains with MNNG, monitoring reversion of a lacZ allele which can revert only by a G. C-to-A. T transition (7): ogt cells were more sensitive to mutation than were ogt+ cells in both ada and ada+ backgrounds (data not shown). The increased sensitivity of ogt ada+ E. coli to mutagenesis by alkylating agents could be due to an altered induction of the adaptive response in these cells. We therefore measured the kinetics of induction of the adaptive response in two different ways: ada gene expression was monitored by Western blot analysis using Ada polyclonal antiserum, and alkA expression was monitored by the induction of 3-galactosidase activity in an alkA-lacZ fusion strain. The immunoreactive 39-kDa Ada MTase was induced in both wildtype (Fig. 4A, top panel) and ogt E. coli (Fig. 4A, bottom panel) within 20 min of MNNG addition (1 ,ug/ml) and continued to accumulate over time (the difference in band intensities between blots is due to slightly different colordeveloping conditions). Likewise, the kinetics of alkA induction were similar in MNNG-treated wild-type and ogt cells (Fig. 4B); as expected (55), alkA was not induced in ada E. coli (Fig. 4B). Furthermore, the kinetics of alkA induction during alkylation exposures ranging from 0.1 to 1.0 ,ug of MNNG per ml were similar in wild-type versus ogt cells (data not shown). In summary, Ogt-deficient E. coli does not appear to be altered in its ability to induce the adaptive response.

Since O6MeG and 04MeT do not appear to block DNA replication (44), they are not thought of as highly lethal

FIG. 3. MNNG-induced mutagenesis. his E. coli were grown to log phase in LB, treated with MNNG at the indicated doses for 15 min, concentrated in M9 salts, and plated on minimal plates lacking histidine (for mutants) and on minimal plates with histidine (for survivors). (A) BS23 (Aada-alkB; 0) and GWR109 (Aada-alkB ogt; 0); (B) F26 (wild type; *) and GWR110 (ogt; Oi).

lesions in E. coli (10, 41, 54). Thus, we did not expect ogt mutants to be much more susceptible than wild-type cells to killing by MNNG or MMS. Indeed, the ogt-J mutation did not sensitize ada+ E. coli to killing by MNNG (Fig. 5A) or MMS (Fig. 5B). However, in an ada genetic background, the ogt-l mutation did sensitize cells to killing by both MNNG (Fig. 5C) and MMS (Fig. 5D). Thus, the Ogt MTase provides ada cells with protection against alkylation-induced cell

killing. It has been reported that O-alkyl repair by the Ada MTase does not provide E. coli with protection against alkylationinduced cell killing and that Ada provides killing resistance only via the induction of the AlkA 3-methyladenine DNA glycosylase (10, 17, 41, 54). For example, overexpression of the Ada MTase in alkA E. coli does not confer resistance to alkylation-induced killing (10, 41, 54). To confirm this result and to determine whether protection by the Ogt MTase also requires a functional alkA gene, we measured MNNGinduced killing in various alkA E. coli strains. Surprisingly, we found that ada alkA E. coli was more sensitive than ada+ alkA E. coli (Fig. 6A), implying that the Ada MTase might provide killing resistance directly as well as through the induction of alkA. The sensitivity of ada alkA E. coli to MNNG-induced killing was not due to disruption of the alkB gene, since overexpression of AlkB from a plasmid did not restore resistance (data not shown). Expression of the Ada MTase or the Ogt MTase from plasmids did not provide ada+ alkA E. coli with greater resistance to killing (Fig. 6B), consistent with the findings of the earlier studies (10, 41, 54). Only when Ada or Ogt was overexpressed in an ada genetic background was increased resistance to killing observed (Fig. 6C). Sensitivity of ogt mutants to spontaneous mutation. It has been suggested that endogeneous cellular compounds such

2072

J. BACTERIOL.

REBECK AND SAMSON

A

B

39

900

. Ln

_z

600 39

I-

300 O"

0_

0 0

1 0 20 30 45 60 75

0

30

60

90

L

CD

Minutes after 1 jig/ml MNNG FIG. 4. Induction of the adaptive response. (A) F26 (wild type; top panel) and GWR110 (ogt; bottom panel) E. coli were treated with 1 ,ug of MNNG per ml, and samples were removed at the indicated times. Cell extract proteins were separated by 12% polyacrylamide gel electrophoresis, transferred to nitrocellulose, and probed with Ada antiserum. The positions of the 39- and 19-kDa forms of Ada are indicated at the left. (B) MV1902,/n alkA::XpSG1 E. coli strain (E), and the same strain with an ogt (O) and an ada (0) mutation were treated with 1 ,ug of MNNG per ml and monitored for P-galactosidase production.

as S-adenosylmethionine and nitrosamines produce a continual flux of DNA alkylation damage in E. coli (1, 35, 52). We therefore determined whether decreased O-alkyl repair in ogt mutants caused an increase in spontaneous mutation. E. coli B/r his cells were plated on minimal media containing 10 ,uM histidine and incubated at 37°C. Cells could divide until the histidine was depleted and thus produced a faint lawn of cells; his' revertants subsequently formed distinct colonies on the faint lawn of his E. coli. Wild-type, ada, and ogt strains produced about the same number of mutant colonies in this plate assay. In contrast, the MTase-deficient ada ogt strain produced over three times this number of mutant colonies (Fig. 7A). Similarly, lacZ E. coli K cells were plated with lactose as the only carbon source to

meastre spontaneous mutation to lacZ+ by a G C-to-A T transition (7); again, ada ogt double mutants accumulated significantly more mutants than did wild-type, ada, and ogt strains (10a). These data suggest that there is indeed an endogenous source of mutagenic DNA alkylation damage in E. coli and that this damage is normally repaired by either -

-

100

50

._-

c

0) L

(D

0 100

U, U, 0 L 0

50

4L 0)

0

C

U) c

L.0)

0 100

L

L

0. 50

0

MNNG Gradient 0

10 30 0 20 Minutes in MNNG

20 40 60 Minutes in MMS

FIG. 5. Alkylation-induced killing. E. coli were grown to log phase in LB, treated with an alkylating agent for indicated times, diluted in M9 salts, and plated on LB plates. F26 (wild type; *) and GWR110 (ogt; O) were treated with 20 ,ug of MNNG per ml (A) or 0.20% MMS (B); BS23 (Aada-alkB; 0) and GWR109 (Aada-alkB ogt; 0) were treated with 5 ,ug of MNNG per ml (C) or 0.05% MMS (D).

(Mg/ml)

FIG. 6. MNNG gradient plates. alkA E. coli strains containing pBR322, pBRogt, or pDS400 (carrying the ada gene) were spread across MNNG gradient plates made with agar containing 0 to 10 pug of MNNG per ml. Survival was recorded as percent growth across the gradient plate. (A) GWR162 (ogt)(pBR322) (0) and GWR163 (Aada ogt)(pBR322) ([1); (B) GWR162(pBR322) (0), GWR162 (pBRogt) (-), and GWR162(pDS400) (0) (some symbols overlap); (C) GWR163(pBR322) (O), GWR163(pBRogt) (D), and GWR163 (pDS400) (0).

IN VIVO ROLE OF Ogt DNA METHYLTRANSFERASE

VOL. 173, 1991

1 2 3 4 5

10 1 100E

6

CL

a

e

40

r

20

7 Spntneu mttin.()PaesayAprxmtl

Ca.

,u\itdn0adicbtdt3°.hs100

platscnann 1-0

0

0

plte Days

at

10

MC

37

Mutants per

107 his E.

Spontaneous mutation. (A) Plate ccli (F26 [wild type; (], GWR110

alkBi;

and GWR1o9 [Aada-alkB 10 p.M histidine

FIG.

7.

0,

Approximately

assay.

[ogt; E], BS23 [Aadaogt; 0]) were plated on minimal

plates containing revertants

were

average

three plates.

cultures

culture

and

incubated

at

his'

37C.

day after plating. Values reflect the of (B and C) Liquid assay. Thirty separate of approximately BS23 his E. ccli (F26 [wild type; 01, and GWR1o9 [Aada-alkB ogt; 0]) were grown to

scored

every

3-ml

t]),

10o

[Aada-alkB;

concentrated and 108 cells per ml in M9 media. Cells plated on M9 plates lacking histidine (B) or LB plates containing 100

about 2

x

were

Bg of rifampin per ml (C); colonies the

the Ada or

spontaneous

Ogt

MTase.

were

scored after 3

days

at

37tC.

Interestingly, the number of

cgt cells steadily increased days (Fig. 7A), suggesting that revertants arose from his cells which were not actively we not accurately determine the histidine plates, we times the cells doubled on the could not accurately measure spontaneous mutation rates or the possibility that the ada cgt strain simply produced a faint lawn that contained more viable cells and hence more mutants. Therefore, we took a second approach to measuring spontaneous mutations. Ten to thirty 3-ml liquid cultures of each of the strains were grown from 10 to 6 x 108 cells, concentrated, and transferred to plates lacking over a

mutants

period

from

ada

of 6

dividing.

Since

number of

could

limiting

2073

same spontaneous mutation rate at the his locus, but the rate was two- to threefold higher in ada ogt E. coli (Table 2). The distribution of his' mutants in ada ogt cultures, in contrast to ada and wild-type cultures (Fig. 7B), did not follow the distribution predicted if mutants only arose during cell growth (15). The distribution of mutants in ada ogt E. coli indicated that mutants arose both before and after plating in the absence of histidine, and therefore, one should calculate spontaneous mutation rates for both dividing and nondividing cells. Using models that include the possibility that mutants could arise after plating (5), we estimate that ada ogt E. coli has the same rate of spontaneous mutation as wild-type E. coli during growth phase and that excess mutants arise only after the cells are plated. The conclusion that mutants arise in nondividing ada ogt cells is consistent with the results of the plate assay, which showed his' mutants continued to arise in ada ogt cells for several days after plating, presumably after the cells had ceased to divide (Fig. 7A). To confirm that the additional spontaneous mutants in ada ogt E. coli arose only in populations of nondividing cells and not in growing cells, we measured the spontaneous rate of forward mutations to rifampin resistance, known to be inducible by O-alkyl DNA damage (48). Cells that were not resistant to rifampin at the time of plating would be killed; thus, only preexisting antibiotic-resistant mutants that arose during the growth of the cultures could be scored. ada ogt cells did not produce rifampin-resistant mutants at a higher rate than did ada single mutants (Fig. 7C; Table 2). Similarly, the rate of spontaneous forward mutation to Tetr at a plasmid-borne locus (34) also showed no increase in ada ogt cells (data not shown). These data support the conclusion that the elevated rate of spontaneous mutation in ada ogt E. coli is confined to nondividing cells.

eliminate

histidine. A distribution of the tures of

wild-type, ada,

7B; the number of

his'

and ada

revertants

among cul-

cgt E. ccli is shown in

mutants per culture

ranged from

0

or

Fig. 1 to

greater than 100, reflecting both the number of mutations per culture and when the mutation occurred in the culture (19). Mutation rates per cell

were

growth of the

determined from

the median number of mutants and the average number of cells per culture (see Materials and Methods). Within each

experiment, wild-type, ada,

TABLE

2.

and

ogt

strains had about the

Mutation rates

Mutations (10-9)/cell

F26 GWR11O BS23 GWR109

Rifr

His+

Strain

Expt 1

Expt 2

Expt 3a

Expt 4

Expt 5

Expt 6

1.6 2.1 1.9

4.0 4.1 3.0

NDb

3.4C

9.1c

ND ND 11 11

ND ND 12 15

ND ND 71 65

ND 2.3 6.0

a Calculated from the distribution among 10 cultures; all others were from 30 cultures. bND, Not determined. c Significantly different (P < 0.01) from F26, GWR1l0, and BS23 within the same experiment.

DISCUSSION The isolation of ogt E. coli strains and the study of their resistances to alkylating agents has enabled us to reach the following conclusions about the roles that the Ogt MTase (DNA MTase II) plays in protecting cells from alkylation damage. Ogt provides resistance to alkylation-induced mutations via the repair of O"MeG and 04MeT lesions, but since Ogt is a suicide enzyme and is not inducible by DNA alkylation damage (29, 32), this resistance is limited. Ada can provide greater resistance to alkylation damage, but its induction is relatively slow (16), requiring at least 20 min, and in wild-type E. coli Ogt efficiently prevents mutation during this vulnerable period. In ada E. coli, the Ogt MTase provides resistance to mutation by low doses of alkylating agents. At higher doses (above 0.3 p.g of MNNG per ml), the molecules of Ogt are depleted and ada E. coli becomes essentially MTase deficient and thus extremely sensitive to mutation induction. Our finding that ada+ ogt cells are considerably more resistant than ada ogt to MNNG-induced mutation at less than 0.3 ,ug of MNNG per ml indicates that the Ada MTase provides protection even when cells are exposed to alkylation doses that are not optimal for the induction of the adaptive response (12, 39). However, these low levels of alkylation partially induce the adaptive response (4; data not shown), and in wild-type cells Ada and Ogt presumably compete for the repair of 06MeG and 04MeT lesions even at low levels of DNA alkylation. Both MTases have similar affinities for 06MeG lesions in vitro, but Ogt has a much higher affinity for 04MeT lesions than does Ada (38a, 56),

2074

REBECK AND SAMSON

suggesting that low levels of O4MeT are probably repaired predominantly by the Ogt MTase. The Ogt MTase also plays a role in protecting ada E. coli from alkylation-induced killing. Recently, it has been shown that DNA MTases from bacterial (48), yeast (37a), and mammalian (36, 49) cells can provide ada E. coli with killing resistance. Thus, it appears that, contrary to earlier reports (10, 41, 54), 06MeG or 04MeT is lethal in some contexts in E. coli. Killing of ada E. coli occurs at doses of MNNG approximately 10-fold higher than the dose required to inactivate the constitutive Ogt molecules (33; Fig. 3A), which are present at about 30 molecules per cell (24). Assuming that the production of DNA lesions is linear with dose, we estimate that at lethal doses of MNNG each treated cell is accumulating hundreds of 06MeG/04MeT lesions. Cell death could result from 0-alkyl lesions leading to lethal mutations or from the inhibition of DNA replication when the 0-alkyl lesions occur in certain parts of the genome, for example, the replication origin (2). In either case, as the cells incubate on the plate, the 0-alkyl lesion has the potential to be repaired to regenerate the original sequence in at least one DNA strand, after which at least one daughter could go on to survive. ogt+ ada cells are probably much more resistant to killing than are ogt ada cells because they continue to constitutively synthesize the Ogt DNA MTase long after treatment with alkylating agent, and this ongoing 0-alkyl repair could produce a survivor long after the initial alkylation treatment. In contrast, for protection against mutagenesis, repair of 0-alkyl lesions by MTase must occur before DNA replication, or the mutation will be fixed in one of the daughters. Thus, the mutational resistance of ogt+ cells must be provided by the 30 molecules of Ogt MTase present before alkylation of the genome plus any extra MTase molecules produced before the lesions are replicated. Finally, the Ogt MTase protects cells from spontaneous mutations which presumably result from 06MeG and 04MeT produced by endogenous alkylating agents. There are several possible sources of DNA alkylation damage in E. coli. S-Adenosylmethionine, a methyl donor for the enzymatic methylation of DNA, RNA, and protein, can act as a weak alkylating agent under physiological conditions (1, 35). Lipid peroxidation reactions can cause DNA alkylation damage through the generation of alkyl radicals (53). Endogenous nitrosation of amines can generate reactive methylating agents in anaerobically growing (52) or stationary (6) E. coli. Whatever the source, the levels of 06MeG and 04MeT formed seem to be relatively low because ogt mutants, which express extremely low levels of active Ada MTase, display the same spontaneous mutation rate as wild-type cells. However, numerous Ada molecules could have been consumed in the repair of endogenous alkylation damage, and these would not be counted among the very few active Ada molecules. Three lines of evidence suggest that the endogenous source of 06MeG and 04MeT becomes significant only in nondividing E. coli: (i) ada ogt cells continue to produce his+ revertants on minimal plates for several days after the plated bacteria cease to divide; (ii) the distribution of his+ mutants over a large number of independent ada ogt cultures has a significant Poisson component which would not be predicted if all of the mutants arose before plating (5); and (iii) ada ogt mutants do not have a higher rate of spontaneous mutation in two forward mutation assays that do not allow nonmutant nondividing cells to survive on selective media. Most spontaneous mutation is usually considered to be a consequence of rare errors in nucleotide insertion during DNA replication and thus to be dependent on cell

J. BACTERIOL.

growth. DNA damage in stationary E. coli could be the cause of these spontaneous mutations. If damaged bases were transcribed incorrectly, and the products of these mutant transcripts allowed cell growth and thus DNA replication, then DNA damages could be permanantly fixed as mutations (47). This mechanism could explain the generation of spontaneous mutants from O6MeG and 04MeT lesions in nondividing ada ogt E. coli. Overexpression of the Ogt MTase (data not shown) or the Ada MTase (42) in wild-type cells does not lower the rate of spontaneous mutation. Thus, while O6MeG and 04MeT lesions appear to be a significant source of spontaneous mutation in MTase-deficient E. coli, they do not appear to contribute to spontaneous mutation in wild-type E. coli. Wild-type levels of DNA repair MTases appear to be sufficient to protect cells from endogenous alkylating agents spontaneously produced by E. coli. ACKNOWLEDGMENTS Akatsu for her work on cloning of the ogt gene. We thank Haruko We thank John Cairns for his many helpful discussions and suggestions. We also thank Pat Foster, Mike Volkert, Carol Gross, Diane Shevell, and Graham Walker for providing strains and plasmids. This work was supported by research grant NP448 from the American Cancer Society and grant 1-PO1-ES03926 from the National Institute of Environmental Health Sciences. L.S. was supported by an American Cancer Society Faculty Research Award. G.W.R. was supported by a Pharmaceutical Manufacturers Association Foundation Advanced Predoctoral Fellowship in Pharmacol-

ogy/Toxicology. REFERENCES 1. Barrows, L. R., and P. N. Magee. 1982. Nonenzymatic methylation of DNA by S-adenosylmethionine in vitro. Carcinogenesis

3:349-351.

2. Bignami, M., and D. P. Lane. 1990. 06-methylguanine in the SV40 origin of replication inhibits binding but increases unwinding by viral large T antigen. Nucleic Acids Res. 18:3785. 3. Bogden, J. M., A. Eastman, and E. Bresnick. 1981. A system in mouse liver for the repair of 06-methylguanine lesions in methylated DNA. Nucleic Acids Res. 9:3089-3103. 4. Cairns, J. 1980. Efficiency of the adaptive response of Escherichia coli to alkylating agents. Nature (London) 286:176-178. 5. Cairns, J., J. Overbaugh, and S. Miller. 1988. The origin of mutants. Nature (London) 335:142-145. 6. Calmels, S., H. Ohshima, M. Crespi, H. Leclerc, C. Cattoen, and H. Bartsch. 1987. N-nitrosamine formation by microorganisms isolated from human gastric juice and urine: biochemical studies on bacteria-catalysed nitrosation, p. 391-395. In H. Bartsch, I. O'Neill, and R. Schulte-Hermann (ed.), The relevance of N-nitroso compounds to human cancer. IARC Scientific Publications, Lyon, France. 7. Cupples, C. G., and J. H. Miller. 1989. A set of lacZ mutations in Escherichia coli that allow rapid detection of each of the six base substitutions. Proc. Natl. Acad. Sci. USA 86:53455349. 8. Demple, B., B. Sedgwick, P. Robins, N. Totty, M. D. Waterfield, and T. Lindahl. 1985. Active site and complete sequence of the suicidal methyltransferase that counters alkylation mutagenesis. Proc. Natl. Acad. Sci. USA 82:2688-2692. 9. Elledge, S. J., and G. C. Walker. 1985. Phasmid vectors for identification of genes by complementation of Escherichia coli mutants. J. Bacteriol. 162:777-783. 10. Evensen, G., and E. Seeberg. 1982. Adaptation to alkylation resistance involves the induction of a DNA glycosylase. Nature (London) 296:773-775. 10a.Foster, P. Personal communication. 11. Green, D. A., and W. A. Deutsch. 1983. Repair of alkylated

VOL. 173, 1991

12. 13.

14. 15.

IN VIVO ROLE OF Ogt DNA METHYLTRANSFERASE

DNA: Drosophila have DNA methyltransferases but not DNA glycosylases. Mol. Gen. Genet. 192:322-325. Jeggo, P. 1979. Isolation and characterization of Escherichia coli K-12 mutants unable to induce the adaptive response to simple alkylating agents. J. Bacteriol. 139:783-791. Karran, P., T. Hjelmgren, and T. Lindahl. 1982. Induction of a DNA glycosylase for N-methylated purines is part of the adaptive response to alkylating agents. Nature (London) 296: 770-773. Kaufman, A., and G. C. Walker. 1990. A constitutive o6_ methylguanine-DNA methyltransferase of Rhizobium meliloti. Mutat. Res. 235:165-169. Lea, D. E., and C. A. Coulson. 1949. The distribution of the numbers of mutants in bacterial populations. J. Genet. 49:264-

285. 16. LeMotte, P. K., and G. C. Walker. 1985. Induction and auto-

17. 18. 19. 20. 21. 22. 23. 24. 25.

26.

27. 28. 29. 30.

31.

32.

33.

regulation of ada, a positively acting element regulating the response of Escherichia coli K-12 to methylating agents. J. Bacteriol. 161:888-895. Lindahl, T., B. Sedgwick, M. Sekiguchi, and Y. Nakabeppu. 1988. Regulation and expression of the adaptive response to alkylating agents. Annu. Rev. Biochem. 57:133-157. Loveless, A. 1969. Possible relevance of 0-6 alkylated deoxyguanosine to the mutagenicity and carcinogenicity of nitrosamines and nitrosamides. Nature (London) 233:206-207. Luria, S. E., and M. Delbruck. 1943. Mutations of bacteria from virus sensitivity to virus resistance. Genetics 28:491-511. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. McCarthy, T. V., P. Karran, and T. Lindahl. 1984. Inducible repair of 0-alkylated DNA pyrimidines in Escherichia coli. EMBO J. 3:545-550. McCarthy, T. V., and T. Lindahl. 1985. Methyl phosphotriesters in alkylated DNA are repaired by the Ada regulatory protein of E. coli. Nucleic Acids Res. 13:2683-2699. Miller, J. H. 1972. Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Mitra, S., B. C. Pal, and R. S. Foote. 1982. 06-methylguanineDNA methyltransferase in wild-type and ada mutants of Escherichia coli. J. Bacteriol. 152:534-537. Morohoshi, F., and N. Munakata. 1987. Multiple species of Bacillus subtilis DNA alkyltransferase involved in the adaptive response to simple alkylating agents. J. Bacteriol. 169:587592. Nakatsuru, Y., N. Nemoto, K. Nakagawa, P. Masahito, and T. Ishikawa. 1987. 06-methylguanine DNA methyltransferase activity in liver from various fish species. Carcinogenesis 8:11231127. Olsson, M., and T. Lindahl. 1980. Repair of alkylated DNA in Escherichia coli. J. Biol. Chem. 255:10569-10571. Potter, P. M., L. Harris, and G. P. Margison. 1989. Mapping of OGT in the E. coli chromosome. Nucleic Acids Res. 17:10505. Potter, P. M., K. Kleibl, L. Cawkwell, and G. P. Margison. 1989. Expression of the ogt gene in wild-type and ada mutants of E. coli. Nucleic Acids Res. 17:8047-8060. Potter, P. M., M. C. Wilkinson, J. Fitton, F. J. Carr, J. Brennand, D. P. Cooper, and G. P. Margison. 1987. Characterization and nucleotide sequence of ogt, the 06-alkylguanineDNA-alkyltransferase gene of E. coli. Nucleic Acids Res. 15:9177-9193. Preston, B. D., B. Singer, and L. A. Loeb. 1986. Mutagenic potential of 04-methylthymine in vivo determined by an enzymatic approach to site-specific mutagenesis. Proc. Natl. Acad. Sci. USA 83:8501-8505. Rebeck, G. W., S. Coons, P. Carroll, and L. Samson. 1988. A second DNA methyltransferase repair enzyme in Escherichia coli. Proc. Nati. Acad. Sci. USA 85:3039-3043. Rebeck, G. W., C. M. Smith, D. L. Goad, and L. Samson. 1989. Characterization of the major DNA repair methyltransferase activity in unadapted Escherichia coli and identification of a similar activity in Salmonella typhimurium. J. Bacteriol. 171: 4563-4568.

2075

34. Rewinski, C., and M. G. Marinus. 1987. Mutation spectrum in Escherichia coli DNA mismatch repair deficient (mutHf) strain. Nucleic Acids Res. 15:8205-8215. 35. Rydberg, B., and T. Lindahl. 1982. Nonenzymatic methylation of DNA by the intracellular methyl group donor S-adenosyl-Lmethionine is a potentially mutagenic reaction. EMBO J. 1:211216. 36. Rydberg, B., N. Spurr, and P. Karran. 1990. cDNA cloning and chromosomal assignment of the human 06-methylguanine-DNA methyltransferase. J. Biol. Chem. 265:9563-9569. 37. Samson, L., and J. Cairns. 1977. A new pathway for DNA repair in Escherichia coli. Nature (London) 267:281-283. 37a.Samson, L., and B. Derfler. Unpublished data. 38. Sassanfar, M., and L. Samson. 1990. Identification and preliminary characterization of an 06-methylguanine DNA repair methyltransferase in the yeast Saccharomyces cerevisiae. J. Biol. Chem. 265:20-25. 38a.Sassanfar, M., et al. J. Biol. Chem., in press. 39. Schendel, P. F., M. Defais, P. Jeggo, L. Samson, and J. Cairns. 1978. Pathways of mutagenesis and repair in Escherichia coli exposed to low levels of simple alkylating agents. J. Bacteriol. 135:466-475. 40. Schendel, P. F., and P. E. Robins. 1978. Repair of 06-methylguanine in adapted Escherichia coli. Proc. Natl. Acad. Sci. USA 75:6017-6020. 41. Sedgwick, B. 1983. Molecular cloning of a gene which regulates the adaptive response to alkylating agents in Escherichia coli. Mol. Gen. Genet. 191:466-472. 42. Sedgwick, B., and P. Robins. 1980. Isolation of mutants of Escherichia coli with increased resistance to alkylating agents: mutants deficient in thiols and mutants constitutive for the adaptive response. Mol. Gen. Genet. 180:85-90. 43. Sedgwick, B., P. Robins, N. Totty, and T. Lindahl. 1988. Functional domains and methyl acceptor sites of the Escherichia coli Ada protein. J. Biol. Chem. 263:4430-4433. 43a.Shevell, D. E. 1990. Ph.D. thesis. Massachusetts Institute of

Technology, Cambridge. 44. Shevell, D. E., A. M. Abou-Zamzam, B. Demple, and G. C. Walker. 1988. Construction of an Escherichia coli K-12 ada deletion by gene replacement in a recD strain reveals a second methyltransferase that repairs alkylated DNA. J. Bacteriol. 170:3294-3296. 45. Singer, M., T. A. Baker, G. Schnitzler, S. M. Deischel, M. Goel, W. Dove, K. J. Jaacks, A. D. Grossman, J. W. Erickson, and C. A. Gross. 1989. A collection of strains containing genetically linked alternating antibiotic resistance elements for genetic mapping of Escherichia coli. Microbiol. Rev. 53:1-24. 46. Snow, E. T., R. S. Foote, and S. Mitra. 1984. Base pairing properties of 06-methylguanine in template DNA during in vitro DNA replication. J. Biol. Chem. 259:8095-8100. 47. Stahl, F. 1988. A unicorn in the garden. Nature (London) 335:112-113. 48. Takano, K., Y. Nakabeppu, and M. Sekiguchi. 1988. Functional sites of the Ada regulatory protein of Escherichia coli. J. Mol. Biol. 201:261-271. 49. Tano, K., S. Shiota, J. Collier, R. S. Foote, and S. Mitra. 1990. Isolation and structural characterization of a cDNA clone encoding the human DNA repair protein for 06-alkylguanine. Proc. Natl. Acad. Sci. USA 87:686-690. 50. Teo, I., B. Sedgwick, B. Demple, B. Li, and T. Lindahl. 1984. Induction of resistance to alkylating agents in E. coli: the ada+ gene product serves both as a regulatory protein and as an enzyme for repair of mutagenic damage. EMBO J. 3:21512157. 51. Teo, I., B. Sedgwick, M. W. Kilpatrick, T. V. McCarthy, and T. Lindahl. 1986. The intracellular signal for induction of resistance to alkylating agents in E. coli. Cell 45:315-324. 52. Tsimis, J., and D. B. Yarosh. 1990. Adaptive response induction by bacterial catalysis of nitrosation. Environ. Mol. Mutagen. 15:69-70. 53. Vaca, C. E., J. Wilhelm, and M. Harms-Ringdahl. 1988. Interaction of lipid peroxidation products with DNA. A review. Mutat. Res. 195:137-149.

2076

REBECK AND SAMSON

54. Volkert, M. R., and D. C. Nguyen. 1984. Induction of specific Escherichia coli genes by sublethal treatments with alkylating agents. Proc. Natl. Acad. Sci. USA 81:4110-4114. 55. Volkert, M. R., D. C. Nguyen, and K. C. Beard. 1986. Escherichia coli gene induction by alkylation treatment. Genetics 112:11-26.

J. BACTERIOL. 56. Wilkinson, M. C., P. M. Potter, L. Cawkwell, P. Georgiadis, D. Patel, P. F. Swann, and G. P. Margison. 1989. Purification of the E. coli ogt gene product to homogeneity and its rate of action on 06-methylguanine, 06-ethylguanine and 04-methylthymine in dodecadeoxyribonucleotides. Nucleic Acids Res. 17:84758484.