Download pdf - Quasi-Adaptive Response to Alkylating Agents in Escherichia coli: A New Phenomenon

1022-7954/05/4105- © 2005 Pleiades Publishing, Inc.0484

Russian Journal of Genetics, Vol. 41, No. 5, 2005, pp. 484–489. Translated from Genetika, Vol. 41, No. 5, 2005, pp. 607–613.Original Russian Text Copyright © 2005 by Vasilieva, Moschkovskaya.

INTRODUCTION

Potentially mutagenic alkylating agents occur bothin the environment and in the living cell, as metabolicproducts. An important source of endogenous alkyla-tion is bacterial nitrosation of amino acids and peptideswith endogenous nitric oxide [1, 2]. All organisms—eubacteria, archaebacteria, and eukaryotes—havemechanisms protecting cells from toxic and genotoxiceffects of alkylation [3]. In

Escherichia coli

, the protec-tive mechanisms involve the

tag

and

ogt

genes, whichare expressed constitutively, and the

ada, alkA

,

alkB

,and

aid

genes, expression of which is associated withinduction of an adaptive response (Ada) [4]. The prod-ucts of the

tag

and

alkA

genes act as glycosylases and,although slightly differing in specificity, both exciseN

3

-meA from DNA to yield apurinic sites [5].The

alkB

gene belongs to the same operon as the

ada

gene, and its expression is similarly controlled bythe

ada

promoter [6]. The AlkB protein is involved inrepairing lesions induced by SN2-methylating agents insingle-stranded DNA, mostly in the replication forkand in transcribed regions [7]. The AlkB protein is iron-dependent and utilizes a unique mechanism of oxidativedemethylation to directly repair N

1

-meA and N

3

-meC,eliminating the methyl group in the form of formalde-hyde [8].

The multifunctional protein AidB is regulated viatwo pathways: one is Ada-dependent, while the other isAda-independent and realized in response to acidifica-tion of the medium [9, 10]. The

aidB

gene is an alarmgene of the

ada

operon: it is expressed to a lower levelin response to nitrosomethylurea (NMU) as compared

with the other genes and is the only gene providing forrepair of interstrand crosslinks induced in DNA bychlorine-substituted nitrosoalkylurea (ACNU) [11].

Among all alkylation products of DNA bases, O

6

-meG occupies a special place because of its preferentialpairing with T rather than C during DNA replication[12, 13]. Though a minor adduct, O

6

-meG plays a cru-cial role in mutagenesis, carcinogenesis, apoptosis, andthe clustogenic effect of methylating and chlorethylat-ing agents. Moreover, O

6

-meG is a lethal lesion, and itselimination depends absolutely on the postreplicativemismatch repair system. Although O

6

-meG is the mostdangerous cyto- and genotoxic adduct, only one or twoproteins repair it without mistakes via a unique irrevers-ible stochiometric reaction [3]. In mammals, this func-tion is performed by O

6

-meG DNA-methyltransferase(MGMT), whose amount is strongly controlled andmaintained at 10

5

molecules per cell [14].

A nonadapted

E. coli

cell contains approximately60 molecules of the Ada protein. However, completeinduction of this protein takes place within 1 h of incu-bation in the presence of an alkylating agent used in anontoxic dose [15]. In addition,

E. coli

cells constitu-tively express the product of the

ogt

gene to approxi-mately 30 molecules per cell. This protein is an alterna-tive of MGMT, is similar to the Ada protein in biochem-ical characteristics, but is not induced by alkylatingagents [16]. The Ada sensory protein, which regulatestranscription of the Ada regulon, is not an enzyme: Adadirectly repairs O

6

-meG and O

4

-meT via an irreversiblestochiometric reaction [3]. Since direct structural stud-ies of the protein–DNA complex are still unfeasible,

Quasi-Adaptive Response to Alkylating Agents in

Escherichia coli

: A New Phenomenon

S. V. Vasilieva and E. Ju. Moschkovskaya

Emanuel Institute of Biochemical Physics, Russian Academy of Sciences, Moscow, 117334 Russia;fax: (095)137-41-01; e-mail: [email protected]

Received April 30, 2004

Abstract

—An original hypothesis of a quasi-adaptive response to nitrosomethylurea (NMU) in

Escherichiacoli

cells was verified experimentally. In contrast to the true Ada response, which is induced in cells pretreatedwith a sublethal dose of NMU, a quasi-adaptive response was induced using NO-containing dinitrosyl ironcomplex with glutathione (DNIC

glu

). Quasi-adaptation increased expression of the Ada regulon and cell resis-tance to the cytotoxic and mutagenic effects of NMU. The levels of

alkA, alkB,

and

aidB

gene expression inquasi-adaptation were higher than in the true Ada response. Thus, experimental evidence was obtained for thealternative mechanism regulating the function of the Ada sensory protein in controlling expression of the Adaregulon during the adaptive response. The free iron—chelating agent

o

-phenanthroline (OP) facilitated degra-dation of DNIC

glu

(by electron paramagnetic resonance (EPR) spectra) and considerably or completely inhib-ited gene expression in the quasi-adaptive response. The new phenomenon extends the functional range of NOcompounds to include a role in genetic signal transduction within the Ada response system in addition to similarroles in the SoxRS, SOS, and OxyR systems in

E. coli.

GENETICS OF MICROORGANISMS

RUSSIAN JOURNAL OF GENETICS

Vol. 41

No. 5

2005

QUASI-ADAPTIVE RESPONSE TO ALKYLATING AGENTS 485

several models have been advanced to explain the con-version of alkylated Ada into a factor regulating tran-scription of the Ada regulon genes [3] and repair ofO

6

-meG [17].

Today, a crystal structure is available and known indetail only for one Ada protein,

E. coli

Ada [18]; struc-tural analysis of its mammalian counterpart is close tocompletion [3]. Since adaptive response proteins ofmammals and

E. coli

are similar in general structure,have homologous C-terminal domains, and are regu-lated similarly [19], bacterial cells can be used as amodel to study the common features of the molecularmechanisms underlying the interaction of adaptiveresponse proteins with potential inductors and to iden-tify new promising regulators of adaptive processes.Since cysteine methylation at SH groups is the crucialfactor of Ada activation, we assumed that the Ada activ-ity can be alternatively regulated using NO compounds,with S-nitrosylcysteine functioning in place of S-meth-ylcysteine in key position Cys-69. In this work, weexperimentally verified our assumption.

MATERIALS AND METHODS

Bacterial strains.

We used

E. coli

strains MV1571

(

alkA

51

::

Mu

-

dl

) (

bla

lac

),

MV

1601 (

alkB

52

::

Mu

-

dl

)(

bla

lac

)

, and MV2176

(

aidB

1

::

Mu

-

dl

) (

bla

lac

)

, whichwere constructed and kindly provided by M. Volkert(University of Massachusetts, United States). Thesestrains contain the

lacZ

structural gene for

β

-galactosi-dase under the control of the

ada

promoter, while thechromosomal

lac

operon is deleted. Thus, expression ofthe genes under study was assayed indirectly by the

β

-galactosidase activity, which was measured colorimet-rically. Genetic characteristics of the strains weredescribed in [20].

Reagents.

We used

o-

nitrophenyl

β

-D-galactopyra-noside (ONPG), cysteine, glutathione, and HEPES(Sigma, United States). A dinitrosyl iron complex withglutathione (DNIC

glu

) was used as a dimer and wasobtained by treating 5.4 mM FeSO

4

and 10.8 mM glu-tathione (iron–thiol ratio 1 : 2 (w/w)) with gaseousnitric oxide in solution (15 mM HEPES, pH 7.6) in aTunberg flask at 200–300 mm Hg. NMU was synthe-sized at the Institute of Chemical Physics and dissolvedin a phosphate buffer (pH 6.0). In addition, we used

E. coli

β

-galactosidase (5000 units, Sigma).

Gene expression.

Expression of the

alkA

,

alkB

, and

aidB

genes was assayed according to a published pro-tocol [21]. Bacterial cells were treated with inductorsfor 30 min. The cell suspension was diluted 50-foldwith LB, supplemented with the chromogen ONPG,and incubated at 37

°

C for 2 h. The buffer for

β

-galac-tosidase activity assays was as described in [8]. The

β

-galactosidase activity was measured colorimetrically at420 nm, using a Specord UV–VIS spectrophotometer(Germany), and computed as

E

= 1000

×

D

420

/

t

, where

D

is the absorbance at 420 nm and

t

is the time of incu-bation in the presence of the chromogen.

Adaptive response

to NMU was induced by treatingcells with

10

–5

M NMU (a sublethal concentration) inthe liquid L medium with aeration at 37

°

C for 1 h. Toinduce a quasi-adaptive response, cells were pretreatedwith DNIC

glu

used at a sublethal concentration in thedark. In some experiments, cells were exposed at thesublethal NMU concentration before treatment withDNIC

glu

.

UV absorption spectra

of NMU and DNIC

glu

solu-tions and their mixtures were recorded with theSpecord UV–VIS spectrophotometer.

Electron paramagnetic resonance (EPR) study.

Cells were grown aerobically in the L medium toOD

6

00

≈

0.4

. To prepare a sample, a culture was centri-fuged at 7000

g

and concentrated to 5 ml after 30-min pre-treatment with DNIC

glu

and, when indicated,

o

-phenan-throline (OP). Cells were collected by centrifugation,resuspended in 0.3 ml of the L medium, and quicklyfrozen in calibrated ampules for EPR examination. EPRspectra were recorded in the X range, using a Radiopanspetrometer (Poland) at modulation amplitude 5 mT,temperature 77 K, and microwave power 5 mW.

β

-Galactosidase activity assays with a cell-free sys-tem

were performed with the chromogenic substrateONPG in the presence of NMU and DNIC

glu

. A sample(2 ml) was incubated at 37

°

C for 1 h, combined with0.4 ml of buffer B containing ONPG [21, 22], and incu-bated for 20 min until it yellowed. The reaction was ter-minated by adding 1 M Na2CO3 [21]. All samples freefrom DNICglu were tested for D420 spectrophotometri-cally against the buffer. Samples containing DNICglu,which is yellowish in solution, were tested against0.2 mM DNICglu.

RESULTS

At the first step to verification of our hypothesis ofquasi-adaptation, we compared the level of inducedexpression of the adaptive response genes for the trueadaptive and quasi-adaptive responses to NMU. Cellswere exposed at sublethal concentrations of NMU orDNICglu. Preliminary experiments showed that DNICgluhad a relatively low cytotoxic effect on E. coli: its sub-lethal concentrations upon a 30-min exposure wereestimated at 0.1 mM for the alkA mutant, 0.2 mM forthe aidB mutant, and 0.5 mM for the alkB mutant. Anal-ysis of the NMU dose dependences showed that thelevel of expression of the Ada regulon genes increasedconsiderably both in the classical and in the quasi-adap-tive response. In the presence of 1.0 mM NMU, the lev-els of gene expression increased twofold in the trueadaptive response and 3.5-fold in the quasi-adaptiveresponse in alkA::lacZ and alkB::lacZ E. coli cells(Figs. 1, 2). The highest quasi-adaptive response wasobserved in cells with a high content of the Ada proteinafter preliminary adaptation at a sublethal concentra-

486

RUSSIAN JOURNAL OF GENETICS Vol. 41 No. 5 2005

VASILIEVA, MOSCHKOVSKAYA

tion of NMU. The level of alkB gene expression in suchcells was more than fivefold higher than in control cellsand more than twofold higher than in true adaptation toNMU (Fig. 2).

Regardless of adaptation, the aidB gene is usuallyexpressed to a far lower level than the alkA and alkBgenes in experiments with NMU [11]. As shown in Fig. 3,the highest level of aidB expression was observed at0.2 mM NMU in intact cells and at 1 mM, a fivefoldhigher concentration, in adapted cells. Still expressionof the aidB gene increased in all variants with preadap-tation.

By definition, the phenotypic expression of the Adaresponse to alkylating agents in E. coli involves anincrease in the resistance of adapted cells to toxic andmutagenic affects of such agents [4]. Experiments onquasi-adaptation were carried out with the E. colialkB::lacZ strain and showed that cell resistanceincreased more than in true adaptation, at least in termsof the parameters under study (Table 1).

The cytotoxic effects of NO and its compounds aremostly associated with inactivation of enzymes in thecell. Since we used the inducible β-galactosidase activ-ity to indirectly characterize the level of gene expres-sion, control experiments were performed to estimatethe effects of MNU and DNICglu on the β-galactosidaseactivity in a cell-free system. The inductors were usedat concentrations allowing the maximum levels of geneexpression (Table 2). Neither NMU nor DNICgluaffected the β-galactosidase activity in these experi-ments.

To achieve our aim, cells were exposed to a com-bined effect of the genotoxic agents MNU and DNICgluin most experiments. In view of this, we studied theinteraction of the two agents by UV absorption spectraof their solutions and mixtures (Fig. 4). Present in a

solution, DNICglu had no effect on the UV absorptionspectrum of NMU. It is of interest that NMU andDNICglu had similar absorption peaks in the region of200–250 nm. EPR spectroscopy was used to check thepermeability of E. coli cells to DNICglu and to study thechanges in DNICglu concentration and structure in thecell. DNICglu is converted from the dimeric into themonomeric form and displays paramagnetic propertiesin the presence of free thiols in solution or upon bindingwith SH groups of proteins. Cells incubated withDNICglu showed an EPR spectrum characteristic of thisagent, with anisotropic g-factors g| = 2.03 and g|| =

30

25

20

15

10

5

0.2 0.4 0.6 0.8 1.0

β-Galactosidase activity, MU

NMU, mM0

1

2

3

4

5

Fig. 1. NMU-dependent induction of alkA gene expressionin E. coli MV 1571. Cells were used (1) before and afterpretreatment with (2) 0.01 mM NMU; (3) 0.1 mM DNICglu;(4) 0.01 mM NMU 0.1 mM DNICglu; or (5) OP variants 1, 3, and 4.

30

25

20

15

10

5

0.2 0.4 0.6 0.8 1.0


NMU, mM0

1

2

3

4

5

40

35

Fig. 2. NMU-dependent induction of alkB gene expressionin E. coli MV 1601. Cells were used (1) before and afterpretreatment with (2) 0.01 mM NMU; (3) 0.5 mM DNICglu;(4) 0.01 mM NMU 0.5 mM DNICglu; or (5) OP variants 1, 3, and 4.

30

25

20

15

10

5

0.2 0.4 0.6 0.8 1.0


NMU, mM0

1

2

34

35

Fig. 3. NMU-dependent induction of aidB gene expressionin E. coli MV 2176. Cells were used (1) before and afterpretreatment with (2) 0.01 mM NMU; (3) 0.2 mM DNICglu;(4) 0.01 mM NMU 0.2 mM DNICglu.



2.014 (Fig. 5). A single washing of the cells with water(variant 2) halved the signal intensity on average (Fig. 5),testifying that a considerable portion of the donor wasabsorbed on the cell surface and lost as a result of cellwashing. Preincubation of cells with 0.1 mM OP for30 min reduced tenfold the intensity of the EPR signalat g| = 2.03 in cells exposed to DNICglu as comparedwith the initial level. The shape of the spectrum alsochanged, suggesting decomposition of DNICglu. Whencells pretreated with OP were washed once with water(variant 4), the intensity of the EPR signal at g| = 2.03increased twofold, to the level characteristic of variant 2.OP decreased expression of the alkA gene to the controllevel in all variants of treatment with DNICglu (Fig. 1).Expression of the alkB gene was suppressed by OP inall variants (Fig. 2).

DISCUSSION

The adaptive response to alkylating agents is uniquein the total set of intricate repair processes in E. coli.The Ada protein acts as a chemosensor in genetic signaltransduction, with its cysteine residues accepting thealkyl groups from alkylated DNA. On the other hand,the Ada protein is directly involved in removing thealkyl group from O6-alkG and restoring the native DNAstructure. Studies of the crystal structure of MGMT andanalysis of models of MGMT binding with alkylatedDNA have made it possible to assume that O6-alkG, themajor target of this enzyme, is flipped out from theDNA chain and positioned in the binding pocket, whichharbors the acceptor cysteine. This residue—Cys-321in E. coli and Cys-145 in human MGMT—is involvedin an extended hydrogen-bonding network, including ahighly organized water molecule and several aminoacid residues [17]. The cysteine residue is thereby acti-vated and attacks O6-alkG, producing alkylcysteine andrestoring the DNA structure. The S-alkylcysteine resi-due formed in the active site of MGMT is not trans-formed back into cysteine; i.e., every MGMT moleculefunctions only once. Alkylated MGMT changes its con-formation and is rapidly degraded by the ubiquitin–pro-teasomal system [3, 17].

When molecular genetic models were constructedfor O6-metG repair by MGMT (Ada) and the processproved to be conserved among various organisms [23],studies were initiated to search for potential agents reg-

ulating the MGMT activity, in particular, in bacterialcells. The most potent inhibitor of human MGMT,O6-benzylG, was constructed on the basis of theoreticalcomputations as a pseudosubstrate of MGMT [24, 25].The inhibitor is efficiently taken up by mammaliancells. Used at 0.2 mM, O6-benzylG inactivates humanMGMT in the culture medium by 90% in 10 min [26].This agent has come into clinical use as a component ofchemotherapy with alkylating carcinolytics, notwith-standing its high toxicity for marrow cells [23]. MGMTinhibitors are partly specific; e.g., the E. coli Ada pro-tein is far less sensitive to O6-benzylG than humanMGMT. Yet O6-allylG, another derivative of guanine, ishighly effective both with MGMT and with the Adaprotein [26].

Interaction with NO usually leads to inactivation ofvarious cell proteins and enzymes, including cysteineproteases, alcohol dehydrogenase, protein kinases, andhuman MGMT [27]. In E. coli, NO acts similarly to thesuperoxide anion , although through another mech-anism, to activate the SoxRS and OxyR regulons [28].Early activation of the SoxR protein proceeds throughtwo steps and involves generation of low-molecular-weight DNIC primers, which interact with thiol groupsof the protein as components of an iron–sulfur complexand facilitate disintegration of [2Fe–2S] clusters in theSoxR protein [29]. Mass spectrometry and titration ofthiol disulfides showed that the intramolecular S=Sbond is formed between Cys-199 and Cys-208 of theOxyR tetramer upon its oxidation [30]. Thus, it is directoxidation of the OxyR protein that underlies the controlof cell sensitivity to oxidative stress by the OxyR regu-lon. In addition to activation of the SoxRS and OxyRregulons, DNICglu induces a quasi-adaptive response toNMU in E. coli cells when used in place of NMU for

O2–

Table 1. Increase in NMU resistance of E. coli alkB::lacZ cells in adaptation

MNU in testexposure, M

Cell survival, % Arg+ revertants per 106 surviving cells

controlpretreatment

controlpretreatment

MNU DNICglu NMU DNICglu

0.01 1 15.5 19.5 4.6 1.8 1.3

0.02 6 12.0 16.75 40 9.5 5.0

0.04 0.16 2.3 2.5 380 21 16.5

Table 2. Effects of NMU and DNICglu on the β-galactosi-dase activity in vitro

Sample composition D420

Buffer B 0.25

Buffer B + β-galactosidase 1.8

Buffer B + β-galactosidase + 0.2 mM DNICglu 1.6

Buffer B + β-galactosidase + 1 mM NMU 1.9

488


VASILIEVA, MOSCHKOVSKAYA

pretreating cells. This finding suggests a greater rangeof functions played by nitric oxide in genetic signaltransduction and, on the other hand, a new mechanismfunctionally activating the Ada chemosensor to controlexpression of the Ada regulon. There is still no evi-dence that the Ada function in repairing O6-alkG is acti-vated in quasi-adaptation, although we did observe adecrease in the frequency of Arg+ revertants (Table 1).Generation of Arg+ revertants in E. coli strain AB1157argE3-oc, an ancestor of E. coli MV1601 alkB::lacZ[9], is due to at least four alternative mechanisms, whileonly rare suppressor conversion from supE44-am tosupB-oc result from transition GC AT, which ischaracteristic of the adaptive response [31]. Note that,in general, it has not been proved so far that NMU-induced mutagenesis is associated with predominantinduction of GC AT transitions, which are oftencaused by other alkylating agents, in particular, methylmethanesulfonate.

The aidB gene is selectively activated to a consider-able extent by chloro derivatives of nitrosourea [11].However, this gene showed only weak expression inexperiments with NMU regardless of the mode of celladaptation (Fig. 3). It is possible that the Ada proteinmust be accumulated to a higher concentration to bindto the aidB promoter than to the alkA or alkB promoter;i.e., that Ada has low affinity for the Ada-binding site ofthe aidB gene [20].

Iron-chelating agents modulate (increase ordecrease) the toxic effect of free radicals in biologicalsystems, testifying to an important role of iron ions in

toxicity [15, 16]. In our experiments, low concen-trations of OP already prevented quasi-adaptation in thealkA strain possibly because of complex decomposi-tion. Inhibition of alkB expression by OP can beexplained by Fe2+ dependence of the AlkB protein [8].

O2–

Partial decomposition of DNICglu was also evidentfrom the EPR spectra of DNICglu-treated cells exposedto OP (Fig. 5, curves 3, 4). The EPR signal intensity atg = 2.03 decreased tenfold and the shape of the spec-trum changed upon cell exposure to OP (curve 3). Incontrast, the intensity of this signal slightly increasedand a characteristic peak at g = 2.014 became undetectablewhen cells were washed with water (Fig. 5, curve 4). Fur-ther studies are necessary to establish the structure ofthe resulting product.

To summarize, we grounded theoretically and werethe first to demonstrate experimentally a new mecha-nism regulating the activity of the Ada sensory proteinin the quasi-adaptive response by NO-containing com-plex of [Fe2(NO2+)2] with glutathione. While S-methyl-cysteine generated from Cys-69 in the acceptor site ofthe Ada protein initiates adaptation in the true adaptiveresponse, S-nitrosylcytosine formed from Cys-69 playsthe key role in quasi-adaptive response. First, this testi-fies that the Ada response is not selectively induced by

1.6

1.0

0.6

0.4

0.2

200 250 300 350 400

Adsorption, D

Wavelength λ, nm

01

23

4

1.4

1.2

0.8

150

Fig. 4. UV adsorption spectra of aqueous solutions of(1) 0.01 mM NMU, (2) 0.2 mM DNICglu, (3) 0.04 mMDNICglu, and (4) 0.01 mM NMU + 0.04 mM DNICglu (1 : 1)after 30-min incubation.

1

2

3

4

g| = 2.03

g|| = 2.014

5 mT

Fig. 5. EPR spectra of E. coli MV 1601 cells incubated with(1, 2) 0.2 mM DNICglu (1) without or (2) with a singlewashing or (3, 4) 0.1 mM OP 0.2 mM DNICglu(3) without or (4) with a single washing.



alkylating agents in E. coli. Second, our finding extendsthe functional range of NO as a signal molecule ingenetic signal transduction and DNA repair.

ACKNOWLEDGMENTS

We are grateful to M. Volkert for the E. coli strains.This work was supported by the Russian Foundationfor Basic Research (project no. 04-04-48234a).

REFERENCES1. Taverna, P. and Sedgwick, B., Generation of an Endoge-

nous DNA-Methylating Agent by Nitrosation in Escher-ichia coli, J. Bacteriol., 1996, vol. 178, pp. 5105–5111.

2. Garcia-Santos, M.P., Calle, E., and Casado, J., AminoAcid Nitrosation Products As Alkylating Agents, J. Am.Chem. Soc., 2001, vol. 123, pp. 7506–7510.

3. Pegg, A.E., Repair of O6-Alkylguanine by Alkyltrans-ferases, Mutat. Res., 2000, vol. 462, pp. 83–100.

4. Samson, I. and Cairns, J., A New Pathway for DNARepair, Nature, 1977, vol. 267, pp. 281–283.

5. Teo, I., Sedgwick, B., Kilpatrik, M., et al., The Intracel-lular Signal for Induction of Resistance to AlkylatingAgents in E. coli, Cell (Cambridge, Mass.), 1986, vol. 45,pp. 315–324.

6. Lindahl, T., Sedgwick, B., Sekiguchi, M., et al., Regula-tion and Expression of the Adaptive Response to Alky-lating Agents, Annu. Rev. Biochem., 1988, vol. 57,pp. 133–157.

7. Dinglay, S., Trewick, S.C., Lindahl, T., et al., DefectiveProcessing of Methylated Single-Stranded DNA byE. coli alkB Mutants, Genes Dev., 2000, vol. 14,pp. 2097–2105.

8. Falnes, P.O., Johansen, R.F., and Seeberg, E., A ThirdMechanism for DNA Damage Reversal Catalysed by theAlkB Protein in E. coli, Abstracts of the 32nd AnnualMeeting of EEMS, Poland, 2002, p. 115.

9. Landini, P., Hajec, L.I., and Volkert, M.R., Structure andTranscriptional Regulation of the Escherichia coliAdaptive Response Gene aidB, J. Bacteriol., 1994,vol. 276, pp. 6583–6589.

10. Smirnova, G.V., Oktyabrsky, O.N., Moshonkina, E.V.,et al., Induction of the Alkylation-Inducible aidB Geneof Escherichia coli by Cytoplasmic Acidification and N-Ethylmaleimide, Mutat. Res., 1994, vol. 314, pp. 51–56.

11. Vasil’eva, S.V., Makhova, E.V., and Moshkovskaya, E.Yu.,Expression and Functions of Adaptive Response Genesin Escherichia coli Treated with Mono- and BifunctionalAlkylating Agents: Interference with SOS Response,Rus. J. Genet., 1999, vol. 35, no. 4, pp. 364–369.

12. Loveless, A., Possible Relevance of O6-Alkylation ofDeoxyguanosine to the Mutagenicity and Carcinogenic-ity of Nitrosamines and Nitrosamides, Nature, 1969,vol. 223, pp. 206–207.

13. Singer, B. and Essigmann, J.M., Site-Specific Mutagen-esis: Retrospective and Properties, Carcinogenesis,1991, vol. 12, pp. 949–955.

14. Foote, R.S., Pal, B.C., and Mitra, S., Quantitation of O6-Methylguanine-DNA Methyltransferase in HeLa Cells,Mutat. Res., 1983, vol. 119, pp. 221–228.

15. Samson, L., Thomale, J., and Rajewsky, M.F., Alterna-tive Pathways for the In Vivo Repair of O6-Methylgua-nine and O4-Alkylthymine in E. coli: The AdaptiveResponse and Nucleotide Excision Repair, EMBO J.,1988, vol. 7, pp. 2261–2267.

16. Potter, P.M., Kleibl, K., Cawkwell, L., et al., Expressionof the ogt Gene in Wild-Type and ada Mutants of E. coli,Nucleic Acids Res., 1989, vol. 17, pp. 8047–8060.

17. Daniels, D.S. and Tainer, J.A., Conserved StructuralMotifs Governing the Stoichiometric Repair of Alky-lated DNA by O6-Alkylguanine-DNA Alkyltransferase,Mutat. Res., 2000, vol. 460, pp. 151–163.

18. Moore, M.H., Gulbus, J.M., Dodson, E.J., et al., CrystalStructure of a Suicidal DNA Repair Protein: The AdaO6-Methylguanine-DNA Methyltransferase from E. coli,EMBO J., 1994, vol. 13, pp. 1495–1501.

19. Kleibl, K., Molecular Mechanisms of AdaptiveResponse to Alkylating Agents in Escherichia coli andSome Remarks on O6-Methylguanine-DNA Methyl-transferase in Other Organisms, Mutat. Res., 2002,vol. 512, pp. 67–84.

20. Volkert, M., Adaptive Response of Escherichia coli toAlkylation Damage, Environ. Mol. Mutagen., 1988,vol. 11, pp. 241–255.

21. Quillardet, P. and Hofnung, M., The SOS Chromotest, aColorimetric Bacterial Assay for Genotoxins: Proce-dures, Mutat. Res., 1985, vol. 147, pp. 65–78.

22. Miller, J.H., Experiments in Molecular Genetics, ColdSpring Harbor, New York: Cold Spring Harbor Lab.,1972.

23. Pegg, A.E., Boosalis, M., Samson, L., et al., Mechanismof Inactivation of Human O6-Alkylguanine-DNA Alkyl-transferase by O6-Benzylguanine, Cancer Res., 2002,vol. 62, pp. 3037–3043.

24. Dolan, M.E., Chae, M.Y., Pegg, A.E., et al., Metabolismof O6-Alkylguanine-DNA Alkyltransferase, CancerRes., 1994, vol. 54, pp. 5123–5130.

25. Dolan, M.E. and Pegg, A.E., O6-Benzylguanine and ItsRole in Chemotherapy, Clin. Cancer Res., 1997, vol. 3,pp. 837–847.

26. Pegg, A.E., Chung, L., and Moschel, R.C., Affect ofDNA on the Inactivation of O6-Alkylguanine-DNA Alkyl-transferase by 9-Substituted O6-Benzylguanine Deriva-tives, Biochem. Pharmacol., 1998, vol. 53, pp. 5265–5271.

27. Liu, L., Xu-Welliver, M., Kanugula, S., et al., Inactiva-tion and Degradation of O6-Alkylguanine-DNA Alkyl-transferase after Reaction with Nitric Oxide, CancerRes., 2002, vol. 62, pp. 3037–3043.

28. Gaudu, P., Moon, N., and Weiss, B., Regulation of theSoxRS Oxidative Stress Regulon, J. Biol. Chem., 1997,vol. 272, no. 8, pp. 5082–5086.

29. Vasil`eva, S.V., Stupakova, M.V., Lobysheva, I.I., et al.,Activation of the Escherichia coli SoxRS Regulon byNitric Oxide and Its Physiological Donors, Biochemistry(Moscow), 2001, vol. 66, no. 9, pp. 984–988.

30. Storz, G. and Imlay, J.A., Oxidative Stress, Curr. Opin.Microbiol., 1999, vol. 2, pp. 188–194.

31. Vasil’eva, S.V. and Davnichenko, L.S., Selective Effectof p-Aminobenzoic Acid on Mutagenesis: PhenotypicAnalysis of Arg+ Revertants Induced by N-Nitrosometh-ylurea in Escherichia coli K-12 AB1157, Genetika(Moscow), 1983, vol. 19, no. 11, pp. 1916–1919.