Cm

Da

b

a

ARRAA

KSSSC

1

biyipi

0d

Mutation Research 686 (2010) 84–89

Contents lists available at ScienceDirect

Mutation Research/Fundamental and MolecularMechanisms of Mutagenesis

journa l homepage: www.e lsev ier .com/ locate /molmutCommuni ty address : www.e lsev ier .com/ locate /mutres

atabolite repression of SOS-dependent and SOS-independent spontaneousutagenesis in stationary-phase Escherichia coli

.G. MacPheea, M. Ambroseb,∗

School of Microbiology, La Trobe University, Bundoora, Victoria, AustraliaSchool of Biological Sciences, Faculty of Science, The University of Auckland, Private Bag 92019, Auckland 1142, New Zealand

r t i c l e i n f o

rticle history:eceived 24 September 2009eceived in revised form 27 January 2010ccepted 27 January 2010vailable online 6 February 2010

eywords:pontaneous mutagenesisOS responsetationary-phaseatabolite repression

a b s t r a c t

Previous work in our laboratory established that a spontaneous mutagenesis process operating instationary-phase Escherichia coli cells undergoing selection is subject to regulation by the global reg-ulatory mechanism known as catabolite repression (formerly also called glucose-repression). Here, weset out to determine the identity of this hitherto unknown catabolite-repressible spontaneous mutationgeneration mechanism(s). We used two different spontaneous mutation detection assays, reversion of aLac− (lacI33�lacZ) frameshift marker and forward mutation to valine-resistance, and tested the effectsof varying the nature of the carbon source(s) present in the selective plating medium on the mutabil-ity of bacterial cells carrying known defects in the recA, umuDC and dinB genes, three well-known SOSresponse genes, whose products are important for mutagenesis in E. coli. Consistent with the results ofour previous Lac− → Lac+ assay using otherwise SOS-proficient bacterial cells, we found that the over-all numbers of spontaneous Lac+ E. coli revertants were highest when the selective medium containedlactose and lowest when it contained lactose plus the non-utilizable but strongly catabolite-repressingglucose analogue, methyl-�-d-glucopyranoside (�MG). In contrast, we found that the numbers of Lac+

revertants appearing on the lactose and lactose + �MG selection plates were greatly diminished and notsignificantly different when the bacterial cells concerned carried either a �recA or �dinB mutation. Fur-thermore, introducing the �dinB mutant allele into bacterial cells over-expressing the recA gene reducedthe numbers of Lac+ mutations to those being recovered with the �dinB cells. These results appear tosuggest that (i) the DinB-dependent mutation generation pathway is alone responsible for spontaneousreversion of the lacI33�lacZ frameshift marker, and (ii) the varying numbers of Lac+ colonies that werecover on the lactose and lactose + �MG plates provide a direct measure of the differential effects of

these particular carbon compounds on the overall expression of the dinB gene. Interestingly, the yields ofspontaneous ValR mutations arising in wild-type, �recA, �dinB and �umuDC cells were found to be sim-ilar, but always tended to be highest when the medium contained only a non-repressing carbon source(glycerol) and lowest when it had been supplemented with a strong catabolite repressor such as glucose or�MG. Together, our results would seem to establish that stationary-phase E. coli cells exposed to strongselection pressures can accumulate spontaneous mutations via SOS-dependent and SOS-independent

ways

mutation generation path

. Introduction

In our attempts to understand how the physiological state ofacterial cells might influence their spontaneous mutability dur-

ng exposure to strong selection pressures, we found that the

ields of spontaneous mutations which could be recovered by plat-ng stationary-phase Escherichia coli cultures on minimal mediumlates were often greatly reduced when glucose was present

nstead of glycerol as a carbon source [1]. In addition, we found that

∗ Corresponding author. Tel.: +64 9 373 7599x84428; fax: +64 9 373 7416.E-mail address: m.ambrose@auckland.ac.nz (M. Ambrose).

027-5107/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.mrfmmm.2010.01.022

whose levels of expression are regulated by catabolite repression.© 2010 Elsevier B.V. All rights reserved.

the yields of spontaneous mutants that could be recovered on min-imal plates containing both glycerol and glucose were lower thanthe yields obtained on plates containing glycerol alone, as werethe yields on glycerol plates additionally supplemented with eitherglucose-6-phosphate or methyl-�-d-glucopyranoside (�MG). Onepossible explanation for these and similar previously reported find-ings is that glucose and certain related carbon compounds are ableto reduce spontaneous mutation yields by depressing intracellu-

lar concentrations of the small effector molecule cyclic adenosine3′:5′-monophosphate (cAMP) in cells of E. coli [2–6]. Based on theseand other experimental observations, we have proposed that somepart of the process leading to the generation and fixation of spon-taneous mutations in stationary-phase E. coli is likely to involve

utatio

asi[

poruw4silbaatctttee[ateDombapmd[Dmuo[

cn

TE

D.G. MacPhee, M. Ambrose / M

n inducible error-prone DNA repair mechanism(s) whose expres-ion is exquisitely sensitive to intracellular cAMP levels, and hences most likely moderated via cAMP-mediated catabolite repression1].

The most extensively studied inducible error-prone DNA repairathways in E. coli are those that are part of the SOS response of thisrganism. Expression of the SOS response is controlled by two keyegulatory proteins, LexA and RecA. LexA is the repressor, whichnder normal conditions binds the operators of all the SOS genesith high affinity, thereby ensuring that expression of more than

0 unlinked SOS genes, including the recA and lexA genes them-elves, is maintained at basal levels [7–9]. The SOS response isnduced upon damage to DNA that causes DNA replication arresteading to the generation of single-stranded DNA (ssDNA). RecAinds these regions of ssDNA (forming a so-called ssDNA-RecA fil-ment) and subsequently undergoes a conformational change tocoprotease, RecA*. The activated RecA* complex then facilitates

he self-cleavage of the LexA repressor, which in turn leads to aoordinated derepression of all the SOS genes and to the stimula-ion of a diverse set of cellular processes [8–10]. Whilst some ofhese damage responsive cellular events are related to the inhibi-ion of cell division (filamentation), recombination and so-calledssentially error-free DNA repair (nucleotide excision repair), oth-rs are involved in translesion DNA synthesis and mutagenesis8,9]. DNA damage which is not easily repaired and thereby actss a persistent block to DNA replication can and often will causehe induction of two SOS genes, namely, umuDC and dinB, whichncode the error-prone DNA polymerases, DNA polymerase V andNA polymerase IV, respectively [11,12]. Depending on the typef DNA damage and its sequence context, both of these SOS poly-erases are capable of facilitating DNA synthesis past replication

locking damage, the outcome of which may involve the gener-tion and fixation of mutations [11,12]. Thus for example, DNAolymerase V is well known for its participation in UV-inducedutagenesis by inserting random nucleotides across from pyrimi-

ine dimers, 6,4-photoproducts and noninstructional (abasic) sites8,9,12]. By contrast, DNA polymerase IV introduces mutations inNA by processing misaligned (primer/template) replication inter-ediates, and also appears to be involved in the phenomenon of the

ntargeted mutagenesis of phage � DNA which is not infrequently

bserved when undamaged � DNA infects UV-irradiated E. coli cells11,13].

In the present study we sought to determine whether theatabolite-repressible spontaneous mutation generation mecha-ism that we have previously found to be capable of operating

able 1scherichia coli K-12 strains.

Strain Description

FC40 ara �(lac-proB)�III thi RifR [F′lacI33�lacZ proAB+]SZ624 FC40 �(srl-recA)306::Tn10RSH61 FC40 recAo281 srlR310::Tn10

RW82 F−�− uvrA6 thi-1 thr-1 araD14 leuB6(am) lacY1 ilv-323(Ts) �(gpt-proA)62 mtl-1 xylA5 rpsL31(strR) tsx-33 supE44 glnV44 galK2(oc)hisG4(oc) rfbD1 kdgK51 sulA211 �(umuDC)595::cat

YG7207 F−�− thi-1 thr-1 araD14 leuB6(am) lacY1 ilv-323(Ts) �(gpt-proA)62mtl-1 xylA5 rpsL31(strR) txs-33 supE44 glnV44 galK2(oc) mgl-51Rac rpoS396 hisG4(oc) argE3(oc) rfbD1 kdgK51 �dinB::kan

DG2430 FC40 �(umuDC)595::catDG2431 FC40 �dinB::kanDG2432 FC40, pYG751 (dinB+)DG2434 RSH61 �dinB::kanPlasmidspYG751 dinB+, AmpR

a See Section 2 for details of strain construction.

n Research 686 (2010) 84–89 85

in stationary-phase E. coli cells is part of (or at the very leastresembles) an SOS-type inducible mutation generation pathway.To test this idea directly, we used two different spontaneousmutation detection assays: (i) reversion of a lac− (lacI33�lacZ)frameshift marker, and (ii) forward mutation to valine-resistance(ValR), and investigated how changing the nature of the primarycarbon source(s) present in the selective plating medium impactsthe mutability of bacterial cells carrying known defects in their recA,umuDC and dinB genes. First, we show that the overall numbersof spontaneous Lac+ E. coli revertants is highest when the selec-tive medium contains lactose and lowest when it contains lactoseplus the strong catabolite-repressing but non-utilizable glucoseanalogue, methyl-�-d-glucopyranoside (�MG). Second, we provideevidence showing that the DinB (DNA polymerase IV)-dependentmutation generation pathway is solely responsible for spontaneousreversion of the lacI33�lacZ frameshift marker, and infer that lac-tose and �MG are exerting their effects on the final numbers ofLac+ revertants appearing under the conditions of selection via theirability to differentially regulate expression of the dinB gene. Third,we show that ValR mutations in E. coli are generated by an SOS (andhence RecA)-independent mutation generation pathway whoseactivities also appear to be subject to cAMP-mediated cataboliterepression. These results imply that spontaneous mutations aris-ing in bacteria undergoing selection can be generated during theoperations of more than one inducible mutation generation pro-cess, one or more of which is SOS-dependent and one or more ofwhich is SOS-independent, and that their full and balanced expres-sion is likely to be determined in some important way(s) by subtlechanges in the physiology of the bacterial cells themselves.

2. Materials and methods

2.1. Bacterial strains

The E. coli strains used are listed in Table 1, and strain constructions are describedbelow. P1 transduction was performed using standard techniques [14]. StrainDG2430 was constructed by transducing �(umuDC)595::cat from RW82 into FC40and selecting for chloramphenicol resistance. Strains DG2431 and DG2434 wereconstructed by transducing �dinB::kan from YG7207 into FC40 and RSH61, respec-tively, and selecting for kanamycin resistance. Strain DG2432 was constructed bytransforming the multicopy plasmid pYG751 containing a functional dinB gene intoCaCl2-competent FC40 by the heat-shock method [15] and selecting for ampicillin-

resistant colonies.

2.2. Media

Bacterial strains were maintained on Oxoid Nutrient agar (Oxoid, West Heidel-berg, Victoria). When required, ampicillin (50 �g/ml), chloramphenicol (15 �g/ml),

Source or construction

P.L. Foster (Indiana University, Bloomington, IN)S.M. Rosenberg (Baylor College of Medicine, Houston, TX)S.M. Rosenberg

R. Woodgate (National Institutes of Health, Bethesda, MD)

T. Nohmi (National Institute of Health Sciences, Setagaya-ku, Tokyo, Japan)

This study, FC40 × P1(RW82)a

This study, FC40 × P1(YG7207)a

This studya

This study, RSH61 × P1 (YG7207)a

T. Nohmi

8 utation Research 686 (2010) 84–89

kspBwc(

2

Ma

2

obmpgrtmtptpbVtnoamtsc

3

tgfdutsso(alad

wmcbsiwoftairi

Fig. 1. Effects of lactose and �MG on yields of spontaneous Lac+ revertants of E. colistrain FC40. Samples (0.1 ml) of stationary-phase cells grown in nutrient broth were

6 D.G. MacPhee, M. Ambrose / M

anamycin (25 �g/ml), or tetracycline (10 �g/ml) was added. Bacteria were grown totationary-phase with shaking in Oxoid Nutrient broth at 37 ◦C (18–20 h). The com-osition of the defined minimal medium was Vogel and Bonner E salts [16], Difcoacto Agar (1.5%, w/v) and thiamine (0.5 �g/ml). Where appropriate the mediumas supplemented with glucose, glycerol or lactose (Ajax, Australia) to a final con-

entration of 0.2% (w/v). In the spontaneous ValS → ValR mutation assay, valine40 �g/ml) was added to the medium.

.3. Chemicals

Antibiotics and methyl-�-d-glucopyranoside were from Sigma (St. Louis, MO).ethyl-�-d-glucopyranoside was dissolved in distilled water prior to autoclaving,

nd when required it was added to media to give a final concentration of 0.2% (w/v).

.4. Mutation assays

The selection of spontaneous Lac+ revertants and ValR mutants was carriedut as previously described [1] and involved plating 0.1 ml samples of unwashedroth-grown stationary-phase cultures on pre-weighed (±1 g) defined minimaledium plates. ValR mutants were recovered on duplicate selective medium

lates containing valine and either glycerol or glucose, glycerol + glucose, orlycerol + methyl-�-d-glucopyranoside (�MG). The selection of spontaneous Lac+

evertants was carried out in a similar fashion, except that samples were plated inriplicate and on minimal lactose plates with and without �MG. The numbers of ValR

utants and Lac+ revertants appearing under the appropriate conditions of selec-ion were scored after incubation at 37 ◦C for 3 and 4 days respectively. The resultsresented are typical of those obtained in at least three separate experiments. Therace amount of broth present in the samples of unwashed broth-grown stationary-hase cultures and carried over to the selective plating medium will support theackground growth of the Lac− cells (in the Lac− → Lac+ reversion assay) and thealS cells (in the ValS → ValR assay) until such time as the broth is exhausted; in

urn, it is also during this growth that it becomes possible to determine whether orot the primary carbon source(s) present in the plating medium has any influencen the number of mutational events occurring in the cells [17–20]. In our mutationalssays, no determination is made of the number(s) of pre-plating and post-platingutational events that might be responsible for the ValR mutants and the Lac+ rever-

ants since we are interested only in the effects of the nature of primary carbonource(s) on the final yields of mutants appearing on the selective plates under theonditions of selection.

. Results and discussion

In our first series of experiments we examined the effect of lac-ose and of the non-utilizable but strongly catabolite-repressinglucose analogue, �MG, on the overall yields of spontaneousrameshift mutations able to be generated in wild-type and SOS-eficient stationary-phase E. coli cells. The experimental system wesed involved the reversion of a lacIZ fusion +1 frameshift muta-ion in cells of the E. coli FC40 strain which has been widely used intudies concerned with stationary-phase mutagenesis [21]. Cells oftrain FC40 are phenotypically Lac− because they carry a deletionf the chromosomal lac-pro region. They also harbour an F′ episomeF′128) that carries pro+ along with an in-frame fusion of the lacInd lacZ genes (lacI33�lacZ) which has a +1 frameshift mutation inacI that is polar on lacZ; in turn, stationary-phase Lac− FC40 cellsre able to revert to Lac+ by any mutation (mostly −1 base-paireletions) that restores the reading frame [21–23].

The data (Fig. 1) shows the results obtained in experiments inhich stationary-phase cells of strain FC40 were plated on definedinimal medium containing either lactose or lactose + �MG as a

arbon source(s). It is evident from the data that the final num-ers of Lac+ revertants arising on the lactose + �MG plates wereignificantly lower (by ∼six-fold) than those recovered on the min-mal lactose-alone plates, a finding which is entirely consistent

ith our earlier reported observations [1] (While on occasion webserved that Lac+ revertants continue to appear even after theourth day of incubation they still number significantly fewer on

he lactose + �MG plates; data not shown). Given that �MG did notppear to be operating by reducing the numbers of Lac− viable cellsn the background, or by inhibiting the growth of potential Lac+

evertants into full-sized colonies (data not shown; and Ref. [1]),t seemed that the profound inhibitory effect of �MG on sponta-

plated on minimal plates containing excess thiamine in addition to (i) lactose and(ii) lactose plus �MG, all at a concentration of 0.2% (w/v). Plates were incubated at37 ◦C for 4 days. Each point represents the average of triplicate plates.

neous lac+ mutation yields in bacterial cells might well stem fromits well-documented capacity to reduce intracellular cAMP levels[2].

If, as we suspect, the Lac+ colonies being recovered under theconditions of selection represents the outcome of an SOS-typeerror-prone mutation generation system (but only when cAMP lev-els are relatively high), it follows that the sorts of Lac+ revertantswhich fail to arise on lactose plates supplemented with �MG (i.e.,when cAMP levels are depressed) will also fail to arise on lactose-alone plates when the cells concerned carry a recA (null) mutation,and hence are SOS-deficient. To test this idea directly, we examinedthe spontaneous reversion of the lacI33�lacZ frameshift marker inwild-type (recA+) and �recA cells of strain FC40 plated on lactoseand lactose + �MG plates. It is clear from the results (Fig. 2A) thatintroduction of the �recA306 allele into FC40 greatly diminishedLac+ reversion yields, and that the numbers of Lac+ mutations gen-erated in the �recA cells incubated on lactose and lactose + �MGplates are not significantly different as compared to those obtainedwith wild-type (recA+) cells incubated under similar conditions.These findings are consistent with those of Cairns and Foster [21]showing that spontaneous reversion of the lacI33�lacZ frameshiftmarker requires a functional RecA protein. More importantly, theresults (Fig. 2A) add considerable weight to our working hypoth-esis that spontaneous mutations arising in stationary-phase E. colicells represent the outcome of an SOS (recA+)-dependent mutationgeneration pathway whose expression is sensitive to intracellu-lar cAMP levels, and can be easily reconciled with those resultsobtained in a biochemical study by Taddei et al. [24] showing thatmaximal expression of the recA gene required elevated cAMP levels.

However, since the expression of all SOS genes in E. coli isrecA-dependent, it was important for us to consider the intriguingpossibility that the very different numbers of Lac+ revertants recov-ered on the lactose and lactose + �MG plates reflects the differentialimpact of these two carbon compounds on the expression of someother SOS gene(s), and not solely recA, a phenomenon which mightotherwise be masked in the �recA background. To test this ideafurther, we first examined the generation of Lac+ mutations in�umuDC FC40 cells. As shown in Fig. 2B, Lac+ reversion appears tooccur normally in �umuDC cells, insofar as the overall numbers ofLac+ colonies able to be recovered with the wild-type and �umuDCcells are quite similar and still higher on lactose plates and signifi-cantly lower on lactose + �MG plates. In addition, these results are

in agreement with those of Cairns and Foster [21] and McKenzie etal. [25] showing that UmuDC (DNA Polymerase V) contributes little,if anything, to the reversion of the lacI33�lacZ frameshift markerin stationary-phase bacterial cells.

utatio

�fgbtrta−tanscstfCmeotGenripvacw

F((p0

D.G. MacPhee, M. Ambrose / M

Fig. 2C shows the results of our experiments with wild-type,dinB and DinB-overproducing cells of FC40. The most notable

eature of the data (Fig. 2C) is that yields of Lac+ mutations arereatly diminished in the �dinB background, and that the num-ers of Lac+ colonies recovered with the �dinB cells plated onhe lactose and lactose + �MG plates are indistinguishable. Theseesults further support an important role(s) for DinB in the spon-aneous revertibility of the lacI33�lacZ frameshift marker [26,27],nd are entirely consistent with DinB’s predilection for generating1 frameshift mutations [28]. The second most notable feature of

he results shown in Fig. 2C is that overproduction of DinB frommulticopy plasmid (pYG751), in otherwise wild-type cells, sig-ificantly increased Lac+ reversion yields and also suppressed thetrong inhibitory effect �MG was having on the appearance of Lac+

olonies under the conditions of selection. Together, the resultshown in Fig. 2B and C are important for at least two reasons. First,hey seem to suggest that DinB is the only SOS polymerase requiredor reversion of the lacI33�lacZ marker in stationary-phase cells.onsistent with this notion, the introduction of the �dinB::kanutant allele into otherwise wild-type FC40 cells constitutively

xpressing the recA (recAo281) gene reduced the overall numbersf Lac+ revertants able to be recovered under the selective condi-ions to those obtained with the �dinB cells (Fig. 2D). Moreover,alhardo et al. [29] showed for the first time that the constitutivexpression of DinB in a lexA3 (Ind−) strain – which expresses aoncleavable LexA repressor such that induction of all of the SOSesponse genes is prevented – was alone sufficient to restore andncrease the revertibility of the lacI33�lacZ marker in stationary-hase cells. Second, the results (Fig. 2B-C) strongly suggest that the

arying numbers of Lac+ colonies appearing on the minimal lactosend lactose + �MG plates is due to the differential impact of thesearbon compounds on the expression of the dinB gene, a findinghich was otherwise hidden in our experiments with the �recA

ig. 2. Effects of lactose and �MG on yields of spontaneous Lac+ revertants of SOS-profi�recA306, �). (B) Strains are wild-type FC40 ( ) and DG2430 (�[umuDC]595::cat,�dinB::kan, ). (D) Strains are wild-type FC40 ( ), DG2431 (�dinB::kan, ), RSH61 (rechase cells grown in nutrient broth were plated on minimal plates containing excess thi.2% (w/v). Plates were incubated at 37 ◦C for 4 days. Each point represents the average o

n Research 686 (2010) 84–89 87

FC40 cells. Our results (Fig. 2D) showing that the yields of Lac+

mutations arising in the �dinB::kan recAo281 cells incubated onthe lactose and lactose + �MG plates are not significantly differentlends further support to this notion.

In view of the results obtained in our Lac− → Lac+ assay, it seemsreasonable to conclude that if stationary-phase E. coli possess atleast one catabolite-repressible SOS-type spontaneous mutationgenerating process, it most likely involves the DinB (DNA poly-merase IV)-dependent pathway. On this line, it is also worth notingthat expression of the E. coli dinB gene has been shown to be con-trolled by the alternative transcription factor, sigma factor �38

(RpoS), which is usually required for regulating the expressionof a wide variety of stationary-phase genes whose products areimportant for the general stress response of this organism [27,30].Bearing in mind our results presented in Fig. 2C and D, togetherwith those reported by other independent investigators [27,30], itis tempting to speculate that expression of the dinB gene in E. colimight be susceptible to regulation by two different physiologicallysensitive transcriptional activators, one being RpoS and the otherbeing cAMP. However, given that intracellular RpoS concentrationsappear to increase in bacterial cells experiencing relatively highcAMP levels, and that rpoS expression has been shown to be reducedin cya E. coli cells which are devoid of intracellular cAMP [30,31],it may well be that lactose and �MG are differentially effectingthe expression of the dinB gene, and hence the overall numbers ofLac+ revertants appearing under the conditions of selection in thepresent study, by moderating intracellular RpoS concentrations viatheir varying effects on cAMP levels.

Encouraged by the results we obtained in our Lac− → Lac+ assay,

we decided to test whether other types of spontaneous muta-tions in E. coli cells are generated in much the same way(s). Inour second series of experiments, we studied the accumulation ofvaline-resistant mutations in stationary-phase E. coli cells. Growth

cient and SOS-deficient E. coli cells. (A) Strains are wild-type FC40 ( ) and SZ624). (C) Strains are wild-type FC40 ( ), DG2432 (pYG751[dinB+], ), and DG2431Ao281, ) and DG2434 (recAo281 �dinB::kan), ). Samples (0.1 ml) of stationary-amine in addition to (i) lactose and (ii) lactose plus �MG, all at a concentration off triplicate plates.

88 D.G. MacPhee, M. Ambrose / Mutation Research 686 (2010) 84–89

Fig. 3. Effects of glycerol, glucose and �MG on yields of spontaneous ValR mutants of SOS-proficient and SOS-deficient E. coli cells. (A) Strains are wild-type FC40 ( ) andS 95::c( lates� (40 �o

osbacbohechp

FnadSstmcmmpcgfipec

miSi

Z624 (�recA306, �). (B) Strains are wild-type FC40 ( ) and DG2430 (�[umuDC]50.1 ml) of stationary-phase cells grown in nutrient broth were plated on minimal pMG, all at a concentration of 0.2% (w/v), in addition to excess thiamine and valinef duplicate plates.

f E. coli K12 strains is strongly inhibited in the presence of evenmall concentrations of the amino acid valine because of a feed-ack inhibition of valine on the acetohydroxyacid-synthase enzymectivities involved in the biosynthesis of valine and isoleucine. E.oli K12 cells incubated in the presence of valine are thereforeeing starved for isoleucine [32]. However, because developmentf resistance to the toxic effects of valine: (i) occurs at relativelyigh frequencies (ca. 5 × 10−7), (ii) can arise as a result of differ-nt mutational events in a number of different genes on the E. colihromosome, and (iii) is expressed immediately, valine-resistanceas been used by us and others to study spontaneous mutagenesisrocesses in stationary-phase E. coli cells [1,32–34].

The results of our spontaneous ValS → ValR assay are shown inig. 3A–C. Much to our surprise, we found that the yields of sponta-eous ValR mutations generated in the wild-type, �recA, �dinBnd �umuDC backgrounds were very similar. As a result, theseata (Fig. 3A–C) would seem to rule out a significant role for anOS-dependent mutation generation system(s) in the formation ofpontaneous ValR mutations in stationary-phase E. coli cells. Impor-antly, it is perfectly clear from Fig. 3A–C that the numbers of ValR

utations which did arise in wild-type, �recA, �dinB and �umuDCells were also very strongly dependent upon the nature of the pri-ary carbon compound(s) made available to them in the platingedium. Thus for example, when glucose or its analogue �MG was

resent in the plating medium, the numbers of ValR colonies whichould be recovered were significantly lower than they were whenlycerol was the sole available carbon source. Taken together, thesendings tend to suggest that stationary-phase E. coli cells mightossess an inducible SOS-independent spontaneous mutation gen-ration pathway(s) whose activities also happen to be subject toAMP-mediated catabolite repression.

The idea that E. coli might possess an inducible SOS-independentutation generation system(s) is by no means new, even although

t appears not to have been as comprehensively studied as theOS-dependent mutagenesis mechanisms operating in this organ-sm. Since 1994, Humayun and co-workers have been reporting on

at, ). (C) Strains are wild-type FC40 ( ) and DG2431 (�dinB::kan, ). Samplescontaining (i) glycerol, (ii) glucose, (iii) glycerol plus glucose, and (iv) glycerol plusg/ml). Plates were incubated at 37 ◦C for 3 days. Each point represents the average

experiments showing that prior UV-irradiation of �recA cells (andhence SOS-defective cells) enhanced mutagenesis at a site-specificethenocytosine (�C) residue borne on a transfected single-strandedM13 phage DNA vector; in turn, these results were interpreted asindicating the existence of an SOS-independent inducible muta-genic phenomenon which was subsequently termed UVM, for UVmodulation of mutagenesis [35,36]. Since that time, UVM mutage-nesis has been shown to be inducible by other damaging agentsin addition to UV, including by alkylating and oxidative DNAdamaging agents [37,38]. However, as long ago as 1985, Thomasand MacPhee [39,40] demonstrated that 9-aminoacridine (9AA)-induced mutagenesis in E. coli also involved a recA+-independentmechanism(s). More importantly for present purposes, these sameauthors also showed that the expression of the recA+-independentmechanism(s) responsible for generating and fixing 9AA-inducedmutations in E. coli was sensitive to glucose-repression [40].Whether the catabolite-repressible SOS-independent mutagene-sis pathway generating spontaneous valine-resistance mutationsin stationary-phase E. coli cells bears some resemblance to theSOS-independent type mechanism(s) driving UVM mutagenesisand 9AA-induced mutagenesis in this organism remains to be fullydetermined.

In summary, this study provides evidence showing that E.coli responds to selection pressures by accumulating spontaneousmutations via SOS-dependent and SOS-independent mutation gen-eration mechanisms whose levels of expression are susceptibleto environmental changes that alter cAMP concentrations in thisorganism. It will be interesting to determine in the future how theactivities these multiple spontaneous mutation generation systemsare coordinated in E. coli and other related organisms in responseto environmental stresses.

Conflict of interest

The authors declare that there are no conflicts of interest.

utatio

A

rNw

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

D.G. MacPhee, M. Ambrose / M

cknowledgements

We are grateful for the advice and comments provided by oureferees. We also thank Patricia Foster, Susan Rosenberg, Takehikoohmi and Roger Woodgate for providing bacterial strains. Thisork was supported by the Australian Research Council.

eferences

[1] M. Ambrose, D.G. MacPhee, Catabolite repressors are potent antimutagensin Escherichia coli plate incorporation assays: experiments with glucose,glucose-6-phosphate and methyl-�-d-glucopyranoside, Mutat. Res. 398 (1998)175–182.

[2] M.H. Saier Jr., B.U Feucht, Coordinate regulation of adenylate cyclase and car-bohydrate permeases by the phosphoenolpyruvate: sugar phosphotransferasesystem in Salmonella typhimurium, J. Biol. Chem. 250 (1975) 7078–7080.

[3] W. Epstein, L.B. Rothman-Denes, J. Hesse, Adenosine 3′:5′-cyclic monophos-phate as mediator of catabolite repression in Escherichia coli, Proc. Natl. Acad.Sci. U.S.A. 72 (1975) 2300–2304.

[4] L. Castro, B.U. Feucht, M.L. Morse, M.H. Saier Jr., Regulation of carbohydratepermeases and adenylate cyclase in Escherichia coli, J. Biol. Chem. 251 (1976)5522–5527.

[5] J.P. Harwood, C. Gazdar, C. Prasad, A. Peterkofsky, Involvement of the glu-cose enzymes II of the sugar phosphotransferase system in the regulation ofadenylate cyclase by glucose in Escherichia coli, J. Biol. Chem. 251 (1976) 2462–2468.

[6] V. Dumay, A. Danchin, M. Crasnier, Regulation of Escherichia coli adenylatecyclase activity during hexose phosphate transport, Microbiology 142 (1996)575–583.

[7] J.W. Little, D.W. Mount, C.R. Yanisch-Perron, Purified LexA protein is a repressorof the recA and lexA genes, Proc. Natl. Acad. Sci. U.S.A. 78 (1981) 4199–4203.

[8] J.W. Little, D.W. Mount, The SOS regulatory system of Escherichia coli, Cell 29(1982) 11–22.

[9] E.C. Friedberg, G.C. Walker, W. Siede, DNA Repair and Mutagenesis, ASM Press,Washington, DC, 1995, pp. 407–464.

10] L.L. Lin, J.W. Little, Isolation and characterization of noncleavable (Ind-) mutantsof the LexA repressor of Escherichia coli K-12, J. Bacteriol. 170 (1988) 2163–2173.

11] J. Wagner, P. Gruz, K. Su-Ryang, M. Yamada, K. Matsui, R.P.P. Fuchs, T. Nohmi,The dinB gene encodes a novel E. coli DNA polymerase, DNA pol IV, involved inmutagenesis, Mol. Cell 4 (1999) 281–286.

12] M. Tang, X. Shen, E.G. Frank, M. O’Donnell, R. Woodgate, M.F. Goodman,UmuD′(2)C is an error-prone DNA polymerase, Escherichia coli pol V, Proc. Natl.Acad. Sci. U.S.A. 96 (1999) 8919–8924.

13] A. Brotcorne-Lannoye, G. Maenhaut-Michel, Role of RecA protein in untargetedUV mutagenesis of bacteriophage lambda: evidence for the requirement for thedinB gene, Proc. Natl. Acad. Sci. U.S.A. 83 (1986) 3904–3908.

14] J.H. Miller, A Short Course in Bacterial Genetics, Cold Spring Harbor LaboratoryPress, Plainview, New York, 1992.

15] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: A Laboratory Manual,second ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York,1989.

16] H.J. Vogel, D.M. Bonner, Acetylornithinase of Escherichia coli: partial purificationand some properties, J. Biol. Chem. 218 (1956) 97–106.

17] M. Demerec, E. Cahn, Studies on mutability in nutritionally deficient strains ofEscherichia coli, J. Bacteriol. 65 (1953) 27–36.

18] M.H.L. Green, W.J. Muriel, Mutagen testing using Trp+ reversion in Escherichiacoli, Mutat. Res. 38 (1976) 3–32.

[

[

n Research 686 (2010) 84–89 89

19] K.E. Mortelmans, B.A.D. Stocker, Ultraviolet light protection, enhancement ofultraviolet light mutagenesis, and mutator effect of plasmid R46 in Salmonellatyphimurium, J. Bacteriol. 128 (1976) 271–282.

20] D.G. MacPhee, Indications that mutagenesis in Salmonella may be subject tocatabolite repression, Mutat. Res. 151 (1985) (1985) 35–41.

21] J. Cairns, P.L. Foster, Adaptive reversion of a frameshift mutation in Escherichiacoli, Genetics 128 (1991) 695–701.

22] S.M. Rosenberg, S. Longerich, P. Gee, R.S. Harris, Adaptive mutation by deletionsin small mononucleotide repeats, Science 265 (1994) 405–407.

23] P.L. Foster, J.M. Trimarchi, Adaptive reversion of a frameshift mutation inEscherichia coli by simple base deletions in homopolymeric runs, Science 265(1994) 407–409.

24] F. Taddei, I. Matic, M. Radman, cAMP-dependent SOS induction and mutage-nesis is resting bacterial populations, Proc. Natl. Acad. Sci. U.S.A. 92 (1995)11736–11740.

25] G.J. McKenzie, R.S. Harris, P.L. Lee, S.M. Rosenberg, The SOS response regulatesadaptive mutation, Proc. Natl. Acad. Sci. U.S.A. 97 (2000) 6646–6651.

26] G.J. McKenzie, P.L. Lee, M.J. Lombardo, P.J. Hastings, S.M. Rosenberg, SOSmutator DNA polymerse IV functions in adaptive mutation and not adaptiveamplication, Mol. Cell 7 (2001) 571–579.

27] J.C. Layton, P.L. Foster, Error-prone DNA polymerase IV is controlled by thestress-response sigma factor, RpoS, in Escherichia coli, Mol. Microbiol. 50 (2003)549–561.

28] S.R. Kim, G. Maenhaut-Michel, M. Yamada, Y. Yamamoto, K. Matsui, T. Sofuni,T. Nohmi, H. Omori, Multiple pathways for SOS-induced mutagenesis inEscherichia coli: an overexpression of dinB/dinP results in strongly enhancingmutagenesis in the absence of any exogenous treatment to damage DNA, Proc.Natl. Acad. Sci. U.S.A. 94 (1997) 13792–13797.

29] R.S. Galhardo, R. Do, M. Yamada, E.C. Friedberg, P.J. Hastings, T. Nohmi, S.M.Rosenberg, DinB upregulation is the sole role of the SOS response in stress-induced mutagenesis in Escherichia coli, Genetics 182 (2009) 55–68.

30] R. Hengge-Aronis, Signal transduction and regulatory mechanisms involved incontrol of the �S (RpoS) subunit of RNA polymerase, Microbiol. Mol. Biol. Rev.66 (2002) 373–395.

31] M.P. McCann, C.D. Fraley, A. Matin, The putative � factor KatF is regulated post-transcriptionally during carbon starvation, J. Bacteriol. 175 (1993) 2143–2149.

32] S.W. Glover, Valine-resistant mutants of Escherichia coli K-12, Genet. Res. 3(1962) 448–460.

33] J. Cairns, J. Overbaugh, S. Miller, The origin of mutants, Nature (London) 335(1988) 142–145.

34] R.M. Schaaper, R.L. Dunn, The antimutator phenotype of E. coli mud is onlyapparent and results from delayed appearance of mutants, Mutat. Res. 480–481(2001) 71–75.

35] V.A. Palejwala, G.A. Pandya, O.S. Bhanot, J.J. Solomon, H.S. Murphy, P.M.Dunman, M.Z. Humayun, UVM, an ultraviolet-inducible RecA-independentmutagenic phenomenon in Escherichia coli, J. Biol. Chem. 269 (1994)27433–27440.

36] V.A. Palejwala, G.E. Wang, H.S. Murphy, M.Z. Humayun, Functional recA, lexA,umuD, umuC, polA, and polB genes are not required for the Escherichia coli UVMresponse, J. Bacteriol. 177 (1995) 6041–6048.

37] G. Wang, V.A. Palejwala, P.M. Dunman, D.H. Aviv, H.S. Murphy, M.S. Rahman,M.Z. Humayun, Alkylating agents induce UVM, a recA-independent induciblemutagenic phenomenon in Escherichia coli, Genetics 141 (1995) 813–823.

38] G Wang, M.Z. Humayun, Induction of the Escherichia coli UVM response by

oxidative stress, Mol. Gen. Genet. 251 (1996) 573–579.

39] S.M. Thomas, D.G. MacPhee, Frameshift mutagenesis by 9-aminoacridine andICR191 in Escherichia coli: effects of uvrB, recA and lexA mutations and of plasmidpKM101, Mutat. Res. 151 (1985) 49–56.

40] S.M Thomas, D.G. MacPhee, RecA-independent mutagenesis in Escherichia colimay be subject to glucose repression, Mutat. Res. 180 (1987) 67–73.