JOURNAL OF BACTERIOLOGY, Aug. 1994, p. 5011-5021 Vol. 176, No. 160021-9193/94/$04.00+0Copyright © 1994, American Society for Microbiology

Isolation and Characterization of Novel Plasmid-EncodedumuC Mutants

ROGER WOODGATE,1* MAHIPAL SINGH,1 OLGA I. KULAEVA,1 EKATERINA G. FRANK,'ARTHUR S. LEVINE,1 AND WALTER H. KOCH2

Section on DNA Replication, Repair and Mutagenesis, National Institute of Child Health and Human Development,Bethesda, Maryland 20892,1 and Molecular Biology Branch, Food and Drug Administration, Washington, D.C. 202042

Received 16 March 1994/Accepted 2 June 1994

Most inducible mutagenesis in Escherichia coli is dependent upon the activity of the UmuDC proteins. Therole of UmuC in this process is poorly understood, possibly because of the limited number of geneticallycharacterized umuC mutants. To better understand the function of the UmuC protein in mutagenic DNArepair, we have isolated several novel plasmid-encoded umuC mutants. A multicopy plasmid that expressedUmuC at physiological levels was constructed and randomly mutagenized in vitro by exposure to hydroxyl-amine. Mutated plasmids were introduced into the umu tester strain RW126, and 16 plasmids that were unableto promote umuC-dependent spontaneous mutator activity were identified by a calorimetric papillation assay.Interestingly, these plasmid mutants fell into two classes: (i) 5 were expression mutants that produced eithertoo little or too much wild-type UmuC protein, and (ii) 11 were plasmids with structural changes in the UmuCprotein. Although hydroxylamine mutagenesis was random, most of the structural mutants identified in thescreen were localized to two regions of the UmuC protein; four mutations were found in a stretch of 30 aminoacids (residues 133 to 162) in the middle of the protein, while four other mutations (three of which resulted ina truncated UmuC protein) were localized in the last 50 carboxyl-terminal amino acid residues. These newplasmid umuC mutants, together with the previously identified chromosomal umuC25, umuC36, and umuC104mutations that we have also cloned, should prove extremel useful in dissecting the genetic and biochemicalactivities of UmuC in mutagenic DNA repair.

The ability of Escherichia coli DNA polymerase III holoen-zyme to replicate across bulky misinstructional lesions in DNA,such as cyclobutane pyrimidine dimers or 6-4 pyrimidine-pyrimidone photoproducts, is dependent upon the functionalactivity of the RecA and UmuD'C proteins (4, 14, 43, 50, 58).In addition, studies with a defined lesion in phage M13 DNAhave shown that the extent and mutagenic consequences oftranslesion DNA synthesis vary and depend upon the specifictype of DNA lesion encountered (3, 4, 25, 32-34).Both genetic and biochemical experiments have shown that

the RecA and UmuDC proteins are induced as part of thecellular SOS response to DNA damage (53, 58). Furthermore,RecA mediates the processing of UmuD to a shorter butmutagenically active form, UmuD' (8, 40, 46). Mutants ofumuD or recA that are unable to promote cleavage arerendered phenotypically nonmutable (2, 18, 31, 46). RecA alsoparticipates directly in the mutagenic process by targeting theUmu-like mutagenesis proteins to DNA (13, 20). Althoughrecent studies have shown that DNA polymerase III holoen-zyme can bypass a synthetic abasic lesion in vitro only in thepresence of the highly purified UmuD'C and RecA proteins(43), the biochemical action of these proteins remains to beresolved. In particular, very little is known about UmuC's rolein this process. To some extent this is due to the limitednumber of umuC mutants that have been characterized genet-ically. The E. coli umuC gene consists of 1,269 base pairs, andto date, only eight umuC mutants have been reported: fourmissense mutants (three chromosomal and one plasmid) (28,31, 36, 48), two frameshift mutants that result in nonsense

* Corresponding author. Mailing address: Building 6, Room 1A13,NICHD, NIH, 9000 Rockville Pike, Bethesda, MD 20892. Phone:(301) 496-6175. Fax: (301) 402-0105. Electronic mail address:woodgate@helix.nih.gov

mutations (both plasmid encoded) (41), and two insertionalmutants (both chromosomal) (1, 16).

In an attempt to better characterize UmuC function bothgenetically and biochemically, we attempted to isolate novelumuC mutants. We have recently developed an Escherichia colistrain that is particularly useful in identifying Umu activity byway of a calorimetric papillation assay (24). We have used thisassay to screen for randomly mutated umuC plasmids thatwere unable to promote normal SOS mutagenesis functions. Inthis report, we present a molecular and genetic characteriza-tion of some of these mutants. Our findings suggest that theintegrity of certain regions of the UmuC protein may be criticalfor activity and that changes in the levels of wild-type UmuCcan also affect the efficiency at which mutagenesis functions arepromoted.

MATERIALS AND METHODS

Bacterial strains. The E. coli strains used in this study arelisted in Table 1. All have been reported elsewhere exceptRW222, which was constructed by a series of P1 transductionsas follows. The suL4211 mutation was introduced into TK603(28) by initially transducing in the sul4100::TnS and closelylinked pyrD mutations from DE1242 (24) and then replacingthem with suLA211 pyrD' from DE1776 (18). The A(umuDC)595::cat, recA4718 srlC300::TnlO, and lexA71::Tn5 mutationswere subsequently transduced from RW82 (55), DE1918 (24),and JL1047 (35) into this strain by selecting for chloramphen-icol resistance, tetracycline resistance and UV sensitivity, andkanamycin resistance, respectively.

Plasmids. The E. coli umuD and umuC genes overlap by 1base pair (30, 42), and mutations in either umuD or umuCrender bacteria phenotypically nonmutable. To facilitate thespecific identification of umuC mutants, we have manipulated

5011

on April 6, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

5012 WOODGATE ET AL.

TABLE 1. E. coli strains used in this study

Strain Relevant genotype reference

RW126 A(umuDC)595::cat recA718 srlC300::TnlO 24lexA71::Tn5 suIA211 hisG4 argE3 galK2supE44

RW222 A(umuDC)595::cat recA718 srlC300::TnlO This paperlexA71::TnS suL4211 uvrA6 hisG4supE44

TK610 umuC36 uvrA6 hisG4 supE44 28DE406 lexIA5(Def) 56DE274 lexAMl(Def) recA730 17

the umuD and umuC genes to physically separate them andplace them on different but compatible plasmids.A low-copy-number spectinomycin-resistant plasmid, pRW66,

that expresses UmuD', the mutagenically active form ofUmuD, was constructed by ligating an appropriately modified1-kb umuD' BglII fragment from pGW2123 (40) and a Sail-digested, pSC101-derived vector, pGB2 (10).The starting vector for the UmuC plasmid was pDS10 (11),

a -10 kb ColEI-based plasmid that expresses UmuDC andconfers resistance to kanamycin and ampicillin. Donnelly andWalker (11) engineered this plasmid to create a novel EcoRIrestriction site between the LexA binding site and the Shine-Dalgarno ribosome binding site of the umuDC operon (Fig. 1).Plasmid pRW116 was constructed by replacing a 4-kb EcoRIfragment from pDS110, encoding kanamycin resistance andthe structural umuDC operon, with a 2-kb EcoRI fragmentcarrying the genetically engineered umuD'C operon fromplasmid pEC42 (20) (Fig. 1). Although this plasmid does notcontain a bona fide ribosome binding site, expression ofUmuD'C from this plasmid was sufficient to fully restoremutagenesis functions to AumuDC strains (data not shown).

LexARBS

pDS110 -

VIUC

R R

pRW116 V

RN MR

pRW124 I UmvC 3YM

FIG. 1. Construction of a plasmid that expresses UmuC at physio-logical levels. The figure depicts a schematic representation of the E.coli umuDC operon. The LexA and ribosomal binding sites are

indicated above each construct, while relevant restriction enzyme sitesare indicated below the construct; R, EcoRI; N, NcoI; M, MluI.Plasmid pDS110 had previously been engineered so that an EcoRIrestriction enzyme site was placed between the LexA binding site andthe Shine-Dalgarno ribosome binding site (RBS) of the umuDCoperon (11). pRW116 was constructed by replacing the umuDCEcoRIfragment of pDS110 with an EcoRI fragment containing a geneticallyengineered umuD'C operon. The UmuC plasmid, pRW124, was

constructed by replacing the Nco-Mlu umuD'-containing fragment ofpRW116 with synthetic oligonucleotides that reconstructed the start ofthe umuC gene. Expression of UmuC from this plasmid fully restoresinducible mutagenesis functions to umuC36 mutants (Table 2).

The UmuC plasmid pRW124 was constructed by removing theNcoI-MluI umuD' interval from pRW116 and replacing it withtwo synthetic oligonucleotides which, when annealed together,reconstructed the start of the umuC gene (Fig. 1). LikepRW1 16, this construct does not contain any obvious ribosomebinding site, yet expression of UmuC from pRW124 wassufficient to fully restore mutagenesis functions to umuC36mutants (see below).Both the medium-copy-number plasmid pRW30 (55) and

the low-copy-number plasmid pRW154 (24) that express theUmuDC proteins from their natural promoter have beendescribed previously.

In vitro mutagenesis of pRW124 and screening for umuCmutants. Randomly mutated pRW124 was obtained followingthe hydroxylamine mutagenesis procedure described by Isack-son and Bertrand (26). Briefly, 2 Rg of purified plasmid DNAwas incubated in 100 RI of 800 mM hydroxylamine HCl(Sigma)-50 mM sodium phosphate, pH 6.0-1 mM EDTA for48 h at 37°C. The mutagenized DNA was recovered in TE (10mM Tris HCl, pH 8.0, 1 mm EDTA) by repeated dilution andconcentration cycles in a Centricon 30 ultrafiltration cartridge(Amicon). Potential umuC mutants were identified by intro-ducing the mutagenized plasmid DNA into the umu testerstrain RW126 harboring the low-copy-number umuD' plasmidpRW66. Transformants were selected on Luria-Bertani (LB)plates supplemented with spectinomycin (50 ,ug/ml) and am-picillin (100 jig/ml) and then restreaked with a sterile tooth-pick onto MacConkey galactose (1%) indicator plates supple-mented with spectinomycin and ampicillin. After 10 days ofincubation at 37°C, colonies that produced less than half of thenumber of Gal' papillae produced by the control strain wereanalyzed further by the qualitative His' mutagenesis assay.

Qualitative mutagenesis assay. Bacterial cultures weregrown overnight in LB medium containing the appropriateantibiotics. Aliquots (1 ml) were centrifuged and resuspendedin an equal volume of SM buffer (38). The ability of particularplasmid-bearing strains to promote Umu-dependent SOS mu-tator activity in the absence of exogenous DNA damage wasjudged by plating 100-,ul aliquots on Davis and Mingioliminimal agar plates supplemented with a trace amount ofhistidine (1 ,g/ml) (7). Umu-dependent chemically inducedmutagenesis was determined essentially as described above,except that 5 RI of a 1:5 dilution of methyl methanesulfonate(MMS) (Sigma) in dimethyl sulfoxide (Sigma) was applied to asmall sterile disk in the center of the plate. Both spontaneouslyarising and MMS-induced His' mutants were scored after 4days of incubation at 37°C.

Quantitative mutagenesis assays. Plasmids that partiallyrestored mutagenesis functions based upon the qualitativeassay were further analyzed by quantitative UV- and MMS-induced mutagenesis assays. Bacterial cultures were grown to acell density of -1 X 108/ml in LB broth with appropriateantibiotics, harvested, and resuspended in SM buffer. For UVmutagenesis assays, cells were irradiated and appropriatedilutions were plated on the low-histidine agar plates asdescribed above. For the MMS assay, cultures were split into1-ml aliquots and incubated with a 5-,ul solution of dimethylsulfoxide containing the indicated volume of MMS for 30 minat 37°C, after which cells were pelleted and washed in SMbuffer; the appropriate dilutions were then plated. His' mu-tants were scored after 4 days of incubation at 37°C, andinduced mutation frequencies were calculated as described bySedgwick and Bridges (45).UV survival. Early-log-phase bacteria grown in LB medium

plus appropriate antibiotics were harvested, resuspended inSM buffer, and irradiated with 254-nm UV light. Dilutions

J. BACTERIOL.

on April 6, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 176, 1994

were plated in triplicate on LB agar plates, and survivingcolonies were counted after overnight incubation at 370C.

Steady-state levels of wild-type and mutant UmuC proteins.Polyclonal antibodies were raised in rabbits to affinity-purifiedUmuC protein (57) by Hazleton Laboratories (Washington,D.C.). UmuC protein was detected in whole-cell E. coliextracts by the Western-light chemiluminescent assay (Tropix)essentially as described previously (23, 56). Briefly, whole-cellextracts from - 108 cells were separated by electrophoresis in a15% acrylamide-sodium dodecyl sulfate gel. Proteins weretransferred to a nitrocellulose support membrane and incu-bated overnight with a 1:10,000 dilution of UmuC antisera.After appropriate washes, the membranes were incubated witha 1:3,000 dilution of a goat anti-rabbit antibody conjugatedwith alkaline phosphatase (Bio-Rad). UmuC was visualizedafter the membrane was incubated with chemiluminescentsubstrate and exposed to X-ray film for an appropriate periodof time. The relative level of UmuC protein was obtained bydensitometric analysis of films. In most cases, values wereobtained from multiple exposures so that the desired bandswere in the linear range of the film (56).DNA sequence analysis. The umuC gene in plasmid

pRW124 (and mutant derivatives) was sequenced by LarkSequencing Technologies Inc. (Houston, Tex.) by standarddideoxy sequencing protocols (44) and with seven uniqueoligonucleotide primers spaced about 200 bp apart.

Subcloning multiple umuC plasmid mutations and cloningof the previously identified chromosomal umuC25, umuC36,and umuC104 mutations. DNA sequence analysis revealedthat plasmid pRW124-73 contained three mutations in umuC.A plasmid that contained only the most 5' mutation (pRW124-73A; AT3) was constructed by replacing the MluI-PstI frag-ment of pRW124-73 with a similar fragment from an undam-aged plasmid. The second mutation (pRW124-73B; GS24) wasisolated as an MluI-HindIII fragment and was cloned into asimilarly digested wild-type vector. The most 3' mutation(pRW124-73C; QX399) was isolated as a BamHI-PstI frag-ment and was cloned into a similarly digested wild-type vector.

Plasmid pRW124-139 contained two mutations in umuC,one of which was silent. It also exhibited elevated levels ofUmuC, suggesting that it may contain an additional mutationthat affected plasmid copy number (data not shown). Themutation resulting in the structural change in umuC (pRW124-139B; QX372) was obtained as a HindIII-BamHI fragment andwas cloned into the similarly digested wild-type vector.The chromosomal umuC25 mutation was obtained from

strain DE1878 (31) by PCR amplification of a 952-bp fragmentwith two oligonucleotide primers, 5'-GAGAAATTCGCGCAACGGT-3' (781s) and 5'-TCCATTCGGCGCTCCTG-3'(1733a) (5' nucleotides of the primers are shown in parenthe-ses; s and a refer to sense and antisense strands, respectively[42]). Amplified DNA was digested with HindlIl and BamHIand cloned into HindIII-BamHI digested plasmid pRW124.The presence of the umuC25 allele was confirmed by allele-specific oligonucleotide colony hybridization analysis using a22-mer 32P-labelled oligonucleotide probe (5'-GTTATCTCCAAG1T1'ATTAAGA-3') as previously described, omittingthe psoralen cross-linking step (9).umuC104 was obtained from DE1882 (31) as a 685-bp PCR

fragment by using primers 5'-GTGAAGATACGCTGGATG-3' (362s) and 5'-CTGTATCCGCCAAATCG-3' (1047a), andcloned into pRW124 after digestion with MluI and HindIII.The presence of the umuC104 mutation was confirmed by thepresence of an extra AseI restriction site that results from theG-*A transition found in umuC104 (31).umuC36 was obtained from TK610 (28, 31) as a 477-bp PCR

lexA (Def)umu +

NOVEL umuC MUTANTSIexA (Def) recA 718

AumuDC

C: LO N+ N T-

q: < Q

% ... 0. 0.

5013

.-o Em

X3: _

cr 'aa I

UmuC

FIG. 2. Levels of plasmid and chromosomally expressed UmuC.UmuC protein was detected in whole-cell extracts as described inMaterials and Methods. The level of chromosomal UmuC was ana-lyzed from DE406 [recA' lexA(Def)] and DE274 [recA4730 lex4(Def)].Plasmid-encoded UmuC was analyzed in strain RW126 [lexA(Def)recA718 AumuDC]. pRW154 is a low-copy-number plasmid thatexpresses UmuDC. pRW30 is a medium-copy-number plasmid ex-pressing UmuDC. pRW124 is a medium-copy-number plasmid thatexpresses only UmuC. [Since UmuD(D') stabilizes UmuC (12), thepRW124 extract was made from a strain that harbors the UmuD'-producing plasmid pRW66]. The position of wild-type UmuC isindicated on the right. Densitometric analysis of this gel reveals thatthe level of chromosomal UmuC in the recA4730 lexA(Def) strain isapproximately 4.4-fold higher than in the recA+ lexA(Def) strain-and isconsistent with our earlier findings (56). Compared with the recA730lexA(Def) strain, plasmids pRW154, pRW124, and pRW30 producedapproximately 3.6-, 2.5-, and 28-fold-higher levels of UmuC, respec-tively.

fragment by using primers 5'-ATGGGCGATCCCTGG-3'(570s) and 5'-CTGTATCCGCCAAATCG-3' (1047a) and wascloned as a 78-bp BglII-SalI fragment into pRW134 (18). TheumuC36 mutation was subsequently subcloned from pRW134into pRW124 as anAgel-BamHI fragment. The presence of theumuC36 mutation was confirmed by the presence of an extraHindIII restriction site that results from the G--A transitionfound in umuC36 (31).

All PCR-amplified regions were subjected to DNA sequenceanalysis to confirm that no additional unwanted mutationswere inadvertently cloned.

RESULTS

Isolation of plasmid-encoded umuC mutants. The aim ofthis study was to isolate new umuC mutants, characterizationof which might provide some insight into the activities ofUmuC. Since mutations in either umuD or umuC render E.coli nonmutable, our first step was to clone the umu genes ontoseparate but compatible plasmids. The mutagenically activeUmuD' protein was expressed from the low-copy-numberplasmid, pRW66, whereas UmuC was expressed from themedium-copy-number plasmid pRW124 (Fig. 1). Both plas-mids contain the same 5' umu regulatory regions, and as aresult, both the UmuD' and UmuC proteins are regulated byLexA protein and are expressed from the natural umuDCoperon promoter. The exception is that pRW124 lacks anobvious ribosome binding site (Fig. 1). As a result, the amountof UmuC produced from the medium-copy-number plasmid,pRW124, is dramatically reduced compared with the amountof UmuC expressed from the umuDC operon cloned into aplasmid with a similar copy number (cf. pRW124 and pRW30;Fig. 2). In fact, the amount of UmuC produced from pRW124

on April 6, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

5014 WOODGATE ET AL.

TABLE 2. Plasmid-encoded umuC mutants

Mutagenesis functions (His' mutants/plate)' Mutation in umuC"umuC uualee Relative amtplasmid Spontaneous MMS induced umuC36 Nucleotide Amino acid umuC allele of Umuac

complementation

Noned 3 7 14 NAW NA NA NApRW124(wt) 123 444 250 NA NA NA 1.0

pRW124-10 14 248 167 933C- )T QX(Am)172 umuC251 NWpRW124-34 3 5 16 873G--A AT152 umuC2529 ND

1"01G->A EK228

pRW124-40 2 2 20 NA NA NA 8.0pRW124-65 6 28 26 1254c-T RC279 umuC253 1.2

1256c ,T

pRW124-73 3 5 11 426G- )A AT3 umuC254 ~ 0.06489G---)-A GS241614C-,T QX(Oc)399

pRW124-77 3 9 16 1284T, SP289 umuC255 0.27pRW124-92 7 116 29 837G--,A AT140 umuC256 0.12

pRW124-100 5 7 20 NA NA NA 4.5pRW124-111 3 10 17 1642c- T SL408 umuC257 0.2

pRW124-120 3 10 21 1629c-__T RX(Op)404 umuC258 0.5pRW124-126 35 297 94 816C--,T LF133 umuC259 0.4

pRW124-132 84 163 137 NA NA NA 2.0pRW124-134 12 84 20 903G- A GR162 umuC260 0.35

pRW124-138 6 6 25 NA NA NA 2.4pRW124-139 6 12 19 845C__T Silent umuC261J 2.3

1533c-_-T QX(Oc)372

pRW124-140 2 5 18 591C-*T QX(Oc)58 umuC262 ND1574c+T Silent

aThe data presented are the mean values from at least three separate cultures, with three plates per culture. RW126 was used to determine the ability of plasmidsto promote spontaneous and MMS mutagenesis, while TK610 was used to assess MMS-induced umuC36 complementation.

b The numbering system used is based upon that previously described by Perry et al. (42). Wild-type umuC consists of 1,269 bp or 422 amino acids. Amino acidchanges are noted by the single-letter code for the original residue followed by the residue in the mutant protein. X denotes a termination codon. The specifictermination codon is indicated in parentheses, Am, amber; Oc, ochre; Op, opal.'The data presented are representative and are based upon the densitometric analysis of multiple exposures of Fig. 4.d The data presented for spontaneous and MMS mutagenesis are in the presence of the low-copy-number umuD' plasmid pRW66.NA, not applicable.

fND, not detectable.g The allele number refers to the combination of mutations found in each particular plasmid. Where subcloned, the individual mutations have been given separate

allele designations.

was comparable to that seen with the umuDC operon on thelow-copy-number plasmid pRW154 and only 2.5-fold higherthan that expressed chromosomally in a recA730 lex4(Def)strain (Fig. 2). On the basis of our previous calculations of 700UmuC molecules per recA730 1exA(Def) cell (56), we estimatethat the level of UmuC produced from pRW124 in strainRW126 corresponds to -1,750 molecules per cell. By compar-ison, we have previously estimated that there are -2,500UmuD' molecules per recA 730 lexA(Def) cell. Plasmid pRW66produces roughly fourfold more UmuD' than that expressedfrom the chromosome (data not shown) and would thereforebe expected to give rise to -10,000 UmuD' molecules per cell.Given that UmuD(D') is a dimer under physiological condi-tions (6, 57), this equates to a ratio of approximately threeUmuD' dimers to one UmuC molecule. Previously, we havenoted that when expressed from the natural umuDC operon,UmuD(D') is produced in an approximately sixfold excess overUmuC (56). The two-plasmid system employed here, there-fore, maintains close to physiological levels and ratios of the

UmuD'C proteins and avoids most, if not all, of the problemsthat are commonly associated with overproducing the Umuproteins from multicopy plasmids (37, 52).By physically separating the umu genes on different plas-

mids, we were able to randomly mutate the umuC plasmid inthe absence of umuD'. Potential mutants were easily identifiedby introducing the mutagenized plasmid into RW126/pRW66and screening for SOS-dependent mutator activity with acolorimetric galactose papillation assay (24). A total of 1,500separate colonies were assayed, and those that produced lessthan half of the normal number of Gal' papillae were chosenfor further study. Of 140 potential mutants analyzed, 16appeared to be either partially or completely deficient in SOSmutator activity as judged by their ability to revert thehisG4(0c) mutation (Table 2). Generally, most plasmids wereunable to promote both spontaneous and MMS-induced mu-tagenesis. The exceptions were pRW124-132, which appearedproficient at spontaneous mutagenesis but had a reducedcapacity for MMS mutagenesis; pRW124-126, which was par-

J. BACTERIOL.

on April 6, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

NOVEL umuC MUTANTS 5015

(#73A) (#73B) umuCI25 (#140) umuC3

PVD\ G ~ SFU

. E P T

3I 4.A.T.

.R.D.R]Q A .

R. NPT.SEQS. WI.DTDIR.

LERTVREL

........

........

........

........

K(#34B)

C

I[.

ETP

q....1%.....:-uI 1.

I

_xA. . I E 90......FAA%7v 3. AV. .

_E_ .7. ...T_

,6

L.N.. R. AL180--DGA. LK..4. ILAEI.E IL. . G.Q. .. DE ...V. Q

S. LEORl.AL

Term(#l0)

F JDY v AJ.270

.ETLT. ..TG

.. K D . .VQ.. T. E......CQ

.ff . . P GVLa36cs

Mi.|M ... GVPVJN 3 60

C P K Term(#65) (#77) umuC25 umuCl22

'.DIP.AR.K 425

Term Term Term L(#139) (#73C) (#120)(#111)

FIG. 3. Alignment of five UmuC-like proteins and position ofE. coli umuC mutants. The five previously characterized UmuC-like proteins werealigned by using the program GeneWorks (IntelliGenetics, Inc.). Areas that are boxed indicate regions that are 100% conserved in all five of thepreviously described UmuC-like proteins. Residues that are found in the consensus protein are indicated with a period (.) Residues that are

different from the consensus are indicated by the single-letter code for that residue. umuC mutations and the plasmid from which they were

identified are indicated below the alignment. Although mutations AT3 and GS24 cause only a slight loss of function (Table 3), they are noted inaddition to the position of the four previously identified chromosomal umuC mutants (umuC25, umuC36, umuC104, and umuC122) and one

plasmid-encoded umuC mutant (umuC125) for the sake of completeness.

tially proficient at both spontaneous and induced mutagenesis;and plasmids pRW124-10, pRW124-92, and to a lesser extentpRW124-65 and pRW124-134, which were defective for spon-taneous mutagenesis but partially proficient for MMS-induced-mutagenesis (Table 2).We were interested in determining if any of the plasmid

mutants could complement a missense umuC mutation. Sincethe umuC36 allele is the best characterized of the threechromosomal umuC mutants, all 16 plasmids were isolated andintroduced into TK610 (umuC36 uvrA6) and analyzed for their

ability to promote MMS-induced mutagenesis. Most plasmidsfailed to complement umuC36, and only plasmids pRW124-10,pRW124-126, and pRW124-132 showed any appreciable in-crease in inducible mutagenesis (Table 2).

Nucleotide changes in pRW124 derivatives unable to pro-mote mutagenesis functions. Our strategy to isolate umuCmutants was based upon the assumption that any plasmid thatfailed to promote mutagenesis functions would probably con-tain a mutation in umuC. Although DNA sequence analysisproved this to be the case for 12 of the 16 plasmids (Table 2),

Consensus

MucBIrrpBSamBlirllC (St)UmnuC (Ec)

Consensus

MucBIrrpBSamBthlmC (St)UEmuC (Ec)

Consensus

MucBLrrpBSamBUmuc (St)UmulC (Ec)

Consensus

MucBIrrpBSamBUnuC (St)UmuC (Ec)

Consensus

MucBTrTpBSamBUmuC (St)UmkuC (Ec)

VOL. 176, 1994

on April 6, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

5016 WOODGATE ET AL.

u=T.EZ).< 0.

T- 0 c N co 00COo T- CN cqCY) CY) XV dr~ ~ ~ T 0LDOCD h N

q- XV CD r*. r** 05)

*- UmuC

FIG. 4. Steady-state levels of wild-type and mutant UmuC proteins. The level of UmuC protein expressed from pRW124 and mutant derivativesunable to promote mutagenesis was analyzed by the chemiluminescent immunoassay. The control lanes at either end of the figure are from strainRW126/pRW66 alone or from the same strain that harbors the wild-type UmuC plasmid, pRW124. The numbers above the individual lanes referto the derivative of plasmid pRW124 from which the extract was obtained. The position of wild-type UmuC is indicated on the right.

to our surprise, four plasmids that resulted in poorly mutableor nonmutable phenotypes (pRW124-40, pRW124-100,pRW124-132, and pRW124-138) did not contain any changesin the structural umuC gene. We have classified these plasmidsas expression mutants.

In vitro, hydroxylamine almost exclusively produces C--Ttransitions. With the exception of pRW124-77, which con-tained a T->C transition, all of the mutations identified in the12 plasmids with structural mutations in umuC could beattributed to a C-*T transition in either the coding or noncod-ing strand of umuC (Table 2). (It is possible that the mutationidentified in pRW124-77 occurred in vivo as a consequence ofthe normal spontaneous mutator activity exhibited by RW126/pRW66/pRW124 [Table 2].)Although many of the 12 plasmids with a structural umuC

mutation contained only a single-base-pair substitution, fiveplasmids, pRW124-34, pRW124-65, pRW124-73, pRW124-139, and pRW124-140, contained multiple changes. However,the two changes in pRW124-65 were in the same codon, andone of the two mutations in pRW124-139 and pRW124-140,respectively, was silent. Although hydroxylamine mutagenesisis generally thought to be random, most of the structuralmutants identified in our screen for nonmutable umuC plas-mids were localized to two regions of the UmuC protein; fourmissense mutations identified in plasmids pRW124-126,pRW124-92, pRW124-34, and pRW124-134 were found in astretch of 30 amino acids (residues 133 to 162) in the middle ofthe protein (Table 2; Fig. 3), whereas four other mutationsencoded by plasmids pRW124-139, pRW124-73C, pRW124-120, and pRW124-111 (the first three of which resulted in atruncated UmuC protein) were localized in the last 50 carbox-yl-terminal amino acid residues (Table 2; Fig. 3).Of particular interest was the C-*T transition mutation

found in pRW124-10 that changed the glutamine residue atposition 172 to a TAG amber stop codon. Strain RW126carries the amber suppressor mutation supE44 that encodes fora glutamine-inserting tRNA rather than a termination signal(22). The efficiency of extragenic suppression varies and, in thecase of supE44, is estimated to occur somewhere between 0.8

and 20% of the time (15, 22, 39). Thus, since the wild-typeresidue in pRW124 already encodes for glutamine, if sup-pressed by SupE44, pRW124-10 would produce wild-typeUmuC. Although this plasmid does contain a bona fidemutation in the structural umuC gene, it was considered aconditional expression mutant since it produces wild-typeUmuC in an amber-suppressing background, albeit at some-what lower levels. Interestingly, as noted above, the plasmidwas unable to promote efficient spontaneous mutagenesis butwas partially proficient for MMS-induced mutagenesis (Table2; and below).

Steady-state levels of plasmid-encoded UmuC protein.Upon analysis of the UmuC protein expressed from all 16plasmids, all four that lacked structural mutations exhibitedelevated steady-state levels of UmuC compared to the wildtype (Table 2; Fig. 4). The increase varied between twofold forplasmid pRW124-132 and eightfold for plasmid pRW124-40.We believe that the increased expression occurs through anincrease in the copy number of the plasmid since the amountof plasmid DNA isolated from these derivatives was muchhigher than normal and was comparable to that of high-copy-number plasmids (data not shown). In contrast, plasmids withstructural changes in umuC exhibited somewhat lower steady-state levels than wild type (Tables 2 and 3; Fig. 4 and 5). Theexception was pRW124-139, which gave approximately 2.3-foldhigher UmuC. However, like the other plasmids with increasedUmuC levels, this probably occurs through an additionalplasmid copy number mutation (see below). As might beexpected from the nucleotide changes, plasmids pRW124-139,pRW124-73, and pRW124-120 produced truncated UmuCproteins. With the exception of the UmuCs encoded bypRW124-10, pRW124-34, pRW124-73, and pRW124-140,steady-state levels of all the plasmid encoded mutant UmuCproteins were comparable to, or greater than, the chromo-somally expressed UmuC in a recA+ lexA(Def) strain (cf. Fig.2 and 4). Since this amount is clearly sufficient to supportcellular mutagenesis, these plasmid mutants are presumablyphenotypically Umu- because of a loss of function rather thana simple reduction in UmuC protein (cf. pRW124-10).

J. BACI'ERIOL.

on April 6, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

NOVEL umuC MUTANTS 5017

TABLE 3. Separation of the multiple mutations identified in pRW124-73 and pRW124-139 and comparison with chromosomal umuCmissense mutations cloned into pRW124

Mutagenesis functions (His' mutants/plate)a Mutation in umuCbumuC Relative amtplasmid Spontaneous MMS induced umuC36 Nucleotide Amino acid of UmuCc

complementation Nueoie Anoad

Noned 3 5 14 NAW NA NA NApRW124(wt) 110 431 267 NA NA NA 1.0pRW124-73A 62 389 223 426G-*A AT3 umuC263 0.27

pRW124-73B 86 402 235 489G-*A GS24 umuC264 0.57pRW124-73C 3 5 15 1614c-*T QX(Oc)399 umuC265 0.28pRW124-139B 6 9 15 1533c-)T QX(Oc)372 umuC266 1.2

pRW124-C25 3 5 10 1288c-)A TK290 umuC25 0.06pRW124-C36 8 54 24 642G-),A EK75 umuC36 0.18pRW124-C104 12 57 15 720G-+A DN101 umuC1O4 0.9

a The data presented are the mean values from at least three separate cultures, with three plates per culture. RW126 was used to determine the ability of plasmidsto promote spontaneous and MMS mutagenesis, while TK610 was used to assess MMS-induced umuC36 complementation.

b Amino acid changes are noted by the single-letter code for the original residue followed by the residue in the mutant protein. X denotes a termination codon. Bothof the truncated UmuC proteins occurred as a result of an Ochre (Oc) termination codon. The nucleotide and amino acid changes in the chromosomal umuC25,umuC36, and umuC104 mutants were obtained from the results of Koch et al. (31).

' The data presented are representative and are based upon the densitometric analysis of multiple exposures of Fig. 6.d The data presented for spontaneous and MMS mutagenesis are in the presence of the low-copy-number umuD' plasmid pRW66.eNA, not applicable.

Ability of plasmid umuC mutants to promote IRR. Inaddition to their role in inducible mutagenesis, the Umuproteins (or functional homologs) are required for inducedreplisome reactivation (IRR) in recA718 and recA4727 strains(24, 29, 49, 54). The ability to promote IRR can easily bemonitored in vivo by the dramatic increase in UV resistanceconferred by umu-complementing plasmids to a recA718AumuDC strain (24, 54). Although the recA718 AumuDClexA71::TnS strain RW126 was not as sensitive as its lexA+

100

0~

U.

z

10-2

0 4 8 12 0 4 8 12UV (Jm-2)

FIG. 5. UV survival curves of RW126/pRW66 and the same strainharboring pRW124 (wild-type) and mutant derivatives. Left panel,expression mutants; right panel, structural mutants; all strains carry thelow-copy-number UmuD' plasmid pRW66. 0, pRW66 alone; 0, pluspRW124 (wild type); A, plus pRW124-132; *, plus pRW124-138; *,plus pRW124-100; A, plus pRW124-40; O, plus pRW124-10; V, pluspRW124-126; O. plus pRW124-92. All of the other structural plasmidmutants failed to restore significant UV resistance and gave resultssimilar to those with pRW124-92 (data not shown).

counterpart (24), there was a significant increase in UVresistance in the presence of pRW66 and pRW124 (Fig. 5).Two of the expression mutants, pRW124-132 and pRW124-138, that exhibited 2- and 2.4-fold increases in UmuC protein,respectively, yielded intermediate survival (Fig. 5, left panel),whereas plasmids that produced much higher (pRW124-40 andpRW124-100) or lower (pRW124-10) levels of UmuC had noeffect on UV survival (Fig. 5, left panel). Of the structuralumuC mutants, only pRW124-126 was partially proficient atrestoring UV resistance (Fig. 5, right panel).

Separation of complex umuC mutations and comparison ofchromosomal umuC mutants cloned into pRW124. As notedabove and in Table 2, some of the plasmid mutants that weidentified contained multiple mutations. The most notable waspRW124-73, which had three separate mutations. We havesubcloned these mutations onto separate plasmids and com-pared their phenotypes. Both of the 5'-most mutations(pRW124-73A [AT3] and pRW124-73B [GS24]) resulted inonly a slight loss of mutagenesis functions (Table 3). The AT3change is in a block of residues that are 100% conserved inUmuC homologs, while the GS24 residue is more variable(Fig. 3). We conclude that either the conservative amino acidchanges are readily accommodated at these positions or that atleast the first 24 amino acids of UmuC are not critical forUmuC's mutagenesis functions. In contrast, the most 3' muta-tion (pRW124-73C; QX399) was completely defective forUmuC functions (Table 3). The original parental plasmid

A wt 73A 73B 73C 139B C25 C36 C104

UmuC

FIG. 6. Steady-state levels of plasmid and chromosomal umuCmutations cloned into pRW124. The level of UmuC protein expressedfrom pRW124 and its mutant derivatives was analyzed by the chemi-luminescent immunoassay. The numbers above the individual lanesrefer to the derivative of plasmid pRW124 from which the whole-cellextract was obtained. The position of wild-type UmuC is indicated onthe left.

VOL. 176, 1994

on April 6, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

5018 WOODGATE ET AL.

>0 100

00

+ G)

coC0.

0 2 4 6 8UV (Jm-2)

FIG. 7. UV-induced mutagenesis conferred on RW126/pRW66 bypRW124 and mutant derivatives. Neither pRW124-10 nor pRW124-92reproducibly gave any His+ mutants over the strain lacking pRW124and for the sake of simplicity, results for all three strains are showntogether (A); 0, plus pRW124 (wild type); 0, plus pRW124-126.

pRW124-73 yielded a low steady-state level of UmuC protein(Table 2; Fig. 4); however, this seems to have arisen as acombination of the three different mutations since the individ-ual mutants showed a somewhat increased level of the mutantUmuC protein (Fig. 6; Table 3).

Like pRW124-73C, the structural mutation in pRW124-139B (QX372) resulted in complete loss of mutagenic activity.This mutation results in a UmuC protein lacking the last 50carboxyl amino acids, and although the level of the truncatedUmuC was lower than that of the original parent plasmid,expression (and stability) of the truncated UmuC protein wassimilar to that of wild-type UmuC (Fig. 6; Table 3). These data,together with the nonmutable phenotypes of plasmidspRW124-120 and pRW124-111 (Table 2), suggest that thecarboxyl terminus of UmuC is critical for mutagenic activity.Three chromosomal missense umuC mutants have been

reported (28, 31, 48), and we were interested in comparing theeffects of the chromosomal mutants in our plasmid expressionsystem. Neither umuC36 nor umuC104 promoted any signifi-cant spontaneous mutator activity; however, both exhibited avery limited ability to promote MMS-induced mutagenesis in

150

wC

m 8 100

2

I

0

MMS (RI)FIG. 8. MMS-induced mutagenesis conferred on RW126/pRW66

by pRW124 and mutant derivatives. U, pRW66 alone; 0, pluspRW124 (wild type); A, plus pRW124-10; O, plus pRW124-92; 0, pluspRW124-126.

120

- 100

Co

cn

CD

co 80

c 6

0 400

C 60

w

2 2

0.5 1.0 1.5

UV(Jm-2)

FIG. 9. UV-induced mutagenesis conferred on RW222/pRW66 bypRW124 and mutant derivatives. *, pRW66 alone; *, plus pRW124(wild type); 0, plus pRW124-10; V, plus pRW124-126; A, pluspRW124-92.

the qualitative plate assay (Table 3). When these plasmidswere assayed for their ability to complement the chromosomalumuC36 mutation in TK610, plasmid-encoded umuC104 failedto complement, but the plasmid-encoded umuC36 produced amodest but reproducible twofold increase in the number ofHis' mutants (Table 3). This finding is consistent with aprevious report suggesting that umuC36 may have a leakyphenotype and is able to promote mutagenic activity undercertain conditions (5). By comparison, the umuC25 mutationappeared to be completely defective for all mutagenic activity(Table 3).With respect to the steady-state levels of the mutant pro-

teins, UmuC104 protein was similar to wild-type UmuC,UmuC36 was slightly reduced, and UmuC25 had the loweststeady-state level of the mutant UmuC proteins (Table 3; Fig.6).

Further characterization of plasmids pRW124-10, pRW124-92, and pRW124-126. On the basis of the qualitative His'reversion assay, the UmuC proteins encoded by plasmidspRW124-10, pRW124-92, and pRW124-126 were able to pro-mote limited mutagenesis functions (Table 2). To determinethe efficiency of mutagenic activity more accurately, we per-formed quantitative UV mutagenesis assays (Fig. 7). To oursurprise, neither plasmids pRW124-10 nor pRW124-92 yieldedany UV-induced mutants that were reproducibly above back-ground (Fig. 7). In contrast, pRW124-126 yielded approxi-mately one-fifth of the mutants seen with the wild-type plas-mid. Since the qualitative mutagenesis experiment used thechemical mutagen MMS rather than UV light, we consideredthe possibility that the differences were due to the differenttypes of lesion generated by each treatment. However, quan-titative mutagenesis assays with MMS gave essentially similarresults (Fig. 8).The difference in the qualitative and quantitative assay may

J. BACTERIOL.

on April 6, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

NOVEL umuC MUTANTS 5019

be explained by the fact that in the qualitative assay, growingcells are exposed to mutagen (from the disk in the center of theplate) for the duration of the experiment, whereas in thequantitative assay, cells are exposed to mutagen (UV light or

MMS) for only a short period of time. Since RW126 islexA(Def) uvr' and therefore expresses nucleotide excisionrepair constitutively, we hypothesized that lesions might berapidly removed via error-free excision repair before the Umuproteins could act to promote error-prone repair. To test thishypothesis, plasmids were introduced into RW222 (1exA71::Tn5 recA718 AumuDC uvrA6). Indeed, in the excision-defec-tive background, both pRW124-10 and pRW124-92 promotedmuch higher levels of mutagenesis than in the excision-proficient background, giving approximately one-third to one-

quarter of the level of mutagenesis seen with the wild-typeplasmid (Fig. 9). By comparison, plasmid pRW124-126 (whichwas partially proficient for mutagenesis functions in the exci-sion-proficient strain) yielded only slightly more mutants in theexcision-defective strain, with approximately one-quarter toone-third of the His+ mutants seen with the wild-type plasmid(cf. Fig. 7 and 9).

DISCUSSION

As part of our studies into the mechanisms of induciblemutagenesis in E. coli, we have isolated several umuC plasmidsthat have a reduced ability to promote UmuC's functions inSOS-dependent spontaneous and induced mutagenesis andalso appear unable to promote IRR in a recA 718 mutant.Interestingly, these plasmids fell into two classes: those mu-

tants that produced either too little or too much wild-typeUmuC and those mutants that contained nucleotide changes inthe structural umuC gene.

Expression mutants. We isolated five plasmids (pRW124-10, pRW124-40, pRW124-100, pRW124-132, and pRW124-138) that caused either a partial or complete loss of UmuCactivity yet did not contain any structural change in the UmuCprotein. Analysis of the UmuC protein expressed from theseplasmids revealed that four produced elevated levels of UmuC,while the fifth [pRW124-10; QX(Am)172] produced signifi-cantly less and was not detectable.The phenotype observed when a limited amount of UmuC is

synthesized is relatively easy to rationalize. SOS mutagenesis isnot a passive process but, rather, occurs after a series ofinefficient reactions. For example, the umuDC operon is tightlyregulated by LexA and would be expected to be one of the lastoperons induced by DNA damage (30). Likewise, posttransla-tional cleavage of UmuD to mutagenically active UmuD' isalso inefficient (8, 56). As a consequence of these inefficientreactions, the cell potentially has time to repair DNA damagevia error-free pathways such as nucleotide excision repair or

recombinational repair before being committed to error-prone

translesion DNA synthesis (8, 56). Reducing the amount ofavailable UmuC protein may simply further shift the balancetowards error-free transcription-coupled nucleotide excisionrepair rather than error-prone repair (Fig. 7 and 8). When thispathway is genetically inactivated, the limited amount ofUmuC can once more promote error-prone repair (Fig. 9).The phenotypes produced by overexpression of UmuC are

not as easy to explain. Plasmids pRW124-40 and pRW124-100,which produced 8-fold and 4.5-fold more UmuC than wildtype, respectively, were phenotypically UmuC- in all of theassays tested. In contrast, pRW124-138, which produced -2.4-fold more UmuC, was defective for mutagenesis functions butpartially proficient at IRR. Plasmid pRW124-132, whichyielded only about twofold more UmuC, was partially profi-

cient at IRR and induced mutagenesis and almost fully profi-cient at spontaneous mutagenesis. If our earlier calculationsare correct, RW126/pRW66/pRW124-132 will contain approx-imately 3,500 UmuD'C molecules per cell. This amount there-fore appears to be close to the uppermost limit of UmuD'Cthat the cell can tolerate before expressing an aberrant phe-notype. Overproduction of the UmuDC proteins from multi-copy plasmids has previously been shown to cause cold sensi-tivity in lexA(Def) cells (37). However, none of the fourplasmids that expressed elevated UmuC exhibited a cold-sensitive phenotype (data not shown). Evidently, the cold-sensitive phenotype requires overproduction of both UmuDand UmuC proteins.Marsh et al. (36) previously reported a plasmid umuC

mutant (umuC125) that resulted in increased sensitivity to UVlight. At the present time, we have not directly determined ifour overproducing plasmids cause UV sensitization. Sincestrain RW126 is already UV sensitive, it is difficult to deter-mine if the UV survival of the strain is simply due to aninability of the umuC plasmid to restore UV resistance or isdue to UV sensitization (Fig. SA). It is possible, therefore, thatthe UmuC- phenotype may occur as a result of increased celllethality.UmuC is thought to physically interact with RecA protein

(21), and very recently Sommer et al. (47) have presentedevidence that overproduction of the UmuD'C proteins inhibitsthe ability of RecA to promote homologous recombination. Itis conceivable that the elevated level of UmuC produced fromour mutant plasmids not only inhibits RecA recombinationalactivities but also compromises its roles in SOS mutagenesis.

It has been difficult to test many of the hypotheses presentedabove because of the nature of the overproduction. All of theUmuC overproducers appeared to be expressed from plasmidswith an increased copy number. These plasmids were some-what unstable, and when introduced into different geneticbackgrounds, many of the survivors had reverted back to alower copy number and a UmuC' phenotype (data notshown). The simple conclusion that can be drawn from thestudies with the expression mutants is that either too much ortoo little UmuC can result in the same loss of functionphenotype.

Structural umuC mutants. Our initial screen for umuCplasmid mutants identified 11 plasmids with structural changesin the UmuC protein. One of these plasmids, pRW124-73, wascomplex and contained three separate changes. We subclonedthese mutations onto individual plasmids and cloned thechromosomal umuC25, umuC36, and umuC104 mutations intopRW124, giving us a total of 17 plasmids expressing a mutantUmuC protein. While we have clearly not identified all of thepossible changes that can lead to a loss of UmuC function,most of the mutations in our plasmid mutants were localized toparticular regions of the protein. For example, four mutationswere found in a stretch of 30 amino acids (residues 133 to 162).Three of these plasmids (pRW124-92, pRW124-126, andpRW124-134) exhibited somewhat similar phenotypes in thatthey were partially leaky for some, or all, of UmuC activities(Table 2). The fourth mutation (AT152/pRW124-34A) mayalso exhibit a similar phenotype but is masked by the secondmutation in pRW124-34 (EK228).Another region where mutations were localized was the

carboxyl terminus of the protein. Plasmid pRW124-111, with aSer-> Leu change just 15 residues from the end of UmuC,resulted in a complete loss of function, as did those thatresulted in truncated UmuC, lacking 50, 24, and 18 residues(plasmids pRW124-139, pRW124-73C, and pRW124-120, re-

VOL. 176, 1994

on April 6, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

5020 WOODGATE ET AL.

spectively). We conclude, therefore, that the carboxyl-terminalregion is critical for function.

If we include the chromosomal umuC25 mutation togetherwith those found in plasmids pRW124-65 and pRW124-77,then a third smaller region (residues 279 to 290) also appearsimportant for activity.

In contrast, we found few mutations in the very aminoterminus of UmuC. Mutations were recovered at residues 3(AT3) and 24 (GS24) but were only identified because theyoccurred in a plasmid (pRW124-73) with a mutation in thecarboxyl terminus (QX399). When these mutations were sub-cloned into undamaged vectors, both AT3 and GS24 causedonly a slight loss of activity. Indeed, had these mutationsoccurred independently, their modest change in phenotypewould not have been detected in the initial screen for umuCmutants. On the basis of these observations, it would appearthat amino acid substitutions in the first few residues of theprotein may not be as critical as those located in the carboxylterminus.Although we believe that we may have identified potential

functional domains of UmuC, the biochemical defect of eachof these putative domains remains to be resolved. Umu-dependent error-prone translesion DNA synthesis is believedto occur as part of a multiprotein complex that requires proteininteractions between UmuC and UmuD' (57), RecA (2, 20, 21,50), and probably subunits of DNA polymerase III holoenzyme(19, 27, 43, 51). It is conceivable that the umuC mutantsisolated here are defective in some, if not all, of the aforemen-tioned interactions. Further characterization of these mutantsshould therefore provide insights into the functional domainsof UmuC as well as its biochemical activities.

ACKNOWLEDGMENTS

We thank Don Ennis and John Little for bacterial strains, GrahamWalker for plasmid pDS110, Konstantin Chumakov for synthesizingthe synthetic oligonucleotide used to construct pRW124, and MaryTrucksess and John Bond for synthesizing the umuC25 allele-specificoligonucleotide.

REFERENCES1. Bagg, A., C. J. Kenyon, and G. C. Walker. 1981. Inducibility of a

gene product required for UV and chemical mutagenesis inEscherichia coli. Proc. Natl. Acad. Sci. USA 78:5749-5753.

2. Bailone, A., S. Sommer, J. Knezevic, M. Dutreix, and R. Devoret.1991. A RecA protein mutant deficient in its interaction with theUmuDC complex. Biochimie 73:479-484.

3. Banerjee, S. K., A. Borden, R. B. Christensen, J. E. LeClerc, andC. W. Lawrence. 1990. $OS-dependent replication past a singletrans-syn T-T cyclobutane dimer gives a different mutation spec-trum and increased error rate compared with replication past thislesion in uninduced cells. J. Bacteriol. 172:2105-2112.

4. Banerjee, S. K., R. B. Christensen, C. W. Lawrence, and J. E.LeClerc. 1988. Frequency and spectrum of mutations produced bya single cis-syn thymine-thymine dimer in a single-stranded vector.Proc. Natl. Acad. Sci. USA 85:8141-8145.

5. Bates, H., B. A. Bridges, and R. Woodgate. 1991. Mutagenic DNArepair in Escherichia coli. XX. Overproduction of UmuD' proteinresults in suppression of the umuC36 mutation in excision defec-tive bacteria. Mutat. Res. 250:199-204.

6. Battista, J. R., T. Ohta, T. Nohmi, W. Sun, and G. C. Walker. 1990.Dominant negative umuD mutations decreasing RecA-mediatedcleavage suggest roles for intact UmuD in modulation of SOSmutagenesis. Proc. Natl. Acad. Sci. USA 87:7190-7194.

7. Bridges, B. A., and R Woodgate. 1985. Mutagenic repair inEscherichia coli: products of the recA gene and of the umuD andumuC genes act at different steps in UV-induced mutagenesis.Proc. Natl. Acad. Sci. USA 82:41934197.

8. Burckhardt, S. E., R Woodgate, R H. Scheuermann, and H.Echols. 1988. UmuD mutagenesis protein of Escherichia coli:

overproduction, purification and cleavage by RecA. Proc. Natl.Acad. Sci. USA 85:1811-1815.

9. Cebula, T. A., and W. H. Koch. 1990. Analysis of spontaneous andpsoralen-induced Salmonella typhimurium hisG46 revertants byoligodeoxyribonucleotide colony hybridization: use of psoralens tocross-link probe to target sequences. Mutat. Res. 229:79-87.

10. Churchward, G., D. Belin, and Y. Nagamine. 1984. A pSC101-derived plasmid which shows no sequence homology to othercommonly used cloning vectors. Gene 31:165-171.

11. Donnelly, C. E., and G. C. Walker. 1989. groE mutants ofEscherichia coli are defective in umuDC-dependent UV mutagen-esis. J. Bacteriol. 171:6117-6125.

12. Donnelly, C. E., and G. C. Walker. 1992. Coexpression of UmuD'with UmuC suppresses the UV mutagenesis deficiency of groEmutants. J. Bacteriol. 174:3133-3139.

13. Dutreix, M., B. Burnett, A. Bailone, C. M. Radding, and RDevoret. 1992. A partially deficient mutant, recA41730, that fails toform normal nucleoprotein filaments. Mol. Gen. Genet. 232:489-497.

14. Dutreix, M., P. L. Moreau, A. Bailone, F. Galibert, J. R Battista,G. C. Walker, and R Devoret. 1989. New recA mutations thatdissociate the various RecA protein activities in Escherichia coliprovide evidence for an additional role for RecA protein in UVmutagenesis. J. Bacteriol. 171:2415-2423.

15. Eggertsson, G., and D. So6l. 1988. Transfer ribonucleic acid-mediated suppression of termination codons in Escherichia coli.Microbiol. Rev. 52:354-374.

16. Elledge, S. J., and G. C. Walker. 1983. Proteins required forultraviolet light and chemical mutagenesis. Identification of theproducts of the umuC locus of Escherichia coli. J. Mol. Biol.164:175-192.

17. Ennis, D. G., B. Fisher, S. Edmiston, and D. W. Mount. 1985. Dualrole for Escherichia coli RecA protein in SOS mutagenesis. Proc.Natl. Acad. Sci. USA 82:3325-3329.

18. Ennis, D. G., A. S. Levine, W. S. Koch, and R Woodgate. Analysisof recA mutants with altered SOS functions. Mutat. Res., in press.

19. Foster, P. L., and A. D. Sullivan. 1988. Interactions betweenepsilon, the proofreading subunit of DNA polymerase III, andproteins involved in the SOS response of Escherichia coli. Mol.Gen. Genet. 214:467-473.

20. Frank, E. G., J. Hauser, A. S. Levine, and R. Woodgate. 1993.Targeting of the UmuD, UmuD' and MucA' mutagenesis proteinsto DNA by RecA protein. Proc. Natd. Acad. Sci. USA 90:8169-8173.

21. Freitag, N., and K. McEntee. 1989. "Activated"-RecA proteinaffinity chromatography of LexA repressor and other SOS-regu-lated proteins. Proc. Natl. Acad. Sci. USA 86:8363-8367.

22. Garen, A. 1968. Sense and nonsense in the genetic code. Science160:149-159.

23. Hauser, J., A. S. Levine, D. G. Ennis, K. M. Chumakov, and RWoodgate. 1992. The enhanced mutagenic potential of the Mu-cAB proteins correlates with the highly efficient processing of theMucA protein. J. Bacteriol. 174:6844-6851.

24. Ho, C., 0. I. Kulaeva, A. S. Levine, and R Woodgate. 1993. A rapidmethod for cloning mutagenic DNA repair genes: isolation ofumu-complementing genes from multidrug resistance plasmidsR391, R446b, and R471a. J. Bacteriol. 175:5411-5419.

25. Horsfall, M. J., and C. W. Lawrence. 1994. Accuracy of replicationpast the T-C (6-4) adduct. J. Mol. Biol. 235:465471.

26. Isackson, P. J., and K. P. Bertrand. 1985. Dominant negativemutations in the TnlO tet repressor: evidence for use of theconserved helix-turn-helix motif in DNA binding. Proc. Natd.Acad. Sci. USA 82:6226-6230.

27. Jonczyk, P., L. Fualkowska, and Z. Ciesla. 1988. Overproduction ofthe e subunit of DNA polymerase III counteracts the SOSmutagenic response ofEscherichia coli. Proc. Natl. Acad. Sci. USA85:9124-9127.

28. Kato, T., and Y. Shinoura. 1977. Isolation and characterization ofmutants of Escherichia coli deficient in induction of mutations byultraviolet light. Mol. Gen. Genet. 156:121-131.

29. Khidhir, M. A., S. Casaregola, and I. B. Holland. 1985. Mecha-nism of transient inhibition of DNA synthesis in ultraviolet-irradiated E. coli: inhibition is independent of RecA whilst recov-

J. BAc-IERIOL.

on April 6, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

NOVEL umuC MUTANTS 5021

ery requires RecA protein itself and an additional, inducible SOSfunction. Mol. Gen. Genet. 199:133-140.

30. Kitagawa, Y., E. Akaboshi, H. Shinagawa, T. Horii, H. Ogawa, andT. Kato. 1985. Structural analysis of the umu operon required forinducible mutagenesis in Escherichia coli. Proc. Natl. Acad. Sci.USA 82:4336-4340.

31. Koch, W. H., D. G. Ennis, A. S. Levine, and R. Woodgate. 1992.Escherichia coli umuDC mutants: DNA sequence alterations andUmuD cleavage. Mol. Gen. Genet. 233:443-448.

32. Lawrence, C. W., S. K. Banerjee, A. Borden, and J. E. LeClerc.1990. T-T cyclobutane dimers are misinstructive rather thannon-instructive, mutagenic lesions. Mol. Gen. Genet. 222:166-168.

33. Lawrence, C. W., A. Borden, S. K. Banerjee, and J. E. LeClerc.1990. Mutation frequency and spectrum resulting from a singleabasic site in a single-stranded vector. Nucleic Acids Res. 18:2153-2157.

34. LeClerc, J. E., A. Borden, and C. W. Lawrence. 1991. Thethymine-thymine pyrimidine-pyrimidone(6-4) ultraviolet lightphotoproduct is highly mutagenic and specifically induces 3'thymine to cytosine transitions in Escherichia coli. Proc. Natl.Acad. Sci. USA 88:9685-9689.

35. Lin, L.-L., and J. W. Little. 1988. Isolation and characterization ofnoncleavable (Ind-) mutants of the LexA repressor of Escherichiacoli K-12. J. Bacteriol. 170:2163-2173.

36. Marsh, L., T. Nohmi, S. Hinton, and G. C. Walker. 1991. Newmutations in cloned Eschenichia coli umuDC genes: novel pheno-types of strains carrying a umuC125 plasmid. Mutat. Res. 250:183-197.

37. Marsh, L., and G. C. Walker. 1985. Cold sensitivity induced byoverproduction of umuDC in Escherichia coli. J. Bacteriol. 162:155-161.

38. Miller, J. H. 1972. Experiments in molecular genetics. Cold SpringHarbor Laboratory, Cold Spring Harbor, N.Y.

39. Miller, J. H. 1992. A short course in bacterial genetics: a labora-tory manual and handbook for Escherichia coli and related bacte-ria. Cold Spring Harbor Laboratory Press, Cold Spring Harbor,N.Y.

40. Nohmi, T., J. R Battista, L. A. Dodson, and G. C. Walker. 1988.RecA-mediated cleavage activates UmuD for mutagenesis: mech-anistic relationship between transcriptional derepression and post-translational activation. Proc. Natl. Acad. Sci. USA 85:1816-1820.

41. Nowicka, A., M. Kanabus, E. Sledziewska-G6jska, and Z. Ciesla.Different UmuC requirements for generation of different kinds ofUV-induced mutations in Escherichia coli. Mol. Gen. Genet., inpress.

42. Perry, K. L., S. J. Elledge, B. Mitchell, L. Marsh, and G. C.Walker. 1985. umuDC and mucAB operons whose products arerequired for UV light and chemical-induced mutagenesis: UmuD,MucA, and LexA products share homology. Proc. Natl. Acad. Sci.USA 82:4331-4335.

43. Rajagopalan, M., C. Lu, R Woodgate, M. O'Donnell, M. F.Goodman, and H. Echols. 1992. Activity of the purified mutagen-esis proteins UmuC, UmuD' and RecA in replicative bypass of an

abasic DNA lesion by DNA polymerase III. Proc. Natl. Acad. Sci.USA 89:10777-10781.

44. Sanger, F., S. Nicklen, and A. R Coulson. 1977. DNA sequencingwith chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA74:5463-5467.

45. Sedgwick, S. G., and B. A. Bridges. 1972. Survival, mutation andcapacity to repair single strand DNA breaks after gamma irradi-ation in different exr- strains of Escherichia coli. Mol. Gen. Genet.119:93-102.

46. Shinagawa, H., H. Iwasaki, T. Kato, and A. Nakata. 1988. RecAprotein-dependent cleavage of UmuD protein and SOS mutagen-esis. Proc. Natl. Acad. Sci. USA 85:1806-1810.

47. Sommer, S., A. Bailone, and R Devoret. 1993. The appearance ofthe UmuD'C protein complex in Escherichia coli switches repairfrom homologous recombination to SOS mutagenesis. Mol. Mi-crobiol. 10:963-971.

48. Steinborn, G. 1978. Uvm mutants of Escherichia coli K12 deficientin UV mutagenesis. I. Isolation of uvm mutants and their pheno-typical characterization in DNA repair and mutagenesis. Mol.Gen. Genet. 165:87-93.

49. Sweasy, J. B., and E. M. Witkin. 1993. Novel SOS phenotypescaused by second-site mutations in the recA4430 gene of Eschenrchiacoli. Biochimie 73:437-448.

50. Sweasy, J. B., E. M. Witkin, N. Sinha, and V. Roegner-Maniscalco.1990. RecA protein of Escherichia ccli has a third essential role inSOS mutator activity. J. Bacteriol. 172:3030-3036.

51. Tadmor, Y., R Ascarelli-Goell, R Skaliter, and Z. Livneh. 1992.Overproduction of the 1B subunit of DNA polymerase III holoen-zyme reduces UV mutagenesis in Eschenchia coli. J. Bacteriol.174:2517-2524.

52. Thomas, S. M. 1993. Extreme cold sensitivity of Salmonellatyphimurium umu clones, effects of the umu region and flankingsequences. Mutat. Res. 285:95-99.

53. Walker, G. C. 1984. Mutagenesis and inducible responses todeoxyribonucleic acid damage in Eschenichia coli. Microbiol. Rev.48:60-93.

54. Witkin, E. M., V. Roegner-Maniscalco, J. B. Sweasy, and J. 0.McCall. 1987. Recovery from ultraviolet light-inhibition of DNAsynthesis requires umuDC gene products in recA718 mutantstrains but not in recA+ strains of Escherichia coli. Proc. Natl.Acad. Sci. USA 84:6804-6809.

55. Woodgate, R 1992. Construction of a umuDC operon substitutionmutation in Escherichia coli. Mutat. Res. 281:221-225.

56. Woodgate, R., and D. G. Ennis. 1991. Levels of chromosomallyencoded Umu proteins and requirements for in vivo UmuDcleavage. Mol. Gen. Genet. 229:10-16.

57. Woodgate, R., M. Rajagopalan, C. Lu, and H. Echols. 1989. UmuCmutagenesis protein of Eschenichia coli: purification and interac-tion with UmuD and UmuD'. Proc. Natl. Acad. Sci. USA 86:7301-7305.

58. Woodgate, R., and S. G. Sedgwick. 1992. Mutagenesis induced bybacterial UmuDC proteins and their plasmid homologues. Mol.Microbiol. 6:2213-2218.

VOL. 176, 1994

on April 6, 2020 by guest

http://jb.asm.org/

Dow

nloaded from