Mol Gen Genet (1986) 205:234-239 © Springer-Verlag 1986

Different efficiency of UmuDC and MucAB proteins in UV light induced mutagenesis in Escherichia coli

Manuel Blanco, Guadalupe Herrera, and Vicente Aleixandre Instituto de Investigaciones Citol6gicas, Obra Social de la Caja de Ahorros de Valencia, Amadeo de Saboya 4, E-46010 Valencia, Spain

Summary. Two multicopy plasmids carrying either the umuDC or the mueAB operon were used to compare the efficiency of UmuDC and MucAB proteins in UV mutagen- esis of Eseheriehia coli K12. It was found that in recA + uvr + bacteria, plasmid pIC80, mucAB + mediated UV muta- genesis more efficiently than did plasmid pSE117, umuDC + . A similar result was obtained in lexA51(Def) cells, exclud- ing the possibility that this was due to a differential regula- tion by LexA of the umuDC and mucAB operons. We con- clude that some structural characteristic of the UmuDC and MucAB proteins determines their different efficiency in UV mutagenesis. This characteristic could be also re- sponsible for the observation that in the reeA430 mutant, pIC80 but no pSE117 can mediate UV mutagenesis. In the reeA142 mutant, pIC80 also promoted UV mutagenesis more efficiently than pSEll7. In this mutant, the recombi- nation proficiency, the protease activity toward LexA and the mutation frequency were increased by the presence of adenine in the medium. In recA + uvrB5 bacteria, plasmid pSEI 17, umuDC caused both an increase in UV sensitivity as well as a reduction in the mutation frequency. These negative effects resulting from the overproduction of UmuDC proteins were higher in reeA142 uvrB5 than in reeA + uvrB5 cells. In contrast, overproduction of MucAB proteins in excision-deficient bacteria containing pIC80 led to a large increase in the mutation frequency. We suggest that the functional differences between UmuDC and Mu- cAB proteins might be due to their different dependence on the direct role of RecA protease in UV mutagenesis.

Key words: Ultraviolet mutagenesis - umuDC and mueAB operons - RecA protease - reeA mutants

Introduction

In Escherichia coli, ultraviolet light induced mutagenesis is dependent upon the induction of the umuD and umuC genes (abbreviated as umuDC), which constitute an operon. This induction is included in the cellular SOS response to DNA damage which results from the proteolytic cleavage of the LexA repressor by activated RecA protein (for re- view, see Walker 1984).

A role for RecA protein in UV mutagenesis is to dere- press the umuDC operon, so that a high level of UmuDC

Offprint requests to: M. Blanco

proteins is present in the cell (Bagg et al. 1981). A second role for RecA protein in UV mutagenesis has been sug- gested (Blanco et al. 1982; Ennis et al. 1985) by the absence of UV mutability in lexA(Def) strains bearing the recA430 allele, formerly known as lexB30 (Blanco et al. 1975). It was assumed that the non-mutability of these strains, in which the umuDC operon was derepressed, resulted from the weak protease activity of the RecA430 protein (Roberts and Roberts 1981 ; Devoret et al. 1983).

The plasmid pKMI01 that carries the mueAB operon, an analogue of umuDC (Perry et al. 1985), restores the UV mutability of reeA430 cells (Waleh and Stocker 1979; Blanco and Rebollo 1981). Therefore, it appears that the presence of the inefficient RecA430 protease prevents the UmuDC proteins but not the MucAB proteins from per- forming their function in the mutagenic process. This sug- gests that in spite of the great homology shared by UmuDC and MucAB proteins, they are, at least in part, functionally different.

To further analyse this we have performed a compara- tive study of UV mutagenesis in cells bearing different reeA alleles, in which either the umuDC or the mucAB operon was introduced on a multicopy plasmid. The present report suggests that MucAB proteins are more efficient than UmuDC proteins in mediating UV mutagenesis. Moreover, whereas over-production of MucAB proteins greatly in- creases UV mutagenesis in bacteria lacking excision repair, over-production of UmuDC proteins in such cells reduces both repair and mutagenesis.

Materials and methods

Escherichia coli K12 strains, phages and plasmids. Bacterial strains used are shown in Table 1. The phage used was 2eIindl reeA: :laeZ + (2GE272; Weisemann etal. 1984). Plasmid pIC80 was constructed as follows: Plasmid pGW1700 (Perry and Walker 1982) was partially digested with HincII and the 2 kb fragment containing the mueAB operon was cloned into HpaI-cleaved pLM203, a derivative of pSEl l7 (Marsh and Walker 1985) that has a deletion of the HpaI fragment containing the umuDC operon (see Fig. 1). The ligation mixture was transformed into IC1657, recA430 and colonies were selected for ampicillin resistance and screened for UV mutability.

Purification of plasmids, digestion with restriction endo- nucleases, ligation and transformation procedures were per- formed as described by Maniatis et al. (1982). For the

Table l . Bacterial strains

Strain Genotype Source or reference

235

ICt637

IC1656 IC1657 IC/65t ICll30

IC1638

IC1654 IC1655 IC1652 IClt88

ICt622 ICl198

ICt626 CSH62 CSH74

thr leu thi proA argE3 ilv t~ galK sup-37 strA31 sfiB recA + As IC1637 but srlC300::Tnl0 As IC1637 but srlC300::Tnl0 recA430 As IC1637 but srlC300::Tnl0 recA142 thr thi proA argE3 his-4 ilv tS gaIK sup-37 strA31 sfiB reeA441 uvrB5 thr thi proA argE3 ilv tS galK sup-37 strA31 sfiB recA + uvrB5 As IC1638 but srlC300::Tnl0 As IC1638 but srlC300::Tnl0 reeA430 As ICt638 but srlC300::Tnl0 recA142 thr leu thi proA his-4 argE3 ilv ts galK sup-37 strA31 s f iB recA441 lexA51 malB: :Tn9 zae: :Tn5 As IClt88 but his + recA + thr thi proA his-4 argE3 ilv ts galK sup-37 strA31 sfiB recA441 uvrB5 lexA51 malB: :TmO zjb: :Tn5 As IC1198 but his + reeA + HfrH HfrKL16

his + reeA + derivative of GC3220 (George et al. 1975) Tet ~ transduced into IC1637 Tet r recA430 transduced into IC1637 TeV recA142 transduced into IC1637 recA441 sfiB derivative of ABtt85 (Howard-Flanders et al. 1966) his + recA + derivative of IC1130

Tet ~ transduced into IC1638 Tet r recA430 transduced into IC1638 Tet ~ recA142 transduced into IC1638 lexA51 derivative of GC3220

CSH74 x ICl188 (His + Str r selec) lexA51 derivative of IC1 t 30

CSH74 x ICI198 (His + Str r selec) Miller (1972) Miller (1972)

l k b i i

b l a , ~ oj_i

pSE117

I I /Jl i JK[ Hp Bg BgSH3 SmSmBHp

I npt ', ~ , pLM203

S B C K IHp S Sm Bg C H3 E 1

m u c B ~1

r l

J ) I / .,c o H2 Bg H2 SmH2

Fig. 1. Restriction map of the recombinant plasmids pSEII7 and pIC80 which carry respectively the umuDC and the mucAB oper- ons. pLM203 is a derivative of pSEt 17 in which the HpaI fragment containing the umuDC operon has been deleted (Marsh and Walker 1985). pIC80 was constructed by cloning the HincII fragment con- taining the mucAB operon from pGW1700 (Perry and Walker 1982) into HpaI-cleaved pLM203. Sites of restriction enzymes are abbreviated as follows: E, EcoRI; S, SalI; B, BamHI; C, ClaI; K, KpnI; Hp, HpaI; Bg, BglII; H3, HindIII; Sm, Sinai; H2, HincII. Ori, origin of replication; npt, neomycin phosphotransferase gene (kanamycin resistance); bla, /Mactamase gene (ampicillin resis- tance)

lexA51 (Def) strains selection for t ransformants was carried out at 42 ° C.

Media. Medium Y M 9 C contained 400 mg casamino acids (vitamin-free; Difco), 2 g glucose and 2 mg thiamine per litre of YM9 buffer (Blanco and Rebol lo 1981).

Arginine-l imit ing solid medium A9 contained 15 g Difco agar, 2 g glucose, 60 mg each threonine, leucine, proline, histidine, isoleucine and valine and 0.9 mg arginine per litre of YM9 buffer. When indicated, adenine at 100 gg/ml was

added to the A9 medium. Viable bacter ia were scored on plates of A9 medium supplemented with 200 gg/ml argi- nine.

U V irradiation. UV ir radia t ion was per formed with a germi- cidal lamp with a maximal output at 254 nm. Doses were measured with a Latar je t dosimeter.

Bacterial mutagenesis. The reversion of the argE3 ochre muta t ion was measured as follows: Bacteria grown in YM9C at 42°C to about 108 cells/ml, were centrifuged, resuspended at the same concentrat ion in YM9 and 0.1 ml samples were spread onto A9 plates. Arg + colonies were scored after 3 days o f incubat ion at 42 ° C.

Measurement offl-galactosidase. This was performed as de- scribed by Weisemann et al. (1984). f l -Galactosidase activity was determined by the method of Mil ler (1972).

Results

Plasmid pIC80, m u c A B increased the frequency of UV in- duced Arg ÷ mutants in IC1656, recA + uvr ÷ bacteria (Ta- ble 2). In the same host, p lasmid p S E l l 7 , u m u D C caused a smaller increase in UV mutagenesis (Table 2). The differ- ence does not result f rom a differential repression by LexA of both operons since it was also observed in IC1622, recA ÷ uvr ÷ lexA51(Def ) cells lacking LexA prote in (Table 2). In these cells, the increase in UV mutabi l i ty p romoted by pIC80 was lower than in l exA 4, whereas the increase depen- dent on pSE117 was undetectable.

In l exA(Def ) hosts plasmids pSEI17 and pIC80 both p romoted cold sensitivity of the cells so that they grew at 42°C but not at 30°C (see Marsh and Walker 1985). This p rompted us to per form all the experiments at 42 ° C. Thus, any negative influence upon cell growth due to over- product ion of the mutagenic proteins was diminished.

Plasmid pSE117, u m u D C increased UV sensitivity in ex-

236

Table 2. Influence of the plasmids pSE117, u m u D C and plC80, m u c A B on UV-induced mutability of recA + strains

Strain Genotype UV dose (J/m 2) Without plasmid With plasmid With plasmid pSE117, u m u D C plC80, m u e A B

S" M/p b MI c S a M/p b MI ° S" M/p b MI °

IC1656 reeA + & x A + uvr + 0 100 3 - 100 4 - 100 12 - 6 80 40 25 83 188 142 88 612 425

18 58 240 208 65 495 300 76 800 633

IC1622 recA + &xA51 uvr + 0 100 5 -- 100 3 - 100 25 -- 6 95 41 21 90 21 16 94 185 80

18 90 110 62 70 93 73 75 346 202

IC1654 recA + N x A + uvrB5 0 100 4 - 100 10 - 100 26 - 1.5 95 528 340 90 1260 1145 96 1990 1110 3 77 1200 1140 55 680 971 90 2500 1250 6 33 1700 2500 7 112 980 50 2100 1900

IC1626 recA + &xA51 uvrB5 0 100 10 -- 100 21 -- 100 82 -- 1.5 92 640 267 71 442 520 98 1480 1060 3 80 1400 1093 60 821 1140 71 2150 1536 6 31 918 1828 19 603 2400 60 2000 1950

The number given in each case is the average of three experiments " S, percent survival b M/p, number of Arg + revertants per plate c MI, number of induced Arg + revertants per 10 7 survivors

Table 3. Plasmid plC80, m u c A B but not pSE117, u m u D C can mediate u v mutagenesis in recA430 strains

Strain Genotype UV dose (J/m 2) Without plasmid With plasmid pSEII7, u m u D C

With plasmid pIC80, m u c A B

S" M/p b MI c S" M/p b MI c S a M/p b MI c

IC1657 recA430 uvr + 0 100 2 -- 100 2 -- 6 27 3 <1 32 2 <1

12 5 1 <1 12 1 <1

IC1655 reeA430 uvrB5 0 100 3 -- 100 2 -- 1.5 22 4 < 1 28 1 < I 3 4 1 <1 5 2 <1

100 5 - 67 345 223 46 144 136

100 15 - 74 968 691 28 1050 2100

The number given in each case is the average of three experiments " S, percent survial b M/p, number of Arg ÷ revertants per plate c MI, number of induced Arg ÷ revertants per 107 survivors

cision-defective IC1654 bacteria and produced a reduction in the number of mutants per plate (Table 2). Consequently, the mutat ion frequency did not increase in propor t ion to the UV dose: This was in contrast to the dose response o f IC1654 without plasmids or carrying pIC80, m u e A B .

This suggests that the lethal effect preferentially affects the potentially mutant cells. Since this lethal effect was not seen when T n l O 0 0 was inserted into plasmid p S E l l 7 , we con- clude that lethality was due to over-product ion of U m u D C proteins.

Note that IC1626, r e c A + u v r B 5 l e x A 5 1 ( D e f ) bacteria were less sensitive to the UmuDC-dependen t lethal effect (Table 2). This suggests that the derepression of SOS genes has a protective effect against the U m u D C - p r o m o t e d letha- lity.

The existence of functional differences between U m u D C and MucAB proteins was previously suggested by the ob- servation that plasmid pKM101, which carries the m u c A B operon, restored UV mutabili ty to r e c A 4 3 0 bacteria which are otherwise not mutable, and in which the u m u D C operon

appears to be inactive (Blanco and Rebol lo 1981; Blanco et al. 1982). As expected, we found that plasmid pIC80, m u e A B (but not pSE117, u m u D C ) mediated U V mutagene- sis in r e c A 4 3 0 strains (Table 3). Moreover , pSE117 was un- able to promote U V mutagenesis in r e e A 4 3 0 l e x A 5 1 ( D e f ) bacteria (data not shown) al though like pIC80 it did pro- duce cold sensitivity. Note, however, that pSE117 did not render r e c A 4 3 0 u v r B 5 bacteria more UV-sensitive, in con- trast to its effect on r e e A + u v r B 5 cells (Table 3). This would be consistent with an absolute inactivity of U m u D C pro- teins in UV-irradiated r e e A 4 3 0 cells.

The RecA430 protein has a weak protease activity to- wards the LexA repressor (Devoret et al. 1983), which can be seen by the low level of fl-galactosidase activity in- duced by U V irradiat ion in r e e A 4 3 0 cells containing a r e c A : : l a e Z + fusion (Fig. 2A; see Elledge and Walker 1983). This protease defect may prevent the direct role of the RecA protein in SOS mutagenesis, leading to the non- mutabil i ty of r e c A 4 3 0 bacteria. It is clear, however, that the low protease activity of the RecA430 protein would

237

u 8 0 0 0

v

6 0 0 0 -g

4 0 0 0

~'~ 2 0 0 0

A / ~ A I A

• . . t o ~ . 0 " o • •

I I I 0 1 2

x

/ []

l i m e , h

Fig. 2A, B. Kinetics of UV induction of recA-lacZ + fusion. Cells were grown at 37 ° C in YM9C up to OD 0.45 (10 a cells/ml). They were then exposed to UV light (15 J/m2), diluted five-fold in YM9C and incubated at 37 ° C. Samples were removed and assayed for fl-galactosidase specific activity. A lC1656: recA + (2GE272) non- irradiated (A) or UV-irradiated (•); ICt657:recA430 (2GE272) non-irradiated (o) or UV-irradiated (o). B ICt651:recA142 (2GE272) non-irradiated (m); UV-irradiated and incubated in YMgC (n); UV-irradiated and incubated in YM9C supplemented with adenine (100 gg/ml) ( x )

Table 4. Recombination proficiency in recA142 bacteria in Hfr x F- crosses

Strain Genotype Adenine Pro + Str' Relative recombinants/ml recombination

frequency

IC1651 r e c A 1 4 2 - 1.6 × 103 0.01 + 1.9 x 104 0.2

IC1657 r e c A 4 3 0 - 3.5x 104 0.3 + 6.0 x 10 4 0.6

IC1656 r e c A + - 1.2 x IO s 1 + 1.0 x lO s 1

Recombination proficiency was measured in Hfr x F- crosses. The HfrH bacterium CSH62 was mixed with the F- recipient in a ratio of 1 : I0. The mixture was incubated for 1 h at 37 ° C and then plated on selective plates either with (+) or without ( - ) adenine at 100 p,g/ml

be sufficient to allow the MucAB proteins to function in the mutagenic process. The MucAB function still depends on RecA since pIC80 was unable to mediate UV mutagene- sis in recA(DeO bacteria deficient in RecA protease (data not shown).

We have also analyzed UV mutagenesis in bacteria car- rying the recA142 mutation; these bacteria are recombina- tion deficient (Horii and Clark 1973) and have a reduced protease activity as measured by the low induction of 2 prophage (Dutreix et al. 1985). The recA142 phenotype probably results from an altered interaction of RecAll42 protein with DNA and ATP (Roberts and Roberts 1981).

We have showed previously that UV mutagenesis can occur in recA142 bacteria and that it was increased when the selective plates were supplemented with adenine (Blanco et al. 1982). That adenine exerts a similar effect upon re- combination proficiency and LexA cleavage is shown in respectively Table 4 and Fig. 2B. Therefore, the recA142 mutation, which renders both the recombinase function and the protease function of the RecA protein adenine-depen-

dent, enabled a study in the same strain of the influence of changes in these functions upon UV mutagenesis.

In Table 5 we show that in IC1651, recA142 uvr ÷ bacte- ria carrying pIC80, mucAB, UV mutagenesis was greatly increased by the presence of adenine in the selective medi- um, confirming the dependence of MucAB proteins upon RecA protease. In IC1651 bacteria, plasmid pSEI17, urnuDC enhanced mutagenesis to a lesser degree. In con- trast, pSEll7 sensitized IC1652, recA142 uvrB5 bacteria to UV killing, causing a great reduction in the mutant yield. Note that no increase in the mutation frequency was ob- served for IC1652, recA142 uvrB5 with pSEI17 except at 1.5 J/m 2 in the presence of adenine (Table 5). This defect in repair and mutagenesis was dependent on over-produc- tion of UmuDC proteins since it was not seen when mutant derivatives of plasmid pSE117 were used (data not shown). It was therefore similar to that observed in recA-- uvrB5 bacteria containing pSE117 (compare Tables 2 and 5).

Discuss ion

In this paper we have shown that the frequency of UV- induced Arg" revertants in bacteria containing the plasmid pIC80, mucAB was higher than in those bearing pSE117, umuDC. Thus, UV mutagenesis occurs more efficiently when mediated by MucAB proteins rather than UmuDC proteins.

This had been suggested previously by the increased UV mutagenesis observed in wild-type bacteria carrying pKM101, mucAB (see Walker 1984). However, the en- hanced effect of the MucAB analogue of the bacterial UmuDC proteins could have been simply due to a gene dosage effect, since the mucAB operon was present in sever- al copies in contrast with the single chromosomal urnuDC operon. In our experimental system, as each operon is pres- ent on a multicopy plasmid a gene dosage effect could be excluded. In addition, since the differences in the level of UV mutagenesis were also observed in recA" lexA(Def) bacteria bearing pSEll7 or pIC80, they could not result from a differential regulation by LexA of both operons. Although we cannot exclude the possibility that transcrip- tion or translation of the mucAB operon is higher than that of umuDC, the cold-sensitive phenotype of lexA(Def) strains containing pSE117 or pIC80 indicates that UmuDC and MucAB proteins are indeed overproduced by these plasmids. We assume then that some structural characteris- tics of the UmuDC and MucAB proteins determine their different functional efficiency in the SOS mutagenic pro- cess.

This different efficiency of the two similar sets of muta- genic proteins that is observed in recA + bacteria would correlate with the observation that MucAB proteins, but not UmuDC proteins, can mediate UV mutagenesis in the presence of the weak RecA protease coded by the recA430 allele (Table 3). That is, in recA430 cells, the difference in the mutagenic efficiency of UmuDC and MucAB would be maximal. The MucAB proteins appear to be less depen- dent on RecA protease, although as shown by the results with the recA142 strains (Table 5 and Fig. 2B), their effi- ciency for UV mutagenesis increases with the protease strength of RecA protein.

We found that overproduction of UmuDC proteins in excision-repair deficient recA" or recA142 bacteria in-

238

Table 5. Effect of the plasmids pSE117, umuDC and plCS0, mucAB on UV-induced mutability of recA142 strains; influence in the selective medium

of adenine

Strain Genotype Adenine" UV dose (J/m 2) Without plasmid With plasmid With plasmid pSE117, umuDC pIC80, mucAB

S b M/p ~ MId S b M/p c MI d S b M/p c MI a

IC1651 recA142 uvr +

IC1652 recA142 uvrB5

0 100 1 - 100 11 - 100 20 - 6 53 11 9 32 16 17 62 101 103

12 17 5 10 6 2 <1 42 85 124 0 100 6 - 100 12 - 100 17 - 6 88 32 17 82 96 57 84 304 292

12 62 42 30 42 150 177 73 624 553

0 100 12 - 100 30 - 100 56 - 1.5 67 45 34 2 5 <1 55 276 275 3 10 6 < 1 0.06 1 <1 14 15 < 1 0 100 12 - 100 29 - 100 52 - 1.5 88 262 130 17 61 134 98 1400 952 3 40 384 413 0.6 8 <1 87 1950 1500

" When indicated, adenine was added to the selective medium b S, percent survival ° M/p, number ofArg ÷ revertants per plate a MI, number of induced Arg + revertants per 107 suvivors

at 100 gg/ml

creases lethality after UV, causing a reduction in the mutan t yield. In these bacteria, except at very low UV doses, the frequency of induced mutants per survivor was lower than in cells that did not contain plasmid and in bacteria over- producing MucAB proteins (Tables 2 and 5). This suggests that the UmuDC-dependen t lethality preferentially affects the potentially mutan t cells.

Over-production of U m u D C proteins in l e x A ( D e f ) bac- teria produces cold sensitivity (growth at 42°C but not at 30 ° C) as a consequence of a replication block (Marsh and Walker 1985). We believe that this replication inhibi- t ion is related to the UmuDC-dependen t negative effect on repair and mutagenesis which is observed in UV-irradi- ated bacteria that lack excision repair. Although we found this effect in l e x A + cells growing at 42 ° C, we think that lesions in the DNA, by acting as preferential binding sites for U m u D C proteins, could enhance the ability of U m u D C proteins to block D N A replication. In addition, since RecA protein also binds preferentially to UV lesions in vitro (Lu et al. 1986), we can imagine an interaction between RecA and U m u D C proteins targeted at D N A lesions. We further- more suggest that RecA protein, by means of its protease activity, could be involved in overcoming the UmuDC-de- pendent replication inhibit ion in order to allow D N A elon- gation. When this step cannot take place, for example due to a limiting level of RecA protease, the persistence of the replication block should lead to cell death.

The U m u D C proteins would play, therefore, a crucial role both in the block of D N A replication promoted by D N A lesions and in the bypass of these lesions. This is consistent with the role at tr ibuted to these proteins in the two-step model for UV mutagenesis proposed by Bridges and Woodgate (1985).

The MucAB proteins might have a similar role to that performed by U m u D C proteins in the mutagenic process. Then, if in the presence of MucAB proteins the bypass step were less stringent in its requirement for RecA pro- tease, D N A lesions would be better tolerated and the muta- genic process would occur with increased efficiency.

Acknowledgements. We are grateful to Dr. Graham Walker for the gift of plasmids and to Dr. Raymond Devoret for critical read- ing of the manuscript. This work was supported by a grant from the Comisi6n Asesora de Investigaci6n Cientifica y T6cnica (0297/81). V. Aleixandre benefited from a fellowship from the Di- putaci6n Provincial de Valencia.

References

Bagg A, Kenyon CJ, Walker GC (1981) Inducibility of a gene product required for UV and chemical mutagenesis in Escherichia coli. Proc Natl Acad Sci USA 78 : 5749-5753

Blanco M, Rebollo JE (1981) Plasmid pKM101-dependent repair and mutagenesis in Escherichia coli ceils with mutations lexB30, t/f and zab-53 in the recA gene. Mutat Res 81:265-275

Blanco M, Levine A, Devoret R (1975) lexB: a new gene governing radiation sensitivity and lysogenic induction in Escherichia coli K12. In: Hanawalt PC, Setlow RB (eds) Molecular mechanisms for repair of DNA. Plenum Press, New York, pp 379-382

Blanco M, Herrera G, Collado P, Rebollo JE, Botella LM (1982) Influence of RecA protein on induced mutagenesis. Biochimie 64:633-636

Bridges BA, Woodgate R (1985) The two-step model of bacterial UV mutagenesis. Mutat Res 150:133-139

Devoret R, Pierre M, Moreau PL (1983) Prophage ~80 is induced in Escherichia coli KI2 recA430. Mol Gen Genet 189:199-206

Dutreix M, Bailone A, Devoret R (1985) Efficiency of induction of prophage 2 mutants as a function of recA alleles. J Bacteriol 161 : 1080-1085

Elledge SJ, Walker GC (1983) The muc genes of pKMI01 are induced by DNA damage. J Bacteriol 155:1306-1315

Ennis DG, Fisher B, Edmiston S, Mount DW (1985) Dual role for Escherichia coli RecA protein in SOS mutagenesis. Proc Natl Acad Sci USA 82:3325-3329

George J, Castellazzi M, Buttin G (1975) Prophage induction and cell division in E. coil III. Mutations sfiA and sfiB restore division in t/fan Ion strains and permit the expression of muta- tor properties of t/f. Mol Gen Genet 140 : 30%332

Horii ZI, Clark AJ (1973) Genetic analysis of the RecF pathway to genetic recombination in Escherichia coli K12: isolation and characterization of mutants. J Mol Biol 80:327-344

Howard-Flanders P, Boyce RP, Theriot L (1966) Three loci in Escherichia c'oli K12 that control the excision of pyrimidine

239

dimers and certain other mutagen products from DNA. Genet- ics 53:1119-1136

Lu C, Scheuermann RH, Echols H (1986) Capacity of RecA pro- tein to bind preferentially to UV lesions and inhibit the editing subunit of DNA polymerase III: a possible mechanism for SOS-induced targeted mutagenesis. Proc Natt Acad Sci USA 83 : 61%623

Maniatis T, Fritsch EF, Sambrook J (1982) Molecular cloning, a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York

Marsh L, Walker GC (1985) Cold sensitivity induced by overpro- duction of UmuDC in Escherichia coli. J Bacteriol 162:155-161

Miller JH (1972) Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York

Perry KL, Walker GC (1982) Identification of plasmid (pKM101) coded proteins involved in mutagenesis and UV resistance. Na- ture 300:278 281

Perry KL, Elledge SJ, Mitchell BB, Marsh L, Walker GC (1985) umuDC and mucAB operons whose products are required for UV light- and chemical-induced mutagenesis: UmuD, MucA,

and LexA proteins share homology. Proc Natl Acad Sci USA 82: 43314335

Roberts JW, Roberts CW (1981) Two mutations that alter the regulatory activity of E. eoli recA protein. Nature 290 : 422424

Waleh NS, Stocker BAD (1979) Effect of host lex, recA, reeF, and uvrD genotypes on the ultraviolet light-protecting and re- lated properties of plasmid R46 in Eseheriehia coli. J Bacteriol 137:830-838

Walker GC (1984) Mutagenesis and inducible responses to deoxyri- bonucleic acid damage in Escherichia coli. Microbiol Rev 48 : 60-93

Weisemann JM, Funk C, Weinstock GM (1984) Measurement of in vivo expression of the recA gene of Escherichia eoli by using lacZ gene fusions. J Bacteriol 160:112-121

Communicated by R. Devoret

Received November 16, 1985 / June 6, 1986