Mol Gen Genet (1995) 247:735-741 © Springer-Verlag 1995

Noel Doyle • Peter Strike

The spectra of base substitutions induced by the impCAB, mucAB and umuDC error-prone DNA repair operons differ following exposure to methyl methanesulfonate

Received: 13 September 1994/Accepted: 24 January 1995

Abstract We have used the lacZ reversion assay to study the mutation spectra induced by the Escherichia coli chromosomal umuDC operon and of its two plas- mid-borne analogues impCAB and mucAB following exposure of cells to UV light and methyl methane- sulfonate (MMS). We have shown that the impCAB, mucAB and umuDC operons all produce a similar re- sponse to UV light which results almost exclusively in AT ~ G C transitions. However, we found that the three operons produced different responses to alkylat- ing agents. We found that with MMS the chromosomal umuDC operon produced almost exclusively AT --, GC transitions, whilst both mucAB and impCAB produced predominantly transversions. In the case of the impCAB operon the mutation spectrum contained more AT ~ TA than GC ~ TA transversions; this bal- ance was reversed with mucAB. The effect of the copy number of the error-prone DNA repair operons upon the mutagenic spectra was also studied. The results obtained suggest that the copy number of the imp operon does not greatly affect the specificity of base substitutions observed. However, an increase in the copy number of the umuDC operon greatly affected the specificity of base substitution, such that virtually no transitions were produced and the spectrum was dom- inated by GC/AT ~ TA transversions. It appears that the three error-prone DNA repair operons impCAB, mucAB and umuDC, despite showing strong structural and functional homologies, can display major differ- ences in the spectrum of base changes induced during mutagenesis. We propose that the type of misincor- poration/chain extension which DNA polymerase III is allowed to synthesize on a damaged DNA template is extremely sensitive to both the amount and type of

Communicated by R. Devoret

N. Doyle. P. Strike University of Liverpool, Department of Genetics and Microbiology, Donnan Laboratories, P.O. Box 147, Liverpool, UK, L69 3BX

error-prone repair proteins present. The modulation of these events by the different proteins can result in widely different mutagenic changes in the repaired DNA.

Key words impCAB • mucAB ' umuDC • Error-prone DNA repair • mutagenic spectra

Introduction

The mutagenic response of Escherichia coli to ultra- violet light and to a wide variety of DNA damaging agents, is a function of the umuDC operon (reviewed by Walker 1984) and homologues of the E. coli genes have been detected in many, but not all, enterobacteria (Sedgwick and Goodwin 1985; Sedgwick et al. 1991a). The Salmonella typhimurium homologue has been parti- cularly well studied (Thomas and Sedgwick 1989; Thomas et al. 1990; Smith et al. 1990), and functional analogues of these genes are also carried by several naturally occurring conjugative plasmids.

Currently, three plasmid-encoded operons have been characterised; the mucAB operon of pKM101, which appears to be restricted to plasmids of incompatibility groups N and M; the imp CAB operon of TP 110 which is carried by several members of incompatibility groups I1, B and FIV (Strike and Lodwick 1987; Lodwick and Strike 1991) and is present on plasmids isolated in the pre-antibiotic era (Sedgwick et al. 1989); and the samAB operon, which has recently been characterised from the cryptic plasmid of S. typhimurium LT2 (Nohmi et al. 1991).

The three plasmid-encoded operons and the chro- mosomal umu-- operon each produce two proteins of similar sizes: UmuD, MucA, ImpA and SamA of ap- proximately 16 kDa which show homology of 42-61%; and UmuC, MucB, ImpB and SamB of approximately 47.5 kDa which show homology of 57-70%, (Kitagawa et al. 1985; Perry et al. 1985; Lodwick et al. 1990;

736

Nohmi et al. 1991). The UmuD and MucA proteins have been shown to undergo a post-translational, RecA-mediated autocleavage reaction similar to that of LexA (Little 1984; Slilaty and Little 1987) in which the 2 kDa NH2-terminal fragment is removed; the remain- ing 14 kDa COOH-terminal fragment is thought to be the active protein product (Burckhardt et al. 1988; Nohmi et al. 1988; Shinagawa et al. 1988; Shiba et al. 1990; Woodgate et al. 1989; Woodgate and Ennis 1991). The cleavage product UmuD' has been shown to form homodimers (Woodgate et al. 1989); evidence also suggests that in vitro the UmuD' protein will complex with UmuC and/or RecA (Freitag and McEntee 1989; Frank et al. 1993), and this complex of UmuD', UmuC, and RecA will interact with the DNA polymerase III holoenzyme to permit translesion synthesis (Woodgate et al. 1989; Rajagopalan et al. 1992). It has been pro- posed that the RecA protein targets the UmuD' protein to damaged DNA, which consequently positions the UmuD'C complex in the correct location for interac- tion with DNA polymerase III, and thus permits trans- lesion synthesis (Frank et al. 1993).

This paper describes experiments to characterise the spectra of base substitutions occurring in the presence of the impCAB, mucAB and umuDC error-prone DNA repair operons. Base substitutions were detected using the in vivo system devised by Cupples and Miller (1989) which is able to detect specific base changes that occur in the Glu-461 codon in the lacZ gene by reversion from Lac- to Lac ÷. Previous studies have suggested that the mutation spectra of umuDC and mucAB might differ in response to the chemical mutagen methyl methanesufonate (MMS). For example, Fowler et al. (1981), using strains bearing the plasmid pKM101 (mucAB), found increased mutation at AT sites with decreased mutation at GC sites in a study using rever- sion of specific trpA alleles as an assay system. Also, Mattern et al. (1985), using the trpA223 allele, found the mutation spectrum induced by MMS to be 92% AT ~ GC transitions and 8 % AT ~ TA/CG transver- sions with umuDC. However, when mucAB was present, again on pKM101, this changed to 45% AT ~ GC and 35% AT--, TA/CG and 20% other mutations. More recently Watanabe et al. (1994) used the same lacZ reversion assay as used here to compare the mutagenic spectra of umuDC, mucAB and samAB following expo- sure to a number of chemical mutagens. Again differ- ences between the operons were noted which agree with those described here. In addition, Urios et al. (1994) showed that in a umu- background different his- alle- les reverted to his ÷ at different rates following exposure to benzo[a]pyrene, aflatoxin B1 and 1-nitropyrene de- pending upon whether umuDC or mucAB was present, again indicating that these operons produce differences in mutation spectra.

However, it is important to note that in most of these previous studies comparing mucAB and umuDC oper- ons, the bacterial hosts for the mucAB-carrying plasmid

pKM101 still possessed a functional umuDC operon. The mutation spectra observed would therefore be a composite of umu and muc effects and interactions. In the experiments described here, the chromosomal umuDC operon was inactivated by the introduction of the umuC122::Tn5 mutation. Although the umuC122::Tn5 mutation gives rise to a truncated UmuC protein (Koch et al. 1992), and the umuD gene continues to produce an active product, both we and others have shown that this defect in the umu ÷ operon cannot be complemented by either muc + or imp + genes to restore any mutagenesis (Perry et al. 1985; Sedgwick et al. 1991b); nor indeed can the introduction of plas- mids carrying individual mucA ÷, mucB +, impA + or impB + genes restore any mutagenesis to strains carry- ing the umuC122::Tn5 mutation (Perry et al. 1985; Sedgwick et al. 1991b; Doyle and Strike, unpublished data). Therefore it seems reasonable to expect that the presence of this mutation will be sufficient to remove the influence of the umuDC operon on the mutation spectra displayed by host cells carrying the plasmid- borne rnucAB and impCAB operons, and allow their effects to be seen in isolation.

This present study therefore describes our work to characterise the specificity of base substitutions follow- ing exposure to UV light and MMS of E. coli cells containing the mucAB and impCAB operons in the absence of mutations caused by umuDC. In addition, experiments were also conducted to assess the effect of copy number of the umuDC and impCAB operons upon the spectra of base substitutions produced.

Materials and methods

Bacterial strains and plasmids

The bacterial strains and plasmids used are shown in Table 1. The six bacterial strains described by Cupples and Miller (1989) each contain the F' plasmid carrying the lacZ gene with a specific base change in the codon for Glu-461, and they are therefore Lac-. These mutations can only be corrected by a specific base substitution which restores a glutamic acid residue at position 461 (Table 1), and the restoration of Glu-461 is easily identified by reversion to Lac +. In order to monitor the base substitutions occurring due to the impCAB and mucAB operons it was necessary to create a com- plementary set of mutant strains defective in the umuDC operon. This was done by transducing the umuC122 :: Tn5 insertional muta- tion, using generalized P1 transduction (Miller 1972), from the strain GW2100 into the strains CC101-CC106 to form ND101-ND106. Therefore CC101 CC106 and the derivatives ND101-ND106, with the exception of the umuC122::Tn5 mutation, are otherwise isogenic.

The plasmid pKG10 contains a 2.5 kb EcoRI-BglII fragment from TP110 carrying the entire imp operon and promoter cloned into the vector pTT-1 (Sedgwick et al. 1989), while R64 is a low copy number IncI1 plasmid which has been shown to carry the imp operon (Lodwick and Strike 1991). The plasmid pIC80 contains a 2 kb HincII fragment carrying the mucAB genes and promoter subcloned into HpaI-cleaved pLM203 (Blanco et al. 1986), and the plasmid pLM207 is a derivative of pSEll7, carrying the umuDC operon (Marsh and Walker 1985).

Table 1 Bacterial strains and plasmids used

737

Genotype Base substitution detected

Reference

Bacterial strains CC101 CC102 CC103 CC104 CC105 CC106

ND101 ND102 ND103 ND104 ND105 ND106

Plasmids pKG10 pIC80 pLM207 R64

ara-, A(lac-proB)xm, [F ' lacI Z , proB +] ara-, A(Iac-proB)xm, IF ' lacI-Z- , proB +] ara , A(Iac-proB)xm, l-F' lacI Z , proB +] ara , A(lac-proB)xln, IF' lac l -Z- , proB +] ara-, A(lac-proB)xm, I-F' l a d - Z - , proB +] ara-, A(lac-proB)xm, IF' lac l -Z- , proB +]

As CC101 but umuC122:: Tn5 As CC102 but umuC122:: Tn5 As CC103 but umuC122:: Tn5 As CC104 but umuC122:: Tn5 As CC105 but umuC122:: Tn5 As CC106 but umuC122:: Tn5

impC+A+B + mucA+B + umuD+C + impC+A+B +

AT --* CG GC ~ AT GC ~ CG GC ~ TA AT --, TA AT ~ GC

AT ~ CG GC --+ AT GC ~ CG GC ~ TA AT ~ TA AT ~ GC

Cuppies and Miller (1989) Cupples and Miller (1989) Cupples and Miller (1989) Cupples and Miller (1989) Cupples and Miller (1989) Cupples and Miller (1989)

This work This work This work This work This work This work

Sedgwick et al. (1989) Blanco et al. (1986) Marsh and Walker (1985) Lodwick and Strike (1991)

Mutat ion assays

Mutagenesis by UV light

Cells were grown in Luria broth with aeration to mid-log phase (OD600 ca. 0.6), pelletted, washed and resuspended in an equal volume of 100 mM MgSO4; 10 mI was placed into a large glass petri dish and irradiated with a UV dose of 60 J /m 2 with constant gentle mixing. Aliquots of 100 gl were taken at suitable time intervals, serially diluted and plated onto Luria agar plates to measure sur- vival; these were incubated overnight at 42°C to prevent any cold sensitivity sometimes associated with these operons (Marsh and Walker 1985). The use of this temperature resulted in a slight increase in the number of mutants observed, compared to incuba- tion at 37 ° C, but did not otherwise affect mutagenesis and main- tained comparability with previous work (Sedgwick et al. 1991b). Mutagenesis was measured by inoculating 10 ml Luria broth in a 100 ml conical flask with 500 gl of the irradiated culture and incubating overnight with constant shaking. Mutants resistant to rifampicin were estimated by plating 100 ~tl of the overnight culture onto Luria agar containing 40 gg/ml rifampicin, the total number of viable cells determined by plating serial dilutions onto Luria agar, and the number of Lac + revertants measured by plating 100 ~tl onto Lactose M56 minimal agar.

Mutagenesis by MMS

A 10 ml portion of Luria broth was inoculated with 200 ~tl of an overnight culture and grown to mid-log phase (OD60 o ca. 0.6). The cells were washed and resuspended in 10 ml M9 buffer. This was divided into two 5 ml aliquots. MMS was added to give a final concentration of 1% and the tubes incubated at 37°C for 1 h. The MMS was inactivated with 1 ml 40% sodium thiosulphate and the cells washed and resuspended in 5 ml M9 buffer. Serial dilutions were prepared and 100 gl plated onto Luria agar to measure sur- vival. A 0.5 mt aliquot was used to inoculate 10 ml Luria broth in a 100 ml flask and incubated overnight. Mutagenesis was measured to rifampicin-resistance and reversion to Lac ÷. The number of rifampicin-resistant mutants was measured by plating 100 gl of the overnight culture onto Luria agar containing 40 gg/ml rifampicin,

the total number of viable cells determined by plating serial dilutions onto Luria agar and Lac + revertants measured by plating 100 gl on to Lactose M56 minimal agar.

Results

The spectra of base substitutions induced by UV light

When UV light was used as the mutagen, the wild-type umuDC ÷ strains (CC101-CC106) showed almost ex- clusively AT ~ GC transitions (Fig. 1); this was as pre- viously reported by Cupples and Miller (1989). A sim- ilar result has also been reported by Shinoura et al. (1983) using a trp reversion assay. When the mucAB and impCAB operons were introduced into the umuC122::Tn5 derivatives (ND101-ND106) using plasmids pIC80 and pKG10, respectively, no signifi- cant difference from this spectrum was noted, and both showed a preponderance of AT --+ GC transitions and lower levels of GC --+ AT transitions (Fig. 1). Therefore, following exposure to UV light, the three operons in- duced similar spectra of base substitutions.

The spectra of base substitutions induced by MMS

Next, the effect of a chemical mutagen upon the spectra of base substitutions was examined. MMS was chosen, as previous reports in the literature have suggested that the mucAB operon produces a MMS mutation spec- trum different from that produced by umuDC (Fowler et al. 1981; Mattern et al. 1985).

The mutation spectra found after treatment with 1% MMS showed differences between the three mutagenic

738

s~

100 - ,~iiiiiiiiiiiiiiiiiiii

8 0 iiiiiiiiiiiiiiiiiiiiii u m u . D C iiiiiiiiiiiiiiiiiiiiil

" iiiiiiiiiiiiiiiiiiiiii 60 i~iiiiiiiNiii~ii~Nii

- i!iii!i!!ii!iiii!ii!ii 4 0 Nili~iiiMili~iiiiiii

- iiiii~hihii~i~ili~i 2 0 ii~iiii~iiiiiiii~iiiii

6 7 {~i!iiiiiiii{ii{ii{iii

100 --

80 i m p C A B

6O

4O

20 85

r j ~ I u

E-~ cA r j r.9

~ 100 ~ O ~=

~ 80

o m

~ 4o

m u c A B

I27

2 7 8

, ~ r b r b ~b , ~

691

598 Kii!iiii!ii!ii!iiiii !!!ii!i!!ii!!!!!!!i!!! ......................

...................... :::::::::::::::::::::

...................... ......................

...................... ......................

......................

iiiiiiiiiiiiiiiiiiiiii ...... p , ..........

672

rb

---7

' - 7

7

Fig. 1 The percentage occurrence of each base substitution and the mean number of Lac ÷ revertants per plate (shown as numbers) when UV light was used as the mutagen. The umuDC operon was present chromosomally on the strains CC101 CC106, and the plasmids pIC80 (rnucAB) and pKG10 (impCAB) were present in the strains ND101-ND106. The numbers above each bar show the mean num- ber of Lac + revertants per plate

operons. The umuDC ÷ strains produced almost exclus- ively AT ~ GC transitions with about 8% GC ~ AT transitions (Fig. 2a). However, strains containing the mucAB operon produced a significantly different mutagenic spectrum in which GC ~ TA transversions were dominant, together with lower levels of AT ~ TA, AT ~ CG, GC ~ CG and AT ~ GC substitutions (Fig. 2c). Strains containing the impCAB operon also produced a mutagenic spectrum dominated by trans- versions, primarily AT ~ T A , with lower levels of GC ~ TA, GC ---, CG, AT --, CG and AT ~ GC base

substitutions (Fig. 2b). The total number of mutations was not greatly increased by mucAB and impCAB, compared to umuDC; the numbers of GC or AT ~ TA transversions occurring in the presence of mucAB (3390 and 1650 Lac ÷ revertants per plate, respectively) or impCAB (1903 and 2426 Lac ÷ revertants per plate, respectively) differing from the number of AT--+GC transitions occurring in the presence of umuDC (1760 Lac ÷ revertants per plate) by only two to threefold. This relatively small increase would not be sufficient for the number of transversions occurring to mask the number of transitions. Nor was there any significant change in the overall mutation rates between the six indicator strains as measured by the number of rifampicin resistant mutations scored (data not shown). Thus, the changes in spectra could be ascribed to differences in the mechanism of mutagenesis rather than an effect on the overall efficiency of the process.

While the two plasmid-borne systems impCAB and mucAB showed some differences between the base sub- stitutions occurring in response to MMS, they were similar to the extent that they both caused high num- bers of GC/AT ~ T A transversions, low levels of G C / A T ~ C G transversions and virtually no transitions. Therefore, the two plasmid-borne operons can be considered to produce essentially similar spectra of base substitutions that are distinct from those pro- duced by the chromosomal umuDC system.

Effect of copy number on MMS mutagenesis

The pKG10 and pIC80 subclones used to introduce the impCAB and mucAB operons, respectively, into the strains N D 1 0 1 - N D 1 0 6 would result in these operons being present in higher copy numbers than the chromo- somal umuDC + genes. It seemed possible that the ob- served difference in spectrum between impCAB/mucAB and umuDC might be due to the increased copy number of the plasmid operons. In order to test this, the naturally occurring conjugative plasmid R64 (copy number of 1-2), which carries the impCAB operon, was introduced into each of the six tester strains. R64 was used to provide imp in low copy number since both the imp parental plasmid T P l l 0 and the ND101 ND106 strains are kanamycin resistant, thus preventing selection for the plasmid. The IncI1 plasmid R64 carries the same imp operon as T P l l 0 (Lodwick and Strike 1991), and carries the selectable tet R marker.

Following exposure to 1% MMS, R64-containing strains produced a similar spectrum of base substitu- tions to those containing the subclone pKG10 used previously (Fig. 2b, e). We conclude that the copy num- ber of the impCAB operon has little effect upon the mutation spectrum. However, when a medium copy number subclone, pLM207, (copy number 10-15;

'~ ~ lO0 1 a 1760 ¢,1 0 / ' : : : : : : '

!iiiiiiiiiiiiiiiiiii! ,iii!iiii! =, 8 0 t l = u m u Z ) C ;;;?;:;:;!;:;.~:;:;:;~ . . . . . . . . . . . . .

60 ;i;;:.;;i:i:i:i:i:i:i: o m iiiiiiiii;i~i:g~i~iiiili?iiii

~ ::::::::::::::::::::::::::

~ 4 0 iiiiii!iii!i!iiiiiiiiiiiiiiiil ~ i~i;i;i;i;iii;;ii!ii;iil ~ =============================

ao ]7~ iiiiiiiiiiiiiiiii;i;i;iiiiiii; :::::::::::::::::::::::::::::

0 ~~.~..9~;;;!:i~!;i:~i:!8i;i:i:i:~;~::i:i:i:!:i:~:i;i:i:i:i:i:i:~!:!:i:~i~i~i:i~i:i;i~iii ::::::::::::::::::::::::::::::

r~

4' * 4 * + 4

U U U ~

O

u ~ O ~

[9 E~ L9 < < U U ,~ U E~ E~ L9

~ 6 0 0

r , ~ :20

U c.)

4 a * 4, A ~.

' 7

739

U ~g r~

70 -- / d

60

40

20

/

865

[9 ~ [9 < < u

4 4, * 4. 4 ~ *

~. 70 - - / e ~9~

6 0 p. . - _ ( R 6 4 )

- raN1 o m 40

~ - 14o ~ 4 _ j . ~ ! i : : i ~ . 0 0

0 / ..::i::::i::iii::i::iii!i ~ ' . . : : . ! ~ i ~ : : '::::;:::::~ / [9 ~ [9 < < rJ r j '~ U E-4 ~ r~ ÷ * * 4 4 *

Fig. 2a-e The percentage occurrence of each base substitution and the mean number of Lac + revertants per plate (shown as numbers) when methyl methanesulfonate (MMS) was used as the mutagen. The umuDC operon was present chromosomally on the strains CC101-CC106 (a), and the plasmids pKG10 (impCAB) (b) and pICS0 (mucAB) were present in the strains ND101-ND106 (e). The effect of the copy numbers of the umuDC and impCAB operons upon the specificity of base substitutions induced following exposure to MMS. The umuDC operon was present either chromosomally in CC101 CC106 (low copy number; a) or upon the plasmid pLM207 in ND101 ND106 (higher copy number d). The impCAB operon was present in ND101-ND106 either on R64 (low copy number; e) or pKG10 (higher copy number; e). The numbers above each bar show the mean number of Lac + revertants per plate

Marsh and Walker 1985) carrying the umuDC operon was introduced into the six tester strains, the mutation spectrum was very different from that of the chromo- somal urnuDC operon following exposure to 1% MMS (Fig. 2a, d). The mutagenic spectrum resembled that obtained with the two plasmid systems imp and muc in that there were high levels of GC/AT ~ TA transver- sions and virtually no transitions. Therefore, for the umuDC operon, copy number appears to have a major effect on the base substitutions induced. It is important to note that these effects can only be due to copy number as, in the plasmid constructs, both the umu and imp genes are still expressed from their own promoters, and so effects cannot be ascribed to differences in the efficiency of expression.

Discussion

In this paper we have compared the spectra of base substitutions occurring in the presence of the umuDC, mucAB and impCAB operons following exposure to UV light and MMS, using reversions of specific muta- tions in the Glu-461 codon of the lacZ gene as an assay system (Cupples and Miller 1989). The results show that the impCAB, mucAB and umuDC error-prone DNA repair operons can each produce different muta- tion spectra in response to different mutagens. While each of the three operons tested produced a similar mutation spectrum in response to UV light, they differ- ed in their responses to MMS and the spectrum of base substitutions produced was dependent upon the error- prone DNA repair operon present. For example, fol- lowing exposure to MMS, umuDC ÷-containing strains

740

showed almost exclusively AT ~ GC transitions, while strains containing either of the plasmid-borne mucAB or impCAB operons produced spectra containing high levels of GC/AT ~ T A transversions, together with lower levels of GC/AT ~ CG transversions. These data are compatible with the earlier observations of Mattern et al. (1985) and the more recent study by Watanabe et al. (1994), which were conducted in the presence of umuDC ÷ and therefore displayed mixed spectra. The somewhat higher number of mutants observed with the mucAB operon has been suggested to be a consequence of the greater efficiency of cleavage of MucA compared to UmuD (Hauser et al. 1992) and may result in a wider range of mutations occurring (Blanco et al. 1986).

Since the plasmid borne-operons were present upon plasmids which would have increased their normal copy number, this raised the question as to whether the differences in the spectra of base substitutions from umuDC were due to a copy number effect. To resolve this question experiments were conducted in which the impCAB operon was introduced into the umuC122:: Tn5 strains (ND101-ND106) either in high copy number, upon pKG10, or upon the low copy number naturally occurring conjugative plasmid R64. The results obtained showed that the spectra of base substitutions induced by MMS in the presence of the impCAB operon were unaffected by its copy number. However, this was not the case with the umuDC operon, in that the presence of umuDC upon a medium copy number plasmid, pLM207, resulted in a change in the spectrum of base substitutions from those obtained with umumC in a low-copy-number chromosomal loca- tion. Thus, when present in higher copy number, the spectrum of base substitutions of umuDC resembles that of the plasmid-borne impCAB and mucAB oper- ons. This suggests that despite the similarities that exist between the error-prone DNA repair operons, the mechanisms of mutagenesis of the plasmid-borne oper- ons can differ from that of the chromosomal umuDC, and that the mechanism of mutagenesis of the umuDC operon can change as its copy number increases, to resemble that of the plasmid-borne operons.

A possible model for mutagenesis

The data presented in this paper demonstrate clearly that the type of misincorporation event resulting from damage to DNA is particularly sensitive to the quantity and type of error-prone repair proteins present. We believe that the data can be accommodated in a simple model based on the known preference for polymerases to insert a purine residue opposite a non-coding lesion (Kunkel 1984), and on the two-stage model for mutagenesis proposed by Bridges and Woodgate (1985). In the context of this two-stage model, there is now very good in vitro evidence that DNA polymerase III can achieve the insertion of a base opposite a

non-coding or mis-coding lesion, but that the sub- sequent chain extension cannot take place without the presence of the error-prone repair proteins (Ra- jagopalan et al. 1992; Belguise-Valladier et al. 1994).

We propose therefore that the initial preference of the DNA polymerase on encountering a damaged tem- plate is to insert a purine, perhaps with a bias towards adenine. In the case of UV-damaged DNA, if adenine is inserted, this will result in the placement of an A oppo- site either a damaged 3' T in a T-T dimer, or a damaged 3' C in a C-C or T-C dimer. Only in the latter cases will this result in a mutation, a GC ~ AT transition. In the case of the T-T dimer, the correct sequence will be restored. If, instead of adenine, the polymerase inserts guanine opposite the damaged base, then a mutation will be created with the major photoproduct, the T-T dimer. As the major mutation product observed follow- ing UV is the AT ~ GC transition, such an event ap- pears to be fairly frequent.

The preferential insertion of purines opposite damaged bases can also explain the spectra of changes observed with MMS mutagenesis. In contrast to UV, MMS primarily damages purine bases. If the poly- merase now attempts to insert its preferred purine opposite the damaged base, a transversion event must occur. This is indeed what happens with impCAB, mucAB, and high copy number umuDC, but it is not what is seen with chromosomal umuDC. We propose therefore that at low concentrations of the UmuDC proteins, chain extension following a transversion mis- incorporation is not permitted. Such extension can only occur in the presence of the more effective MucAB or ImpCAB protein combinations, or if large quantities of UmuDC proteins are present. If a transversion mis- incorporation takes place in the presence of limiting quantities of UmuDC, extension is delayed and proof- reading must occur. Only when a misincorporation of the transition type has occurred is extension permitted by normal levels of UmuDC. Thus we suggest that the observed change in mutation spectrum is a reflection not of altered polymerase specificity with respect to the base inserted, but of the type of base insertion which must occur if extension is to be permitted. The basic mechanisms of misincorporation and extension pro- moted by the three error-prone repair operons can therefore be considered to be essentially the same, ex- cept that UmuDC at normal levels is not able to permit extension following a transversion mispairing. The two plasmid operons show slightly greater homology to each other than to umuDC and are at least partially interchangeable for mutagenesis (Doyle and Strike, manuscript submitted), and it is perhaps therefore not suprising to find that they are functionally more similar to each other than they are to umuDC.

Acknowledgements This work was supported by a SERC post- graduate research studentship to Noel Doyle. We would like to thank Dr. J. H. Miller for providing us with bacterial strains.

741

References

Belguise-Valladier P, Maki H, Sekiguchi M, Fuchs RP (1994) Effect of single DNA lesions on in vitro replication with DNA poly- merase III hotoenzyme: comparison with other polymerases. J Mol Biol 236:151-164

Blanco M, Herrera G, Aleixandre V (1986) Different efficiencies of UmuDC and MucAB proteins in UV light induced mutagenesis of Escherichia coli. Mol Gen Genet 205 : 234-239

Bridges BA, Woodgate R (1985) Mutagenic repair in Escherichia coli: products of the recA and umuDC genes act at different steps in UV-induced mutagenesis. Proc Nati Acad Sci USA 82: 4193-4197

Burckhardt SE, Woodgate R, Scheuermann RH, Echols H (1988) UmuD mutagenesis protein of E. coIi: overproduction, purifica- tion and cleavage by RecA. Proc Natl Acad Sci USA 85:1811 1815

Cupples CG, Miller JH (1989) A set of lacZ mutations in Escherichia coil that allow rapid detection of each of the 6 base substitutions. Proc Natl Acad Sci USA 86 : 5345-5349

Fowler RG, McGinty L, Mortelmans KE (1981) Mutational specifi- city of ultraviolet light in Escherichia coli with and without the plasmid pKM101. Genetics 99 : 25 40

Frank EG, Hauser JH, Levine AS, Woodgate R (1993) Targeting of the UmuD, UmuD' and MucA' mutagenesis protein to DNA by RecA protein. Proc Natl Acad Sci USA 90:8169-8173

Freitag N, McEntee K (1989) "Activated" RecA protein affinity chromatography of LexA repressor and other SOS regulated proteins. Proc Natl Acad Sci USA 86 : 8363-8367

Hauser J, Levine AS, Ennis DG, Chumakov KM, Woodgate R (1992) The enhanced mutagenic potential of the MucAB proteins correlates with the highly efficient processing of the MucA pro- tein. J Bacteriol 174:6844-6851

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

Koch WH, Ennis DG, Levine AS, Woodgate R (1992) Escherichia coIi umuDC mutants DNA sequence alterations and UmuD cleavage. Mol Gen Genet 233 : 443-448

Kunkel TA (1984) Mutational specificity of depurination. Proc Natl Acad Sci USA 81 : 1494-1498

Little JW (1984) Control of the regulatory system by the level of RecA protease. Biochimie 64 : 585-589

Lodwick D, Strike P (1991) Distribution of sequences homologous to the impCAB operon of TPl l0 among bacterial plasmids of different incompatibility groups. Mol Gen Genet 229 : 27-30

Lodwick D, Owen D, Strike P (1990) DNA sequence analysis of the imp UV protection and mutation operon of the plasmid TPll0: identification of a third gene. Nucleic Acids Res 18 : 5045-5050

Marsh L, Walker GC (1985) Cold sensitivity induced by over- production of umuDC in Escherichia coll. J Bacteriol 162: 155-161

Mattern IE, Olthoff FP, Jacobs-Meijsing BLM, Enger-Valk BE, Pouwels PH, Lohman PHM (1985) A system to determine base pair substitutions at the molecular level based on restriction enzyme analysis; influence of the muc genes of pKM101 on the specificity of mutation induction in E. coli. Mutat Res 148 : 35-45

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

Nohmi T, Battista JR, Dobson LA, Walker GC (1988) RecA- mediated cleavage activates UmuD for mutagenesis: mechanistic relationship between transcriptional depression and posttransla- tional activation. Proc Natl Acad Sci USA 85 : 1816-1820

Nohmi T, Hakura A, Nakai Y, Watanabe M, Murayama SY, Sofuni T (1991) Salmonella typhimurium has two homologous but differ- ent umuDC operons. Cloning of a new umuDC-like operon (sa- mAB) present on a 60 megadalton cryptic plasmid of Salmonella typhimurium. J Bacteriol 173 : 1051-1063

Perry KL, Elledge SJ, Mitchell B, Marsh L, Walker GC (1985) umuDC and mucAB operons whose products are required for UV light and chemical-induced mutagenesis: UmuD, MucA, and LexA products share homology. Proc Natl Acad Sci USA 82:4331-4335

Rajagopalan M, Lu C, Woodgate R, O'Donnell M, Goodman MF, Echols H (1992) Activity of the purified mutagenesis proteins UmuC, UmuD', and RecA in the replicative bypass of an abasic DNA lesion by DNA polymerase III. Proc Natl Acad Sci USA 89:10777 10781

Sedgwick SG, Goodwin P (1985) Differences in mutageuic and recombinational DNA repair in enterobacteria. Proc Natl Acad Sci USA 82:4172-4176

Sedgwick SG, Thomas SM, Hughes VM, Lodwick D, Strike P (1989) Mutagenic DNA repair genes on plasmids from the pre-anti- biotic era. Mol Gen Genet 218:323 329

Sedgwick SG, Ho C, Woodgate R (t991a) Mutagenic DNA repair in enterobacteria. J Bacteriol 173:5604-5611

Sedgwick SG, Lodwick D, Doyle N, Crowne H, Strike P (1991b) Functional complementation between chromosomal and plas- mid mutagenic DNA repair genes in bacteria. Mol Gen Genet 229 : 428-436

Shiba T, Iwasaki H, Nakata A, Shinagawa H (1990) Proteotytic processing of MucA protein in SOS mutagenesis: both processed and unprocessed MucA may be active in the mutagenesis. Mol Gen Genet 224:169-176

Shinagawa H, Iwasaki H, Kato T, Nakata A (1988) RecA protein dependent cleavage of UmuD protein and SOS mutagenesis. Proc Natl Acad Sci USA 85:1806-1810

Shinoura Y, Ise T, Kato T, Glickman BW (1983) umuC-mediated misrepair mutagenesis in Escherichia coli: extent and specificity of SOS mutagenesis. Mutat Res 111:51-59

Slilaty SN, Little JW (1987) Lysine-156 and serine-ll9 are needed for LexA repressor cleavage: a possible mechanism. Proc Natl Acad Sci USA 84:3987-3991

Smith CM, Koch WH, Franklin SB, Foster PL, Cebula TA, Eisen- stadt E (1990) Sequence analysis of the Salmonella typhimurium LT2 umuDC operon. J Bacteriol 172:4964-4978

Strike P, Lodwick D (1987) Plasmid genes affecting DNA repair and mutation. J Cell Sci Suppl 6:303-321

Thomas SM, Sedgwick SG (1989) Cloning of Salmonella typhimurium DNA encoding mutagenic DNA repair. J Bacteriol 171 : 5776-5782

Thomas SM, Crowne HM, Pidsley SC, Sedgwick SG (1990) Struc- tural characterisation of the Salmonella ~yphimurium LT2 umu operon. J Bacteriol 172:17 24

Urios A, Herrera G, Aleixandre V, Sommer S, Blanco M (1994) Mutability of Salmonella tester strains TA1538 (hisD3502) and TA1535 (hisG46) containing the UmuD' and UmuC proteins of Escherichia coli. Environ Mol Mutagen 23 : 63 67

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

Watanabe M, Nohmi T, Ohta T (1994) Effects of the umuDC, mucAB and samAB operons on the mutational specificity of chemical mutagenesis in Escherichia coli: II. Base substitution mutagen- esis. Mutat Res 314:39 49

Woodgate R, Ennis DG (1991) Levels of chromosomally encoded Umu proteins and requirements for in vivo UmuD cleavage. Mol Gen Genet 229:10-16

Woodgate R, Rajagopalan M, Lu C, Echols H (1989) UmuC muta- genesis protein of Escherichia coIi: purification and interaction with UmuD and UmuD'. Proc Natl Acad Sci USA 86:7301-7305