QUARTERLY UV- and MMS-induced mutagenesis of �O(am)8 phage under nonpermissive conditions for phage DNA replication �

Mutagenesis in Escherichia coli, a subject of many years of study is considered to be related to DNA replication. DNA lesions nonrepaired by the error-free nucleotide excision repair (NER), base excision repair (BER) and recombination repair (RR), stop replication at the fork. Reinitiation needs translesion synthesis (TLS) by DNA polymerase V (UmuC), which in the presence of accessory proteins, UmuD', RecA and ssDNA-binding protein (SSB), has an ability to bypass the lesion with high mutagenicity. This enables reinitiation and extension of DNA replication by DNA polymerase III (Pol III). We studied UV- and MMS-induced mutagenesis of lambdaO(am)8 phage in E. coli 594 sup+ host, unable to replicate the phage DNA, as a possible model for mutagenesis induced in nondividing cells (e.g. somatic cells). We show that in E. coli 594 sup+ cells UV- and MMS-induced mutagenesis of lambdaO(am)8 phage may occur. This mutagenic process requires both the UmuD' and C proteins, albeit a high level of UmuD' and low level of UmuC seem to be necessary and sufficient. We compared UV-induced mutagenesis of lambdaO(am)8 in nonpermissive (594 sup+) and permissive (C600 supE) conditions for phage DNA replication. It appeared that while the mutagenesis of lambdaO(am)8 in 594 sup+ requires the UmuD' and C proteins, which can not be replaced by other SOS-inducible protein(s), in C600 supE their functions may be replaced by other inducible protein(s), possibly DNA polymerase IV (DinB). Mutations induced under nonpermissive conditions for phage DNA replication are resistant to mismatch repair (MMR), while among those induced under permissive conditions, only about 40% are resistant.

The mechanisms of mutagenesis induced by UV and ionizing radiation, as well as by chemical mutagens, have long been a subject of great interest.It is now well established that mutations constitute the first step in carcinogenesis, and are the cause of many genetic diseases in man.The most widely studied model for UV-induced and chemical mutagenesis is Escherichia coli.Many mutagens, including UV light, induce mutations in E. coli by the SOS pathway (SOS-mutagenesis) (for reviews see: Friedberg et al., 1995;Koch & Woodgate, 1998;Sutton et al., 2000;2002;Shina & Hader, 2002;Taylor, 2002).According to the current hypothesis on the mechanism of UV-induced mutagenesis, and mutagenesis by some chemicals, mutations are generated during replication of damaged DNA with an involvement of DNA polymerase V (Pol V), which is an error-prone translesion synthesis (TLS) DNA polymerase (for reviews see: Sutton et al., 2000;2002;Goodman & Tippin, 2000;Livneh, 2001;Kobayashi et al., 2002).Pol V encoded by the umuC gene is an extremely weak DNA polymerase without a proofreading activity and with low processivity (about 6 bp), is activated for lesion bypass by the UmuD', RecA and SS DNA-binding protein (SSB) proteins, and stimulated by the processivity subunits of DNA polymerase III, the b subunit sliding clamp and the g complex clamp loader (Tang et al., 1999;Sutton et al., 2001;Becherel et al., 2002).Inactivation of TLS by umuC mutation strongly decreases mutagenesis implying that most mutations are caused by Pol V-depended TLS.Activated Pol V replicates effectively apurinic/apyrimidinic sites (AP-sites), T^T photodimers and 6-4 photoproducts (Rajagopalan et al., 1992;Tang et al., 2000).The proposed model for TLS by Pol V is as follows.As the replication fork is stopped by an unrepaired lesion in DNA, single-stranded DNA (ssDNA) is bound by the SSB protein, RecA displaces it and forms a nucleoprotein filament, covering the ssDNA region including the primer termini near the lesion (Reuven et al., 2001).Pol V binds to the primer-template complex guided by the RecA filament.In loading of Pol V, UmuD' interacts with the RecA filament and with UmuC, and then UmuC interacts with the primer.Pol V holoenzyme starts DNA synthesis and is able to replicate through the lesion.The RecA filament disassembles in a reaction requiring ATP hydrolysis which leads to dissociation of Pol V, so that Pol III can take over and continue replication.It is believed that there is a coordinated action of Pol III with V as well as with other translesion bypass DNA polymerases, e.g.Pol IV (DinB) (Kobayashi et al., 2002).Thus, according to the hypothesis, the mutagenic process is related to the replication of damaged DNA.Such a mechanism of mutagenesis is possible only in actively metabolising cells, but not in resting or nondividing ones, e.g.somatic cells.However, it is well known that mutations may be induced by environmental mutagens, even in nondividing somatic cells, leading to cancer, implying that such mutagenesis would follow a different pathway, most probably one, independent of replication of damaged DNA.So, we asked the question whether mutations may be induced under nonpermissive conditions for DNA replication, and if so, by what mechanism.We have previously observed that UV-induced mutations in E. coli MV1178 recA recF may be formed prior to the resumption of DNA replication in UV-irradiated cells (Pietrzykowska & Felczak, 1991).This suggested that mutations may be generated in a manner independent of replication of damaged DNA.To verify this hypothesis we studied UV-induced mutagenesis of lO(am)8 phage in a host nonpermissive for phage DNA synthesis, E. coli 594 sup + .Phage lO(am)8 with an amber mutation in gene O, is unable to replicate its DNA in a host which does not possess a suppressor of the amber mutation (supE).The product of the O gene -lO protein, is known to be responsible for initiation of phage DNA replication (Zylicz et al., 1984;1988) and lO(am)8 phage is unable to repli-cate its DNA in the suppressor-less host E. coli 594 sup + (Ogawa & Tomizawa, 1968;Shuster & Weissbach, 1969).We now present results indicating that UVand MMS-induced mutations may arise under nonpermissive conditions for DNA replication.

MATERIALS AND METHODS
Media and growth conditions.Luria-Bertani (LB) medium (containing, per litre, 10 g tryptone, 5 g of yeast extract, and 5 g NaCl) (Miller, 1972) was used for growth of E. coli cultures and, when necessary, was supplemented with the appropriate antibiotic at the following final concentrations: chloramphenicol, 35 mg/ml; ampicillin, 100 mg/ml; tetracycline, 50 mg/ml; kanamycin, 50 mg/ml.Solid medium contained 1.2% Bacto-Agar, and all incubations were at 37°C.Liquid cultures were carried out overnight in LB with aeration, diluted 1:50 in fresh LB medium, and incubated to a density of 2-3 ´10 8 cells per ml.Large scale preparation of bacterio- phage lO(am)8 was performed as described by Sambrook et al. (1989).
Bacterial strains and plasmids.The E. coli strains and plasmids used in this study are listed in Table 1.E. coli 594 sup + and C600 supE were used as host strains for lO(am)8 phage and for construction of host mutants used in this work.Bacterial strains were constructed using standard methods of P1 transduction (Miller, 1972) and plasmid transformation (Sambrook et al., 1989).SOS induction.Bacteria grown to a density of 2 ´10 8 cells per ml were centrifuged and resuspended in half the original volume of 10 mM MgSO 4 , and divided into two parts.One part was UV-irradiated (254 nm) with a dose of 70 J/m 2 for E. coli 594 sup + and 40 J/m 2 for C600 supE and their derivatives.The second nonirradiated part was a control.The control and irradiated samples were diluted with an equal volume of 2 ´concentrated LB.The control was left on ice and the irradiated cells were incubated for SOS-induction at 37°C with shaking for 20 min.
Survival of the bacteria after irradiation with UV (254 nm) was 20-30%.
Treatment of lO(am)8 phage with UV or MMS.For both UV-and MMS-induced mutagenesis, lO(am)8 phage was diluted in 100 mM Tris/HCl (pH 7.0) supplemented with 10 mM MgSO 4 to a final concentration of 1-2 ´10 9 phages per ml.Then, a sample of the diluted phage was treated with a mutagen, the rest was left as a control.For UV mutagenesis the phage was irradiated on a Petri dish with UV light (254 nm) with a dose of 150 J/m 2 or 25 J/m 2 (for details see Results).For MMS-induced mutagenesis the phage was treated with methyl-methane sulfonate (MMS) at final concentration of 0.33% for 45 or 75 min (for details see Results) at 37°C, and to stop the reaction an equal volume of cold 40% sodium thiosulfate was added.This mixture was incubated at 37°C for 10 min and then cooled on ice (Ebisuzaki et al., 1975).As a control phages were treated for 75 min at 37°C with MMS previously inactivated with sodium thiosulfate.Survival of phages treated with MMS or UV was estimated in C600 supE cells or their derivatives and was between 2-6%.Phages prepared in this way were used for examination of mutagenesis.UV and MMS mutagenesis of lO(am)8 phage.The number of mutated phages was estimated as the number of ineffective centres (IC).lO(am)8 phage treated with a mutagen (UV or MMS) was adsorbed to 594 sup + host cells preirradiated with UV light (+SOS ) or not (-SOS) with multiplicity of infection (m.o.i.) about 0.2.After 15 min of adsorption at 37°C, additional drop of 594 sup + indicator strain and 3 ml of warm soft agar were added, plated and incubated overnight at 37°C.Appropriate dilutions of phage were plated on C600 supE strain to estimate the number of surviving phages.In the case of mutagenesis in the C600 supE strain, infected cells with mutagen treated phages after adsorption and plating were incubated at 37°C for 3 h for expression of mutations, then chloroform was added on a paper disc in the cover of the Petri dish (for 20 min) to kill C600 supE cells.After removal of the paper discs and aeration, a second layer of soft agar with 594 sup + indicator cells for lO + revertants was added and incubated overnight at 37°C.Mutation frequency was calculated as the number of revertants per 10 7 surviving phages.All manipulations with UV-irradiated phages and bacteria were done under yellow light, and overnight incubation of plates was performed in the dark to avoid photoreactivation.

UV-induced reversion of lO(am)8 to lO + in E. coli 594 sup + host
To test the hypothesis that mutations may occur before resumption of DNA replication, we studied reversion of lO(am)8 to lO + in E. coli 594 sup + host cells unable to support growth of the lO(am)8 mutant due to inability to replicate its DNA.In this host, UV-induced reversion of lO(am)8 to lO + could be observed only if the mutation occurred before initiation of DNA replication and enable the phage to form plaques on 594 sup + host.The results presented in Fig. 1 show that lO(am)8 phage efficiently reverts to lO + in E. coli 594 sup + cells nonpermissive for phage DNA replication.The frequency of reversion is dependent on the UV dose to the phage.Preirradiation of the host cells before infection with UV-irradiated phage increases the mutation frequency about 10-fold, indicating that some SOS function(s), induced in the host, stimulate the mutagenic process.It is worth to note that a significant increase in the number of l mutants is seen also when only the phage was irradiated.

UV-induced mutagenesis of lO(am)8 phage in E. coli 594 sup + cells is not dependent on nucleotide excision repair (NER)
There were two reasons to study the effect of the uvrA mutation on UV-induced mutagenesis of lO(am)8 phage in the host nonpermissive for phage DNA replication: (1) to check whether DNA replication of the irradiated phage in the nonpermissive host may be initiated by UvrABC nuclease incision; (2) whether nucleotide excision repair (NER) is involved in the generation of mutations under nonpermissive conditions for phage DNA replication.Cohen-Fix & Livneh (1992) described UV mutagenesis in ColE1 plasmid independent of DNA replication, but dependent on the UvrABC-excinuclease.We therefore examined the effect of the uvrA mutation on the UV-induced mutagenesis of lO(am)8 phage in E. coli 594 sup + uvrA cells.The results (Table 2) show that at a UV-dose giving similar phage survival (about 2%) in uvr + and uvrA hosts (after 150 and 25 J/m 2 , respectively), similar levels of mutants were observed in 594 sup + and 594 sup + uvrA cells.This suggests that: (1) UvrABC-excinuclease is not involved in the mutagenic process occurring under nonpermissive conditions for lO(am)8 phage replication, and (2) that incision by UvrABC can not initiate DNA replication of UV-irradiated lO(am)8 phage in 594 sup + host.If the incision could initiate DNA replication in 594 sup + cells, the level of lO + revertants in 594 sup + should be similar to that observed in C600 supE host permissive for phage DNA replication.As may be seen in Table 5, the efficiency of UV-induced mutagenesis of lO(am)8 is 3-fold lower in the 594 sup + then in C600 supE.Secondly, the deficiency in UvrABC activity in 594 sup + uvrA host does not affect the level of UV-induced mutagenesis of lO(am)8 when UV-doses were used to give the same survival.At similar survival the same levels of mutants were observed in 594 sup + uvrA and 594 sup + hosts (Table 2).On the other hand, at a low UV dose (25 J/m 2 ) given to the phage, a 3-4 times higher level of lO + revertants is observed in 594 sup + uvrA than in 594 sup + uvr + , indicating that UV lesions in the DNA of lO(am)8 phage, which are processed by a mutagenic mechanism operating in the host nonpermissive for DNA replication, may be repaired by the error-free NER.
The effect of dnaQ49 mutation on spontaneous and UV-induced reversion of lO(am)8 to lO + in hosts nonpermissive and permissive for phage DNA replication Looking for additional evidence that lO(am)8 phage DNA is not replicated in the nonpermissive E. coli 594 sup + host, we examined the mutator effect of the dnaQ49 mutation on lO(am)8 phage in E. coli 594 sup + dnaQ49 cells.
The dnaQ gene, coding for the e-subunit of DNA polymerase III holoenzyme, is responsible for the proofreading activity of the DNA replicating complex.Mutation dnaQ49 leads to a high mutator effect, related to DNA replication (Horiuchi et al., 1978;Piechocki et al., 1986).We introduced the dnaQ49 mutation into 594 sup + and C600 supE cells and studied its effect on spontaneous and UV-induced mutagenesis of lO(am)8 phage at 37°C, a temperature nonpermissive for the dnaQ49 mutant.
Introduction of the dnaQ49 mutation into 594 sup + and C600 supE cells increased 100-200-fold spontaneous and UV-induced mutations Rif S to Rif R in the bacterial cells, but had no mutator effect on UV-induced or spontaneous mutagenesis of lO(am)8 phage in the nonpermissive host, 594 sup + dnaQ49 (Table 3).By contrast, the dnaQ49 mutation in the permissive, C600 supE dnaQ49 cells, leads to a high mutator effect in lO(am)8 phage, increasing the mutation frequency by about two orders of magnitude.The results suggest that in the nonpermissive host, E. coli 594 sup + , replication of lO(am)8 DNA does not occur.This is in accordance with the data reported by Ogawa & Tomizawa (1968), Shuster & Weissbach (1969), Wyatt & Inokuchi (1974) as well as with in vitro studies on the role of lO protein in the initiation of phage DNA replication (Zylicz et al., 1984;1988;Dodson et al., 1989).This supports the hypothesis that UV-induced mutagenesis of lO(am)8 in the E. coli 594 sup + host is not related to replication of damaged DNA, and sug- gests the existence in E. coli cells of more than one mechanism by which UV-induced mutations may be generated; one related to error-prone replication of damaged DNA, and another independent of DNA replication.

Effect of UmuD' overproduction on UV-induced reversion of lO(am)8 to lO + in 594 sup + cells
According to the current hypothesis on the mechanism of UV-induced mutagenesis, and mutagenesis by some chemicals, mutations are generated during replication of damaged DNA with an involvement of TLS by Pol V. DNA Pol V, the product of umuC gene, is activated by the UmuD', RecA and SSB proteins for TLS (see introduction).The requirement for the UmuD/D' and C proteins in mutagenesis induced by various mutagens has been known for years, but only lately have their functions in TLS elucidated (Tang et al., 1998;1999;2000;Ruven et al. 1999;Maor-Shoshani et al., 2000;Sutton et al., 2000).To learn more about the possible mechanism of mutagenesis induced under nonpermissive conditions for phage DNA replication, we studied the role of the UmuD/D' and C proteins in this mutagenic process.As mentioned above, induction of SOS functions stimulated UV-induced reversion of lO(am)8 to lO + in 594 sup + cells (Fig. 1), pointing to the role of inducible functions in mutagenesis occurring under nonpermissive conditions for DNA replication.We examined the role of UmuD/D' and C proteins in this mutagenic pathway by studying the effect of overproduction of these proteins on the efficiency of UV-induced reversion of lO(am)8 phage to lO + in the nonpermissive host cells, E. coli 594 sup + and the permissive C600 supE.The results (Table 4) show that overproduction of UmuD/D' and C proteins significantly increases the frequency of UV-induced reversion of phage lO(am)8 to lO + in non-induced for SOS 594 sup + host (-SOS).Overproduction of UmuDC from pSE117, UmuD'C from pGW2123 and UmuD' from pGW2122 increased the frequency of mutations 7-, 24-and 12-fold, respectively.On the other hand, the enhancing effect of the overproduction of UmuD/D' and C proteins in the C600 supE cells permissive for phage DNA replication is much lower, being 1.5-fold for pSE117 (UmuDC) and pGW2122 (UmuD') and 3-fold for pGW2123 (UmuD'C) (Table 4).Preirradiation of the 594 sup + cells to induce the SOS functions (+SOS) reduces the effect of overproduction of UmuD/D' and C proteins (Table 4).This is most probably due to deliv-Table 3. Influence of dnaQ49 mutation on UV-induced mutagenesis of lO(am)8 in the 594 sup + host and its derivatives ery of chromosomal UmuD/D' and C. The enhancing effect in 594 sup + host is seen with both the low copy plasmid pSE117 (UmuDC) and the high copy pGW2123 (UmuD'C) (about 20 and 100 copies per cell, respectively).The difference in the enhancing effect by these two plasmids, 7-and 24-fold, respectively, may be related either to the amount of UmuD and C proteins overproduced by the low copy pSE117 and the high copy pGW2123 plasmids, or to the fact that the protein UmuD encoded by pGW2123 is in the UmuD' form, known to be active in mutagenesis (Nohmi et al., 1988).It is most interesting that overproduction of UmuD' alone in 594 sup + (pGW2122) strain so efficiently (12-fold) stimulates UV-induced reversion of lO(am)8 to lO + in non-SOS-induced host cells.This suggests that a high level of UmuD', and low, non induced, of UmuC are necessary and sufficient for UV-induced mutagenesis of lO(am)8 in the nonpermissive host and points to an important role of UmuD' protein in this mutagenic pathway.

Effect of deletion of umuDC operon and overproduction of UmuD' or UmuD'C proteins on UV-induced mutagenesis of lO(am)8 in a 594 sup + DumuDC host
To examine in more detail the role of the UmuD' and C proteins in the mutagenic pro-cess occurring under nonpermissive conditions for phage DNA replication, we constructed the E. coli 594 sup + DumuDC strain, with a deleted umuDC operon.Plasmids overproducing either UmuD' (pGW2122) or UmuD'C (pGW2123) were introduced into this strain, and UV-induced reversion of lO(am)8 to lO + was monitored.As shown in Table 5, deletion of the umuDC operon in the 594 sup + DumuDC host reduces about 10 times (from 180 ´10 -7 to 20 ´10 -7 of mutants) UV-induced mutagenesis of lO(am)8 phage to the level observed in non-SOS-induced cells.Introduction of plasmid pGW2122 (UmuD') into this strain does not restore UV-mutagenesis of the phage.However, overproduction of both proteins, UmuD' and C, in 594 sup + DumuDC (pGW2123) cells does restore UV-mutagenesis of the l phage even in non SOS-induced cells, enhancing the mutation frequency about 200-fold.It follows that UmuD' protein, although important, as shown above, alone is not sufficient for the mutagenic process occurring under nonpermissive conditions for phage DNA replication.The UmuC protein is also required, albeit at a high level of UmuD' the constitutive level of UmuC seems to be sufficient (see Table 4).Preirradiation of the host to induce the SOS functions does not increase the number of mutants in the 594 sup + DumuDC or in 594 sup + DumuDC (pGW2122) strains, show-  5), the enhancing effect observed with 594 sup + DumuDC (pGW2123) seems to be related to improvement of survival of the mutants, due to induction of DNA repair systems, e.g.NER or RR, rather than to induction of some other mutagenic functions increasing the level of mutagenesis.

Different requirement for SOS functions and UmuD' and C proteins in UV-induced mutagenesis of lO(am)8 phage in the hosts nonpermissive and permissive for phage DNA replication
As shown above, in UV-induced mutagenesis of lO(am)8 phage in 594 sup + cells, nonpermissive for phage DNA replication, UmuD' and C proteins are required, and induction of other SOS-function(s) seems to be non essential.
To compare the mechanisms of UV mutagenesis under permissive and nonpermissive conditions for phage DNA replication, we also examined the effect of deletion of the umuDC operon on UV-mutagenesis of lO(am)8 phage in the E. coli C600 supE strain permissive for phage DNA replication.In contrast to the 594 sup + DumuDC, deletion of the umuDC operon in the C600 supE DumuDC host does not abolish SOS-dependent mutagenesis of l phage, and even the level of mutants is two times higher than in C600 supE (Table 5).Moreover, overproduction of UmuD' or of UmuD' and C proteins leads to some reduction of the number of mutants in SOS-induced C600 supE DumuDC cells.This might be related to the fact that a high level of UmuC protein inhibits mutagenesis (Woodgate et al., 1994) and DNA replication (Opperman et al., 1996;Frank et al., 2000) (see also Discussion).
The foregoing points to different requirements for UV-induced mutagenesis under permissive and nonpermissive conditions for phage DNA replication and suggests the involvement of different mechanisms.While UV-doses to phage and bacteria as in Table 4.
UmuD' and C proteins are indispensable for UV-mutagenesis in the nonpermissive host, they do not seem to be essential in the mutagenic process occurring under permissive conditions for phage DNA replication, so that some other inducible SOS-function(s) may replace the UmuDC proteins.This may be Pol IV (DinB), another DNA polymerase involved in TLS (Wagner et al., 1999;Kim et al., 2001).Moreover, in contrast to 594 sup + DumuDC, overproduction of UmuD'C in C600 supE DumuDC does not restore mutagenesis of lO(am)8 phage in non-SOS-induced cells (Table 5).

Different effects of deficiency in mismatch repair on mutagenesis of lO(am)8 phage induced under nonpermissive and permissive conditions for phage DNA replication
The results presented above suggest involvement of different mechanisms for UV-induced mutagenesis of lO(am)8 phage under permissive and nonpermissive conditions for phage DNA replication.To learn more about the mechanisms, we compared the effect of a deficiency in MutHLS-dependent mismatch repair (MMR) on UV-induced reversion of lO(am)8 phage to lO + , in permissive and nonpermissive hosts.MutHLSdependent MMR is known to repair mismatches formed during DNA replication to avoid mutations (Friedberg et al., 1995;Modrich et al., 1996).We constructed mutL deficient strains 594 sup + mutL and C600 supE mutL and compared the effect on UV-induced reversion of lO(am)8 phage to lO + .The results presented in Table 6 show that MMR deficiency does not affect either spontaneous or UV-induced mutagenesis of lO(am)8 phage in the 594 sup + mutL strain nonpermissive for phage DNA replication.In contrast, in the C600 supE mutL host, a high level of spontaneous mutations and a two-fold enhancement of UV-induced mutagenesis are observed.This clearly indicates that all UV-induced mutations in the O gene of lO(am)8 phage in the 594 sup + host are resistant to MMR, and support the notion that they are not a result of DNA replication.The mismatch repair deficiency did not affect the reversion of lO(am)8 phage to lO + induced by MMS in 594 sup + mutL, either (not shown).On the other hand, in the permissive host C600 supE, at least half of the UV-induced mutations in the lO(am)8 phage are subject to the MMR, implying that they are related to replication of the UV-damaged phage DNA.In addition, the absence of a mutator effect on lO(am)8 phage in 594 sup + mutL cells supports the finding that, in the 594 sup + host, DNA replication of lO(am)8 phage does not occur.4.

MMS-induced reversion of lO(am)8 to lO + under nonpermissive conditions for phage DNA replication. The effect of UmuD' protein
We were wondering which of the photoproducts in UV-irradiated phage DNA may be the premutagenic lesion leading to reversion of lO(am)8 phage to lO + .UV irradiation of DNA leads to the formation of several photoproducts, the most frequent being pyrimidine photodimers and 6-4 photoproducts (Friedberg et al., 1995).Among the minor photoproducts in UV-irradiated DNA are pyrimidine photohydrates and thymine glycols (Tg) (Yamane et al., 1967;Fisher et al., 1976;Demple & Linn, 1982;Boorstein et al., 1990;Doetsch et al., 1995).Cytosine photohydrates dehydrate readily and may also be converted by deamination to the more stable uracil photohydrates in irradiated DNA.Cytosine photohydrates, as well as uracils and Tg, are recognised by specific DNA-glycosylases (Demple & Linn, 1980), leading to generation of AP-sites, which are known to be efficient premutagenic lesions (Loeb & Peterson, 1986;Woodgate & Levine, 1996;Rothwell & Hickson, 1997;Kow, 2002), as well as thymidylyl-(3'-5')-deoxyadenosine (TA*), reported to be a potential premutagenic lesion (Zhao & Taylor, 1996;Otoshi et al., 2000).Sequence analysis of the O gene of lO(am)8 phage (data unpublished) shows that in the coding strand the TAG stop codon, the amber mutation in lO(am)8 phage, is in the sequence 5' GTG TAG ATC 3' in which Tg or the TA* photoproduct could be formed as the main photoproduct, and could be responsible for mutagenic events.So the possibility that AP-sites, a result of DNA glycosylase(s) activity, are the premutagenic lesions, can not be excluded.MMS is known to induce mutations mainly by formation of AP-sites as a result of the activity of the two 3-mA-DNA glycosylases recognising methylated purines in DNA (Friedberg et al., 1995).It has been reported that MMS-induced mutagenesis in E. coli is highly dependent on the level of UmuDC proteins (Doyle & Strike, 1995;Grzesiuk & Janion, 1996).Assuming that, in UV-induced mutagenesis of lO(am)8 phage, AP-sites might be the main premutagenic lesions we compared the effect of overproduction of UmuD' on MMS-and UV-induced mutagenesis of the phage.
The effect of overproduction of UmuD' on MMS-and UV-induced reversion of lO(am)8 to lO + in non-SOS-induced E. coli 594 sup + (pGW2122) host was studied.As shown in Table 7 treatment of phage with UV light (150 J/m2 ) or 0.33% MMS for 75 min leads to some induction of lO + revertants in 594 sup + cells.However, overproduction of UmuD' in the strain 594 sup + (pGW2122) increases the Since MMS induces mutations mainly by formation of AP-sites, the fact that overproduction of UmuD' in 594 sup + (pGW2122) affects in a similar way the mutagenesis of lO(am)8 induced by UV and MMS, suggests that AP-sites might be the premutagenic lesions in UV-and MMS-induced mutagenesis of the phage.The primary mutagenic lesion could be Tg or 3-methyloadenine (3-mAmet)known to be recognized by DNA-glycosylases.The sequence surrounding the amber codon in which mutation occurs also suggests the possibility that Tg or 3-mAmet may be formed.This supports and underlines the role of UmuD' protein in the mutagenic process occurring under nonpermissive conditions for DNA replication, and suggests that the premutagenic lesions processed by UmuD' might be AP-sites.

DISCUSSION
The current hypothesis on the mechanism of mutagenesis induced in E. coli by UV light and many chemicals assumes that mutations are a result of translesion synthesis and replication of damaged DNA.It is now well documented that UmuD' 2 C complex possesses an intrinsic DNA polymerase activity (DNA Pol V) that facilitates error-prone translesion DNA synthesis at replication forks stalled by a lesion, which is followed by resumption and continuation of replication by DNA polymerase III complex (Tang et al., 1998;1999;2000;Reuven et al., 1999;Goodman & Tippin, 2000;Livneh, 2001).Thus, the mutagenic process is considered to be related to replication of damaged DNA.Our results indicate that SOS-dependent mutagenesis induced by UV or MMS may occur under nonpermissive conditions for DNA replication, and support our previous finding that UV-induced mutagenesis in E. coli MV1178 uvrA recF may occur before resumption of DNA replication in UV-irradiated cells (Pietrzykowska & Felczak, 1991).Our present findings show that UV-induced mutagenesis of lO(am)8 phage may occur in a host nonpermissive for phage DNA replication, E. coli 594 sup + .lO(am)8 phage has the amber mutation in gene O encoding the lO protein which is responsible for initiation of phage DNA replication.Lack of lO(am)8 replication in the 594 sup + host was first shown by Ogawa & Tomizawa (1968) and Shuster & Weisbach (1969) and the role of the gene O product in initiation of phage DNA replication is now well established (Wyatt & Inokuchi, 1974;Zylicz et al., 1984;1988;Dodson et al., 1989;Stephens & McMacken, 1997).The absence of lO(am)8 phage DNA replication in the 594 sup + host is also supported by our experiments with mutants 594 sup + dnaQ49 and 594 sup + mutL.Both mutations are known to cause a high mutator effect related to DNA replication (Horiuchi et al., 1978;Piechocki et al., 1986;Fridberg et al., 1995;Rasmussen et al., 1998).Inactivation of the proof-reading activity of the 3'®5' exonuclease in a dnaQ49 mutant at 37°C markedly increases the number of errors introduced during DNA replication.On the other hand, the mutation mutL leads to a high mutator effect due to deficiency in postreplicative MutHSL-dependent MMR (Friedberg et al., 1995;Rasmussen et al., 1998).The absence of mutator effects of these two mutations on the lO(am)8 phage in 594 sup + , and high mutator effect in C600 supE support the notion that phage DNA replication does not occur in 594 sup + host cells.The possibility that DNA replication of UV-irradiated lO(am)8 phage might be initiated by UvrABC nuclease seems to be excluded by the experiments with the E. coli 594 sup + uvrA host.If incision by UvrABC nuclease could initiate replication of UV-irradiated DNA, one would expect the absence of UV-induced mutagenesis of lO(am)8 phage in the excision-repair-defective 594 sup + uvrA host.Our experiments show that this is not the case.In addition, if UvrABC nuclease could initiate replication of irradiated phage DNA, the level of UV-induced mutagenesis of lO(am)8 phage in 594 sup + uvr + should be comparable to that in C600 supE, a permissive host.As can be seen in Table 5 the efficiency of mutagenesis in 594 sup + uvr + constitutes about 30% of that in C600 supE, permissive for phage DNA replication.The absence of lO(am)8 phage DNA replication in the E. coli 594 sup + host is also supported by the fact that untargeted mutagenesis (UTM) of the phage, known to be related to the replication of undamaged l phage DNA in preirradiated cells (Brotcorne-Lannoye & Maenhaut-Michel, 1986;Kim et al., 1997;2001) does not occur in this host.It may be seen in Table 2 that when lO(am)8 phage is not irradiated with UV light, the level of lO + revertants is the same in nonirradiated (-SOS) and irradiated (+SOS) 594 sup + cells.This is in agreement with in vitro studies on the crucial role of the lO protein in initiation of l phage DNA replication (Zylicz et al., 1984;Stephens & McMacken, 1997).This is good evidence that DNA replication of lO(am)8 phage in the 594 sup + host does not occur, and strongly suggests that the observed UV-and MMS-induced mutations of lO(am)8 phage in 594 sup + cells are generated before initiation of DNA replication.Thus, we are dealing with a new mutagenic pathway in E. coli, independent of DNA replication.Another mutagenic pathway independent of DNA replication was described by Cohen-Fix & Livneh (1992) under in vitro conditions.It differs from that presented here in several respects, in that it requires the UvrABC excinuclease and DNA polymerase III, but not the functions of umuC + and recA + gene products.In contrast, the mutagenic pathway we describe does require the UmuD' and C proteins and is independent of the UvrABC excinuclease.What may be the mechanism of UV-induced mutagenesis independent of DNA replication?We consider that UV-induced mutation of lO(am)8 in the 594 sup + host is not a result of recombination.It has been reported that recombination is not involved in UV mutagenesis of l phage (Miura & Tomizawa, 1970), and the m.o.i. of 0.2 used in our experiments was too low to promote recombination.In general, recombination repair is considered to be error-free.This mutagenic pathway clearly differs from that occurring under permissive conditions for phage DNA replication.It is absolutely dependent on the UmuD' and C proteins, since we showed that deletion of the umuDC operon significantly reduces UV-mutagenesis of lO(am)8 in the 594 sup + DumuDC strain.The requirement for both proteins is illustrated by the fact that UV-mutagenesis of lO(am)8 in the 594 sup + DumuDC host may be recovered by delivery, from plasmid pGW2123, of both UmuD' and C proteins, but not by UmuD' alone.On the other hand, overproduction of UmuD' from pGW2122 in the wild type 594 sup + strain increases 12-15 times UV-and 18 times MMS-induced frequency of lO + revertants in the non-SOS-induced cells (Table 4 and 7).This suggests that for the mutagenic pathway a high level of UmuD' and a low, noninduced level of UmuC protein are necessary and sufficient, and implies an important role of UmuD' protein in the mutagenic process.A possible role of UmuC (Pol V) on the survival of mutants, in 594 sup + rather than in the mutagenic process itself, can not be excluded.We observed that overproduction of UmuD' and C proteins in the C600 (pGW2123) strain enhances 2-6 times (depending on the UV-dose) survival of UV-irradiated lO(am)8 phage (Table 8).How can one explain that under nonpermissive conditions the mutagenic process needs a high level of UmuD' and low (non-induced) level of UmuC?If UmuD' plays an important role in the generation of the mutations in lO(am)8 phage in the 594 sup + host, reversion of the amber mutation in gene O of the phage would enable initiation of DNA replication.However, phage irradiated with a dose of 150 J/m 2 has a number of unrepaired lesions which will stop the replication fork.This needs TLS synthesis with involvement of Pol V (UmuD' 2 C) for survival of the phage.The increasing effect of overproduction of UmuD'C on survival of lO(am)8 phage in C600 supE host seen in Table 8 seems to be in agreement with the hypothesis.
The absence of the UmuD' and C protein functions in the deletion mutant 594 sup + DumuDC can not be replaced by any other SOS-inducible function(s) in mutagenesis of lO(am)8 phage since no effect of preirradiation of this host cells on the level of mutants was seen (Table 5).In contrast to the 594 sup + DumuDC host, deletion of the umuDC operon in C600 supE DumuDC did not abolish UV-induced mutagenesis of lO(am)8 phage, indicating that a different mutagenic processes independent of UmuD'C may occur in the host permissive for phage DNA replication.It requires preirradiation of the host cells to induce some SOS function(s) that are not UmuD and C proteins.Overproduction of UmuD', and to a lesser extent of UmuD'C, reduces the level of mutants (Table 5), which is in agreement with the reported inhibition of mutagenesis by a high level of UmuC protein (Woodgate et al., 1994).What could be the SOS function(s) responsible for the UmuD'Cindependent mutagenesis in C600 supE DumuDC host?One possible candidate is the product of the dinB gene, a homologue of UmuC protein, reported to be Pol IV (Wagner et al., 1999;Tang et al., 2000).Overproduction of the dinB gene product was reported to highly enhance UTM of l phage (Kim et al., 1997;2001).In the last few years the roles of several DNA polymerases controlled by SOS in E. coli has been elucidated: Pol II (PolB), Pol IV (DinB) and Pol V (UmuC) (for a review see: Goodman & Tippin, 2000;Livneh, 2001;Wagner et al., 2002).Which of them may be responsible for the UV-and MMS-induced mutagenesis of lO(am)8 in the permissive C600 supE supE DumuDC cells?Pol II, being known to be involved in frameshift mutagenesis in the case of base substitution mutagenesis studied in lO(am)8 phage, seems to be excluded.On the other hand Pol IV, able to produce frameshifts and base substitutions in l phage (Kim et al., 1997;2001)  merase which can replace Pol V in the C600 DumuDC host.It was shown in vitro that both Pol IV and Pol V may bypass photochemical lesions and AP-sites, but Pol V is much more efficient than Pol IV (Tang et al., 2000;Napolitano et al., 2000).In the absence of Pol V, Pol IV may become active in mutagenic TLS.Studies on the role of SOS-controlled DNA polymerases in induced mutagenesis in lO(am)8 phage under nonpermissive and permissive conditions are under way.
Our results have shown that in E. coli two different mutagenic pathways may operate, one dependent and the other independent of DNA replication.This is also corroborated by the fact that deficiency in MutHLS-dependent mismatch repair has no effect on UV-and MMS-induced mutagenesis of lO(am)8 phage in the 594 sup + mutL host, but increases it about 2.5-fold in C600 supE mutL (Table 6).This implies that about 60% of the mutations induced in the host permissive for DNA replication undergo postreplicative mismatch repair, and about 40% are resistant to MMR.By contrast, all mutations generated in 594 sup + cells are not subject to mismatch repair.This is in agreement with the observation that the efficiency of UV-induced mutagenesis of lO(am)8 in the 594 sup + host is 30-40% of that observed in C600 supE (see Table 5).Moreover, sequence analysis of lO + revertants showed that most mutations induced in the 594 sup + host are transversions, and in C600 supE transitions prevail (data unpublished).These two pathways could not be separated in C600 supE cells able to replicate phage DNA, but it proved possible with the use of the 594 sup + host unable to replicate lO(am)8 phage DNA.We believe that our studies on mutagenesis independent of DNA replication may be useful for understanding the mechanism of generation of mutations in nondividing cells e.g.somatic cells.Studies on DNA repair and mutagenesis in eukaryotic cells, including man, indicate that general rules may be common for both Prokaryota and Eukaryota (Sutton & Walker, 2001;Goodman, 2002;Stary et al., 2003).Most significant is the identification in eukaryotic cells, including humans, of DNA polymerases belonging to the Y family (earlier called the UmuC/Rad30/DinB family) which promote translesion replication (for reviews see: Woodgate, 1999;Goodman & Tippin, 2000;Livneh, 2001;Wang, 2001;Woodgate, 2001;Boudsocq et al., 2002).
We thank Professor David Shugar for critical reading of the manuscript.
of overproduction of UmuD and C proteins on UV-induced reversion of lO(am)8 to lO + in non-SOS-and SOS-induced 594 sup + and C600 supE hosts UV doses used: 150 J/m 2 to phage and 70 J/m 2 or 40 J/m 2 to induce SOS in 594 sup + and C600 supE hosts, respectively (+SOS); nonirradiated hosts (-SOS); numbers in the table are ratios of mutation frequency in the host with a plasmid overproducing Umu proteins to the host with a control plasmid (pBR322 for pSE117) or without plasmid for pGW2123 and pGW2122.ing that no other induced SOS function(s) can replace the UmuD and C proteins in the nonpermissive host.On the other hand, preirradiation of 594 sup + DumuDC (pGW2123) cells overproducing both UmuD' and C proteins enhances two-fold the level of mutants in comparison to that in non-preirradiated 594 sup + DumuDC (pGW2123), indicating that some inducible function(s), other than the UmuD' and C proteins, slightly enhances the level of mutants.However, since there was no effect of preirradiation of 594 sup + DumuDC (pGW2122) cells on the level of mutants (Table

Table 6 .
Effect of mutL on UV-induced mutagenesis of lO(am)8 phage in the host nonpermissive (594 sup + mutL) and permissive (C600 supE mutL) for phage DNA replication Mutation frequencies and (±S.D.) are average values from at least six independent experiments, each in duplicate.UV doses to the hosts used for induction SOS as in Table may be the translesion DNA poly-Table 8. Effect of overproduction of UmuD' and UmuD'C proteins on survival of UV-irradiated lO(am)8 phage (%) Survival (%) of lO(am)8 phage irradiated with UV-dose (J/m 2 ) and (±S.D.) are average values from at least five independent experiments, each in duplicate.

Table 2 . Comparison of UV-induced mutagenesis of lO(am)8 phage in E. coli 594 sup + and its uvrA derivative Mutation
frequencies and (±S.D.) are average values from at least four independent experiments, each in duplicate.

Table 5 . UV-induced reversion of lO(am)8 phage to lO + in 594 sup + DumuDC and C600 supE DumuDC hosts and the effect of overproduction of UmuD' and C proteins Mutation
frequencies and (±S.D.) are average values from at least five independent experiments, each in duplicate.

Table 7 . Effect of overproduction of UmuD' protein on UV-and MMS-induced reversion of lO(am)8 to lO + in non-SOS-induced 594 sup + strain
Mutation frequencies and (±S.D.) are average values from at least five independent experiments, each in duplicate.Phage was irradiated with UV light (150 J/m mutation frequency 15-fold for UV and 18-fold for MMS mutagenesis relative to 594 sup + cells without the plasmid.It follows that in both UV-and MMS-induced mutagenesis, overproduction of UmuD' and the basal noninduced level of UmuC are sufficient for lO(am)8 phage reversion under nonpermissive conditions for phage DNA replication.