Mutant forms of the Escherichia coliβ sliding clamp that distinguish between its roles in replication and DNA polymerase V-dependent translesion DNA synthesis

Authors

  • Mark D. Sutton,

    Corresponding author
    1. Department of Biochemistry, School of Medicine and Biomedical Sciences, University at Buffalo, SUNY, 3435 Main Street, 140 Farber Hall, Buffalo, NY 14214, USA.
      E-mail mdsutton@Buffalo.edu; Tel. (+1) 716 829 3581; Fax (+1) 716 829 2661.
    Search for more papers by this author
  • Jill M. Duzen,

    1. Department of Biochemistry, School of Medicine and Biomedical Sciences, University at Buffalo, SUNY, 3435 Main Street, 140 Farber Hall, Buffalo, NY 14214, USA.
    Search for more papers by this author
  • Robert W. Maul

    1. Department of Biochemistry, School of Medicine and Biomedical Sciences, University at Buffalo, SUNY, 3435 Main Street, 140 Farber Hall, Buffalo, NY 14214, USA.
    Search for more papers by this author

E-mail mdsutton@Buffalo.edu; Tel. (+1) 716 829 3581; Fax (+1) 716 829 2661.

Summary

The Escherichia coliβ sliding clamp is proposed to play an important role in regulating DNA polymerase traffic at the replication fork. As part of an ongoing effort to understand how organisms manage the actions of their multiple DNA polymerases, we examined the ability of several mutant forms of the β clamp to function in DNA polymerase V- (pol V-) dependent translesion DNA synthesis (TLS) in vivo. Our results indicate that a dnaN159 strain, which expresses a temperature sensitive form of the β clamp, was impaired for pol V-dependent TLS at the permissive temperature of 37°C. This defect was complemented by a plasmid that expressed near-physiological levels of the wild-type clamp. Using a dnaN159 mutant strain, together with various plasmids expressing mutant forms of the clamp, we determined that residues H148 through R152, which comprise a portion of a solvent exposed loop, as well as position P363, which is located in the C-terminal tail of the β clamp, are critically important for pol V-dependent TLS in vivo. In contrast, these same residues appear to be less critical for pol III-dependent replication. Taken together, these findings indicate that: (i) the β clamp plays an essential role in pol V-dependent TLS in vivo and (ii) pol III and pol V interact with non-identical surfaces of the β clamp.

Introduction

Despite the high degree of efficiency with which base excision and nucleotide excision pathways repair DNA lesions (Friedberg et al., 1995), circumstances arise in which damaged bases either evade repair or cannot be repaired. In these cases, lesions presumably persist in the DNA until such time as the replication machinery encounters them. However, because replicative DNA polymerases are by necessity high fidelity enzymes, they are unable to bypass most non-instructive or distorting DNA lesions (Friedberg et al., 1995; Baker and Bell, 1998). Therefore, when the replication fork encounters a lesion that was not repaired, the lesion must be tolerated via a specialized form of DNA replication referred to as translesion DNA synthesis (TLS). TLS is inherently error-prone, in part resulting from the relaxed fidelity of the specialized DNA polymerases that catalyse it, as well as from the fact that many lesions are either non-coding or miscoding. Nevertheless, TLS contributes significantly to cell survival following DNA damage by allowing the completion of DNA replication.

The process of TLS is best understood in Escherichia coli, where it has been studied for more than 30 years (reviewed in Sutton et al., 2000). Although three of the five E. coli DNA polymerases (pol II, pol IV and pol V) are SOS regulated (reviewed in Sutton and Walker, 2001), and are capable of participating in TLS, most TLS following ultraviolet light (UV) irradiation or exposure to chemical carcinogens requires the umuDC-encoded DNA pol V (Kato and Shinoura, 1977; Steinborn, 1978). In addition to regulation by SOS, pol V polymerase activity is further regulated post-transcriptionally by a RecA-mediated self-cleavage event that serves to remove the N-terminal 24 residues of the umuD gene product, yielding UmuD′ (Burckhardt et al., 1988; Nohmi et al., 1988; Shinagawa et al., 1988). The UmuD′2 homodimer, together with UmuC, functions as pol V to enable most TLS.

Although the intact form of UmuD, acting together with UmuC, is inactive as a DNA polymerase, it is suggested to play a role in a primitive DNA damage checkpoint control (Opperman et al., 1999; Murli et al., 2000). Modest overexpression of these proteins, or a non-cleavable form of the intact UmuD protein together with UmuC (but not UmuD′ together with UmuC), delayed the resumption of DNA synthesis following replication blocking DNA damage (Opperman et al., 1999). Furthermore, expression of a non-cleavable form of UmuD, together with UmuC, conferred a modest (i.e. approximately fivefold) protection against killing following UV irradiation. This protective effect was dependent on functional nucleotide excision repair, and was completely independent of the TLS function of pol V (Opperman et al., 1999). Furthermore, UmuD and UmuD′ both interact with components of pol III, but do so differently: intact UmuD interacts more strongly with the β processivity clamp subunit of pol III than does UmuD′, while UmuD′ interacts more strongly with the α catalytic subunit of pol III than does intact UmuD (Sutton et al., 1999). Thus, taken together, these findings suggest a model in which intact UmuD, together with UmuC, helps to regulate DNA replication following induction of the SOS response, presumably via direct physical interaction with components of pol III, thereby allowing additional time for nucleotide excision repair to remove lesions from the DNA before the resumption of DNA replication. A fascinating aspect of this model is that cleavage of UmuD to UmuD′ releases the checkpoint, while simultaneously activating the intrinsic DNA polymerase activity of pol V to effect TLS (Opperman et al., 1999; Sutton et al., 1999).

The E. coliβ sliding clamp, which is encoded by the dnaN gene, was initially identified based on its ability to confer processivity on pol III (Burgers et al., 1981; Stukenberg et al., 1991). It has since been reported that β interacts with a variety of proteins involved in different aspects of DNA metabolism, including the other four E. coli DNA polymerases (Bonner et al., 1992; Sutton et al., 1999; Tang et al., 1999; Lopez de Saro and O’Donnell, 2001; Sutton et al., 2001; Lenne-Samuel et al., 2002; Sutton et al., 2002). Based on a recent bioinformatics study, a consensus sequence (QL[S/D]LF), referred to as the eubacterial clamp-binding motif, has been identified that appears to mediate, at least in part, interactions of certain partner proteins with β (Dalrymple et al., 2001). Peptides homologous to sequences resembling this motif present in four of the five different E. coli DNA polymerases have been shown to interact with βin vitro (Dalrymple et al., 2001; Lopez de Saro et al., 2003; Wijffels et al., 2004), presumably through contact with a hydrophobic cleft in the clamp (Jeruzalmi et al., 2001). Furthermore, mutations impairing the predicted eubacterial clamp-binding motif present in pol II, pol IV and pol V have been constructed (Becherel et al., 2002; Lenne-Samuel et al., 2002). In each case, the respective DNA polymerase appeared to loose its ability to promote TLS in vivo, suggesting that these polymerases must interact with the clamp in order to gain access to the replication fork. Consistent with this conclusion, in vitro reconstitution of polymerase V- (pol V-) dependent TLS using purified components suggested that the β sliding clamp protein plays an important role in stimulating pol V processivity by stabilizing the pol V-DNA complex (Pham et al., 2001). Finally, the δ subunit of γ complex, which acts to load the clamp onto DNA, also interacts with the hydrophobic cleft of β (Jeruzalmi et al., 2001). Taken together, these findings suggest that competition between different partners for binding to β might help to co-ordinate both the nature and the order of the events that take place at a replication fork (Sutton et al., 1999; 2000; Walker et al., 2000; Sutton and Walker, 2001; Sutton, 2004).

Despite the fact that pol III and pol V appear to compete with each other for binding to the clamp, our previous results suggest that pol V and pol III bind overlapping yet discrete surfaces on β (Sutton et al., 2001; 2004; Duzen et al., 2004). Therefore, as part of an ongoing effort to understand how organisms manage the actions of their different DNA polymerases to co-ordinate DNA replication with DNA repair and damage tolerance, we set out to identify mutant forms of the β clamp that were impaired for pol V-dependent TLS in vivo, yet retained an ability to participate in pol III-dependent DNA replication. Our results discussed below indicated that although a dnaN159 mutant is able to grow at 37°C, it is nevertheless impaired for pol V-dependent TLS at this temperature. Furthermore, using a dnaN159 mutant strain bearing plasmids expressing near-physiological levels of different mutant β clamp proteins, we have identified amino acid residues of the β clamp required for pol V-dependent TLS in vivo. By contrast, these same mutant clamp proteins retained an ability to grow under normal laboratory conditions, indicating that they were able to function in pol III-dependent DNA replication in vivo.

Results

The dnaN159 mutant strain is impaired for pol V-dependent TLS at 37°C

The E. coliβ sliding clamp is proposed to play an important role in co-ordinating DNA polymerase traffic at the replication fork (reviewed in Lopez de Saro and O’Donnell, 2001; Sutton and Walker, 2001). As part of an ongoing effort to understand how E. coli manages the actions of pol V to co-ordinate TLS with ongoing replication, we set out to establish conditions whereby we could evaluate the ability of plasmid-expressed mutant β clamp proteins to function in TLS in vivo. The dnaN159 allele encodes a mutant form of the β clamp that bears two amino acid substitutions (G66E and G174A) and displays a temperature sensitive growth phenotype. The ability of the dnaN159 mutant strain MS101 to form colonies on solid media is quantitatively similar at temperatures ranging from 30 to 37°C, while growth is severely impaired at higher temperatures (Grompone et al., 2002; Sutton, 2004). We hypothesized that these mutations would also impair pol V-dependent TLS at elevated temperatures, thereby providing a means to characterize the ability of plasmid-expressed mutant clamp proteins to catalyse pol V-dependent TLS.

To examine this possibility, we compared the efficiency of SOS mutagenesis at 30 and 37°C using an argE3(Oc)→Arg+ reversion assay: pol V-dependent TLS serves as the mechanistic basis for SOS mutagenesis (reviewed in Sutton et al., 2000). Briefly, cultures of isogenic dnaN + (MS100) and dnaN159 (MS101) strains were grown in liquid media at the permissive temperature of 30°C before irradiation with UV light. Aliquots of UV irradiated cultures were then immediately spread onto minimal plates supplemented with trace arginine, followed by incubation for three days at 30 or 37°C before counting the number of Arg+ revertants. Although the dnaN159 (MS101) and dnaN+ (MS100) strains were indistinguishable from each other at 30°C with respect to their proficiencies in SOS mutagenesis (Fig. 1A), proficiency of the dnaN159 mutant at 37°C was consistently reduced by a factor of three- to fourfold relative to the dnaN + strain (Fig. 1B).

Figure 1.

Proficiency of dnaN+ and dnaN159 strains in SOS mutagenesis. Respective proficiencies in SOS mutagenesis of strains MS100 (relevant genotype: dnaN+) and MS101 (relevant genotype: dnaN159) were measured at 30°C (A and C) and 37°C (B and D) using the standard argE3(Oc)→Arg+ reversion assay (A and B), or a rifampicin resistance assay (C and D) as described in Experimental procedures. Shown for argE3(Oc)→Arg+ reversion is the average of duplicates from two independent experiments ± range. Shown for RifR is the average of four separate experiments ± the standard deviation. Mutability is expressed relative to the isogenic dnaN + strain MS100, which was set equal to 100%. This strain produced 398 or 654 Arg+ revertants per 107 survivors at 30°C, 561 and 743 Arg+ revertants per 107 survivors at 37°C, 69, 40, 18 and 17 RifR colonies per 107 survivors at 30°C, and 60, 61, 548 and 783 RifR colonies per 107 survivors at 37°C.

Respective proficiencies in SOS mutagenesis of the dnaN+ (MS100) and dnaN159 strains (MS101) were also measured using a rifampicin resistance assay. In this analysis, strains were grown to mid-exponential phase at 30°C, either irradiated or mock-irradiated with UV, then cultured overnight at 30 or 37°C before being plated onto solid Luria–Bertani (LB) medium either containing or lacking rifampicin. As observed with the Arg+ reversion assay, the dnaN159 (MS101) and dnaN+ (MS100) strains were similarly proficient in SOS mutagenesis at 30°C (Fig. 1C). However, at 37°C the frequency of rifampicin resistance for the dnaN159 strain was reduced by more than 18-fold relative to the isogenic dnaN+ strain (Fig. 1D).

Both the temperature sensitive growth phenotype (Table 1) and the SOS mutagenesis defect of the dnaN159 strain MS101 were restored by transformation with plasmid pJD100, which expresses wild-type β clamp, but not the control plasmid pWSK29 (Fig. 2). In contrast, transformation of strain MS101 with plasmid pJD109, which expresses dnaN159, did not suppress the temperature sensitive growth phenotype of the dnaN159 mutant strain (Table 1), although it did suppress the SOS mutagenesis defect of this strain at 37°C (Fig. 2). Based on quantitative Western blot analysis, the steady-state level of the β clamp was roughly threefold higher than normal physiological levels in strain MS101 bearing the dnaN+-expressing plasmid pJD100, while it was comparable to normal physiological levels in the same strain bearing the dnaN159-expressing plasmid pJD109 (Table 2). These results, suggesting that the β159 mutant was more susceptible to proteolysis than is the wild-type clamp, are consistent with our finding that the steady-state level of β159 was reproducibly ∼30 to ∼40% lower than that of the wild-type clamp among three separate pairs of isogenic strains (Table 2).

Table 1.  Abilities of plasmid-expressed dnaN + and dnaN159 to complement the temperature sensitive growth phenotype of E. coli MS101.a
Plasmid dnaN alleleColony-forming units per millilitre of MS101 transformant at
30°C42°CRatio (42/30°C)
  • a. 

    Representative transformants of strain MS101 (relevant genotype: dnaN159) bearing the indicated plasmids were grown overnight at 30°C, and appropriate dilutions were plated onto LB agar plates supplemented with 150 µg ml−1 ampicillin and 50 µM IPTG. Colony-forming units (cfu) were counted after overnight incubation at 30 or 42°C. Values shown for cfu at 30 and 42°C, as well as for the ratio (42/30°C), are the average of two separate experiments ± the range.

pWSK29None1.5 (±0.4) × 109<1.0 × 104<6.7 (±0.15) × 10−6
pJD100 dnaN + 1.7 (±0.2) × 109  1.4 (±0.4) × 1090.82 (±0.13)
pJD109 dnaN159 1.7 (±0.2) × 109  7.5 (±1.6) × 105 4.5 (±1.5) × 10−4
Figure 2.

Expression of wild-type dnaN or dnaN159 from a plasmid suppresses the pol V-dependent SOS mutagenesis defect of a dnaN159 mutant at 37°C. Respective proficiencies in SOS mutagenesis of strain MS101 (relevant genotype: dnaN159) bearing either pWSK29 (control), pJD100 (dnaN+) or pJD109 (dnaN159) were measured at 37°C using the standard argE3(Oc)→Arg+ reversion assay as described in Experimental procedures. Shown is the average of duplicates from two independent experiments ± range. Mutability is expressed relative to the strain bearing the dnaN+-expressing plasmid pJD100, which was set equal to 100%. This strain produced 104 or 115 Arg+ revertants per 107 survivors at 37°C. Based on the one-way Anova with Dunnett test, the difference between the means of the 3 strains at the 95% confidence interval was statistically significant (= 0.018). In pair-wise comparisons, the difference between the means of the pJD109 and the pWSK29 transformants was statistically significant at the 95% confidence interval (t-value = −5.10, P = 0.036), while that between the pJD100 and pWSK29 transformants was not (t-value = −4.22, P = 0.052).

Table 2.  Steady-state levels of chromosome-expressed and various plasmid-expressed β clamp proteins in different E. coli strains.
E. coli strainaPlasmidaβ clamp protein expressed fromMolecules per cell
(as dimer)b
Relative
steady-state levelc
ChromosomePlasmid
  • a. 

    E. coli strains and plasmids used in this analysis are described in Table 5.

  • b. 

    The number of molecules of β clamp (as dimer) present in each strain was measured using quantitative Western blot analysis as described in Experimental procedures. The numbers represent the average of the sum of chromosomally expressed as well as plasmid-expressed (if relevant) clamp from two independent experiments ± the range. Values reported above are consistent with previous reports suggesting that E. coli strain C600 possesses between 350 and 460 clamps cell−1 (Leu et al., 2000).

  • c. 

    The steady-state level of the clamp in each dnaN159 strain bearing the indicated plasmid is expressed relative to the level observed in the corresponding wild-type dnaN parent strain, which was arbitrarily set equal to 1.0 (i.e. the steady-state level of β observed in each MS101 transformant is expressed relative to the level of wild-type β observed in strain MS100, the level observed in each MS107 transformant is expressed relative to MS106, and levels observed in each MS120 transformant are expressed relative to that observed in MS119).

  • d. 

    Other mutant β clamp proteins cloned into pSU38 were expressed at similar levels based on qualitative Western blot analysis (data not shown).

MS100None dnaN + None 402 (±60) 1.0
MS101pWSK29 dnaN159 None 281 (±90)  0.70
pJD100  dnaN + 1144 (±532)  2.85
pJD109  dnaN159  226 (±50)  0.56
MS106None dnaN + None 542 (±83) 1.0
MS107pWSK29 dnaN159 None 364 (±31)  0.67
pJD100  dnaN + 1218 (±354)  2.25
pJD101 Q61K1291 (±562)  2.38
pJD102 S107L 927 (±500)  1.71
pJD103 D150N1375 (±771)  2.54
pJD104 G157S1010 (±542)  1.86
pJD105 V170M1156 (±510)  2.13
pJD106 E202K1114 (±635)  2.06
pJD107 M204K1010 (±219)  1.86
pJD108 P363S1010 (±542)  1.86
MS119None dnaN + None 557 (±37) 1.0
MS120dpSU38 dnaN159 None 325 (±163)  0.58
pACYCdnaN+  dnaN + 1525 (±275)  2.74
pACYCβ363 P363S1518 (±69)  2.73
pACYCβ150 D150N1388 (±113)  2.49
pACYCβ3A β-149–1511238 (±25)  2.22
pACYCβ5A β-148–1521238 (±325)  2.22
pACYCβ148 H148A1688 (±313)  3.03
pACYCβ152 R152A1650 (±538)  2.96

Taken together, results discussed above indicate that although a dnaN159 mutant is capable of pol III-dependent growth at 37°C, it is impaired for pol V-dependent TLS at this temperature, possibly because of limiting levels of active β159 clamp protein. Furthermore, our finding that near-physiological levels of the wild-type clamp expressed from a plasmid could complement the SOS mutagenesis defect of the dnaN159 mutant strain provided us with a method by which we could determine whether other mutations in the clamp impair SOS mutagenesis in vivo.

Proficiency in SOS mutagenesis of mutant β clamp proteins impaired for genetic interaction with pol V

We recently described the identification and preliminary biochemical characterization of eight, novel plasmid-encoded dnaN alleles impaired for genetic interactions with pol V (Sutton et al., 2001). These alleles were identified based on their inability, when overexpressed, to confer a cold-sensitive growth phenotype on a strain expressing pol V at roughly twofold higher than the normal SOS-induced levels (Sutton et al., 2001). Interestingly, these same dnaN alleles were also impaired for exacerbating the cold sensitive growth phenotype conferred by similarly elevated steady-state levels of intact UmuD together with UmuC (Sutton et al., 2001). We have recently purified these mutant β clamp proteins to homogeneity, and characterized their interactions with UmuD, UmuD′ and the α catalytic subunit of pol III in vitro (Duzen et al., 2004); these findings are summarized in Table 3. Based on these analyses, it was unclear whether these mutant clamp proteins are impaired for the checkpoint function of UmuD/UmuC, the TLS function of pol V, or both.

Table 3.  Abilities of different plasmid-encoded dnaN alleles to complement the temperature sensitive growth phenotype of E. coli MS107.a
Plasmid dnaN alleleAbility to interact with (in vitro)bColony-forming units per millilitre of MS107 transformant at
UmuDUmuD′pol III α30°C42°CRatio (42/30°C)
  • a. 

    Representative transformants of strain MS107 (relevant genotype: dnaN159ΔuvrB::cat) bearing the indicated plasmids were grown overnight at 30°C, and appropriate dilutions were plated onto LB agar plates supplemented with 150 µg ml−1 ampicillin and 50 µM IPTG. Colony-forming units were counted after overnight incubation at 30 or 42°C. Values shown for cfu at 30 and 42°C, as well as the ratio (42/30°C), are the average of triplicates ± the standard deviation. NA, not applicable.

  • b. 

    Interactions of each purified mutant β protein with UmuD, UmuD′ or the α catalytic subunit of DNA polymerase III (pol III α) was measured in vitro using chemical cross-linking (UmuD and UmuD′) or Superose-12 gel filtration chromatography (pol III α). These results were reported previously (Duzen et al., 2004) and are included here for comparison to plating efficiencies. Abbreviations are as follows: ++, proficient for interaction; +, moderately impaired for interaction; –, severely impaired for interaction (but interaction is still detectable in vitro); NA, not applicable.

  • c. 

    ND, not determined because of the apparent instability of the MS107(pJD103) strain when grown at 42°C. Growth values at 42°C ranged from <1.0 × 103−6.8 × 108 (data not shown). Moreover, different clones selected following growth at 42°C displayed different phenotypes, consistent with the idea that they had acquired suppressor mutations.

  • d. 

    Based on formaldehyde or glutaraldehyde cross-linking experiments (Duzen et al., 2004), these mutant β clamp proteins appeared to interact more strongly with UmuD and/or UmuD′ (as indicated) than did wild-type β.

pWSK29NoneNANANA2.9 (±0.12) × 109<1.0 × 104<3.5 (±0.17) × 10−6
pJD100 dnaN + ++++++1.5 (±0.35) × 109  1.2 (±0.06) × 109  0.84 (±0.14)
pJD101Q61K++++++2.5 (±0.44) × 109  2.5 (±0.44) × 109  1.0 (±0.04)
pJD102S107L+++++2.9 (±0.12) × 109  2.6 (±0.25) × 109  0.92 (±0.06)
pJD103D150N++++++1.6 (±0.27) × 109NDcNDc
pJD104G157S++d+2.1 (±0.23) × 109  2.2 (±0.38) × 109  1.0 (±0.07)
pJD105V170M++2.8 (±0.75) × 109  1.9 (±0.01) × 109  0.72 (±0.22)
pJD106E202K++d++d++3.0 (±0.30) × 109  2.5 (±0.64) × 109  0.86 (±0.22)
pJD107M204K++d++++2.4 (±0.21) × 109  2.4 (±0.23) × 109  1.0 (±0.11)
pJD108P363S2.3 (±0.20) × 109  2.2 (±0.25) × 109  0.95 (±0.19)

To determine whether any of these eight mutant β clamp proteins were impaired for pol V-dependent TLS in vivo, we subcloned all eight dnaN alleles into the low copy number plasmid pWSK29, which expresses near-physiological levels of the clamp on addition of 50 µM IPTG to the growth medium (Table 2). Based on quantitative Western blot analysis, the steady-state level of each mutant β clamp protein was between 927 and 1375 clamps cell−1, or roughly twofold higher than normal physiological levels. Importantly, under these conditions, each of the eight mutant dnaN alleles except for D150N was able to complement the temperature sensitive growth phenotype of the dnaN159 mutant strain at 42°C (Table 3), indicating that each was proficient for pol III-dependent DNA replication.

To determine whether the seven dnaN alleles that were proficient in pol III replication were impaired for pol V-dependent TLS, we measured their abilities to participate in SOS mutagenesis using a dnaN159 mutant strain. Briefly, cultures were grown at 42°C in the presence of 50 µM IPTG, irradiated with UV light, cultured overnight at 42°C, and then appropriate dilutions were plated onto LB plates with or without rifampicin. Based on results of this analysis (Fig. 3A), the P363S mutant was significantly impaired for SOS mutagenesis, yielding only 8.4% as many rifampicin resistant colonies as the wild-type strain (1.7 induced RifR colony per 107 survivors for P363S, versus 20.3 induced RifR colonies per 107 survivors for dnaN+). The other dnaN mutants were similar to dnaN + with respect to their ability to promote pol V-dependent TLS (Fig. 3).

Figure 3.

Ability of different dnaN mutant strains to function in pol V-dependent SOS mutagenesis.
A. Respective proficiencies in SOS mutagenesis of representative transformants of strain MS107 (relevant genotype: ΔuvrB::cat dnaN159) bearing the indicated pWSK29-encoded dnaN alleles were measured at 42°C using a rifampicin resistance assay as described in Experimental procedures. Shown is the average of duplicates from three independent experiments ± the standard deviation. Mutability is expressed relative to the strain bearing the dnaN+-expressing plasmid pJD100, which was set equal to 100%. This strain produced 26, 12 or 24 RifR colonies per 107 survivors at 42°C.
B. Respective proficiencies in SOS mutagenesis of strain MS120 (relevant genotype: dnaN159) bearing the indicated pSU38-encoded dnaN alleles were measured at 42°C using the standard argE3(Oc)→Arg+ reversion assay as described in Experimental procedures. Shown is the average of duplicates from two independent experiments ± the range. Mutability is expressed relative to the strain bearing the dnaN +-expressing plasmid pACYCdnaN+, which was set equal to 100%. This strain produced 96 or 226 Arg+ revertants per 107 survivors at 42°C. Based on the two sample T-test, the difference between the means for the dnaN+ and the P363S strains in (A) is statistically significant at the 95% confidence interval (t-value = 4.13, P = 0.014), while that in (B) is not (t-value = 1.57, P = 0.26). The lack of statistical significance for the values in (B) might relate to the wide range in values observed for the dnaN+ (control) strain (see above).

In addition to pWSK29, we also subcloned each of the dnaN alleles into the moderate copy number plasmid pSU38. Like plasmid pWSK29, pSU38 also contains the lac promoter. However, presumably because of its higher copy number, IPTG induction was not required in order to achieve near-physiological steady-state levels of the different clamp proteins. In fact, a slightly higher steady-state level of the clamp was observed with strain MS120 bearing each of the different pSU38 derivatives and grown in the absence of IPTG than that observed with MS107 bearing the corresponding pWSK29 derivatives grown in the presence of 50 µM IPTG (Table 2). In contrast to the pWSK29 derivative expressing D150N (pJD103), the pSU38 derivative expressing D150N (pACYCβ150) was able to form colonies on solid media at 42°C, although it did not grow well at 42°C (see Table 4 legend), especially in liquid culture, and hence was not analysed under these conditions (see below). In contrast, the other seven plasmid-borne dnaN alleles were able to complement the temperature sensitive growth phenotype of strain MS120 (Table 4), further indicating that each was proficient for pol III-dependent replication.

Table 4.  Ability of dnaN alleles bearing mutations in the D150 loop region to complement the temperature sensitive growth phenotype of E. coli MS120.a
Plasmidb dnaN allele Colony-forming units per millilitre of MS120 transformant at
30°C42°CRatio (42/30°C)
  • a. 

    Representative transformants of strain MS120 (relevant genotype: dnaN159) bearing the indicated plasmids were grown overnight at 30°C, and appropriate dilutions were plated onto LB agar plates supplemented with 60 µg ml−1 kanamycin. Colony-forming units were counted after overnight incubation at 30 or 42°C. Values shown for cfu at 30 and 42°C, as well as the ratio (42/30°C), are the average of triplicates ± the standard deviation.

  • b. 

    The CmR and ApR pSU38 derivatives expressing either wild-type β or D150 loop mutations (see Table 5) displayed similar growth phenotypes as those reported above for the KmR pSU38 derivatives (data not shown).

  • c. 

    Colony-forming units obtained following incubation at 42°C were small relative to the pACYCdnaN + control strain. The modest temperature sensitivity of strain MS120(pACYCβ150) was further examined using a transformation assay. When plasmid pACYCβ150 was used to transform strain MS120, and the transformation reaction was split and incubated at both 30 and 42°C, 504 normal sized colonies and 668 tiny colonies were observed at 42°C, while 2384 were observed at 30°C. In contrast, transformation of MS120 with pACYCdnaN+ produced 2192 colonies at 30°C, compared to 3056 at 42°C. Transformation with pSU38 produced 2136 colonies at 30°C, compared to 1 at 42°C. Taken together these results indicate that pACYCβ150, like plasmid pJD103 (see Table 3), does not fully suppress the temperature sensitive growth phenotype of a dnaN159 mutant strain at 42°C.

pSU38None2.2 (±0.15) × 109<1.0 × 105<4.5 (±0.03) × 10−5
pACYCdnaN+ dnaN + 1.7 (±0.15) × 109  1.8 (±0.10) × 109  1.0 (±0.08)
pACYCβ61Q61K1.5 (±0.47) × 109  1.9 (±0.25) × 109  1.3 (±0.35)
pACYCβ107S107L2.3 (±0.20) × 109  2.6 (±0.20) × 109  1.2 (±0.16)
pACYCβ157G157S2.1 (±0.20) × 109  2.4 (±0.36) × 109  1.1 (±0.28)
pACYCβ170V170M2.0 (±0.12) × 109  1.7 (±0.17) × 109  0.84 (±0.12)
pACYCβ202E202K2.3 (±0.26) × 109  2.4 (±0.21) × 109  1.0 (±0.06)
pACYCβ204M204K2.2 (±0.26) × 109  2.0 (±0.21) × 109  0.90 (±0.46)
pACYCβ363P363S2.6 (±0.26) × 109  2.7 (±0.25) × 109  1.1 (±0.18)
pACYCβ148H148A2.1 (±0.10) × 109  1.9 (±0.06) × 109  0.89 (±0.07)
pACYCβ150cD150N1.2 (±0.36) × 109  0.83 (±0.29) × 109  0.78 (±0.42)
pACYCβ152R152A0.80 (±0.17) × 109  1.9 (±0.06) × 106  2.5 (±0.52) × 10−3
pACYCβ3Aβ-149–1511.7 (±0.29) × 109  2.0 (±0.21) × 106  1.2 (±0.26) × 10−3
pACYCβ5Aβ-148–1521.2 (±0.06) × 109  3.1 (±0.21) × 105  2.5 (±0.05) × 10−4

Use of these higher copy number plasmids, together with the dnaN159 mutant strain MS120 confirmed that the P363S mutation was impaired for TLS (Fig. 3B). Under these conditions, the P363S mutant was only 30.7% as efficient at UV-induced argE3(Oc)→Arg+ reversion as the same strain bearing the dnaN+-expressing plasmid pACYCdnaN+ (51.5 Arg+ revertants per 107 survivors for P363S compared to 161 Arg+ revertants per 107 survivors for dnaN+). Taken together, these results clearly demonstrate that P363S is impaired for pol V-dependent SOS mutagenesis in vivo.

Proficiency of β clamps bearing amino acid substitutions at position H148 through R152 in pol III-dependent DNA replication and in pol V-dependent TLS in vivo

Because of the fact that D150N was unable to efficiently complement the temperature sensitive growth phenotype of the dnaN159 mutant strain at 42°C when expressed from either pWSK29 (Table 3) or pSU38 (Table 4), we were forced to use a different approach to examine its proficiency in pol V-dependent SOS mutagenesis. Although many dnaN159 mutant strains are able to grow at 37°C, an effect on growth depending on their sulA genotype has been noted (B. Michel, pers. comm. and data not shown), presumably because of the fact that dnaN159 mutants are chronically induced for the global SOS response, and hence may be susceptible to sulA-dependent lethal filamentation when grown at 37°C (Sutton, 2004). Although the dnaN159 mutant strain MS101 grows well at temperatures up to 37°C, MS120 was temperature sensitive at 37°C (Fig. 4C): this difference likely relates to the fact that MS120 is sulA+ while MS101 is sulA211 (see Table 5). Importantly, our finding that D150N could complement the temperature sensitive growth phenotype of MS120 at 37°C (Fig. 4C) provided us with a method by which we could analyse its ability to participate in pol V-dependent SOS mutagenesis.

Figure 4.

Ability of different D150 loop mutations to complement the temperature sensitive growth phenotype of the dnaN159 mutant strain.
A. Amino acid sequence of the D150 loop region in the β clamp. Secondary structural elements (α1, loop and β2 of domain II of the β clamp) based on the crystal structure of β solved by Kong et al. (1992) are indicated. Also indicated are the amino acid changes resulting in each D150 loop mutation. Positions D150 and G157, which were identified previously by the D150N and G157S mutations (Sutton et al., 2001), are labelled.
B. Amino acid positions within the structure of the β clamp that when altered impair pol V-dependent TLS are indicated. One β clamp protomer is in grey, and the other is in black. Residues mutated in β159 (G66 and G174) are indicated as space-filled molecules. Also indicated are positions P363, and residues 148–152 (within the D150 loop region). The hydrophobic cleft that many of the known clamp partners interact with is located between P363 and the D150 loop region. These structure figures were generated using RasMac v2.6 (http://www.umass.edu/microbio/rasmol/getras.htm#rasmac) and the co-ordinates for β (2POL) from the PDB.
C. Fresh overnight cultures of strain MS120 (relevant genotype: dnaN159) bearing the different plasmid-encoded D150 loop mutations grown at 30°C in M9 minimal medium supplemented with 0.2% glucose, 5 µg ml−1 thiamine, 0.4% Casamino acids and 60 µg ml−1 kanamycin were diluted 1:500 into the same prewarmed medium, and growth at 37°C was measured as a function of time by monitoring the change in optical density at 595 nm. Symbols are as follows: pSU38, ◆; pACYCdnaN+ (dnaN+), ○; pACYCβ148 (H148A), ▵; pACYCβ150 (D150N), ▴; pACYCβ152 (R152A), ◊; pACYCβ3A (β-149–151), □; and pACYCβ5A (β-148–152), ▪. Shown is the average of two independent experiments ± the range.

Table 5.  E. coli strains and plasmid DNAs used in this study.
  • a. 

    See Experimental procedures for further description of strain and plasmid constructions.

  • b. 

    The dnaN159 allele was originally named dnaN59 (Yuasa and Sakakibara, 1980; Ohmori et al., 1984), but has since been renamed dnaN159 by the E. coli Genetic Stock Centre (Grompone et al., 2002; Sutton, 2004).

  • c. 

    The Δ(dinB-yafN)::kan allele was constructed by Kim and colleagues (Kim et al., 1997), and was originally named ΔdinB::kan. However, McKenzie and colleagues (McKenzie et al., 2003) have recently referred to it as Δ(dinB-yafN )::kan to distinguish it from other dinB alleles that do not exert a polar effect on yafN. For consistency, we have used the nomenclature suggested by McKenzie and colleagues in this report.

E. coli strains
Strain Genotype Source or construction a
DH5α endA1 hsdR17 (rKmK+) glnV44 thi-1 recA1 gyrA relA Δ(lacZYA-argF)U169  deoR (φ80dlacΔ[lacZ]M15)Laboratory stock
RW118 thr-1 araD139 Δ(gpt-proA)62 lacY1 tsx-33 supE44 galK2 hisG4(Oc)  rpsL31 xyl-5 mtl-1 argE3(Oc) thi-1 sulA211Laboratory stock (Ho et al., 1993)
MS100RW118; dnaN+tnaA300::Tn10Laboratory stock (Sutton, 2004)
MS101bRW118; dnaN159 tnaA300::Tn10Laboratory stock (Sutton, 2004)
MS106MS100; ΔuvrB::catLaboratory stock (Sutton, 2004)
MS107MS101; ΔuvrB::catLaboratory stock (Sutton, 2004)
AB1157 xyl-5 mtl-1 galK2 rpsL31 kdgK51 lacY1 tsx-33 supE44 thi-1 leuB6
 hisG4(Oc) mgl-51 argE3(Oc) rfbD1 proA2 ara-14 thr-1 qsr-9 qin-111
Laboratory stock
MS119AB1157; dnaN +tnaA300::Tn10P1(MS100) × AB1157
MS120AB1157; dnaN159 tnaA300::Tn10P1(MS101) × AB1157
RW120RW118; ΔumuDC595::catLaboratory stock (Ho et al., 1993)
MS122MS120; ΔumuDC595::catP1(RW120) × MS120
FC1237c ara Δ(lac-proB) Δ(dinB-yafN)::kan Laboratory stock
MS123MS120; Δ(dinB-yafN)::kanP1(FC1237) × MS120
MS124MS122; Δ(dinB-yafN)::kanP1(FC1237) × MS122
Plasmid DNAs   
Plasmid Characteristics Source or construction a
pWSK29ApR; pSC101 origin, general cloning vectorLaboratory stock (Wang and Kushner, 1991)
pJD100ApR; pWSK29 derivative that bears dnaN+ on an EcoRI-SalI fragment  downstream of PlacLaboratory stock (Sutton, 2004)
pJD101ApR; as pJD100, but bears Q61KThis work
pJD102ApR; as pJD100, but bears S107LThis work
pJD103ApR; as pJD100, but bears D150NThis work
pJD104ApR; as pJD100, but bears G157SThis work
pJD105ApR; as pJD100, but bears V170MThis work
pJD106ApR; as pJD100, but bears E202KThis work
pJD107ApR; as pJD100, but bears M204KThis work
pJD108ApR; as pJD100, but bears P363SThis work
pJD109ApR; as pJD100, but bears dnaN159 (G66E-G174A)This work
pSU38KmR; p15A origin, general cloning vectorLaboratory stock (Bartolome et al., 1991)
pSU38dnaN+KmR; pSU38 derivative that bears dnaN + coding sequence on a  BglII-SalI fragment downstream of PlacThis work
pACYCdnaN +KmR; pSU38dnaN+ derivative that bears the complete dnaN + gene,  including all known promoters, on an EcoRI-SalI fragmentThis work
pACYCβ61KmR; as pACYCdnaN+, but bears Q61KThis work
pACYCβ107KmR; as pACYCdnaN+, but bears S107LThis work
pACYCβ157KmR; as pACYCdnaN+, but bears G157SThis work
pACYCβ170KmR; as pACYCdnaN+, but bears V170MThis work
pACYCβ202KmR; as pACYCdnaN+, but bears E202KThis work
pACYCβ204KmR; as pACYCdnaN+, but bears M204KThis work
pACYCβ363KmR; as pACYCdnaN+, but bears P363SThis work
pACYCβ148KmR; as pACYCdnaN+, but bears H148AThis work
pACYCβ150KmR; as pACYCdnaN+, but bears D150NThis work
pACYCβ152KmR; as pACYCdnaN+, but bears R152AThis work
pACYCβ3AKmR; as pACYCdnaN+, but bears β-149–151 (G149A, D150A and V151A  substitutions)This work
pACYCβ5AKmR; as pACYCdnaN+, but bears β-148–152 (H148A, G149A, D150A,  V151A and R152A substitutions)This work
pACMdnaN+CmR; pACYCdnaN+ derivativeThis work
pACMβ148CmR; as pACMdnaN+, but bears H148AThis work
pACMβ150CmR; as pACMdnaN+, but bears D150NThis work
pACMβ152CmR; as pACMdnaN+, but bears R152AThis work
pACMβ3ACmR; as pACMdnaN+, but bears β-149–151 (G149A, D150A and V151A  substitutions)This work
pACMβ5ACmR; as pACMdnaN+, but bears β-148–152 (H148A, G149A, D150A,  V151A and R152A substitutions)This work
pAMPdnaN+ApR; pACYCdnaN+ derivativeThis work
pAMPβ5AApR; as pAMPdnaN+, but bears β-148–152 (H148A, G149A, D150A,  V151A and R152A substitutions)This work

Based on results of modified argE3(Oc)→Arg+ and hisG4(Oc)→His+ reversion assays (Fig. 5), D150N was indistinguishable from dnaN+ with respect to pol V-dependent SOS mutagenesis (Fig. 5). Thus, in summary, seven of the eight mutant clamp proteins identified previously based on their impaired ability to interact genetically with the different umuDC gene products retained an ability to promote SOS mutagenesis in vivo (see Figs 3 and 5). We reasoned that one explanation for these results could be that the genetic assay used to identify these dnaN alleles was specific for identifying mutants of the clamp impaired for the checkpoint function of umuDC. Alternatively, positions of amino acid substitutions identified in these mutant clamps might represent sites in β involved in interactions with the different umuDC gene products, but the effects conferred by the amino acid changes might be too modest to impair SOS mutagenesis in a significant way under near-physiological conditions. Consistent with the latter conclusion, it was recently suggested that ∼75% of the binding energy between the β clamp and the α catalytic subunit of pol III was attributable to interactions involving the eubacterial clamp binding motif located at the C-terminal end of α, and the hydrophobic cleft of β (Lopez de Saro et al., 2003).

Figure 5.

Proficiency of D150 loop mutations in SOS mutagenesis. Respective proficiencies in SOS mutagenesis of representative transformants of strain MS120 (relevant genotype: dnaN159) bearing the indicated pSU38-encoded dnaN + or D150 loop alleles were measured at 37°C using the modified argE3(Oc)→Arg+ (black bars) or modified hisG4(Oc)→His+ reversion assay (grey bars) as described in Experimental procedures. UV mutability is expressed relative to the strain bearing the dnaN +-expressing plasmid pACYCdnaN + (34, 68 or 44 induced Arg+ revertants per 107 survivors, and 39, 54 or 64 His+ induced revertants per 107 survivors). Shown is the average of duplicates from three independent experiments ± the standard deviation. Using this assay, strain RW118 (relevant genotype: dnaN +umuD +C +) produced 101 Arg+ revertants per 107 survivors, while strain RW120 (relevant genotype: dnaN +ΔumuDC595::cat) produced 4 Arg+ revertants per 107 survivors when measured at 37°C (data not shown), indicating that reversion in this assay required umuDC-function. Based on the two sample T-test, the difference between the means of the dnaN + and the β-148–152 strains for both argE3(Oc)→Arg+ (t-value = 4.82, P = 0.0085) and hisG4(Oc)→His+ reversion (t-value = 7.38, P = 0.0018) were statistically significant at the 95% confidence interval.

In order to distinguish between these two models, as well as to establish the importance of the D150 loop in pol III-dependent replication and in pol V-dependent SOS mutagenesis, we targeted positions H148 through R152 in various combinations by site directed mutagenesis under the assumption that clustered amino acid substitutions would display a more severe phenotype than single amino acid substitutions (see Fig. 4A and B). We chose to focus on this region, which we will refer to as the D150 loop region, for the following reasons: (i) based on the crystal structure of the β clamp, the D150N mutation is one of only a few of the substitutions identified that is located within a solvent exposed loop (see Fig. 4A), suggesting that amino acid substitutions in this region will not alter the overall structure of the clamp; (ii) of eight dnaN mutations identified using the genetic assay, only D150N and G157S, which is also located within the loop containing D150, were identified more than once each (D150N was identified two independent times, while G157S was identified five independent times; Sutton et al., 2001), suggesting that the region surrounding these residues is particularly important for interaction with the different umuDC gene products and (iii) various lines of evidence suggest that residues within the loop encompassing D150 are important for interaction with both the α catalytic subunit of pol III (Duzen et al., 2004) and the δ subunit of the γ clamp loader complex (Jeruzalmi et al., 2001).

Based on quantitative Western blot analysis, the steady-state level of clamp in each D150 loop derivative was roughly two- to threefold higher than normal physiological levels (Table 2). Our finding that R152A, β-149–151 and β-148–152 were each unable to complement the dnaN159 temperature sensitive growth phenotype at 42°C (Table 4) suggested that positions G149 through R152 of β are important for interaction with one or more components of pol III. However, all five D150 loop mutations were able to complement the temperature sensitivity of the MS120 strain at 37°C, as indicated by their efficient growth in liquid culture at this temperature (Fig. 4C). By comparison, the dnaN159 strain bearing the empty pSU38 vector was unable to grow under these conditions (Fig. 4C). Thus, despite the fact that residues H148 through R152 of β are important for pol III-dependent replication (Table 4), each of the D150 loop mutations retained function for normal DNA replication in vivo at 37°C.

In order to establish whether residues of the clamp in the vicinity of D150 are important for pol V-dependent TLS, we measured the efficiency of SOS mutagenesis at 37°C in representative transformants of strain MS120 bearing each of the different D150 loop mutations. As noted above, D150N was indistinguishable from wild-type β with respect to SOS mutagenesis using a modified version of the argE3(Oc)→Arg+ reversion assay (Fig. 5). Likewise, β-149–151 was similarly proficient for SOS mutagenesis with this assay. However, in striking contrast to these results, the β-148–152 mutant was severely impaired for SOS mutagenesis. Neither the H148A nor the R152A single mutations alone were impaired for SOS mutagenesis (Fig. 5), suggesting that the SOS defect required all five residues to be changed. Essentially identical results were observed with each of the D150 loop mutations using an analogous hisG4(Oc)→His+ reversion assay (Fig. 5). Taken together, these results indicate that residues 148–152 of the β clamp are critically important for pol V-dependent TLS, and are consistent with the idea that the genetic assay did in fact identify positions in β involved in interactions with the different umuDC gene products.

The β-148–152 mutation confers a UV sensitive phenotype that is epistatic withΔumuDC595::cat

A ΔumuDC595::cat mutant is roughly 10-fold more sensitive to UV irradiation than an isogenic umuD +C + strain (Friedberg et al., 1995). UV sensitivity appears to be the result of both the loss of the checkpoint and the TLS functions of the different umuDC gene products (Opperman et al., 1999 and reviewed in Sutton et al., 2000). We predicted that, as a result of their impaired ability to carry out pol V-dependent SOS mutagenesis, β-148–152 and possibly P363S would display a UV sensitive phenotype that is epistatic with ΔumuDC595::cat. To investigate this, we first examined UV sensitivity of the various dnaN mutations using a umuD +C + strain. Sensitivity to UV irradiation at doses ranging from 0 to 7 J m−2 of representative transformants of strain MS107 bearing each of the seven dnaN alleles expressed from pWSK29 that were proficient for growth at 42°C (Q61K, S107L, G157S, V170M, E202K, M204K and P363S) were comparable to the dnaN+ strain (data not shown). Similar results were observed using strain MS120 together with the pSU38 constructs (data not shown). The lack of UV sensitivity observed for P363S is presumably because of the fact that this mutant retains partial function for pol V-dependent SOS mutagenesis (see Fig. 3).

In striking contrast to these results, the strain expressing D150N displayed a severe sensitivity to UV light at 37°C (Fig. 6A). The fact that β-149–151 and β-148–152 were similarly UV sensitive suggests that sensitivity of these mutants is largely attributable to the substitution at position D150. Consistent with this conclusion, H148A was indistinguishable from dnaN+, and R152A was only slightly UV sensitive (Fig. 6A). UV sensitivity of D150N, R152A and β-149–151 was unexpected because these mutants were proficient for pol V-dependent SOS mutagenesis. Moreover, our finding that UV sensitivity of the different D150 loop mutants was not epistatic with ΔumuDC595::cat (data not shown) indicated that sensitivity was not attributable to a umuDC-dependent function.

Figure 6.

Effect of D150 loop mutations on UV sensitivity of E. coli. Respective sensitivities to UV irradiation of representative transformants of strains MS120 (relevant genotype: dnaN159) bearing pACYCdnaN + (dnaN +), pACYCβ148 (H148A), pACYCβ150 (D150N), pACYCβ152 (R152A), pACYCβ3A (β-149–151) or pACYCβ5A (β-148–152) (A), MS123 [relevant genotype: dnaN159Δ(dinB-yafN)::kan] bearing pACMdnaN + (dnaN +), pACMβ148 (H148A), pACMβ150 (D150N), pACMβ152 (R152A), pACMβ3A (β-149–151) or pACMβ5A (β-148–152) (B) and MS124 [relevant genotype: dnaN159Δ(dinB-yafN)::kanΔumuDC596::cat] bearing pAMPdnaN + or pAMPβ5A (C) were measured as described previously (Sutton, 2004). Results shown are averages of duplicates from two independent experiments ± the range. The inset in (B) is the same as shown in the main panel, but lacks the data for β mutants H148A, D150N, R152A, and β-149–151. The dashed lines in (C) are the same as shown in (B). Survival values for dnaN + and β-148–152 in the dnaN159Δ(dinB-yafN)::kan strain (MS123) are included in (C) for direct comparison with values for these same β mutations in the dnaN159Δ(dinB-yafN)::kanΔumuDC595::cat strain (MS124). Based on the two sample T-test, the differences between the means of the dnaN+ and the β-148–152 strains in the dnaN159Δ(dinB-yafN)::kan strain (MS123) were statistically significant at the 95% confidence level (t-value = 6.14, P = 0.025 for 50 J m−2, and t-value = 6.82, P = 0.021 for 100 J m−2), while those for the dnaN159Δ(dinB-yafN)::kanΔumuDC595::cat strain (MS124) were not (t-value = 1.90, P = 0.20 for 50 J m−2).

We recently determined that a dnaN159 mutant strain was sensitive to UV light, and that this sensitivity was suppressed (not epistatic with) by inactivation of pol IV [Δ(dinB-yafN)::kan] (Sutton, 2004). We therefore measured UV sensitivity of the D150 loop mutations using a Δ(dinB-yafN)::kan derivative of MS120 (Fig. 6B). With the exception of β-148–152, which retained UV sensitivity in this strain, sensitivity of the other D150 loop mutants was completely suppressed by (and not epistatic with) inactivation of pol IV [Δ(dinB-yafN)::kan]. We hypothesized that the approximately fivefold sensitivity of β-148–152 was attributable to its pol V-dependent SOS mutagenesis defect. Further analysis using a Δ(dinB-yafN)::kanΔumuDC595::cat derivative of MS120 confirmed that the approximately fivefold sensitization of the β-148–152 mutation to killing by UV in the Δ(dinB-yafN)::kan strain was indeed attributable to a umuDC-dependent function (Fig. 6C). However, despite the fact that UV sensitivity of β-148–152 was epistatic with ΔumuDC595::cat, the extent of β-148–152 UV sensitivity was less severe than that conferred by ΔumuDC595::cat. These results are consistent with the idea that β-148–152 specifically impairs pol V-dependent TLS, and that another function(s) of umuDC that helps to protect E. coli from the lethal effects of UV irradiation, such as the checkpoint, is unaffected by alanine substitution of residues 148–152.

Discussion

Results discussed in this report demonstrate that the β clamp plays an essential role in pol V-dependent TLS, and establish that mutant forms of the clamp distinguish between its roles in pol III-dependent DNA replication and pol V-dependent TLS: not only did we determine that the dnaN159 mutant strain is conditionally impaired for SOS mutagenesis, but we further established that residues H148 through R152, as well as P363 of the β clamp are critically important for proper pol V function in vivo. Although it is currently unclear whether these residues in β interact directly with pol V, or whether the amino acid substitutions at these positions affect the structure of the clamp, thereby impairing interactions with pol V, our findings nonetheless suggest that residues of the clamp in addition to those comprising the hydrophobic cleft are essential for interactions of the clamp with pol V, as well as possibly other partner proteins. Although our findings are consistent with a model in which the β clamp helps to manage the actions of the different E. coli DNA polymerases by forcing them to compete with each other for access to the replication fork, it remains to be determined whether access of pol V to the primer-template junction is additionally regulated by ‘active’ and ‘inactive’ conformations of the β clamp-pol V complex, as has been suggested to occur with the related E. coli Y family polymerase, pol IV (Bunting et al., 2003).

We previously characterized the abilities of many of the mutant β clamp proteins discussed above to interact with UmuD, UmuD′ and the α catalytic subunit of pol III in vitro (Duzen et al., 2004; see Table 3). This work indicated that P363S was impaired for direct interaction with the UmuD′ subunit of pol V in vitro (Duzen et al., 2004). Thus, the inability of the P363S mutant to carry out SOS mutagenesis is seemingly a result of an altered interaction of the mutant P363S clamp with pol V, consistent with the notion that the previously described UmuD′-β interaction is biologically important (Sutton et al., 1999; 2002). In contrast, D150N was proficient for interaction with UmuD′ (Duzen et al., 2004). However, the TLS defect of the mutant β-148–152 clamp suggests that the D150 loop region is critically important for interaction with the UmuD′ and/or UmuC subunit(s) of pol V.

S107L and G157S were found to be marginally impaired for interaction with UmuD′ (Duzen et al., 2004). The fact that these mutant clamp proteins retain the ability to participate in pol V-dependent TLS when expressed at roughly two- to threefold higher than normal physiological levels suggests that they are able to interact productively with pol V in vivo. The remaining five mutant β clamp proteins that we had previously identified using a genetic assay (Q61K, D150N, V170M, E202K and M204K) were able to interact with UmuD′in vitro (Duzen et al., 2004), and as discussed above, were proficient in pol V-dependent TLS in vivo. It is formally possible that all seven of these mutant clamp proteins are impaired for TLS, but that β159 somehow complements their defect (i.e. via formation of heterodimers bearing one β159 protomer and one plasmid-encoded mutant clamp protomer). However, our finding that the P363S mutant is able to fully complement the temperature sensitive growth phenotype of the dnaN159 mutant, yet is nonetheless impaired for TLS, argues against this model. Given that multiple protein–protein and protein–nucleic acid interactions are required for pol V-dependent TLS, it is perhaps not surprising that the genetic assay might have successfully identified mutant forms of the clamp impaired for interactions with pol V under the artificial conditions of the genetic selection used (Sutton et al., 2001), but that despite this, seven of the eight mutant clamps identified retain sufficient affinity for pol V to promote TLS under physiological conditions. Our results concerning the D150 loop mutations are consistent with this conclusion (Figs 4 and 5). Alternatively, the mutant clamps that were proficient for pol V-dependent TLS may be impaired for the checkpoint function of umuDC. However, our findings that: (i) none of these mutant dnaN alleles conferred a significant UV sensitivity and (ii) that β-148–152 (and D150N) may be proficient for the checkpoint (Fig. 6), are inconsistent with this model. Detailed characterization of these mutants with respect to replication following UV irradiation is necessary in order to resolve this question.

Based on Superose-12 gel filtration chromatography, the P363S substitution impaired interaction of the clamp with the α catalytic subunit of pol III in vitro (Duzen et al., 2004). However, our finding that near-physiological levels of P363S complemented the temperature sensitive growth phenotype of the dnaN159 mutant indicates that the mutant P363S clamp protein is able to interact productively with both the α catalytic subunit of pol III, as well as the γ clamp loader complex in vivo. The D150N protein was only modestly impaired for interaction with αin vitro (Duzen et al., 2004). Thus, the inability of near-physiological levels of the D150N clamp protein to fully complement the temperature sensitive growth phenotype of the dnaN159 mutant was unexpected. This, taken together with our finding that the β-149–151, the R152A and the β-148–152 mutations were unable to completely suppress the temperature sensitive growth phenotype of the dnaN159 mutant at 42°C when expressed at higher than physiological levels indicates that the D150 loop region of the clamp performs at least one essential function in vivo. R152 of β reportedly forms a salt bridge with E48 of the δ subunit of clamp loader (Jeruzalmi et al., 2001). Thus, temperature sensitivity of the various D150 loop mutations might be resulting from an impaired ability of these mutant clamps to interact with clamp loader. However, given that clamps must be loaded onto the DNA in order to participate in pol III-dependent DNA replication, the D150 loop mutations must nevertheless retain an ability to interact with clamp loader under the conditions in which they were analysed in this study.

Positions H148 through R152 are located adjacent to the hydrophobic cleft in β that is proposed to interact with the eubacterial clamp-binding motif (see Fig. 4B). Position P363 is located within the C-terminal tail of the clamp, which is located on the far side of the cleft relative to the D150 loop. The UmuC subunit of pol V bears a reasonable match to the eubacterial clamp-binding motif that is suggested to interact with the cleft (Dalrymple et al., 2001). Thus, it is possible that substitutions in the D150 loop and at P363 may alter the structure of the hydrophobic cleft, thereby perturbing interaction of the clamp with pol V to impair TLS. However, in this case, these mutations would be predicted to also impair interactions involving the clamp and other partner proteins that bind the cleft, including the α catalytic subunit of pol III and the δ subunit of the γ clamp loader complex (Naktinis et al., 1996). Thus, although an effect of the D150 loop region mutations on the structure of the cleft is consistent with our findings that R152A, β-149–151 and β-148–152 were unable to fully complement the temperature sensitive growth phenotype of the dnaN159 mutant, it does not fully explain why the β-148–152 and P363S mutations were more severely impaired for pol V-dependent TLS than they were for normal growth. Consequently, our results suggest that the β-148–152 and P363S mutations do not significantly alter the overall structure of the hydrophobic cleft, or the clamp loading process in vivo, but instead represent separation of function alleles that distinguish between the roles of β in pol V-dependent TLS and pol III-dependent replication.

Our results indicate that, in addition to their affects on pol III and pol V (and possibly the clamp loader), the D150 loop mutations also confer a UV sensitive phenotype that is dependent on the dinB-encoded pol IV. This finding suggests that the D150 loop is not important for interaction of the clamp with pol IV, and that as a result, pol IV is able to compete effectively with pol III for interaction with the clamp and subsequent access to the replication fork. The available structural information regarding the β clamp-pol IV complex is consistent with this conclusion (see Fig. 4B): although residues H148-R152 do not appear to make direct contact with pol IV in the proposed replication competent (i.e. ‘active’) conformation, residues adjacent to these do (Bunting et al., 2003; Burnouf et al., 2004). Finally, although UV sensitivity of β-148–152 was epistatic with ΔumuDC595::cat, the extent of the UV sensitivity conferred by β-148–152 was less severe than that conferred by ΔumuDC595::cat, consistent with the idea that β-148–152 is impaired for pol V-dependent TLS, but nonetheless retains an ability to protect E. coli from the lethal effects of UV irradiation in a umuDC-dependent manner. Further work is required in order to determine if the UV sensitive phenotype of β-148–152 is related to the checkpoint function of umuDC.

In summary, our results suggest important differences in how the β clamp interacts with pol III, pol IV and the different umuDC gene products. We suggest that these differences serve an important role in co-ordinating the actions of these polymerases during DNA replication and TLS in vivo. Through use of a combination of common and unique binding sites on the surface of the clamp, the cell could impose a hierarchy on the different polymerases such that only certain partners are able to replace particular resident partners on a clamp, thereby helping to control both the order and the nature of the events that occur at a replication fork. Further biochemical characterization of P363S and the various D150 loop mutant β clamp proteins, in particular β-148–152, should lead to a more complete understanding of the mechanisms responsible for co-ordinating pol III-dependent DNA replication with pol IV- and pol V-dependent TLS, as well as fundamental aspects of polymerase switching in general.

Experimental procedures

Escherichia coli strains and bacteriological techniques

Escherichia coli strains were routinely grown in LB medium (Miller, 1992). IPTG (50 µM) was added to media in which strains bearing the pWSK29-encoded dnaN alleles were grown. When necessary the following antibiotics were used at the indicated concentrations: ampicillin (Ap), 150 µg ml−1; chloramphenicol (Cm), 20 µg ml−1; kanamycin (Km), 60 µg ml−1; tetracycline (Tc), 12 µg ml−1; and rifampicin (Rif), 50 µg ml−1.

Escherichia coli strains used in this study are described in Table 5. All strains are derivatives of E. coli K-12, and were constructed using P1vir-mediated generalized transduction (Miller, 1992). Strains MS119 and MS120 were constructed by selection for tetracycline resistance conferred by tnaA300::Tn10, which is ∼90% linked to dnaN (or dnaN159). Colony polymerase chain reaction (PCR) and automated nucleotide sequence analysis (RPCI Biopolymer facility, Buffalo NY) were used to confirm that the dnaN159 allele in MS120 contained both the G66E and G174A substitutions, as described previously (Sutton, 2004). Strain MS122 was constructed by selection for chloramphenicol resistance. Successful transduction of the ΔumuDC595::cat allele was confirmed by screening for sensitivity to UV light, and non-mutability (data not shown). Strains MS123 and MS124 were constructed by selection for kanamycin resistance. The presence of the Δ(dinB-yafN)::kan allele was confirmed by colony PCR using primers specific to the region extending from upstream of dinB through yafN (data not shown).

Plasmids

Plasmid DNAs are described in Table 5. pJD100 is a pWSK29 derivative, and has been described previously (Sutton, 2004). Plasmids pJD101 through pJD108 are pJD100 derivatives that bear the indicated mutant dnaN alleles. They were constructed by replacing the 1.1 kb XhoI-SalI dnaN-containing fragment of pJD100 with the corresponding XhoI-SalI fragment bearing the indicated dnaN allele obtained from the respective pJR210 derivatives described previously (Sutton et al., 2001). Plasmid pJD109 bears the dnaN159 allele and was cloned using the Quickchange kit (Stratagene) exactly as per the manufacturer's recommendations. Plasmid pJD100 was used as template in Quickchange PCR reactions, together with the following primers: DnaN G66E top, 5′-CAGCCACACGAGCCAGAAGCGACGACCGT TCCGG-3′; DnaN G66E bottom, 5′-CCGGAACGGTCGTCG CTTCTGGCTCGTGTGGCTG-3′; DnaN G174A top, 5′-GCAACCGA CGCCCACCGTCTG-3′; and DnaN G174A bottom, 5′-CAGACGGTGGGCGTCGGTTGC-3′. Bacterial transformation of DpnI-treated PCR products was performed using calcium chloride treated DH5α as described previously (Sutton and Kaguni, 1995). The complete nucleotide sequence of the dnaN159 allele in pJD109 was confirmed by automated nucleotide sequence analysis (RPCI Biopolymer facility, Buffalo NY) as described (Sutton et al., 2001; Sutton, 2004).

pACYCdnaN + is a pSU38 derivative and was constructed in two steps. First, the 1.2 kb BglII-SalI fragment from pJRC210 was subcloned into BamHI-SalI digested and calf intestinal alkaline phosphatase-treated pSU38 to generate an intermediate construct called pSU38dnaN +. The 1.4 kb EcoRI-XhoI fragment from pACYCdnaA+ (Sutton and Kaguni, 1995), which bears the coding sequence for the N-terminal 49 residues of β, as well as all known dnaN promoters, was then subcloned into the similarly digested and calf intestinal alkaline phosphatase-treated pSU38dnaN + plasmid to yield pACYCdnaN +.

The eight pSU38-derivatives bearing the different dnaN alleles impaired for genetic interactions with the different umuDC gene products were constructed by ligating the EcoRI-SalI fragment from the corresponding pWSK29-derivaitves (pJD101-pJD108) into pSU38. pACYCβ148, pACYCβ152, pACYCβ3A and pACYCβ5A were generated using the Quickchange kit (Stratagene) exactly as per the manufacturer's recommendations. Plasmid pACYCdnaN + was used as template in Quickchange PCR reactions, together with the following primer pairs: H148A top, 5′-CAGTTTTCTATGGCGGCTCAGGACGTTCGCT ATTACTTAA ATGG-3′, and H148A bottom, 5′-CCATTTAAGTAATAGCG AACGTCCTGAGCCGCCATAGAAAACTG-3′; R152A top, 5′-CAGTTTTCTATGGCGCATCAGGACGTTGCCTA TTACTT AA ATGG-3′, and R152A bottom, 5′-CCATTTAAGTAATAGGCAA CGTCCTGATGCGCCATAGAAAACTG-3′; 149–151 top, 5′-CAGTTTTCT A TGGCGCATGCGGCCGCTCGCT ATTACTTA AATGG-3′, and 149–151 bottom, 5′-CCATTTAAGTAATAGCG AGCGGCCGCATGCGCCATAGAAAACTG-3′; and 148–152 top, 5′-CTATGGCGGCTGCGGCCGCTGCCTATTACTTAAAT GG-3′, and 148–152 bottom, 5′-CCATTTAAGTAATAGGCAG CGGCCGCAGCCGCCATAG-3′. Bacterial transformation of DpnI-treated PCR products was performed using calcium chloride treated DH5α as described previously (Sutton and Kaguni, 1995). The complete nucleotide sequence of each dnaN allele was confirmed by automated nucleotide sequence analysis (RPCI Biopolymer facility, Buffalo, NY).

Because certain alleles used in some strain constructions conferred resistance to the selectable markers present on the plasmids expressing the D150 loop mutations, derivatives of these plasmids expressing chloramphenicol or ampicillin resistance were constructed. Plasmid pACMdnaN +, and its derivatives (see Table 5), confer resistance to chloramphenicol, and were constructed by subcloning a PCR-generated DNA fragment containing the cat gene from plasmid pKO3 (Link et al., 1997) into the unique BclI site present within the kan gene of pACYCdnaN +, and its derivatives. Plasmids pAMPdnaN + and pAMPβ5A confer ampicillin resistance and were constructed in a similar fashion by subcloning PCR-generated DNA fragment containing the bla gene from plasmid pET16b (Novagen) into the unique BclI site of pACYCdnaN + and pACYCβ5A respectively.

Quantitative Western blot analysis

Strains were grown in LB medium supplemented either with both ampicillin and 50 µM IPTG (for pWSK29-bearing strains) or with kanamycin (for pSU38-bearing strains). Based on plating experiments, the number of viable cells per 1 ml of sample equivalent to OD595 = 1.0 were determined (data not shown). Cells from exponentially growing cultures (OD595∼0.5–0.6) were harvested by centrifugation, resuspended in 50 µl of 0.8% saline, and lysed by addition of an equal volume of 4× SDS-PAGE loading buffer [200 mM Tris-HCL (pH 6.8), 8% SDS, 0.4% bromophenyl blue, 40% glycerol] containing 10%β-mercaptoethanol. A 10 µl aliquot of each sample was electrophoresed through 12% SDS-PAGE, transferred to PVDF membrane (Millipore), and probed with polyclonal antibodies specific to β as described previously (Sutton et al., 2001). Detection of immunoreactive material was achieved using Super Signal Western Dura Extended Chemilumenescence substrate (Pierce). The level of β clamp in each sample (calculated as dimers cell−1) was calculated relative to a standard curve of purified β, the concentration of which was determined by Bradford analysis (Pierce), using the Molecular Analyst software (Bio-Rad) as described previously (Sutton et al., 1999). The number of β clamps per cell for different strains expressing each mutant clamp protein was determined using the following formula:

image

where B = number of β clamp dimers per cell, M = grams of β clamp in the whole cell lysate, N = number of cells (based on plating experiments; data not shown), A = Avogadro's number (6.023 × 1023 mol cell−1), and W = molecular mass of the β clamp, as dimer (81.2 kDa). Values reported for the steady-state level of each plasmid expressed clamp protein represent the average of two independent whole cell lysates, each run on an independent SDS-PAGE/blot with a corresponding standard curve.

SOS mutagenesis assays

For SOS assays utilizing strains MS101 and MS107 bearing the pWSK29-derived plasmids expressing the different dnaN alleles, media was supplemented with 50 µM IPTG: IPTG was not added to cultures of strain MS120 bearing the pSU38-derived plasmids. Proficiency in SOS mutagenesis was measured using either the argE3(Oc)→Arg+ reversion assay (Walker, 1977) or a rifampicin resistance assay. For the latter, overnight cultures grown in M9 minimal medium (Miller, 1992) supplemented with 0.2% glucose, 5 µg ml−1 thiamine and 0.4% Casamino acids were subcultured in the same medium and grown at 30, 37 or 42°C to mid-exponential phase (OD595∼0.5–0.6). Following UV irradiation, 1 ml of each culture, as well as 1 ml of each mock-UV treated culture, was mixed with 9 ml of prewarmed supplemented M9 medium. After overnight growth at 30, 37 and/or 42°C as indicated in the legends to the figures, 100 µl of each culture was plated in duplicate onto solid LB medium supplemented with 50 µg ml−1 rifampicin (Sigma). The number of viable cells in each culture was determined by plating appropriate serial dilutions on solid LB medium followed by overnight incubation at the same temperature. SOS mutagenesis was calculated by subtracting the number of RifR colonies observed on the mock UV-treated plates from those observed on the UV-treated plates, and dividing by the number of viable cells.

Because of their UV hypersensitivity, proficiency in SOS mutagenesis of strain MS120 bearing the different D150 loop mutations was measured using a modified argE3(Oc)→Arg+ reversion assay. Briefly, this assay was similar to that reported previously (Walker, 1977), except cultures were grown overnight at 37°C following UV irradiation as described above for the rifampicin resistance assay before being plated out onto minimal agar plates lacking arginine. We confirmed that the UV-induced Arg+ reversion measured using this modified assay was dependent on pol V [using this assay, strain RW118 (relevant genotype: dnaN +umuD +C +) produced 101 Arg+ revertants per 107 survivors, while strain RW120 (relevant genotype: dnaN +ΔumuDC595::cat) produced 4 Arg+ revertants per 107 survivors when measured at 37°C]. An analogous assay was also performed in which hisG4(Oc)→His+ reversion was measured using supplemented minimal media lacking histidine. This method for measuring Arg+ or His+ reversion is referred to as the ‘modified reversion’ assay in this report.

Acknowledgements

The authors wish to thank the three expert reviewers for their insightful comments, Dr Kenneth Blumenthal (University at Buffalo) for suggestions on the manuscript, as well as members of our lab for helpful discussions. This work was supported by Public Service Health Grant GM066094 (M.D.S.), Howard Hughes Medical Institute Biomedical Research Support Program Grant 53000261 to the School of Medicine and Biomedical Sciences, University at Buffalo, SUNY, and start-up funds from the School of Medicine and Biomedical Sciences, University at Buffalo, SUNY (M.D.S.).

Ancillary