A ΔdinB mutation that sensitizes Escherichia coli to the lethal effects of UV and X-radiation

Abstract

The DinB (PolIV) protein of Escherichia coli participates in several cellular functions. We investigated a dinB mutation, Δ(dinB-yafN)883(::kan) [referred to as ΔdinB883], which strongly sensitized E. coli cells to both UV- and X-radiation killing. Earlier reports indicated dinB mutations had no obvious effect on UV radiation sensitivity which we confirmed by showing that normal UV radiation sensitivity is conferred by the ΔdinB749 allele. Compared to a wild-type strain, the ΔdinB883 mutant was most sensitive (160-fold) in early to mid-logarithmic growth phase and much less sensitive (twofold) in late log or stationary phases, thus showing a growth phase-dependence for UV radiation sensitivity. This sensitizing effect of ΔdinB883 is assumed to be completely dependent upon the presence of UmuDC protein; since the ΔdinB883 mutation did not sensitize the ΔumuDC strain to UV radiation killing throughout log phase and early stationary phase growth. The DNA damage checkpoint activity of UmuDC was clearly affected by ΔdinB883 as shown by testing a umuC104 ΔdinB883 double-mutant. The sensitivities of the ΔumuDC strain and the ΔdinB883 ΔumuDC double-mutant strain were significantly greater than for the ΔdinB883 strain, suggesting that the ΔdinB883 allele only partially suppresses UmuDC activity. The ΔdinB883 mutation partially sensitized (fivefold) uvrA and uvrB strains to UV radiation, but did not sensitize a ΔrecA strain. A comparison of the DNA sequences of the ΔdinB883 allele with the sequences of the Δ(dinB-yafN)882(::kan) and ΔdinB749 alleles, which do not sensitize cells to UV radiation, revealed ΔdinB883 is likely a “gain-of-function” mutation. The ΔdinB883 allele encodes the first 54 amino acids of wild-type DinB followed by 29 predicted residues resulting from the continuation of the dinB reading frame into an adjacent insertion fragment. The resulting polypeptide is proposed to interfere directly or indirectly with UmuDC function(s) involved in protecting cells against the lethal effects of radiation.

1. Introduction

Escherichia coli has five DNA polymerases, including two Y-family translesion synthesis (TLS) polymerases, PolIV or DinB, and PolV or UmuD’₂C, that are induced as part of the SOS response [1, 2]. The biological functions of UmuDC proteins include roles in cellular survival after UV radiation-induced DNA damage via TLS [3, 4], in cellular checkpoint activity [5] and in induced and spontaneous mutagenesis [1, 6-10].

DinB protein is involved in TLS past template cytotoxic DNA alkylation lesions and adducts at the N2-position of deoxyguanosine [11, 12], replication arrest-stimulated recombination [13], stress-induced mutagenesis [14, 15], survival under conditions of nucleotide starvation [16], and it may possibly function as a cellular checkpoint by inhibiting replication fork progression [17, 18]. Some studies have suggested a role for DinB in spontaneous mutagenesis in growing cells [19], while others have not found evidence for DinB in this process [20, 21].

During experiments in our laboratory designed to evaluate the effect of DinB on UV radiation mutagenesis, it appeared that the ΔdinB883 mutant strain was significantly more sensitive to the lethal effects of UV radiation than an isogenic dinB⁺ strain, and that the degree of this sensitivity was dependent on culture growth phase. This observation was not consistent with previously reported results for dinB mutants [3, 22], so we initiated a study to specifically address how the ΔdinB883 allele affects UV radiation sensitivity, including its interaction with a ΔumuDC mutation, which knocks out the other E. coli Y-family translesion polymerase, PolV.

2. Materials and methods

2.1. Bacterial strains

The bacterial strains used in this study are listed in Table 1. Strains were constructed by P1 transduction as described [23]. As this project developed, it became apparent that allele numbers would be helpful to our discussion of the three Δ(dinB) mutations involved in this study. The Coli Genetic Stock Center assigned numbers for our purposes, and these numbers are explained further in Table 1.

Table 1.

Escherichia coli K-12 strains used in this study.

Strains	Genotype	Source or derivation
BW25113	Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) rph-1 Δ(rhaD-rhaB)568 hsdR514 F− ! −	Coli Genetic Stock Center
AB1157	argE3(oc) hisG4(oc) leuB6(amIII) Δ(gpt-proA)62 thr-1 thi-1 ara-14 galK2 lacY1 mtl-1 xyl-5 tsx-33 rfbD1 mgl-51 rpsL31 supE44(amSuII) rac F− ! −	Laboratory stock
EST945	As AB1157, but: Δ(recA-srl)306 srlR301::Tn10 λ(recA+)	E Tessman
JW0221-1	As BW25113, but ΔdinB749::kan [referred to as ΔdinB749]	Coli Genetic Stock Center
RW82	ΔumuDC595::cat uvrA6 hisG4(oc) ilv323ts leuB6(amIII) Δ(gpt-proA)62 thr-1 thi-1 araD14 galK2(oc) lacY1 mtl-1 xyl-5 tsx-33 kdgK51 rfbD1 rpsL31 supE44(amSuII)	[34]
SMR4562	Δ(lac-proAB)XIII thi ara Rifr [F′ proAB+ lacI33! lacZ]	SM Rosenberg
SMR6111	As SM4562, but: [F′ Δ(dinB-yafN)882(::kan) proAB+ lacI33! lacZ] Δ(dinB-yafN)882(::kan) [referred to as ΔdinB882]	SM Rosenberg
SR2227	As AB1157, but: trpA78 rpsE (gpt-proA)+ rpsL+	Laboratory stock
SR2626	uvrB5 gal::Tn10 leuB19 metE70 thyA36 deoC2? lacZ53 rha-5 rpsL151 IN(rrnD-rrnE) F− λ−	Laboratory stock
SR4109	As SR2227, but: Δ(dinB-yafN)883(::kan) [referred to as ΔdinB883]	SR2227 × P1vir■YG7207, Kmr
SR4128	As SR2227, but: ΔumuDC595::cat [Referred to as ΔumuDC]	SR2227 × P1■RW82, Cmr
SR4129	As SR2227, but: ΔdinB883 ΔumuDC	SR4109 × P1■ RW82, Cmr
SR4172	As SR2227, but: uvrA::Tn10	SR2227 × P1vir■SY55, Tcr
SR4173	As SR2227, but: uvrA::Tn10 ΔdinB883	SR4109 × P1vir■SY55, Tcr
SR4175	As SR2227, but: uvrB5 gal::Tn10	SR2227 × P1vir■SR2626, Tcr
SR4177	As SR2227, but: uvrB5 gal::Tn10 ΔdinB883	SR4109 × P1vir■SR2626, Tcr
SR4178	As SR2227, but: Δ(recA-srl)306 srlR301::Tn10	SR2227 × P1vir■EST945, Tcr
SR4179	As SR2227, but: Δ(recA-srl)306 srlR301::Tn10 ΔdinB883	SR4109 × P1vir■EST945, Tcr
SR4385	sulA211 thi-1 Δ(lac-gpt)5 ilv(Ts) mtl-1 rpsL31 lexA51(Def) umuC104	DE1868, R Woodgate
SSR4391	AS SR4385, but ΔdinB883	SR4385 × P1vir■SR4109, Kmr
SY55	uvrA::Tn10	BA Bridges
YG7207	As AB1157, but: Δ(dinB-yafN)::kan	[31]

P1vir indicates strain construction by transduction; cat, kan or Tn10 indicate genes conferring resistance to chloramphenicol (Cm^r), kanamycin (Km^r) or tetracycline (Tc^r), respectively; other genetic nomenclature have been defined [43]. To simplify discussion, we asked the Coli Genetic Stock Center to assign allele numbers for two of the Δ(dinB) mutations discussed in this work. The following designations were assigned as requested:, Δ(dinB-yafN)882(::kan) for strain SMR6111and Δ(dinB-yafN)883(::kan) for strain SR4109, and its derivatives. Abbreviated terminology for these alleles is defined in this table.

2.2. Media

LB (Luria-Bertani) broth was 1% tryptone, 0.5% yeast extract and 1% NaCl [23]. Tryptone agar was 1% tryptone, 0.5% NaCl and 1.5% agar. Transductants were selected on tryptone plates supplemented with kanamycin at 50 μg/ml, chloramphenicol at 20 μg/ml, or tetracycline at 15 μg/ml. Saline was 0.85% NaCl. PB was 67 mM NaK phosphate buffer.

2.3. UV- and X-radiation survival

Cells were grown to saturation in LB broth, diluted 1:250 into fresh broth, and grown at 37°C with aeration in log phase until the optical density at 600 nm (OD₆₀₀) was 0.6. Cells were harvested by centrifugation (6 min at 6,000 x g), washed twice and resuspended in PB. UV irradiation used a Sylvania germicidal lamp (254 nm) while working under GEF40GO Gold 40 W lamps. In two studies, cells were UV-irradiated on tryptone plates to determine survival. In all other studies, cells were UV-irradiated in PB and then plated to determine survival. X-irradiation used a Polaris 160 kV cabinet irradiator (Kimtron) with a 3000 W, Varian NDI-161 tube running at 160 kV and 15 mA, and a dose rate of 56.6 Gy/min. Cells were bubbled with air during X-irradiation. Irradiated cell samples were plated on tryptone agar, and incubated at 37ºC in the dark for 24 h before counting colonies to determine colony-forming units per milliliter (CFU/ml) and cell survival. Statistical analyses relied on SigmaPlot for Windows, v.12.0 (Systat Software, Inc) with significance defined as P≤0.05. Statistical comparisons of paired cell survival curves were performed using a nonparametric analysis of covariance (ANCOVA) with the covariate being radiation dose [24]. Other statistical comparisons relied on ANOVA.

2.4. Genomic DNA isolation, PCR amplification and DNA sequencing

Genomic DNA was isolated from overnight LB broth cultures of the ΔdinB strains, SR4109, SMR6111 and JW0221-1, and a dinB⁺ strain (SR2227) using a commercially available system (Gentra Puregene kit, Gentra Systems Inc), which essentially involves cell lysis using a heated detergent, followed by RNase digestion, protein precipitation via the addition of a concentrated salt solution, centrifugation to pellet the protein and harvest the supernatant. The genomic DNA was then precipitated from the supernatant by the addition of isopropanol. The material was collected by centrifugation and washed with 70% ethanol. The DNA was briefly air dried and resuspended in TE buffer (10 mM Tris –HCL, pH 8.0, 1 mM EDTA). The Gentra protocol is essentially a simple modification of the detergent lysis/salt precipitation protocols of Buffone and Darlington [25] and Miller et al. [26]. The isolated genomic DNA samples were analyzed by NanoDrop spectrophotometry to determine yield and estimate purity. Agarose gel electrophoresis was used to estimate average fragment size and confirm the absence of ethidium bromide-stainable RNA.

DNA sequences of interest from the ΔdinB and dinB⁺ strains were PCR-amplified for DNA sequencing analysis. The PCR primers used are listed (Table S1, Supplementary Data). The PCR products were separated by electrophoresis in Tris acetate EDTA agarose gels and imaged to confirm product size, determine yield and inspect for evidence of non-specific products. With the absence of any detectable non-specific PCR products confirmed, the target PCR fragment was batch purified by adsorption to silica spin columns and eluted in Tris buffer. The isolated target PCR products were analyzed by NanoDrop spectrophotometry to determine yield and estimate purity. Agarose gel electrophoresis was used to confirm product size and absence of detectable degradation fragments. The purified PCR products were sequenced commercially (Sequetech; Mountain View, CA) using single primer extension with an ABI 373OXL capillary sequencer. The “dinB” sequences determined for strains SR4019, SMR6111, SR2227 and JW0221-1 are shown Fig. S1 (Supplementary Data).

3. Results

3.1. The ΔdinB883 mutation sensitizes cells to the lethal effects of UV and X-radiation

For log phase cells, the ΔdinB883 strain was more sensitive to both UV- (Fig. 1A) and X-radiation (Fig. 1B) than the isogenic wild-type strain as measured by colony-forming ability. Sensitivity to X-radiation indicated the protective effect of DinB was not limited to UV radiation-specific DNA lesions, although the sensitization determined at 10^-3 survival in the ΔdinB883 strain was less for X-radiation (88-fold, Fig. 1B) than for UV-radiation (130-fold, Fig. 1A).

Fig 1.

The ΔdinB883 allele reduces cell survival after exposure to radiation. (A) Survival data are shown for logarithmic phase E. coli wild-type (SR2227, ●) and ΔdinB883 (SR4109, ○) strains after UV radiation exposure in buffer. (B) Same format as in A, except strains were irradiated with 160 kV X-rays, aerobically. Data are means ± SD from triplicate experiments.

3.2. The impact of the ΔdinB883 mutation in UV-irradiated cells is growth phase-dependent and is highly dependent on UmuDC function

To more fully characterize the UV radiation sensitivity conferred by the ΔdinB883 mutation, we determined cell survival in ΔumuDC and wild-type strain backgrounds at multiple time points throughout log and stationary phase growth. Samples were taken at 30-min intervals over a 20-h period to determine culture growth (by OD₅₅₀ values) and capacity for UV radiation survival. To reduce complexity, only hourly sample data are shown in Fig. 2, but the omitted 30-min data are consistent with the data shown. Wild-type cells maintained a small, nearly-constant level of UV radiation sensitivity throughout the time course of this experiment. In contrast, both ΔdinB883 and ΔumuDC strains showed growth-phase dependence and were more sensitive to UV radiation than the wild-type strain at every time point with the effect being most pronounced in early to mid-log phase cells (as much as 386-fold) and declining to as little as 1.7-fold as the culture aged (Table 2). The ΔdinB883 strain was less sensitive than the ΔumuDC strain, which was similar in sensitivity to the ΔdinB883 ΔumuDC double-mutant strain (Fig. 2), suggesting DinB protects against lethal DNA damage via direct or indirect interaction with UmuDC.

Fig. 2.

The ΔdinB883 allele does not sensitize a ΔumuDC strain to UV radiation. Growth and survival data are shown for E. coli wild-type (SR2227), ΔdinB883 (SR4109), ΔumuDC (SR4128), and ΔdinB883 ΔumuDC (SR4129) strains incubated overnight, diluted (1:50) into fresh LB broth supplemented with 0.1% glucose and incubated at 37ºC with aeration. Samples were taken at times indicated to determine values for OD₅₅₀ and survival following UV radiation exposure at 25 J m^-2. Cells were irradiated on tryptone plates. Survival data are means ± SD from triplicate experiments. Since OD₅₅₀ values were nearly identical for all four strains, only average values are shown.

Table 2.

Growth phase dependence of UV radiation sensitization due presence of ΔdinB883 and ΔumuDC mutations.

Culture incubation times (min)	Culture OD550 (units)	Relative UV radiation survival for wild type (WT)/mutant strain, mean ± SD (No. of values averaged)
		WT/ ΔdinB883	WT/ ΔumuDC	WT/ ΔdinB883 ΔumuDC
60-240	<1.1	167 ± 57 (4)	334 ± 125 (4)	386 ± 186 (4)
300	1.4	20 (1)	71 (1)	104 (1)
360-640	~3	6.4 ± 2.0 (5)	34 ± 21 (5)	48 ± 23 (5)
1102-1230	~5	1.7 ± 0.3 (2)	11 ± 3 (2)	7 ± 1 (2)

Data are derived from Fig. 2. UV radiation survival for the wild-type strain (SR2227) divided by survival for the mutant strain SR4109 (ΔdinB883), SR4128 (ΔumuDC), or SR4129 (ΔdinB883 ΔumuDC) at the same time and averaged for the listed sample times.

3.3. The ΔdinB883 mutation modestly sensitizes strains to UV radiation that are defective in nucleotide excision repair (NER)

To determine if there is an interaction between NER and DinB activity, ΔdinB883 uvrA and ΔdinB883 uvrB double-mutant strains were constructed and tested for UV radiation sensitivity. As expected, the uvrA and uvrB strains, being defective in NER, were very sensitive to UV radiation (Fig. 3). However, sensitization of the uvrA and uvrB strains by the ΔdinB883 mutation was only about fivefold when determined at the 10^-3 survival level in uvr ΔdinB883 double-mutant strains, which compared to the 130-fold sensitization to UV radiation seen for the ΔdinB883 mutation in wild-type cells at the same survival level (Fig. 1A). Sensitization of uvr strains by ΔdinB883 was only significant (P = 0.006) at UV radiation doses of 9-15 J m-² and was insignificant (P ≥0.05, by ANCOVA) at smaller UV radiation doses (Fig. 1A). It is concluded that ΔdinB883 only modestly sensitizes cells to UV radiation in the absence of NER function.

Fig. 3.

The ΔdinB883 allele sensitizes uvrA and uvrB strains to UV radiation. Survival data are shown for logarithmic phase E. coli uvrA (SR4172, ●), uvrA ΔdinB883 (SR4173, ○), uvrB (SR4175, ▼) and uvrB ΔdinB883 (SR4177, Δ) strains UV-irradiated in buffer. Data are means ± SD from triplicate experiments. The survival curves for ΔdinB883 strains are significantly different from those for paired dinB+ strains over the range of 9-15 J m^-2 (P<0.05).

3.4. The ΔdinB883 mutation does not sensitize a ΔrecA strain to UV radiation

Cells without functional RecA are defective in all forms of DNA repair and tolerance mechanisms relevant to UV radiation-induced DNA damage, except for basal levels of NER [27]. We determined the effect of the ΔdinB883 mutation on the UV radiation sensitivity of a ΔrecA strain (Fig. 4). The ΔrecA strain and the ΔdinB883 ΔrecA double-mutant strain were not statistically different (P ≥0.05, by ANCOVA) in UV radiation sensitivities. Although the ΔdinB883 ΔrecA double-mutant strain consistently showed a greater sensitivity than the ΔrecA strain in multiple experiments, it is concluded that ΔdinB883 had no effect on UV radiation sensitivity in the absence of RecA function.

Fig. 4.

The ΔdinB883 allele does not sensitize a ΔrecA strain to UV radiation. Survival data are shown for logarithmic phase E. coli ΔrecA (SR4178, ●) and ΔrecA ΔdinB883 (SR4179, ○) strains UV-irradiated in buffer. Data are means ± SD from triplicate experiments. The two survival curves shown are not significantly different (P>0.05).

3.5. The ΔdinB883 mutation suppresses the checkpoint function of UmuDC

UmuDC functions as both a DNA polymerase and as a participant in a DNA damage checkpoint [27]. To determine which of these functions is suppressed by ΔdinB883 we measured the UV radiation sensitivity of an umuC104 lexA51 strain in the presence and absence of ΔdinB883 (Fig. 5). The umuC104 allele is devoid of polymerase activity [28, 29], but retains its checkpoint function [30]. The sensitization of the umuC104 lexA51 strain by ΔdinB883, indicates that ΔdinB883 does partially suppress UmuDC checkpoint function. The level of ΔdinB883 suppression of UV radiation resistance in the umuC104 lexA51strain is less than the level of ΔdinB883 suppression in an umuDC⁺ strain at similar UV radiation doses (Fig. 1) suggesting that the TLS activity of umuDC may also be somewhat impaired by the presence of ΔdinB883. However, this comparison is somewhat compromised since the umuDC⁺ strain used in Fig. 1 and the umuC104 strain used in Fig. 5 differ in genotype beyond the umuDC locus.

Fig. 5.

The ΔdinB883 allele sensitizes an umuC104 strain to UV radiation. Survival data are shown for logarithmic phase E. coli umuC104 lexA51 (SR4385, ●) and umuC104 lexA51 ΔdinB883 (SR4391, ○) strains UV-irradiated in buffer. Data are means ± SD from triplicate experiments.

The lexA51 allele produces a defective LexA protein which allows constitutive expression of LexA-repressed genes [27]. UV radiation sensitization by ΔdinB883 of the umuC104 lexA51 strain indicates that ΔdinB883 does not act by suppressing normal LexA activity.

3.6. The UV radiation sensitizing effect of the ΔdinB883 mutation is suggested to be due to a novel fusion polypeptide derived from DinB

Given that the UV radiation sensitivity observed for the ΔdinB883 strain conflicted with previous reports of the UV-radiation phenotype of dinB mutants, we compared the UV radiation sensitivity (under our experimental protocol) of our ΔdinB883 strain with two other ΔdinB mutant strains, SMR6111 and JW0221-1, which are reported to carry dinB null alleles [31, 32]. We found that strains (SMR6111 and JW0221-1, respectively) carrying the Δ(dinB-yafN)882(::kan) [referred to as ΔdinB882] or ΔdinB749::kan [referred to as ΔdinB749] alleles were not different in their UV radiation sensitivity from two dinB⁺ strains (SMR4562 and SR2227) used for comparison (Fig. 6). In contrast, all four strains were significantly more resistant than the ΔdinB883 strain (Fig. 6; P<0.05, by Kruskal-Wallis One Way ANOVA on ranks with pair-wise comparison using Dunn's Method).

Fig. 6.

The ΔdinB883 allele uniquely has the ability to sensitize wild-type cells to UV radiation at 25 J m^-2. Survival data is shown for E. coli strains; dinB⁺ (SMR4562 and SR2227), ΔdinB882 (SMR6111), ΔdinB749 (JW0221-1) and ΔdinB883 (SR4109). Cells were prepared and UV-irradiated to determine survival as in Fig. 2. Data are means ± SD from 5-10 experiments per strain. Strain SR4109 is significantly more sensitive than all other strains (P<0.05).

To test whether another mutation in strain SR4109 might be responsible for its unexpected radiation sensitivity, we moved the ΔdinB883 and ΔdinB882 alleles, which are both derived from the original Δ(dinB-yafN)::kan mutant strain (YG7207), developed by Kim et al. [31] into different dinB⁺ strains by P1 transduction, selecting for kanamycin resistance. In every transductant strain tested, the resulting UV radiation sensitivities remained different for the ΔdinB882 and ΔdinB883 alleles (data not shown). These results suggest it is unlikely the phenotypic difference for radiation sensitivity between strains SR4109 and SMR6111 is due to a gene other than dinB.

To compare the ΔdinB883 and ΔdinB882 alleles at the molecular level, we amplified and sequenced the dinB DNA region from three ΔdinB strains, SR4109, SMR6111, and JW0221-1, and from a dinB⁺ control strain (SR2227). Strain JW0221-1 is from the Keio collection of single-gene knockout mutants [33], and its ΔdinB749::kan [referred to as ΔdinB749] allele is unrelated to the alleles in strains SR4109 and SMR6111. As expected from the method of its construction [33], sequencing the dinB region of strain JW0221-1 showed only the dinB start codon followed by nucleotides consistent with the kan insertion sequence (Fig. S1, Supplementary Data), confirming the null allele status of ΔdinB749. Sequence analysis of the dinB region of strains SR4109 and SMR6111 revealed slightly different ΔdinB alleles (Fig. 7 and Fig. S1, Supplementary Data). Comparison of the nucleotide sequences suggests both strains maintain the wild-type dinB nucleotide sequence for a stretch of 161 nucleotides (Fig. S1, Supplementary Data), and the DinB amino acid sequence for the first 54 amino acids (Fig. 7). The nucleotide sequence differs starting at position 162, the first nucleotide derived from the DNA fragment from the plasmid pUC4K insert, which carries the kan gene [31]. Alleles ΔdinB883 and ΔdinB882 have the same proximal insert nucleotide sequences, except for the run of G’s starting at position 169. For ΔdinB883, the run consists of 12 G’s, while ΔdinB882 shows a run of 11 G’s (Fig. 7 and Fig. S1, Supplementary Data). The different lengths of the run of G’s creates different reading frames and distal DinB amino acid sequences for the two alleles (Fig. 7). We sequenced 270 nucleotides distal to the run of G’s and found the sequence identical in both strains (data not shown). We also sequenced 530 nucleotides proximal to the dinB start codon and found the same wild-type DNA sequence in both strains (data not shown). It is concluded that the single-base difference in this run of G's is the cause of the different UV radiation sensitivity phenotypes associated with alleles ΔdinB883 and ΔdinB882. We also sequenced the ΔdinB allele of YG7207, the strain from which ΔdinB883 was derived in this study, and found that it also had a run of 12 G's.

Fig. 7.

DNA and amino acid (AA) sequences for wild-type (dinB⁺) and ΔdinB mutant strains. DNA and predicted fusion polypeptide sequences are shown for dinB⁺ (wild-type, SR2227, not UV radiation sensitive), ΔdinB883 (SR4109, significantly UV radiation sensitive), and ΔdinB882 (SMR6111, not UV radiation sensitive) strains. The nucleotide sequences are identical for the three strains for the first 53 codons. Codon 54 differs in the ΔdinB883 and ΔdinB882 strains from the dinB⁺ DNA sequence but the AA remains the same. After codon 54 the dinB⁺ DNA sequence and resulting AAs differs from those of ΔdinB883 and ΔdinB882. From codon positions 55 to 61, the AA sequences are identical for ΔdinB883 and ΔdinB882 as are the DNA sequences except for codon 61. After codon 61 the DNA and AA sequences both differ in the two ΔdinB strains. The nucleotides of the kan^r insertion fragment fused to ΔdinB sequence begin opposite AA position 54 (3^rd position), but the AA sequence is not affected until AA position 55.

4. Discussion

4.1. ΔdinB883 is likely a “gain-of-function” mutation

Our initial observation that strain SR4109 (ΔdinB883) was sensitive to radiation was surprising, since earlier studies had clearly indicated mutant dinB alleles had no effect on UV radiation-induced lethality [3, 22]. We confirmed that two other putative ΔdinB strains, SMR6111 and JW0221-1 (ΔdinB882 and ΔdinB749, respectively) were not UV radiation sensitive (Fig. 6).

These disparate results for the UV radiation sensitivity of ΔdinB strains led to further experiments. In one approach, we confirmed that the UV radiation sensitivity phenotype (sensitive and resistant, respectively) was tightly linked (by bacteriophage P1 transduction) for both the ΔdinB883 and ΔdinB882 alleles (data not shown). More significantly, we sequenced the dinB regions in three ΔdinB strains, SR4109, SMR6111, and JW0221-1, and a dinB⁺ control strain (SR2227). This analysis revealed a single-base difference between strains SR4109 and SMR6111 in the run of G’s associated with the insertion of the kan sequence during construction of the parental YG7207 strain (Fig. 7). We suggest this single base difference is the basis for the different UV radiation sensitivities seen for strains SR4109 and SMR6111.

The single base difference (12 G's vs. 11 G's) associated with ΔdinB in strains SR4109 and SMR6111, respectively, is believed to have created different reading frames and different truncated versions of mutant DinB polypeptides (Fig. 7). DinB883 is suggested to be an 83 amino acid fusion polypeptide comprised of 54 wild-type DinB amino acids followed by seven non-wild-type amino acids (shared with DinB882) and then by 22 unique C-terminal amino acids before a stop codon appears in the nucleotide sequence (Fig. 7). The DinB882 fusion polypeptide protein is suggested to be identical to DinB883 protein for its first 61 amino acids, then different in its 14 C-terminal amino acids. DinB882 should be 75 amino acids in length considering the position of its stop codon in the DNA sequence (Fig. 7). Since the ΔdinB882 strain shows the same lack-of-radiation-sensitivity phenotype (Fig. 6) as two dinB⁺ strains (SMR4562 and SR2227) strains and a confirmed complete dinB deletion strain (JW0221-1), we conclude that DinB883 has a novel active function, which reduces cell survival after radiation treatment, presumably by interfering with DNA repair or tolerance mechanisms.

4.2. The DinB883 fusion polypeptide causes UV radiation sensitivity by modulating UmuCD activity, which is dependent upon cell growth phase

When replication forks are blocked in E. coli cells by DNA lesions, such as those created by UV radiation exposure, the SOS response is induced [27]. This response involves the coordinated enhanced expression of a set of SOS genes whose products increase cell survival. The SOS regulon includes dinB, umuDC and the NER genes uvrA and uvrB, among others [27]. In studies that prevent photoreactivation, UV radiation survival in E. coli depends upon NER, DNA recombination processes and, to a lesser extent, UmuDC activity [27]. Generally speaking, once NER is blocked by uvrA or uvrB mutations, additional mutations impacting NER should have little effect. While the effect is modest, the ΔdinB883 mutation does increase the UV radiation sensitivity of uvrA and uvrB strains in late log phase cells (Fig. 3), suggesting DinB883, to a small degree, interferes with DNA repair or tolerance mechanisms other than NER.

The ΔdinB883 allele clearly has no UV radiation sensitizing effect on a ΔumuDC strain (Fig. 2, Table 2), and it should be noted that the ΔumuDC allele used in this study is a deletion of both the umuC and umuD genes and is assumed to be a null allele [34]. No significant differences were observed for UV radiation sensitivity between ΔumuDC and ΔumuDC ΔdinB883 strains over a range of growth conditions from early log phase to early stationary phase (Fig. 2). Additionally, ΔumuDC ± ΔdinB883 strains were much more UV radiation sensitive in their early log growth phases than in late log or early stationary growth phases. In contrast, the wild-type strain maintained the same high level of UV-irradiated cell survival throughout all growth phases (Fig. 2, Table 2). UmuDC function has generally been assumed to be a minor component among the survival mechanisms after UV radiation exposure [27], but most studies on UV radiation sensitivity in E. coli have used cells in late log or early stationary phase. We suggest that UmuDC function is more important for survival from UV radiation exposure in early log than in later growth phases. One explanation for this observation is that in early log phase, rapidly-dividing cells with multiple replication forks may demand a higher level of overall DNA repair and damage tolerance capacity after UV radiation exposure than in later growth phases. Another possible explanation is that early log phase cells exhibit a metabolic state very different from that of late log or stationary phase cells and may accumulate a greater number of DNA lesions due to endogenous DNA damage. Also, the ΔdinB883 strain was less sensitive to UV radiation than a ΔumuDC strain (Fig. 2, Table 2), which suggests DinB883 incompletely suppresses UmuDC function in UV radiation protection.

4.3. Possible interactions of the DinB883 fusion polypeptide with other proteins

Our genetic and DNA sequence data analyses predict ΔdinB883 will encode a fusion polypeptide (DinB883) 83 amino acids in length with the first 54 being the N-terminal portion of the wild-type DinB protein. We propose DinB883 reduces the activity of UmuDC, resulting in reduced survival after UV radiation exposure. DinB883 and DinB882, which is suggested to not interfere with DNA repair, have identical sequences for the first 61 amino acids, so presumably their different phenotypes result from C-terminal residues.

Wild-type DinB binds to several proteins including UmuD and RecA [35]. The umuDC operon encodes the UmuD and UmuC proteins that function in TLS in a highly-regulated fashion [12]. UmuD forms dimers (UmuD₂), which in conjunction with UmuC, are thought to reduce the rate of DNA replication in the presence of DNA damage, allowing time for NER and other error-free forms of repair and damage tolerance to remove or bypass potentially lethal DNA lesions [5]. Eventually, the interaction of UmuD₂ with RecA results in self-cleavage, which removes N-terminal amino acids and converts UmuD₂ to UmuD′₂[36]. UmuD′₂ associates with UmuC to become an active DNA polymerase (PolV) which, along with RecA, effects TLS past remaining lesions in DNA [37].

Both UmuD₂ and RecA have been previously shown to bind to DinB resulting in suppression of DinB's induction of frameshift mutations by enclosing its open site [32]. This close association of these three proteins is suggested to occur during the initial SOS response before significant amounts of TLS activity occur. Two surface DinB amino acids, cysteine 66 (C66) and proline 67 (P67), have been recently found to modulate the binding of DinB to RecA and UmuD₂ [38].

DinB883 presumably has the capacity to bind to some protein(s) in a way that inhibits UmuDC activity although this may not involve binding to UmuD or UmuC themselves. The ΔdinB883 sensitizing of an umuC104 strain (Fig. 5) indicates that the DNA damage checkpoint function of UmuDC is being suppressed. One possibility is that DinB883 binds to UmuD or some other protein and this interferes with the checkpoint activity. DinB883 may also interfere with the TLS activity of UmuDC although we have no data to strongly support this possibility.

It is not clear why DinB883 has radiation-sensitizing activity, whereas DinB882 lacks activity. DinB883 should be eight amino acids longer than DinB882, which may influence their relative stabilities. No proteins that might preferentially bind or interact with DinB883 are obvious. It is also of interest to evaluate the stability of ΔdinB883 compared to ΔdinB882, given that nucleotide runs have been long recognized as frameshift mutation hotspots [39]. So far, in the sequencing of multiple isolates of both strains, we have not observed sequence changes in either allele.

5. Conclusion

We have described a single base (G) difference in the DNA sequence of two commonly used alleles (ΔdinB883 and Δdin882), where one allele displays a novel umuDC-dependent activity that reduces cell survival in the presence of DNA damage, presumably by interfering with repair or tolerance mechanisms. Both alleles are derivatives of the Δ(dinB-yafN) allele originally characterized by Kim et al. [31] that has been used in many ΔdinB studies and presumably initially had a run of 11 or 12G's. The plasmid pUC4K was used in the construction of the Δ(dinB-yafN) allele [31] and it contained a run of 12G's [40]. This would suggest that the Δ(dinB-yafN) allele had 12G's and thus was identical to ΔdinB883. Our copy of YG7207, the strain containing the Δ(dinB-yafN) allele, also has 12G's. However we are reluctant to firmly conclude that the original construction of the dinB deletion contained 12G's since the sequence at the time of construction is not known as far as we are aware of. Since it is unclear historically what the original run number was and when the run either gained or lost one G, this event may have impacted the results of earlier published studies on ΔdinB.

Highlights.

• We describe Δ(dinB-yafN)883(::kan), a novel dinB allele, referred to as ΔdinB883, a deletion that sensitizes E. coli cells to UV irradiation.

• This UV radiation sensitivity is most acute in the early logarithmic phase of culture growth.

• This UV radiation sensitivity is completely dependent upon a functional umuDC operon

• Sequencing reveals ΔdinB883 retains the proximal 161 nucleotides, i.e., 54 amino acids, of the wild-type sequence.

• The ΔdinB883 mutant is hypothesized to produce a peptide of 83 amino acids, DinB883, that compromises UmuDC function.