A Role for the umuDC Gene Products of Escherichia coli in Increasing Resistance to DNA Damage in Stationary Phase by Inhibiting the Transition to Exponential Growth

Abstract

The umuDC gene products, whose expression is induced by DNA-damaging treatments, have been extensively characterized for their role in SOS mutagenesis. We have recently presented evidence that supports a role for the umuDC gene products in the regulation of growth after DNA damage in exponentially growing cells, analogous to a prokaryotic DNA damage checkpoint. Our further characterization of the growth inhibition at 30°C associated with constitutive expression of the umuDC gene products from a multicopy plasmid has shown that the umuDC gene products specifically inhibit the transition from stationary phase to exponential growth at the restrictive temperature of 30°C and that this is correlated with a rapid inhibition of DNA synthesis. These observations led to the finding that physiologically relevant levels of the umuDC gene products, expressed from a single, SOS-regulated chromosomal copy of the operon, modulate the transition to rapid growth in E. coli cells that have experienced DNA damage while in stationary phase. This activity of the umuDC gene products is correlated with an increase in survival after UV irradiation. In a distinction from SOS mutagenesis, uncleaved UmuD together with UmuC is responsible for this activity. The umuDC-dependent increase in resistance in UV-irradiated stationary-phase cells appears to involve, at least in part, counteracting a Fis-dependent activity and thereby regulating the transition to rapid growth in cells that have experienced DNA damage. Thus, the umuDC gene products appear to increase DNA damage tolerance at least partially by regulating growth after DNA damage in both exponentially growing and stationary-phase cells.

Mechanisms which temporarily block DNA replication and cell cycle progression after exposure to DNA-damaging agents have been shown to play an important role in mediating resistance to these agents in eukaryotes (10, 17, 41, 57). The inhibition of growth following DNA damage allows DNA repair to occur prior to continued DNA replication and chromosome segregation, thereby ensuring the fidelity of these processes. Despite the fact that they face similar environmental challenges, the extent to which prokaryotes respond to DNA damage by controlling aspects of their cell cycle is much less well understood (4, 5, 19, 26). Bacterial septation has been shown to be tightly regulated after DNA damage (20), and we have very recently presented evidence supporting a model for a umuDC-dependent prokaryotic DNA damage checkpoint in exponentially growing Escherichia coli cells (38). It seems possible that regulation of growth after DNA damage might also play a role in DNA damage tolerance in cells that have suffered DNA damage while in a quiescent phase but then experience a change in environmental conditions that normally promotes growth, such as an increase in available nutrients.

In E. coli, a set of at least 20 genes are coordinately induced in response to DNA damage in a process known as the SOS response (17, 27, 52). The SOS response is regulated by RecA and the transcriptional repressor LexA (17). Cells sense that they have experienced DNA damage when RecA forms nucleoprotein filaments with single-stranded DNA produced as a consequence of the damage and mediates the cleavage of LexA, thereby inducing the expression of the SOS genes. Gene products regulated as part of the SOS response include those involved in DNA repair, induced mutagenesis, the regulation of cell division, and other functions (17). Interestingly, the SOS response is also induced in quiescent cells, even in the absence of exposure to exogenous DNA-damaging agents (54), suggesting that DNA damage accumulates in quiescent cells which must be repaired prior to resumed growth.

The umuDC genes are regulated as part of the SOS response, and the functions of their gene products are needed for most of the mutagenesis resulting from exposure to DNA-damaging agents such as UV light (17, 27, 52). Posttranslational RecA-mediated proteolytic cleavage of UmuD to UmuD′, the carboxyl-terminal 12-kDa fragment of UmuD (8, 36, 50), is required for DNA damage-induced mutagenesis, while uncleaved UmuD has been implicated in a DNA damage checkpoint (38). The structure of crystallized UmuD′₂ has been solved (42), and the correct interface of the UmuD′₂ dimer in solution has been determined by nuclear magnetic resonance methods (13). Both UmuD and UmuD′ form complexes with UmuC (6, 59). DNA damage-induced mutagenesis results from errors introduced during the process of replicative bypass of a DNA lesion that requires DNA polymerase III, UmuD′, UmuC, and RecA (43, 45, 55, 56).

Constitutive expression of umuDC from a multicopy plasmid in E. coli causes a growth inhibition at 30°C but not at 42°C that is associated with an inhibition of DNA replication at the restrictive temperature (32, 39, 52). Uncleaved UmuD, which is inactive in SOS mutagenesis (36), is the form of the umuD⁺ product that acts in combination with UmuC to confer cold sensitivity for growth (39). These observations suggested the possibility that uncleaved UmuD and UmuC might have a novel role modulating the E. coli cell cycle after DNA damage (38, 39) and stimulated us to undertake the experiments that led to our recent model for a umuDC-dependent prokaryotic DNA damage checkpoint in exponentially growing cells (38).

In this paper, we describe how our studies of the phenomenon of growth inhibition at 30°C caused by overexpression of the umuDC operon resulted in our discovery that physiologically relevant levels of the umuDC gene products modulate the transition to rapid growth of E. coli cells that have suffered DNA damage while in stationary phase and then experience a nutrient upshift. This activity of the umuDC gene products is correlated with an increase in survival after UV irradiation. The increased UV resistance in stationary phase conferred by the umuDC gene products appears to result from counteracting an activity of Fis. The Fis protein, the levels of which increase dramatically upon exposure of quiescent cells to an environment with increased nutrients (1), is a small DNA binding protein which regulates the growth phase transition from stationary phase to exponential growth (40). These data support a model for a specific regulated mechanism in prokaryotes that increases survival of cells that have suffered DNA damage while in stationary phase by temporally inhibiting their growth when they experience a nutritional upshift so that accurate repair can occur.

MATERIALS AND METHODS

Strains and plasmids.

The E. coli strains and plasmids used in this work are listed with their relevant features in Table 1. Genetic markers were transferred between strains by P1(vir) transduction performed as described by Miller (33).

TABLE 1.

List of strains and plasmids

Strain or plasmid	Relevant genotype and/or comments	Source or reference
Strains
GW2771	lexA+ recA+ umuD+C+	39
GW8018	lexA(Def)::spc recA441 sulA11 sfiC2	39
GW8023	GW2771 Δ(umuDC)595::cat	39
GW8025	GW8018 Δ(umuDC)595::cat	39
RJ1802	fis::767 kanr	1
GW8027	lexA(Def)::spc recA430 umuD+C+	39
GW8034	GW8018 fis::767 kanr	GW8018 × P1(RJ1802)
GW8037	GW2771 fis::767 kanr	GW2771 × P1(RJ1802)
GW8038	GW8023 fis::767 kanr	GW8023 × P1(RJ1802)
GW8040	GW8027 Δ(umuDC)595::cat	39
Plasmids
pBR322/kan	Kmr Tcs	39
pSE115	umuD+C+; pSC101	11
pSE117	umuD+C+; pBR322	11
pTO4	umuD(SA60)C+; pBR322	This work

Reagents and media.

Ampicillin, kanamycin, spectinomycin, chloramphenicol, and tetracycline were purchased from Sigma (St. Louis, Mo.). [methyl-³H]thymidine (83 Ci/mmol) was purchased from Amersham Corp. (Arlington Heights, Ill.). Thymidine and 2′-deoxyadenosine were purchased from Sigma. The Western Lights kit for chemiluminescent detection in immunoblot assays was purchased from Tropix (Bedford, Mass.). The anti-Fis antibody (1) was a kind gift from Reid C. Johnson. The anti-UmuD/D′ antibody used in these studies has been described previously (2). Bacteria were grown in Luria-Bertani (LB) or M9 medium supplemented with Casamino Acids as indicated and in LB agar (47). Antibiotics were used at the following concentrations: ampicillin, 150 μg/ml; kanamycin, 25 μg/ml; spectinomycin, 20 μg/ml; chloramphenicol, 30 μg/ml; tetracycline, 12.5 μg/ml.

Cold sensitivity assays and growth curve analyses.

Quantitative transformation assays, growth curve analyses, and immunoblot analyses were performed as previously described (39). β-Galactosidase activity assays were performed at various times during growth curve analyses at 42 and 30°C of strains containing lacZ-gene fusions as described by Miller (33). For experiments involving UV irradiation, cells were grown in M9 medium supplemented with 0.4% Casamino Acids. Five-milliliter samples of the cultures were irradiated with UV light on 60- by 15-mm petri dishes at the indicated dose. To reduce the effects of shading by dense cultures, stationary-phase cells were diluted threefold in saline prior to UV irradiation. In experiments where growth after nutrient upshift was monitored, UV-irradiated and unirradiated stationary-phase cultures were immediately diluted 1:100 into fresh M9 medium and grown in the dark at 37°C. At the indicated times, serial dilutions from each culture were plated and incubated in the dark to determine CFU per ml.

DNA synthesis assays.

All DNA synthesis assays were performed in duplicate. Briefly, 16-h-old overnight cultures were diluted 1:125 in LB medium and grown at 42°C. During lag phase and exponential growth, an aliquot of each culture growing at 42°C was shifted to 30°C. After 10 min and 2 h, duplicate aliquots of the cultures at 42 and 30°C were transferred to tubes at 42 and 30°C, respectively, with continued aeration. DNA synthesis assays were performed essentially as described by Bukau and Walker (7). Incorporation of [methyl-³H]thymidine into DNA at each time point was normalized to the optical density at 600 nm (OD₆₀₀) of the culture at the start of the assay.

RESULTS

Constitutive expression of umuDC inhibits the transition from stationary phase to exponential growth at 30°C.

As part of our effort to understand the basis for our previous observation that constitutive expression of umuDC from a multicopy plasmid leads to growth inhibition and filamentation at 30°C but not at 42°C (32, 38, 39), we examined whether constitutive expression of umuDC causes an equivalent inhibition of bacterial growth at 30°C in exponentially growing and lag-phase cells. Cultures of a strain that constitutively expresses umuD⁺C⁺ from a multicopy plasmid due to the presence of a lexA(Def) mutation [GW8025(pSE117)] and a corresponding strain that lacks umuD⁺C⁺ [GW8025(pBR322kan)] were diluted after 11 h in stationary phase into fresh LB medium and grown at 42°C. At various times during growth, a portion of each culture was shifted to 30°C and subsequent growth was monitored by measuring the OD₆₀₀. Maximal umuDC-dependent growth inhibition at 30°C was observed when the strain that constitutively expresses umuD⁺C⁺ was shifted to 30°C during lag phase, i.e., within 1 h of dilution into fresh medium (Fig. 1A). When this strain was shifted to 30°C during exponential growth, umuDC-dependent growth inhibition was markedly reduced. In contrast, the control strain lacking umuD⁺C⁺ did not exhibit an inhibition of growth when shifted to 30°C during either lag phase or exponential growth (data not shown). These results suggest that cells in lag phase at 30°C are substantially more susceptible to the inhibitory effects on growth of the overexpressed umuDC gene products than are exponentially growing cells at 30°C.

FIG. 1.

umuDC-dependent inhibition of growth at 30°C is growth phase specific. (A) A stationary-phase culture of a lexA(Def) strain carrying a multicopy umuD⁺C⁺ plasmid [GW8025(pSE117)] was diluted 1:125 into fresh LB medium, and growth at 42°C was monitored by measuring OD₆₀₀ (□). A portion of this culture was shifted to 30°C (time of shift indicated by arrows) during lag phase (■) and exponential growth (●), and subsequent growth was monitored by measuring OD₆₀₀. (B) Stationary-phase cultures of strains GW8025(pBR322kan) and GW8025(pSE117) were diluted into fresh LB medium, grown at 42°C for 1 h, and shifted to 30°C. Growth at 30°C was monitored by measuring OD₆₀₀. □, GW8025(pBR322kan), 1:125 dilution; ■, GW8025(pSE117), 1:20 dilution; ●, GW8025(pSE117), 1:125 dilution; ▴, GW8025(pSE117), 1:400 dilution.

The possibility that cells growing exponentially were less susceptible to umuDC-dependent growth inhibition at 30°C because they were at a higher cell density was ruled out by the results of the experiment shown in Fig. 1B. Growth inhibition was monitored after a culture of the strain constitutively expressing umuD⁺C⁺ [GW8025(pSE117)] that had been in stationary phase for 11 h was diluted 1:20, 1:125, or 1:400 in fresh medium, grown for 1 h at 42°C, and shifted to 30°C during lag phase. The degree of growth inhibition at 30°C for all three cultures constitutively expressing umuD⁺C⁺ was essentially the same and markedly distinct from the rapid growth at 30°C of the corresponding strain that lacks umuD⁺C⁺ [GW8025(pBR322kan)]. Taken together, these analyses indicate that high levels of the umuDC gene products interfere with the growth phase transition from lag phase to exponential growth at 30°C in a fashion that is cell density independent.

Constitutive expression of umuDC inhibits DNA synthesis at 30°C.

It had been previously reported that the umuDC-dependent growth inhibition at 30°C is associated with a rapid, reversible inhibition of DNA synthesis upon a shift to the restrictive temperature (32). To test the possibility that the rapid inhibition of DNA synthesis at 30°C is dependent on the growth phase of the culture, we compared the rate of DNA synthesis of a strain that constitutively expresses umuD⁺C⁺ from a plasmid [GW8025(pSE117)] to that of a control strain lacking umuD⁺C⁺ [GW8025(pBR322kan)] (Fig. 2). No umuDC-dependent difference in the rate of DNA synthesis was observed between cultures growing at 42°C either during lag phase or during exponential growth (Fig. 2A and B). When these cultures were shifted to 30°C during the lag phase, a rapid umuDC-dependent decrease in the rate of DNA synthesis (approximately threefold) was observed within 10 min of the shift to 30°C (Fig. 2C). In contrast, no umuDC-dependent decrease in the rate of DNA synthesis was observed in the exponentially growing cultures 10 min after the shift to 30°C (Fig. 2D). In fact, the exponentially growing umuD⁺C⁺ culture [GW8025(pSE117)] had a reproducibly slightly higher rate of DNA synthesis than the control strain which lacks umuDC. Thus, in a striking parallel to the results from the growth curve analyses, a rapid approximately threefold inhibition of DNA synthesis at 30°C conferred by the umuDC gene products was observed only when the shift to the restrictive temperature was made during lag phase.

FIG. 2.

UmuDC-mediated inhibition of DNA synthesis at 30°C. Stationary-phase cultures of GW8025(pBR322kan) (□) and GW8025(pSE117) (▴) were diluted 1:125 in fresh LB medium and grown at 42°C. A portion of these cultures was shifted to 30°C during lag phase and exponential growth. At various times subsequently, the rate of DNA synthesis at 42 and 30°C was determined. Incorporation of [methyl-³H]thymidine was normalized at each time point to the OD₆₀₀ of the culture. The ratios of the rates of DNA synthesis of the ΔumuDC to umuDC⁺ cultures are indicated in the following in parentheses (A) 42°C, lag phase (1.0); (B) 42°C, exponential growth (1.0); (C) 30°C for 10 min after shift in lag phase (2.7); (D) 30°C for 10 min after shift in exponential growth (0.8); (E) 30°C for 2 h after shift in lag phase (5.0); (F) 30°C for 2 h after shift in exponential growth (2.0).

The inhibition of DNA synthesis after 2 h at 30°C increased in the umuD⁺C⁺ culture that was shifted during lag phase to fivefold that of the control, while the inhibition of DNA synthesis in the umuD⁺C⁺ culture shifted during exponential growth was approximately twofold (Fig. 2E and F). Thus, prolonged incubation at the restrictive temperature of 30°C leads to a more general approximately twofold umuDC-mediated inhibition of DNA synthesis which is independent of the growth phase of the culture. However, this latter mode of umuDC-mediated DNA synthesis inhibition could be an indirect consequence of growth inhibition. umuDC-mediated inhibition of DNA synthesis at 30°C therefore appears to occur by at least two different mechanisms: one which is growth phase dependent and another which is growth phase independent.

umuDC-dependent inhibition of growth at 30°C is exacerbated by prolonged starvation.

We then tested whether the time-dependent physiological changes that occur during stationary phase influence the severity of the umuDC-dependent inhibition of the growth phase transition from stationary to exponential phase. Cultures of E. coli strains constitutively expressing umuD⁺C⁺ either from a low-copy-number plasmid [GW8025(pSE115)] or from a higher-copy-number plasmid [GW8025(pSE117)] as well as a control strain lacking umuD⁺C⁺ [GW8025(pBR322kan)] were grown in LB medium at 42°C (Fig. 3A). At various times (8, 16, and 24 h after inoculation), a portion of each culture was diluted 1:125 in fresh medium to measure umuDC-dependent inhibition of growth at 30°C (Fig. 3B to D). This experiment revealed that an increase in the time the cells had spent in stationary phase from approximately 3 to 19 h led to a concomitant increase in the severity of umuDC-dependent inhibition of growth at 30°C. This was most clearly seen in the strain that expressed lower levels of the umuD⁺C⁺ gene products because it contained a lower-copy-number plasmid [GW8025(pSE115)]. The growth at 30°C of this strain [GW8025(pSE115)] and that of the control strain lacking umuD⁺C⁺ were indistinguishable upon nutrient upshift after approximately 3 h in stationary phase (Fig. 3B). However, after approximately 11 h in stationary phase, a slight umuDC-dependent growth inhibition at 30°C was seen upon nutrient upshift (Fig. 3C). By the time the cells had been in stationary phase for 19 h, a marked umuDC-mediated growth inhibition at 30°C was seen upon nutrient upshift (Fig. 3D). This increase in umuDC-dependent growth inhibition at 30°C was not due to a significant decrease in CFU per milliliter during prolonged incubation in stationary phase (data not shown). Furthermore, immunoblot analyses did not reveal any significant changes in the levels of UmuD during the time that the cells became increasingly sensitive to umuDC-dependent growth inhibition at 30°C (data not shown). These observations suggested that prolonged incubation in stationary phase results in physiological changes that increase the susceptibility of cells to the umuDC-dependent inhibition of the transition from stationary phase to exponential growth at 30°C.

FIG. 3.

Effect of prolonged starvation on umuDC-mediated growth inhibition at 30°C. (A) The following strains were grown in LB medium at 42°C: GW8025(pBR322kan) (□), GW8025(pSE115) (●), and GW8025(pSE117) (▴). At various times during growth (as indicated by arrows), a portion of each culture was diluted 1:125 in fresh LB medium, grown at 42°C for 1 h, and shifted to 30°C. Subsequent growth at 30°C was monitored by measuring OD₆₀₀. (B to D) Results are shown for cultures 8 (B), 16 (C) and 24 (D) h old at the time of nutrient upshift.

The umuDC gene products regulate the resumption of rapid growth upon nutrient upshift of UV-irradiated stationary-phase cultures.

The results presented in Fig. 1, 2, and 3 led us to consider the possibility that the umuDC gene products might play a previously uncharacterized role in cellular resistance to killing by UV light by delaying the transition to rapid growth of cells that suffer DNA damage while in stationary phase and then experience a nutrient upshift. Such a delay would allow additional time for DNA repair to occur prior to replication. To test this hypothesis, we examined the effect of the umuDC gene products on the growth kinetics and survival of stationary-phase cultures that had been UV irradiated and then diluted into fresh medium (i.e., a nutrient upshift). In this case, we did not use a strain that overexpressed the umuDC gene products but rather employed a wild-type strain (umuD⁺C⁺ recA⁺ lexA⁺) and compared its behavior to that of an isogenic ΔumuDC mutant. Cultures of the two strains that had been in stationary phase for 11 h were irradiated with UV light at a dose of 50 J/m² and immediately given a nutrient upshift by subculturing 1:100 into fresh medium at 37°C in the dark, and samples were plated to determine CFU per milliliter. This dose resulted in an approximately 60% reduction in viability of the umuD⁺C⁺ strain and an approximately 96% reduction in viability of the ΔumuDC strain when the UV-irradiated cells were plated immediately after subculturing. Subsequently, at various times after the nutrient upshift, samples were plated so that the number of viable cells per milliliter in each culture could be monitored as a function of time (Fig. 4). Because UV irradiation reduced the number of viable cells in the culture, we measured CFU per milliliter (Fig. 4 and 5) instead of OD₆₀₀ so that our measurements of growth would reflect the number of viable cells and would not be masked by the absorbance of the large number of nonviable cells in the irradiated culture. The growth levels, assayed by the change in CFU per milliliter of the unirradiated umuD⁺C⁺ (closed circles) and ΔumuDC (open squares) cultures after nutrient upshift, were identical (Fig. 4). In contrast, there was a clear difference in the growth kinetics of the umuD⁺C⁺ and ΔumuDC cultures after UV irradiation in stationary phase (Fig. 4). The irradiated umuD⁺C⁺ culture showed a pronounced increase in the lag phase upon nutrient upshift compared to that of the unirradiated controls, whereas the irradiated ΔumuDC culture began growth more rapidly. These experiments show that physiologically relevant levels of the umuDC gene products (expressed from the single chromosomal copy of the umuD⁺C⁺ operon in a lexA⁺recA⁺ background) can control the transition from stationary phase to exponential growth of cells that have experienced DNA damage while in stationary phase. The failure of the ΔumuDC strain to properly modulate this transition might account, at least in part, for its increased sensitivity when it is UV irradiated in stationary phase.

FIG. 4.

Effect of the umuDC gene products on survival and growth of UV-irradiated stationary-phase cultures after nutrient upshift. Cultures were grown at 37°C. GW2771 (umuD⁺C⁺) and GW8023 (ΔumuDC) were grown in M9 medium to stationary phase. One milliliter of each culture was exposed to a UV dose of 50 J/m². UV-irradiated and unirradiated cultures were diluted 1:100 in fresh M9 medium and grown in the dark. At various times during growth, serial dilutions were plated to quantify CFU per milliliter. Symbols: ●, GW2771 (umuD⁺C⁺) without UV irradiation; □, GW8023 (ΔumuDC) without UV irradiation; ○, GW2771 (umuD⁺C⁺) with UV irradiation; ■, GW8023 (ΔumuDC) with UV irradiation.

FIG. 5.

Effect of uncleaved UmuD and UmuC on survival and growth of UV-irradiated stationary-phase cultures after nutrient upshift. Cultures were grown at 37°C. GW8027 [recA430 lexA(Def) umuD⁺C⁺] and GW8040 [recA430 lexA(Def) ΔumuDC] were grown in M9 medium to stationary phase. One milliliter of each culture was exposed to a UV dose of 50 J/m². UV-irradiated and unirradiated cultures were diluted 1:100 in fresh M9 medium and grown in the dark. At various times during growth, serial dilutions were plated to quantify CFU per milliliter. Symbols: ●, GW8027 (recA430 umuD⁺C⁺) without UV irradiation; □, GW8040 (recA430 ΔumuDC) without UV irradiation; ○, GW8027 (recA430 umuD⁺C⁺) with UV irradiation; ■, GW8040 (recA430 ΔumuDC) with UV irradiation.

We tested whether the umuDC-dependent inhibition of cellular growth of UV-irradiated stationary-phase cells we had observed correlated with an inhibition of DNA synthesis. We used the same protocol that has been used with UV-irradiated exponentially growing cells (37, 38) but did not observe a detectable effect (data not shown). Although this might appear to suggest that the umuDC gene products are delaying cell growth by affecting some stage of the cell cycle other than DNA replication, this is not necessarily the case. Recent work has indicated that there may be as many as three separate pathways for recovery of DNA replication in UV-irradiated exponentially growing E. coli cells (44), and we had to analyze cells that overexpress the umuDC gene products in order to detect allele-specific umuDC-dependent effects on the recovery of DNA replication in UV-irradiated exponentially growing cells that were proficient in these other pathways (38).

Uncleaved UmuD and UmuC regulate the transition to exponential growth of cultures UV irradiated in stationary phase.

In previous work, we had found that uncleaved UmuD, the form inactive in SOS mutagenesis, confers a significantly higher degree of growth inhibition at 30°C in combination with UmuC than does UmuD′, the form active in SOS mutagenesis (39). Similarly, we had observed that it was uncleaved UmuD together with UmuC that delayed the recovery of DNA replication and growth of cells that had suffered DNA damage while in exponential phase (38). We were therefore interested in determining whether uncleaved UmuD, acting together with UmuC, was capable of regulating the transition of UV-irradiated cells from stationary phase to exponential growth. To test this, we took advantage of the fact that UmuD is not cleaved in a recA430 mutant background (50). Since LexA cleavage is also inefficient in a recA430 strain (12), it was necessary to introduce a lexA(Def) mutation to permit expression of the umuDC operon. We UV irradiated stationary-phase cultures of lexA(Def) recA430 umuD⁺C⁺ and lexA(Def) recA430 ΔumuDC strains, immediately subjected them to a nutrient upshift by dilution in fresh medium, and assayed for viable cells at various times after the upshift as described above. A UV dose of 25 J/m² in stationary phase resulted in an approximately 64% reduction in viability in GW8027 (recA430 umuD⁺C⁺) and an approximately 86% reduction in viability in GW8040 (recA430 ΔumuDC) when the UV-irradiated cells were plated immediately after subculturing. Similar to the situation in the recA⁺ background, in a recA430 background the presence of uncleaved UmuD and UmuC resulted in a distinct increase in the duration of the lag phase upon nutrient upshift of UV-irradiated cultures (Fig. 5), indicating that the cleavage of UmuD to UmuD′ is not required for this activity. Thus, the umuDC-dependent delay in the transition to rapid growth of UV-irradiated stationary-phase cultures appears to be the result of a novel activity of uncleaved UmuD and UmuC. This delay in the onset of rapid growth after nutrient upshift would allow additional time for the repair of DNA damage accumulated in stationary phase.

Fis alleviates umuDC-mediated inhibition of growth at 30°C.

When we discovered the growth phase dependence of the growth inhibition that is caused by overexpression of the umuDC operon at 30°C, we were interested in determining whether the umuDC gene products might be interfering with the expression or function of gene products required during stationary phase or during the transition from stationary phase to exponential growth. We therefore examined whether rpoS⁺, which encodes the alternative sigma factor ς^s responsible for the coordinate regulation of 50 to 100 genes expressed in response to various forms of stress and upon entry into stationary phase (3, 24, 31), plays a role in umuDC-mediated cold sensitivity. However, we found that an rpoS mutation had no effect on umuDC-mediated inhibition of growth at 30°C (data not shown). Another possibility was suggested by the data of Taddei et al. (54), who found that the SOS regulon is induced under starvation conditions in a cyclic AMP (cAMP)-dependent and rpoS-independent manner. Once again, we found that the Δcya mutation, which abolishes the ability to produce cAMP, had no effect on umuDC-mediated growth inhibition at 30°C (data not shown). Thus, umuDC-dependent inhibition of growth at 30°C proceeds by a pathway that is independent of rpoS- and cya-regulated genes.

The best-characterized protein to date known to be involved in the transition from stationary phase to exponential growth is the fis gene product, which undergoes a striking 500-fold induction in the first two cell divisions after nutrient upshift (1). We therefore examined whether Fis plays a role in the phenomenon of umuDC-dependent growth inhibition at 30°C. To do this, we employed a quantitative transformation assay, which has proven to be the most sensitive measure of growth inhibition at 30°C (39). In this assay, either pBR322 or a derivative carrying the umuD⁺C⁺ operon (pSE117) was transformed into lexA(Def) strains which differed in their fis genotype. The data in Table 2 show that there was a 400-fold increase in the cold sensitivity conferred by pSE117 (umuD⁺C⁺; pBR322) in a lexA(Def) fis strain relative to that in a lexA(Def) fis⁺ strain. Reciprocally, overexpression of Fis from the plasmid pRJ400 (62) partially suppressed umuDC-mediated inhibition of growth at 30°C (data not shown). These results suggest that Fis or a Fis-regulated function counteracts the growth inhibition at 30°C conferred by high levels of the umuDC gene products.

TABLE 2.

Effect of fis on the transformation efficiency of umuDC-expressing plasmids in lexA(Def) strains

Strain [relevant genotype]	Plasmid and genotype	Transformants/ml (30°C/42°C) (mean ± SD)
GW8018 [lexA(Def) fis+]	pBR322	1.04 ± 0.01
	umuD+C+; pBR322	0.029 ± 0.001
GW8034 [lexA(Def) fis]	pBR322	0.98 ± 0.02
	umuD+C+; pBR322	0.000069 ± 0.000005

To determine whether high levels of the umuDC gene products directly alter Fis protein levels at 30°C, we analyzed the levels of Fis protein during growth of a culture constitutively expressing umuDC by immunoblot analyses with anti-Fis antibody (1). We also examined whether constitutive expression of umuDC altered Fis transcriptional activity at 30°C by analyzing the expression of two Fis-regulated lacZ fusions, frg-733::TnphoA′-4 and proP-104::TnphoA′-4 (61, 62). No change in Fis protein levels, the pattern of Fis protein induction, or Fis regulation of either of these fusions was conferred by high levels of the umuDC gene products (data not shown). Therefore, the inhibition of the transition to exponential growth mediated by the umuDC gene products was not due to a direct effect of UmuD and UmuC on either Fis protein levels or the regulatory activity of Fis that is involved in controlling frg or proP expression. Thus, the interaction between umuDC and fis that affects growth control at 30°C must occur either through another activity of Fis or indirectly through a gene regulated by Fis.

Counteracting fis is central to the DNA damage tolerance in stationary phase conferred by the umuDC gene products.

Our observations that a Fis-dependent function counteracts the umuDC-dependent inhibition of growth at 30°C (Table 2) suggested the possibility that the umuDC gene products confer UV resistance, at least partially, by counteracting a fis-dependent process and thereby regulating growth in cells that have experienced DNA damage. We therefore examined whether the inhibition of a Fis-dependent activity by the umuDC gene products led to increased survival after UV irradiation of cultures in stationary phase. The increased UV sensitivity of ΔumuDC strains (Fig. 5, GW8023) has been observed previously (25, 58) and has been attributed to the inability of these strains to carry out translesion synthesis, the mechanistic basis of SOS mutagenesis (9, 43). Our recent studies suggest that a component of the resistance to UV killing is indeed provided by umuDC-dependent translesion synthesis (which requires UmuD′) but that a second component is due to a DNA damage checkpoint (which requires UmuD) (38). Interestingly, we found that inactivation of fis largely suppressed the UV sensitivity of a ΔumuDC strain that had been UV irradiated while in stationary phase (Fig. 6, GW8038). This suggests that the UV sensitivity of ΔumuDC strains in stationary phase is due to a fis-dependent activity and that the umuDC gene products increase UV resistance by inhibiting this activity. The fis ΔumuDC strain was nonmutable by UV light, just like a fis⁺ ΔumuDC strain, indicating that the increased survival of the fis ΔumuDC strain was not mediated through the restoration of the translesion synthesis process that is responsible for UV mutagenesis (data not shown). Loss-of-function fis mutants exhibit an increase in the duration of lag phase and a reduction in growth rate following nutrient upshift (40), so that they may not require the growth delay normally imposed by the umuDC gene products. Furthermore, inactivation of fis alone did not have a striking effect on UV resistance (Fig. 5, GW8037) and did not affect SOS mutagenesis (data not shown). These results are consistent with the model that the umuDC gene products inhibit growth and thereby increase survival of stationary-phase cultures exposed to UV irradiation by counteracting a Fis-mediated activity that promotes growth.

FIG. 6.

Effect of fis::767 on survival after UV irradiation in stationary phase. Cultures were grown at 37°C. Symbols: ■, GW2771 (fis⁺ umuD⁺C⁺); ●, GW8023 (fis⁺ ΔumuDC); ▵, GW8037 (fis::767 umuD⁺C⁺); ○, GW8038 (fis::767 ΔumuDC). Survival of cells UV irradiated after 11 h in stationary phase is shown.

Counteracting fis is also central to the umuDC-dependent increase in survival of UV-irradiated exponentially growing cells.

There are certain parallels between the umuDC-dependent resistance to killing observed when UV-irradiated stationary-phase cells are given a nutrient upshift (see above) and the component of the umuDC-dependent resistance to UV killing of exponentially growing cells that is mediated by uncleaved UmuD and UmuC (38). In both bacterial growth states, the increased DNA damage tolerance is correlated with a delay in the resumption of growth after DNA-damaging treatment and the phenomena appear to be the result of a novel activity requiring uncleaved UmuD and UmuC rather than UmuD′ and UmuC. These parallels suggest that similar mechanisms could be responsible for the umuDC-dependent regulation of growth after DNA damage in exponentially growing and stationary-phase cells. We therefore examined whether the inhibition of a Fis-dependent activity is involved in the UV resistance conferred by the umuDC gene products in exponentially growing cultures (Fig. 7). Similar to the result in stationary-phase cultures, the inactivation of fis suppressed the UV sensitivity of ΔumuDC cultures in exponential growth (Fig. 7). This suggests that a conceptually similar mechanism of inhibiting growth by inhibiting a Fis-dependent activity may play some role in umuDC-mediated UV resistance in exponentially growing cells.

FIG. 7.

Effect of fis::767 on survival after UV irradiation of exponentially growing cultures. Cultures were grown at 37°C. Symbols: ■, GW2771 (fis⁺ umuD⁺C⁺); ●, GW8023 (fis⁺ ΔumuDC); ▵, GW8037 (fis::767 umuD⁺C⁺); ○, GW8038 (fis::767 ΔumuDC). Survival of cells UV irradiated in exponential growth is shown.

DISCUSSION

The results presented in this paper lead us to propose that the umuDC gene products play an unexpected role in the resistance of E. coli to killing by UV irradiation by regulating growth after cells have experienced DNA damage in stationary phase. The umuDC gene products, expressed from the single chromosomal copy of the operon and regulated in an SOS-dependent manner, increase the survival of UV-irradiated stationary-phase cells in a fashion which is correlated with a significant umuDC-dependent increase in the length of the lag phase upon nutrient upshift. Such a delay could increase DNA damage tolerance by allowing additional time for DNA repair before the cell attempts to replicate DNA directly at the site of lesions, thereby avoiding possible mutations as well as possibly deleterious secondary consequences resulting from failed attempts to replicate the damaged DNA. Our results suggest that it is UmuC and UmuD, the form inactive in SOS mutagenesis, that function to regulate the transition to rapid growth upon UV irradiation of stationary-phase cells. Such a mechanism to regulate growth upon nutrient upshift could also increase the viability of stationary-phase cells that have accumulated DNA damage while quiescent as a result of endogenous, rather than exogenous, DNA-damaging agents.

In exponentially growing cells, we have suggested that the umuDC gene products play two distinct and temporally separated roles in DNA damage tolerance (38). In our proposed model, the first of these, which similarly requires the uncleaved UmuD protein and UmuC, delays the recovery of DNA replication and cell growth after DNA damage, thereby allowing additional time for accurate repair systems to remove or process the damage before replication is attempted. RecA-mediated cleavage of UmuD to UmuD′ then acts as a molecular switch that permits the umuDC gene products to carry out their second role, in which UmuD′ and UmuC aid in the resumption of DNA replication by participating in translesion synthesis, a potentially mutagenic process that enables a cell to cope with unrepaired or irreparable lesions. The results we have reported here concerning cells that have suffered DNA damage while in stationary phase and are then exposed to a nutrient upshift are consistent with the umuDC gene products similarly playing two distinct and temporally separated roles in DNA damage tolerance, a growth delay role requiring UmuD and a translesion synthesis role requiring UmuD′. Different physiological and biochemical roles for uncleaved UmuD and UmuD′ are supported by our evidence that there are substantial differences in the structural conformations of the UmuD₂ and UmuD′₂ dimers (23, 28–30; A. Guzzo and G. C. Walker, unpublished results). Furthermore, the differential proteolytic susceptibilities of the various forms of the umuD⁺ gene product (UmuD₂ homodimers are susceptible to lon⁺-dependent degradation, UmuD′₂ homodimers are relatively stable, and the UmuD′ that is in UmuD-UmuD′ heterodimers is susceptible to clpX⁺P⁺-dependent degradation [16, 22]) suggest an additional mechanism for how the cell may modulate umuDC-dependent activities. A role for the umuDC gene products in DNA damage tolerance besides participating in the potentially mutagenic process of translesion synthesis is intriguing because of the observation that several bacterial species that carry umuDC homologs are nonmutable by UV light (49, 60).

The ability of the umuDC gene products to increase the survival of cells that have suffered DNA damage while in stationary phase could be very important because the natural environments of prokaryotic cells are often characterized by limited amounts of nutrients in which the opportunity to replicate exponentially, due to an increase in the available nutrients, is experienced only occasionally (24). Thus, the majority of prokaryotic cells in nature exist in a quiescent state that is not unlike stationary phase of cultures grown under laboratory conditions. Stationary-phase cells, although more resistant to DNA-damaging agents such as hydrogen peroxide or alkylating agents than exponentially growing cells (24), nevertheless might accumulate DNA lesions while quiescent due to exposure to endogenous or exogenous DNA-damaging agents. This is consistent with the finding that the SOS response, and presumably the umuDC operon, is induced in a fraction of the cells in a stationary-phase culture that has not been exposed to exogenous DNA-damaging treatments (54). The accumulated DNA damage in stationary-phase cells would pose a significant problem when the bacteria experience a nutrient upshift and attempt to rapidly divide, unless there was a mechanism to inhibit growth until DNA repair has been completed.

Our results suggest that the mechanism by which the umuDC gene products increase UV resistance in stationary-phase and exponentially growing cells involves, at least in part, counteracting a Fis-dependent activity. The Fis protein, the levels of which increase dramatically within the first two cell divisions after nutrient upshift (1), has been implicated in the regulation of the transition from stationary phase to exponential growth (40). Fis, a small DNA-binding protein, plays many roles in the regulation of cell growth (15). For example, Fis is involved in DNA replication (14, 21), the growth phase-specific regulation of the supercoiling level of DNA and of the expression of approximately 40 genes (18, 48, 62), and the regulation of the synthesis of components of the translational machinery in response to growth rate (35, 46). Counteracting one or more Fis-dependent functions such as these by the umuDC gene products could lead to the inhibition of growth.

Mutations that inactivate fis cause an increase in the duration of the lag phase and a reduction in growth rate following nutrient upshift (40). This mutant phenotype parallels the physiological consequence of constitutive expression of umuDC from a multicopy plasmid, which inhibits the transition to exponential growth at 30°C, resulting in the inhibition of growth at the restrictive temperature of 30°C. The fact that the inactivation of fis exacerbates umuDC-mediated cold sensitivity suggests that the umuDC-mediated inhibition of the transition from stationary phase to exponential growth is counteracted by an activity of Fis. In addition, Fis levels have been shown to decrease markedly during prolonged starvation (40). Thus, the time-dependent decrease of Fis during prolonged incubation in stationary phase could account, at least in part, for the increase in umuDC-dependent growth inhibition at 30°C observed when stationary-phase cultures incubated for prolonged periods experience a nutrient upshift. The fis mutant phenotype also parallels the umuDC-dependent delay in the resumption of rapid growth upon nutrient upshift of UV-irradiated stationary-phase cultures (which is conferred by physiologically relevant levels of the umuDC gene products). The observation that a fis mutation suppresses the UV sensitivity of ΔumuDC strains supports the hypothesis that certain mechanisms which delay rapid growth after UV irradiation can thereby function to increase UV resistance.

Our observation that umuDC-overexpressing cells that had previously been in stationary phase exhibit an immediate inhibition of DNA replication upon a shift to 30°C, whereas corresponding exponentially growing cells do not, indicates that the umuDC gene products have an inherent capacity to inhibit, at least partially, the attempts of cells that were previously in stationary phase to carry out DNA synthesis upon a nutrient upshift. The rapidity of the inhibition suggests that it is the elongation phase of DNA synthesis that is being inhibited rather than initiation. Since these cells were not SOS induced, uncleaved UmuD would be expected to be the predominant, if not the exclusive, form of the umuD gene products present in the cells during these experiments. It is not yet clear whether the growth phase dependence of the inhibition of DNA synthesis is due to growth phase-specific proteins, to the presence of multiple partially replicated chromosomes in the exponentially growing cells, or to some other reason. However, we have recently shown that UmuD and UmuD′ have differential abilities to interact with components of E. coli's replicative polymerase, pol III, so it seems reasonable that the inhibition of DNA synthesis that we observed in these experiments resulted from a direct interaction of uncleaved UmuD and UmuC with pol III (53). Our observation that fis mutations exacerbate the cold sensitivity of cells that overexpress umuDC while suppressing the UV sensitivity of ΔumuDC cells is consistent with the umuDC gene products exerting opposing effects on DNA replication, but this possibility still needs to be tested. Further work will also be required to establish whether the umuDC-dependent growth delay observed when UV-irradiated stationary-phase cells experience a nutrient upshift is exerted through a direct effect of uncleaved UmuD and UmuC on pol III. Although we failed to detect a umuDC-dependent effect in DNA synthesis in such cells in our initial experiments, it has recently become clear that E. coli has multiple mechanisms for restarting DNA replication after DNA damage (44). It could be that UmuD and UmuC only inhibit the subset of replication events in which pol III is located directly at the site of the lesion so that it may be necessary to use a mutant background in order to detect such umuDC-dependent effects (44).

The proposed activity of the umuDC gene products of inhibiting the transition to exponential growth of stationary-phase cells exposed to a DNA-damaging treatment has certain similarities to a eukaryotic DNA damage checkpoint (10, 17, 34, 51). The similarity is most strong to that of cells that have suffered DNA damage while in G₀ and are then given the opportunity to reenter the active cell cycle. The recA⁺ gene product acts as the sensor of DNA damage in E. coli, analogous to RAD9 or POL2 in Saccharomyces cerevisiae (34), that initiates the signal transduction cascade that leads to the ultimate effect on the cell cycle. The transduction of the signal by RecA results in the cleavage of LexA, which, in turn, results in the induction of the SOS regulon, including the umuDC operon (17). The umuDC gene products act in stationary-phase cells that have experienced DNA damage to delay growth, i.e., entry into the cell cycle. Counteracting an activity of Fis is central to the DNA damage tolerance conferred by the umuDC gene products upon exposure to DNA-damaging agents in both stationary-phase and exponentially growing cultures. Since bacteria might accumulate DNA damage in stationary phase, mechanisms that delay growth and DNA replication may play vital roles in increasing survival by allowing time for the DNA damage to be repaired prior to the resumption of growth.