Non-growing Escherichia coli cells starved for glucose or phosphate use different mechanisms to survive oxidative stress.

Recent data suggest that superoxide dismutases are important in preventing lethal oxidative damage of proteins in Escherichia coli cells incubated under aerobic, carbon starvation conditions. Here, we show that the alkylhydroperoxide reductase AhpCF (AHP) is specifically required to protect cells incubated under aerobic, phosphate (Pi) starvation conditions. Additional loss of the HP-I (KatG) hydroperoxidase activity dramatically accelerated the death rate of AHP-deficient cells. Investigation of the composition of spent culture media indicates that DeltaahpCF katG cells leak nutrients, which suggests that membrane lipids are the principal target of peroxides produced in Pi-starved cells. In fact, the introduction of various mutations inactivating repair activities revealed no obvious role for protein or DNA lesions in the viability of ahp cells. Because the death of ahp cells was directly related to ongoing aerobic glucose metabolism, we wondered how glycolysis, which requires free Pi, could proceed. 31P nuclear magnetic resonance spectra showed that Pi-starved cells consumed Pi but were apparently able to liberate Pi from phosphorylated products, notably through the synthesis of UDP-glucose. Whereas expression of the ahpCF and katG genes is enhanced in an OxyR-dependent manner in response to H2O2 challenge, we found that the inactivation of oxyR and both oxyR and rpoS genes had little effect on the viability of Pi-starved cells. In stark contrast, the inactivation of both oxyR and rpoS genes dramatically decreased the viability of glucose-starved cells.

Control of the LexA regulon by pH: evidence for a reversible inactivation of the LexA repressor during the growth cycle of Escherichia coli.

The LexA repressor controls the expression of several genes, including lexA, recA, and sfiA, which are induced when exponentially growing bacteria are exposed to DNA-damaging agents. Induction of this so-called SOS response takes place while LexA is cleaved in a reaction that requires the RecA protein and damaged DNA. We have shown that large fluctuations in the cellular concentration of the LexA repressor and in the rate of transcription of the sfiA gene also occur spontaneously during bacterial growth in complex medium such as LB. The possibility that changes in external or internal pH may explain these fluctuations has been explored. A consistent pattern was established whereby conditions leading to either increased or decreased pH were associated with altered expression of the lexA and sfiA genes. These data can be explained by a model in which the LexA repressor exists in either of two forms in equilibrium: a form favoured at homeostatic internal pH, which has a low affinity for the operators of LexA-controlled genes; and a form accumulated in response to a transient decrease in internal pH, which has a high affinity for operators.

Phosphate starvation and low temperature as well as ultraviolet irradiation transcriptionally induce the Escherichia coli LexA-controlled gene sfiA.

The LexA repressor controls the expression of several SOS genes, such as lexA, recA and sfiA, which are induced by DNA damage. Induction results from the activation of the RecA protein that favours the cleavage and thus the inactivation of LexA. It has been shown that the activation of RecA results from its binding to damaged DNA. It is therefore believed that in growing bacteria, in the absence of any DNA-damaging treatment, the intracellular level of LexA remains stable at a high basal level and, hence, SOS genes are expressed at relatively low basal levels. In contrast, we show here that the intracellular level of LexA and the rate of transcription of the sfiA gene may vary markedly throughout the growth cycle of wild-type Escherichia coli. We provide evidence that such changes result from two superimposed processes: proteolytic cleavage of LexA upon dilution of stationary phase bacteria, and increase in strength of the promoters of the lexA and sfiA genes when bacteria approach the stationary phase. We show that a signal which strongly increases the strength of the sfiA gene promoter is starvation for phosphate. Such induction was not significantly affected by mutations either in phoB (encoding the transcriptional regulator for the phosphate regulon) or rpoS (encoding a putative stationary phase-specific sigma factor). However, sfiA induction by phosphate starvation appeared to be markedly inhibited by the presence of the osmZ205 mutation which alters the histone-like protein H-NS, suggesting that changes in the DNA structure may play a role in signal transduction during phosphate starvation. As previously shown for several processes which are controlled by H-NS, induction of sfiA was modulated by growth temperature.

Role of the histone-like proteins OsmZ and HU in homologous recombination.

The HU protein of Escherichia coli has been implicated in various site-specific recombination reactions. Moreover, recent data suggest that HU may also participate in homologous recombination. In particular, it has been shown that P1 transduction is inhibited in the absence of HU [Kano and Imamoto, Gene 89 (1990) 133-137]. In contrast, we found that transductional recombination and conjugational recombination were almost normal in hupA hupB mutants. However, it appeared that the recombination proficiency of hupA hupB mutant bacteria was reduced tenfold in an intrachromosomal recombination assay. Moreover, we found that intrachromosomal recombination was reduced tenfold in a gyrB226 strain and by more than 100-fold in an osmZ205 strain. The gyrB226 mutation affects the DNA gyrase activity, while mutations in osmZ are highly pleiotropic, affecting the expression of a variety of genes and increasing the frequency of site-specific inversion events. Since it has been shown that the hupA hupB mutations, like the gyrB226 mutation, decrease the level of DNA supercoiling, whereas the osmZ205 mutation increases the level of DNA supercoiling, it appears that the histone-like proteins HU and OsmZ may play a key role in intrachromosomal recombination by affecting the DNA topology.

Properties of RecA441 protein reveal a possible role for RecF and SSB proteins in Escherichia coli.

We examined the possibility that the recA441 mutation, which partially suppresses the UV sensitivity of uvr recF mutant bacteria, exerts its effect by coding for an altered RecA protein that competes more efficiently than the RecA+ protein with SSB for ssDNA in vivo. Using an assay measuring recombination between UV-damaged lambda DNA and intact homologous DNA, we found that the introduction of the recA441 mutation partially suppressed the defects in recombination in bacteria lacking RecF activity but not in bacteria with excess SSB, although recombination was affected more in recF mutants than in bacteria overproducing SSB. These results therefore do not support the hypothesis that RecA441 protein, or RecA protein with the help of RecF protein, is required during recombination of UV-damaged DNA to compete with SSB for ssDNA.

Inhibition of cell division in hupA hupB mutant bacteria lacking HU protein.

Escherichia coli hupA hypB double mutants that lack HU protein have severe cellular defects in cell division, DNA folding, and DNA partitioning. Here we show that the sfiA11 mutation, which alters the SfiA cell division inhibitor, reduces filamentation and production of anucleate cells in AB1157 hupA hupB strains. However, lexA3(Ind-) and sfiB(ftsZ)114 mutations, which normally counteract the effect of the SfiA inhibitor, could not restore a normal morphology to hupA hupB mutant bacteria. The LexA repressor, which controls the expression of the sfiA gene, was present in hupA hupB mutant bacteria in concentrations half of those of the parent bacteria, but this decrease was independent of the specific cleavage of the LexA repressor by activated RecA protein. One possibility to account for the filamentous morphology of hupA hupB mutant bacteria is that the lack of HU protein alters the expression of specific genes, such as lexA and fts cell division genes.

New recA mutations that dissociate the various RecA protein activities in Escherichia coli provide evidence for an additional role for RecA protein in UV mutagenesis.

To isolate strains with new recA mutations that differentially affect RecA protein functions, we mutagenized in vitro the recA gene carried by plasmid mini-F and then introduced the mini-F-recA plasmid into a delta recA host that was lysogenic for prophage phi 80 and carried a lac duplication. By scoring prophage induction and recombination of the lac duplication, we isolated new recA mutations. A strain carrying mutation recA1734 (Arg-243 changed to Leu) was found to be deficient in phi 80 induction but proficient in recombination. The mutation rendered the host not mutable by UV, even in a lexA(Def) background. Yet, the recA1734 host became mutable upon introduction of a plasmid encoding UmuD*, the active carboxyl-terminal fragment of UmuD. Although the recA1734 mutation permits cleavage of lambda and LexA repressors, it renders the host deficient in the cleavage of phi 80 repressor and UmuD protein. Another strain carrying mutation recA1730 (Ser-117 changed to Phe) was found to be proficient in phi 80 induction but deficient in recombination. The recombination defect conferred by the mutation was partly alleviated in a cell devoid of LexA repressor, suggesting that, when amplified, RecA1730 protein is active in recombination. Since LexA protein was poorly cleaved in the recA1730 strain while phage lambda was induced, we conclude that RecA1730 protein cannot specifically mediate LexA protein cleavage. Our results show that the recA1734 and recA1730 mutations differentially affect cleavage of various substrates. The recA1730 mutation prevented UV mutagenesis, even upon introduction into the host of a plasmid encoding UmuD* and was dominant over recA+. With respect to other RecA functions, recA1730 was recessive to recA+. This demonstrates that RecA protein has an additional role in mutagenesis beside mediating the cleavage of LexA and UmuD proteins.

KIN, a mammalian nuclear protein immunologically related to E. coli RecA protein.

A polypeptide of about 120 kDa, called KIN, has been identified in rat FR 3T3 cells by immunoblotting using affinity-purified antibodies against the RecA protein of Escherichia coli (38 kDa). The KIN protein as shown by fluorescent light microscopy and electron microscopy is essentially concentrated in the nucleus. Its level is higher in proliferating than in quiescent cells. Cell treatment with mitomycin C increases the level of the KIN protein. We sought similar proteins in other mammalian cells. Proteins with the same electrophoretic mobility were detected in mouse, monkey and human cell lines as well as in rat and mouse embryos.

RecA protein-promoted cleavage of LexA repressor in the presence of ADP and structural analogues of inorganic phosphate, the fluoride complexes of aluminum and beryllium.

Complexes formed from A13+ or Be2+ and fluoride inhibit the single-stranded DNA-dependent ATPase activity of RecA protein. In contrast, poly(dT)-RecA-ADP complexes, which are inactive for cleavage of LexA protein, become fully active in the presence of AlF4- or BeF3- ions. These data suggest that fluoride complexes of aluminum and beryllium (called herein X) convert RecA-ADP complexes, which bind weakly to single-stranded DNA, into RecA-ADP-X complexes, which bind tightly to single-stranded DNA, the ADP-X moiety behaving as a nonhydrolyzable analogue of ATP. We propose that AlF4- and BeF3- ions act as analogues of inorganic phosphate by binding to the site of the gamma-phosphate of ATP on RecA-ADP complexes, hence mimicking the single-stranded DNA-RecA-ADP-Pi transition state. We conclude that the elementary reaction that switches RecA protein from a high affinity single-stranded DNA binding state to a low affinity single-stranded DNA binding state is not ATP hydrolysis per se but Pi release.

Overproduction of single-stranded-DNA-binding protein specifically inhibits recombination of UV-irradiated bacteriophage DNA in Escherichia coli.

Overproduction of single-stranded DNA (ssDNA)-binding protein (SSB) in uvr Escherichia coli mutants results in a wide range of altered phenotypes. (i) Cell survival after UV irradiation is decreased; (ii) expression of the recA-lexA regulon is slightly reduced after UV irradiation, whereas it is increased without irradiation; and (iii) recombination of UV-damaged lambda DNA is inhibited, whereas recombination of nonirradiated DNA is unaffected. These results are consistent with the idea that in UV-damaged bacteria, SSB is first required to allow the formation of short complexes of RecA protein and ssDNA that mediate cleavage of the LexA protein. However, in a second stage, SSB should be displaced from ssDNA to permit the production of longer RecA-ssDNA nucleoprotein filaments that are required for strand pairing and, hence, recombinational repair. Since bacteria overproducing SSB appear identical in physiological respects to recF mutant bacteria, it is suggested that the RecF protein (alone or with other proteins of the RecF pathway) may help RecA protein to release SSB from ssDNA.

[Mutagenesis and responses induced by DNA damage in Escherichia coli: principle of bacterial tests for detecting carcinogenic or antitumoral substances]. / Mutagénèse et réponses induites par l'endommagement de l'ADN chez Escherichia coli: principe des tests bactériens pour la détection des substances cancérogènes ou antitumorales.

Various bacterial tests are used for the detection of potential carcinogenic or antitumoral agents. These so-called short-term tests measure different processes triggered by DNA damage: mutagenesis ("Mutatest"), induction of genes that belong to the recA-lexA regulon ("Umu-test" and "SOS-chromotest"), and induction of provirus ("Inductest"). Although these tests share common properties, each exhibits characteristic properties which are revealed by analyzing the various DNA repair mechanisms brought about in E. coli.

Effects of overproduction of single-stranded DNA-binding protein on RecA protein-dependent processes in Escherichia coli.

Overproduction of single-stranded DNA-binding protein (SSB) in Escherichia coli led to a decrease in the basal level of repressor LexA. Expression of the LexA-controlled genes was increased differentially, depending on the affinity of the LexA repressor for each promoter: expression of the recA and sfiA genes was increased 5-fold and 1.5-fold, respectively. Despite only a slight effect on expression of sfiA, which codes for an inhibitor of cell division, bacteria overproducing SSB produced elongated cells. In fact, the effect on cell shape appeared to be essentially independent of the expression of the sfiA and recA genes. Bacteria overproducing SSB were therefore phenotypically similar to bacteria partially starved of thymine, in which filamentation results from both sfiA-dependent and sfiA-recA-independent pathways. These data indicate that excess SSB acts primarily by perturbing DNA replication, thereby favoring gratuitous activation of RecA protein to promote cleavage of LexA protein. When bacteria overproducing SSB were exposed to a DNA-damaging agent such as ultraviolet light or mitomycin C, the recA and sfiA genes were fully induced. Induction of the sfiA gene occurred, however, at higher doses in bacteria overproducing SSB protein than in bacteria with normal levels of SSB. Whereas the efficiency of excision repair was apparently increased by excess SSB, the efficiency of post-replication recombinational repair was reduced as judged by a decrease in the recombination proficiency between a prophage and ultraviolet-irradiated heteroimmune infecting phage. Following induction of ssb+ bacteria with mitomycin C, the cellular content of SSB was slightly increased. These results provide evidence that SSB modulates RecA protein-dependent activities in vivo. It is proposed that SSB favors the formation of short complexes of RecA protein and single-stranded DNA that mediate cleavage of the LexA and lambda repressors, while it delays the formation of long nucleoprotein filaments, thereby slowing down RecA-promoted recombinational events in uninduced as well as in induced bacteria.

Role of Escherichia coli RecA protein in SOS induction and post-replication repair.

The RecA protein of Escherichia coli plays a central role in DNA repair mechanisms. When it is incubated with single-stranded DNA and a nucleoside triphosphate, the purified RecA protein acts both by promoting cleavage of the LexA protein, the repressor of the SOS genes, and by catalyzing strand exchange between a variety of DNA molecules. A model for the regulation of the activity of the RecA protein in a cell exposed to a DNA damaging treatment is proposed.

A yeast protein analogous to Escherichia coli RecA protein whose cellular level is enhanced after UV irradiation.

In Saccharomyces cerevisiae, a protein was recognized by polyclonal antibodies raised against homogeneous Escherichia coli K 12 RecA protein. The cellular level of the yeast protein called RecAsc (molecular weight 44 kDa, pI 6.3), was transiently enhanced after UV irradiation. Protease inhibitors were required to minimize degradation of the RecAsc protein during cell lysis. The RecAsc protein exhibited similar basal levels and similar kinetics of increase after UV irradiation in DNA-repair proficient (RAD+) strains carrying mitochondrial DNA or not (rho0). This was also true for the following DNA-repair deficient (rad-) strains: rad2-6 rad6-1 rad52-1, a triple mutant blocked in three major repair pathways; rad6-delta, a mutant containing an integrative deletion in a gene playing a central role in mutagenesis; pso2-1, a mutant that exhibits a reduced rate of mutagenesis and recombination after exposure to DNA cross-linking agents.

RecA protein--promoted lambda repressor cleavage: complementation between RecA441 and RecA430 proteins in vitro.

Induction of prophage lambda occurs in recA441 mutant lysogens after a shift to 42 degrees C in the presence of adenine. If the synthesis of RecA441 protein is maintained at a low basal level by the presence of a second mutation in the recA441 gene, recA453, induction of prophage lambda is prevented. The ability to induce prophage lambda is restored by the introduction, on a transducing phage, of a second recA gene carrying the recA430 mutation; by itself, the RecA430 protein is devoid of activity against the lambda repressor (Rebollo et al. 1984). In order to explain how the RecA430 protein might complement the RecA441 protein to provide lambda repressor cleavage in a recA453-441 (recA430) diploid lysogen, we characterized the cleavage reaction catalysed by a mixture of these proteins in vitro. Our results suggest that, in the presence of dATP, the RecA441 and RecA430 proteins form mixed multimers on single-stranded DNA, in which the RecA441 protein molecules enhance the DNA binding affinity of RecA430 protein molecules, but RecA430 protein molecules support no cleavage of the lambda repressor. Although the effects of the RecA430 and single-strand binding (SSB) proteins are similar in vitro, we show that the SSB protein cannot substitute for the RecA430 protein in restoring lambda repressor cleavage in a recA453-441 lysogen. Comparison of the stimulatory effect of long single-stranded DNA with that of (dA)14 oligonucleotides on the RecA441 protein-directed cleavage of the lambda repressor in the presence of various nucleoside triphosphates (NTPs) indicates that the cooperative binding of the RecA441 protein to single-stranded DNA stabilizes the RecA protein-DNA complexes so that they remain intact long enough to support cleavage of the lambda repressor. We conclude that the low basal level of the RecA441 protein in a recA453-441 cell is sufficient to cleave the lambda repressor, under conditions where a normal basal level of RecA430 protein is also present allowing the formation of mixed multimers on single-stranded DNA regions normally present in the cell.

Restoration of RecA protein activity by genetic complementation.

Bacteria carrying either recA430 or recA453-441 mutations are sensitive to UV-irradiation since they amplify the synthesis of RecA protein either poorly or not at all. We show here that, in a recA453-441 (recA430) heterodiploid, UV-resistance and amplification of RecA430 protein were restored, indicating that the cellular level of RecA-associated protease activity was high enough to inactivate LexA repressor. Prophage 434 repressor was also extensively inactivated, whereas RecA430 protein alone cannot cleave this substrate. On the other hand, during growth of the recA453-441(recA430) heterodiploid at 42 degrees C in the presence of adenine, a treatment activating only RecA441 protein, RecA441 protease activity was as high as in a recA441 haploid. In contrast, following this inducing treatment, there was no complementation between RecA441 and RecA+ proteins in a recA453-441(recA+) heterodiploid. These results indicate that multimerization of RecA protein molecules results in a functional interaction that, in some combination between RecA protein subunits, may enhance RecA-associated protease activity.

Prophage phi 80 is induced in Escherichia coli K12 recA430.

In contrast to prophage lambda, wild-type prophage phi 80 was induced by UV-irradiation or thymine deprivation in recA430 mutants of E. coli K12. There was no induction of prophage phi 80 in two recombination-deficient mutants recA13 and recA99. Phage phi 80ind3, a non-inducible derivative in a rec+ was not induced in a recA430 lysogen. Two other lambdoid prophages were tested for UV-induction in recA430 lysogens: in common with lambda prophage, 434 was not induced whereas prophage 21 was induced in 1% of the cells. Induction of RecA430 protein synthesis was 30% of that observed in recA+ bacteria at 30 min of post-irradiation incubation, indicating that LexA repressor had been cleaved by RecA430 protease. In lexA1 recA430 and lexA1 recA+ bacteria, RecA protein synthesis was not amplified, yet, prophage phi 80 was fully induced. If phi 80cI repressor is inactivated by cleavage by RecA430 protease as is LexA repressor, RecA430 protease can inactivate all the molecules of phi 80cI repressor, its basal level being high enough in a recA430 lysogen. In such a lysogen, a fraction only of 21cI and LexA repressors are cleaved but no molecules of either lambda cI or 434cI repressor. We postulate that RecA430 protein has an altered pattern of recognition of repressor molecules and a cleavage efficiency which is more efficient the more remote is the repressor conformation from that of lambda repressor.

A brief consideration of the SOS inducing signal.

SOS functions are induced in E. coli by treatments that damage cellular DNA or interrupt its synthesis. The biochemical basis of induction is activation of the specific proteolytic activity of recA protein, which then inactivates the lexA repressor. We discuss the development of the inducing signal in the cell.

Cleavage of lambda repressor and synthesis of RecA protein induced by transferred UV-damaged F sex factor.

Transfer of a UV-damaged F sex factor to a recipient lambda lysogen induces prophage lambda development. Under these conditions RecA protein synthesis was induced and lambda repressor cleaved, as observed upon direct induction, that is, when the recipient lambda lysogen was directly exposed to UV-light. The efficiency of induction of RecA protein synthesis in recipient bacteria which had received an irradiated F-lac factor was about 80% of that measured upon direct induction. We observed the simultaneous disappearance of lambda repressor and a slight production of cleavage fragments; quantitation by densitometric scanning of the autoradiogram after correction for the efficiency of transfer indicated that 55% of lambda repressor was cleaved. Transfer of UV-damaged Hfr DNA failed to induce RecA protein synthesis. A lambda phage vector carrying oriF, the cloned origin of F plasmid replication, after exposure to UV-light and infection of a recipient lysogen, induced RecA protein synthesis and a moderate but significant cleavage of lambda repressor. Indirect induction by UV-damaged F sex factor or phage lambda oriF resulted in biochemical cellular reactions similar to those observed upon direct induction. LexA repressor that negatively controls RecA protein synthesis appeared more susceptible to cleavage than did lambda repressor.

Cell survival, UV-reactivation and induction of prophage lambda in Escherichia coli K12 overproducing RecA protein.

The effect of the cellular level of RecA protein on the ability of E. coli K12 bacteria to (i) survive UV-irradiation (ii) promote UV-reactivation of UV-damaged phage lambda (iii) induce prophage lambda was determined in bacterial mutants with discrete increasing levels of RecA protein. The various levels of RecA protein were obtained by combining lexA and recA alleles. Except for the double mutant lexA3 recAo98, whose repair ability was 25% less than that observed in wild type bacteria, bacterial survival was proportional to the level of RecA protein measured after 90 min of incubation. In lexA3 recAo98 bacteria, RecA protein, at a constitutive high basal level, failed to compensate totally for the lack of LexA repressor cleavage; UV-reactivation of UV-damaged phage lambda was not restored; yet, prophage lambda was induced with 35% efficiency. Efficient UV-induction of prophage lambda is linked to the induction of lexA-controlled host processes that repair the UV-damaged prophage.