TEXT

Top Abstract Text References

The ruv locus of Escherichia coli consists of three genes, ruvA, ruvB, and ruvC, whose products are known to be involved in homologous genetic recombination and DNA repair (17, 25). Genetic and biochemical studies have revealed important roles for the Ruv proteins in the late, postsynaptic stage of recombination. During this stage, Holliday junctions made by RecA-mediated homologous pairing and strand exchange are processed into mature recombinant molecules. The RuvA protein acts in this process as a molecular matchmaker: by binding to a Holliday junction, it allows the RuvB helicase to bind and catalyze branch migration (22). The RuvC protein is an endonuclease that resolves Holliday intermediates into recombinant products (13). Recent genetic and biochemical data indicate that, in addition to RuvC, functional RuvA and RuvB are also needed for efficient Holliday junction resolution (7, 18, 23). It has therefore been suggested that RuvABC act together as a unique branch migration-resolution complex (7, 18, 23).

Cells carrying mutations in any of the ruv genes display moderate deficiencies in recombination and sensitivity to various DNA-damaging agents (16, 21). This phenotype of ruv mutants is extremely pronounced in a recBC sbcBC background, suggesting that RuvABC are indispensable for recombination and repair in the RecBCD-independent (RecF) pathway (3, 16). Furthermore, recBC sbcBC ruv mutants display severely reduced viability (16, 18), indicating the necessity for Holliday junction processing in normally growing recBC sbcBC cells.

In this work we have studied the cytological effects of introducing a ruvC mutation into the recBC sbcBC background. We have found that the ruvC mutation severely impairs cell growth, chromosome segregation, and cell division in exponentially growing recBC sbcBC mutants.

The E. coli strains used in this study are listed in Table 1. All strains except FC581 and TH446 are derivatives of AB1157. Mutant strains were constructed by P1 transduction, as described by Miller (19). Transductants were isolated on Luria-Bertani (LB) plates (19) supplemented with appropriate antibiotics (tetracycline [Tc], 10 µg/ml; kanamycin [Km], 25 µg/ml; chloramphenicol [Cm], 20 µg/ml) and tested for UV sensitivity. Bacterial cultures were grown in LB medium (19) at 37°C with shaking. Cell growth was monitored by measuring the optical density at 600 nm (OD₆₀₀). For cytological analysis, overnight bacterial cultures were diluted 10⁵-fold in fresh medium and grown until they reached an OD₆₀₀ of 0.2 (corresponding to approximately 14 mass doublings). At this point, cells were fixed with 0.1% OsO₄; their chromosomes were subsequently stained with 4',6-diamidino-2-phenylindole (DAPI) (0.5 µg/ml). The fixation and staining procedures, as well as preparation of slides for microscopy, were performed as described by Woldringh et al. (26). Cells were observed with a Zeiss Axiovert 35 microscope adjusted for combined phase-contrast and fluorescence microscopy. Photographs were taken with Ilford FP4 Plus (125 ASA) film. To determine total cell numbers, a portion of bacterial culture was fixed and stained as mentioned above, diluted in 0.2 M cacodylate buffer (pH 7.0), and then filtered through a polycarbonate membrane filter (0.2-µm pores) (Poretics). The filter (containing homogeneously distributed cells) was placed on a microscope slide; a drop of immersion oil was put on its surface and covered with a coverslip. Cells were visualized and counted with a fluorescence microscope.

As shown in Table 2, the presence of a ruvC mutation in the recBC sbcBC background resulted in severe reduction of cell viability and much slower growth. In addition, the recBC sbcBC ruvC mutant exhibited profound changes in cell morphology (Fig. 1D to G). The population of this mutant contained cells that were unusually variable in length, ranging from normal-sized small cells to exceptionally long filaments. The filaments, which comprised about 50% of the population, displayed striking defects in chromosome segregation. In many filaments, chromosomes were stacked into aggregates, with large regions of the cells completely devoid of DNA. About 20% of filaments contained only one large DNA mass located usually near the cell center. In other elongated cells, DNA was unequally distributed in two or more masses, some of which were several times larger than normal nucleoids. Because of chromosome nondisjunction, cell division produced numerous anucleate cells (Table 2), which were usually pinched off from the DNA-free ends of filaments (Fig. 1D). A low proportion of filaments (about 1%) displayed shape irregularities such as bulges and Y forms (Fig. 1E to G). Bulging was apparently associated with the extreme DNA condensation, which caused local distortion of the cell wall (Fig. 1E). The origin of the branched cells is less clear, since the branching was sometimes observed in cell regions containing no DNA (Fig. 1G).

View larger version (161K): [in this window] [in a new window]   FIG. 1.   Effects of ruvC mutation on chromosome segregation and cell morphology. Exponentially growing cells were fixed with OsO4, and their DNA was visualized with DAPI. (A) AB1157 (wild type). (B) JC7623 (recBC sbcBC). (C) LMM748 (ruvC). (D to G) LMM711 (recBC sbcBC ruvC). Arrows indicate the emergence of an anucleate cell (D) and branched cells (F and G). Bar, 10 µm.

Moderate cell division and chromosome segregation defects were also observed in cells of the parental recBC sbcBC and ruvC strains (Fig. 1B and C). However, the morphology of these strains was less severe than that of LMM711. Also, their viability and growth rates were much closer to those of the wild-type strain (Table 2).

Since ruv mutants are defective in recombination, we surmised that their chromosome segregation defect might have arisen from a recombination process which was properly initiated but then remained blocked due to unresolved Holliday intermediates. To test this hypothesis, we introduced a recA mutation into the recBC sbcBC ruvC strain LMM711. This mutation was expected to abolish completely early steps in recombination (17) and therefore to prevent the formation of Holliday intermediates. As shown in Fig. 2B, the recA mutation significantly improved both chromosome partitioning and cell division of recBC sbcBC ruvC cells. According to all cytological parameters measured, strain LMM832, which carried a ruvC mutation in combination with recA, was indistinguishable from the ruvC⁺ recA strain LMM824 (Table 2). The great majority of cells in both populations were small, with a DNA distribution close to normal. The filaments occasionally observed were on average much shorter than in strain LMM711, and their DNA was packed into small, distinct nucleoids (Fig. 2B). The overall improvement in chromosome segregation was also reflected by a 100-fold drop in the frequency of anucleate cells (Table 2). However, despite the more regular cell morphology of recA mutants, their viability and growth rate remained low.

View larger version (166K): [in this window] [in a new window]   FIG. 2.   Effects of recA, lexA (Ind), and recAo mutations on ruvC-associated chromosome segregation and cell division defects. (A) LMM824 (recBC sbcBC recA). (B) LMM832 (recBC sbcBC ruvC recA). (C) LMM820 [recBC sbcBC lexA (Ind)]. (D) LMM823 [recBC sbcBC lexA (Ind) ruvC]. (E) LMM834 [recBC sbcBC lexA (Ind) recAo]. (F) LMM835 [recBC sbcBC lexA (Ind) recAo ruvC]. Bar, 10 µm.

The RecA protein is involved not only in recombination but also in induction of the SOS response. When activated by DNA damage, RecA coprotease mediates cleavage of the SOS repressor (the LexA protein), thereby inducing some 20 genes involved in DNA repair, mutagenesis, and cell division control (24). Strains carrying mutations in either the sbcB or ruv genes display spontaneous derepression of the SOS response (1, 14, 20, 21). Although this derepression is moderate in single mutants, it could be stronger in an sbcB ruvC double mutant due to the additive effect of mutations. Taken together, these facts suggest that the suppressing effects of the recA mutation described above might not result solely from the abolition of recombination but also from the silencing of the SOS response. To check the possible role of the SOS response in the phenotypes of the recBC sbcBC ruvC mutant, we constructed strain LMM823, carrying in addition a lexA3 (Ind) mutation. This mutation was shown to produce a LexA repressor resistant to RecA coprotease, which therefore prevents induction of the SOS response (15, 24).

It can be seen from Table 2 and Fig. 2D that the lexA (Ind) mutation, like the recA mutation, significantly reduced the defects in chromosome partitioning and cell division typical of the recBC sbcBC ruvC mutant. However, suppression by the lexA (Ind) mutation was less efficient, allowing 10-fold-more anucleate cells and twice as many filaments as in the recA derivative. From these results we inferred that both recombinogenic and SOS-inducing functions of the RecA protein are responsible for the morphological defects observed in strain LMM711.

Since the recA gene is itself regulated as part of the SOS regulon (24), we wanted to check whether the lexA (Ind)-associated improvement of chromosome segregation and cell division is related to the low level of recA gene expression. To accomplish this, we used the operator-constitutive mutation recAo98, which allows constitutive high-level synthesis of RecA protein in both lexA⁺ and lexA (Ind) backgrounds (9). Strain LMM835, carrying a recAo mutation in addition to the lexA (Ind) allele, showed a phenotype quite similar to that of the recA⁺ lexA⁺ strain LMM711 (Table 2, Fig. 2F). Despite the difference in their lexA alleles, the two strains showed the same degree of chromosome nondisjunction, accompanied by frequent production of DNA-less cells. We thus concluded that recA gene overexpression is the principal SOS function which is responsible for the chromosome segregation defect observed in recBC sbcBC ruvC mutants. This finding corroborates our notion that chromosome segregation is affected by unsuccessful recombination initiated by the RecA protein (see above). Most probably, the excess RecA protein in strains LMM711 and LMM835 increases the frequency of recombination events, leading to accumulation of unresolved cross-links between chromosomes. Such a conclusion is in accord with previous findings that the elevated basal level of recA expression in recBC sbcBC mutants contributes to their proficiency in conjugal recombination (20).

Overexpression of the recA gene also restored the cell division defect, although the number of filaments in the recAo lexA (Ind) mutant was somewhat lower than in its recA⁺ lexA⁺ counterpart (Table 2). This difference could arise from LexA (Ind)-dependent repression of sfiA, an SOS gene that codes for a cell division inhibitor (11, 12). Like LMM711, some filaments of strain LMM835 showed aberrant shapes, including bulges and Y forms (not shown).

The results obtained with strain LMM835 show that ruvC-associated filamentation is largely independent of a LexA-controlled cell division inhibitor(s). Therefore, some alternative mechanism must also be involved in cell division control in ruvC mutants. Interestingly, filamentation was present in all strains that exhibited a significant chromosome segregation defect, indicating that unresolved chromosomes could exert a negative effect on the cell division process. This observation is in accord with the nucleoid occlusion model, which predicts control mechanisms that block cell division in the absence of proper chromosome segregation (26). However, this model was called into question by the recent finding that septation is not prevented by the presence of a nucleoid (6). Alternatively, the presence of unresolved recombination intermediates in ruvC cells could impede progression of replication forks (1), thereby inducing both SOS-dependent and -independent mechanisms of cell division inhibition. Such mechanisms have been shown to operate in UV-irradiated E. coli cells in which replication forks are blocked by pyrimidine dimers (5).

The unusual cell morphology, such as bulged and branched cells, was obviously produced by the same conditions that caused cell filamentation, suggesting that aberrant chromosome processing might have a more general effect on cell wall structure. Both types of aberrant cells have been described previously in cultures of certain ftsZ mutants that display altered septum formation (4). In these mutants, morphological anomalies were shown to result from the aberrant septation events. Cell branching was also observed with thy mutants grown under conditions of thymine limitation (27). In the latter case, morphological alterations were suggested to be caused by irregular nucleoid movement that affected the process of septum synthesis. A similar mechanism might explain the aberrant cell shapes observed in our experimental system.

Despite the great differences in their cytological patterns, all recBC sbcBC ruvC derivatives used in this work displayed similarly low levels of viability (Table 2). Therefore, the morphological defects observed in some of these strains cannot be the principal cause of cell death. The poor viability is more likely related to the recombination deficiency common to all recBC sbcBC ruvC strains used. Recombination proficiency is clearly critical for repair of DNA damage arising spontaneously in recBC sbcBC cells during exponential growth.