TEXT

Top Abstract Text References

The Escherichia coli peptide antibiotic microcin J25 (MccJ25) is active against gram-negative bacteria, including E. coli itself, Salmonella, and Shigella (17). Penetration of MccJ25 into E. coli cells is mediated by the outer membrane receptor FhuA, the TonB complex, and the SbmA protein (18, 19). Genetic studies have localized a cluster of three genes (mcjA, -B, and -C) necessary for the synthesis of the toxin and a fourth gene (mcjD) that confers immunity to a continuous 4.8-kb region of plasmid DNA (22). The nucleotide sequence of this region revealed four open reading frames corresponding to the four genes previously found (23). Comparison of the predicted McjA polypeptide with the amino acid sequence of MccJ25 indicated that mcjA encodes the primary structure of MccJ25 as a 58-amino-acid precursor. Subsequent removal of a 37-residue leader peptide and cyclization by head-to-tail linkage leads to the mature 21-amino-acid cyclic MccJ25 (3, 23). These processes could be mediated by the mcjB and mcjC gene products. The predicted mcjD gene product, which is highly similar to several membrane translocator proteins of the ABC exporters family, has been shown to be required for MccJ25 secretion, thus explaining its ability to confer MccJ25 immunity to susceptible cells (23).

We are interested in identifying chromosomal genes affecting the production of MccJ25. In this study, we analyze an E. coli null tolC mutant which is impaired in the production of the antibiotic.

Isolation and characterization of a mutation that abolishes MccJ25 production from the low-copy, natural MccJ25-producing plasmid. A chromosomal Tn5 insertion that eliminated MccJ25 production from strain MC4100 harboring the wild-type, low-copy-number, MccJ25-producing plasmid pTUC100 (17) was obtained by using 467 (2) as previously described (4). The transposon was then introduced by P1 transduction (15) into several Hfr strains. Mating experiments using these Tn5 Hfr derivatives suggested a location of the insertion between 62 and 67 min of the E. coli genetic map. Transduction experiments demonstrated that the transposon was 27% cotransducible with the metC locus, which maps at 65 min on the E. coli chromosome (1). This placement raised the possibility that the mutation under study occurred in tolC, which is at min 66 and is cotransducible with metC at about the same frequency (1, 25). Therefore, we tested the mutant for increased sensitivity to detergents, a well-characterized tolC mutant phenotype (25), and it was found to be hypersensitive to 0.05% deoxycholate. Finally, we introduced well-characterized tolC mutations, such as those of strains SC44 (tolC::Tn5) (24) and CAG12184 (tolC::Tn10) (21), by P1 transduction into strain MC4100 (pTUC100). The transductants failed to show inhibition zones when grown on a lawn of sensitive cells (Fig. 1). On the basis of these observations, we concluded that our mutant had an insertion into tolC. Introduction of plasmid pAX629, which carries a cloned copy of the wild-type tolC gene (12), into the mutants rescued the MccJ25⁺ phenotype (Fig. 1), showing that the effect of the tolC mutation on MccJ25 production is primarily due to the inactivation of the tolC gene itself rather than to a possible polarity of the insertion mutation on the expression of a downstream gene.

View larger version (80K): [in this window] [in a new window]   FIG. 1.   Inhibition halos produced by tolC mutants and their respective parental strains carrying different microcin plasmids. Tests were carried out by stabbing individual colonies into a lawn of the hypersensitive indicator Salmonella newport. Strains: 1, MC4100 (pTUC100); 2, MC4100 tolC::Tn5 (pTUC100); 3, MC4100 tolC::Tn5 (pTUC100, pAX629); 4, MC4100 (pTUC202); 5, MC4100 tolC::Tn5 (pTUC202).

Cell growth and microcin production are affected in TolC strains harboring multiple copies of microcin genes. To test the effect of the tolC mutation on microcin production by a multicopy plasmid, strains MC4100, MC4100 tolC::Tn5, and MC4100 tolC::Tn10 were transformed with pTUC202, a pACYC184 derivative carrying the microcin production and immunity genes cloned in a 6-kb BamHI-SalI fragment (22). Introduction of this plasmid into the tolC mutants resulted in unstable transformants, which grew poorly upon restreaking in Luria-Bertani (LB) medium, forming small and sickly colonies, and frequently lost viability after a few passages. This phenotype was even more severe in minimal medium, where transformants failed to form colonies upon restreaking. This is probably due to the greatest microcin production in this medium (17). In contrast, control MC4100 (pTUC202) transformants gave rise to healthy colonies in both media. The growth-inhibitory phenotype was due to microcin production, because nonproducing derivatives of pTUC202 did not affect the growth of tolC cells.

The growth-inhibitory phenotype is relieved by mutations that reduce MccJ25 synthesis. M9 plates supplemented with chloramphenicol were seeded with MC4100 tolC::Tn5 (pTUC202) cells and incubated for 48 h at 37°C. Several mutants that grew well upon restreaking on the same medium were selected for further analysis. They were chloramphenicol resistant and therefore still retained the plasmid. However, they did not secrete microcin. In addition, they remained sensitive to deoxycholate, and thus were not tolC⁺ revertants (which could arise by precise excision of the Tn5). The apparent suppression of the unhealthy phenotype could result from either a plasmid or a host mutation. If suppression resulted from a mutation in a bacterial gene (e.g., a target mutation or a regulatory mutation which lessened the toxin production), then pTUC202 isolated from the candidate suppressor cells would be expected to display a normal microcin production phenotype when subsequently transformed into fresh MC4100 cells. In this way, we determined that several of our mutants contained host mutations. Additionally, when these mutants were spontaneously cured of the resident plasmid and tested for resistance to exogenous MccJ25, none of the cured derivatives was resistant, suggesting that they were not target mutants but carried regulatory mutations impairing expression of the mcj genes. This was confirmed by introducing plasmid pTUC202 with a mcjA-lacZ translational fusion into one of these mutants as well as into the tolC parental strain. There was a large decrease in mcjA expression in comparison with the control; in fact, the mutant showed a 35-fold reduction in -galactosidase activity (68 versus 2,415 Miller units [15]). Finally, plasmid mutants impaired in microcin production were also obtained. Altogether, these results indicated that a decrease in MccJ25 production is responsible for the suppression of the unhealthy phenotype.

TolC effect on microcin biosynthesis. Reduced MccJ25 release from tolC cells might be due to a deficiency in antibiotic biosynthesis or to a decrease in its secretion. To discriminate between these possibilities, plasmid pTUC202 (mcjA-lacZ) was transformed into MC4100 and MC4100 tolC::Tn10. The levels of -galactosidase were not significantly different in both strains throughout exponential growth. For example, at an optical density at 600 nm of 0.7 the wild-type and mutant strains had -galactosidase readings of 90 and 100 Miller units, respectively. In the stationary phase, after overnight growth, the tolC strain had roughly threefold less -galactosidase activity than the parent strain (5,000 versus 14,000 units). Although this effect may contribute to the overall decrease in MccJ25 production, it seems too weak to explain by itself the 12-fold reduction in the amount of extracellular microcin seen in mutant cultures. Thus, most of the effect of tolC mutations is likely to be at the secretion level.

The effect of tolC mutations is not a consequence of their effect on major porin expression. It has been demonstrated that tolC mutations increase the transcription of micF antisense RNA, leading to a concomitant reduction in the expression of OmpF (16). To test whether the effect of the tolC mutation on microcin production was due to the lack of porin, plasmid pTUC202 was introduced into strains MC4100 ompF::Tn5 and MC4100 ompR101. The typical ompR101 mutation normally results in the absence of both OmpF and OmpC proteins (20). The transformants synthesized microcin normally, indicating that the major porins are not involved in microcin production.

The absence of TolC affects MccJ25 immunity negatively. When the immunity to exogenous microcin of MC4100 (pTUC202) and MC4100 tolC::Tn5 (pTUC202) was titrated by the critical dilution method we found that while the parent strain was fully immune, the microcin preparation used could be diluted 1,280-fold and still inhibit growth of the tolC mutant. This phenomenon was not dependent on the kind of mutation or the genetic background since the same result was observed when another tolC-defective strain, CAG12184 (tolC::Tn10) (21), was used as a host. These results clearly indicated that TolC did affect the expression of immunity. To test the possibility that this deficient immunity phenotype could be alleviated by raising the number of copies of the immunity gene, MC4100 tolC::Tn5 was first transformed with pJS300, a pUC18 derivative carrying only the immunity gene (22), and then with pTUC202. In spite of the supplementary immunity protein, double transformants still exhibited the deficient immunity phenotype and the growth deficiency in minimal medium. We have evidence indicating that endogenously synthesized MccJ25 requires the mcjD gene product to be exported out of the cells (23). Indeed, the McjD protein displays all the structural characteristics common to ABC transporters (7, 23). Thus, the immunity to MccJ25 conferred by McjD seems to be mediated by active extrusion of microcin, which would keep the antibiotic concentration in the cytoplasm below a critical level. Peptide antibiotic systems usually contain a specific immunity gene, not involved in antibiotic production. For example, in the case of the microcin B17 immunity system, although the two proteins McbE-McbF forming the ABC exporter provide a limited immunity by export of the antibiotic, a third component, McbG, is required for full immunity (10). Its counterpart in the MccJ25 system has not yet been found (22, 23). In the present study, we tested the degree of immunity provided by McjD alone by spotting solutions with various MccJ25 concentrations onto an LB plate seeded with MC4100 cells harboring the mcjD gene cloned into low-copy-number plasmids (pACYC184 or pSC101). The transformants were fully immune to added microcin of the highest concentration available (1 mg/ml; 10 µg per spot). Since TolC is implicated in the secretion of several proteins in E. coli (6, 8, 9, 13, 24), it does seem plausible that McjD could form an export complex with TolC for the secretion of MccJ25. If so, the absence of TolC should lead to an impairment of the immunity function. A prediction of this model is that tolC cells harboring pTUC202 should accumulate intracellular active MccJ25. However, as stated before, we have not been able to extract increased amounts of MccJ25 from these cells. It is possible that after transformation with the plasmid there is in fact an accumulation of microcin, which would exert a toxic effect on several metabolic processes, including that of MccJ25 synthesis itself. At the time of lysing the cells, they are so sick that this would lead to a poor recovery of intracellular microcin. In addition, there could be an increased proteolytic degradation of accumulated internal microcin in these cells. A similar situation has been described by Garrido et al. (10) for mcbE and mcbF mutants, which cannot export microcin B17. These authors have been unable to extract microcin B17 from these immune deficient cells, and speculate that this failure is due to the poor growth of the cells and the loss of the microcin-producing plasmid, even under selective pressure. Finally, note that a lack of accumulation following a blockade in export has also been observed for -hemolysin (24).