No abstract

Tetracycline resistance encoded by four genetically different determinants residing on plasmids in Escherichia coli was shown to be associated in each case with an energy-dependent decrease in accumulation of the antibiotic in whole cells in which resistance had been induced. The different class determinants examined were those on plasmids RP1 (class A), R222 (class B), R144 (class C), and RAI (class D). This decrease in accumulation was attributable to an active efflux, because everted (inside-out) membrane vesicles made from tetracycline-induced E. coli cells containing any one of the four plasmids were shown to concentrate tetracycline by an active influx. This active uptake was not seen in inside-out vesicles from sensitive cells or uninduced R222-containing cells. In vesicles from induced R222-containing cells, the efflux appeared to be carrier-mediated with a Km of about 6 ;M. These results demonstrate that active export of tetracycline is a common component of the mechanism for tetracycline resistance encoded by different plasmid-borne determinants in bacteria. The tetracyclines are bacteriostatic antibiotics used to treat a broad spectrum of microbial disease agents in humans, animals, and plants (1). They act by inhibiting protein synthesis, specifically by preventing aminoacyl tRNA from binding to the A site on the ribosome during peptide elongation (2-4). Bacterial resistance to tetracycline is widespread and is caused by at least four different resistance determinants (5) carried on plasmids in the host bacterial cell. In most cases, resistance is inducible by subinhibitory amounts of tetracycline, and for some plasmids it can reach 200 times the resistance of plasmidless cells. The most common tetracycline resistance determinant in Escherichia coli appears to be that borne on transposon TnlO.t Plasmid-mediated tetracycline resistance does not lead to inactivation of the tetracycline molecule (6-8). Rather, resistance is associated with a decrease in tetracycline accumulation (8-11). A tetracycline-inducible inner membrane protein of molecular weight about 37,000, TET, is associated with TnlO-encoded resistance (8, 12), and a repressor activity of TET synthesis has been found (13). A TET protein of similar size encoded by other plasmid-borne resistance determinants has been described (8,12,(14)(15)(16).In previous work we showed two transport systems for tetracycline in sensitive E. coli K-12 cells, only one of which was sensitive to energy inhibitors (17). Both were altered by the tetracycline resistance plasmid R222 (11). The energy-dependent component of uptake in sensitive cells was replaced in resistant cells by a non-energy-requiring uptake of a lesser rate. In addition, the energy-independent uptake system in sensitive cells was decreased to 1/3-1/5 (11). This decrease appeared to be at least partly reversed by energy inhibitors (11), which suggested that energy was required for the lowered tetracycline accumulation. Using everted membrane vesicles, we now demonstrate that resistant...

Chromosomal multiple-antibiotic-resistant (Mar) mutants of Escherichia coli, selected on agar containing low concentrations of tetracycline or chloramphenicol, were 6-to 18-fold less susceptible to the fluoroquinolones than were their wild-type E. coli K-12 or E. coli C parental strains. The frequency of emergence of such mutants was at least 1,000-fold higher than that of those selected by the fluoroquinolone norfloxacin directly. When Mar mutants, but not wild-type cells, were plated on norfloxacin, mutants resistant to high levels of norfloxacin (2 ,ug/ml) appeared at a relatively high (-10-7) frequency. In addition to decreased amounts of OmpF, Mar mutants had other outer membrane protein changes and were four-to eightfold less susceptible to fluoroquinolones than was an ompF::Tn5 mutant lacking only OmpF. Accumulation of [3H]norfloxacin was more than threefold lower in the Mar mutants than in wild-type cells and twofold lower than in the OmpF-deficient derivative. These differences were not attributable to a change in the endogenous active efflux system for norfloxacin in E. coli. Norfloxacin-induced inhibition of DNA synthesis was threefold lower in intact cells of a Mar mutant than in susceptible cells, but this difference was not seen in toluene-permeabilized cells. Insertion of TnS into marA (min 34.05 on the chromosome) led to a return of the wild-type patterns of norfloxacin accumulation, fluoroquinolone and other antimicrobial agent susceptibilities, and outer membrane protein profile, including partial restoration of OmpF. These findings together suggest that marA-dependent fluoroquinolone resistance is linked to decreased cell permeability, only part of which can be accounted for by the reduction in OmpF. Once mutated to marA, cells can achieve high levels of quinolone resistance at a relatively high frequency.

Mar (multiple antibiotic resistant) mutants of Escherichia coli express chromosomally mediated resistance to a variety of structurally unrelated hydrophilic and hydrophobic antibiotics. Insertion of transposon Tn5 into the marA locus at min 34.05 on the chromosome completely reverses the Mar phenotype (A. M. George and S. B. Levy, J. Bacteriol. 155:531-540, 1983). We found that among changes in the outer membrane of Mar mutants, porin OmpF was greatly reduced, although Mar mutants were more resistant than cells lacking only OmpF. Transduction of the marA region from a Mar strain, but not a wild-type strain, led to loss of OmpF. P1 transduction of marA::Tn5 into a Mar mutant partially restored OmpF levels. Therefore, OmpF reduction required a mutation in the marA region. Mar mutants of an ompF-lacZ operon fusion strain expressed 50 to 75% of the beta-galactosidase activity of the isogenic non-Mar parental strain, while Mar mutants of a protein fusion strain expressed less than 10% of the enzyme activity in the non-Mar strain. These changes were completely reversed by insertion of marA::Tn5. The responsiveness of OmpF-LacZ to osmolarity and temperature changes was similar in Mar and wild-type strains. Although some transcriptional control may have been present, OmpF reduction appeared to occur primarily by a posttranscriptional mechanism. The steady-state levels of ompF mRNA were twofold lower and the mRNA was five times less stable in the Mar mutant than in the wild-type strain. Expression of micF, which lowers ompF mRNA levels, was elevated in Mar strains, as revealed by a micF-lacZ fusion. Studies with strains deleted for the micF locus showed that the marA-dependent reduction of OmpF required an intact micF locus. Our findings suggest that the marA locus directly or indirectly increases micF expression, causing a posttranscriptional decrease in ompF mRNA and reduced amounts of OmpF.