crossmark
A Transposon Screen Identifies Genetic Determinants of Vibrio cholerae Resistance to High-Molecular-Weight Antibiotics Tobias Dörr,* Fernanda Delgado, Benjamin D. Umans, Matthew A. Gerding, Brigid M. Davis, Matthew K. Waldor Howard Hughes Medical Institute, Brigham and Women’s Hospital Division of Infectious Diseases and Harvard Medical School Department of Microbiology and Immunobiology, Boston, Massachusetts, USA
Gram-negative bacteria are notoriously resistant to a variety of high-molecular-weight antibiotics due to the limited permeability of their outer membrane (OM). The basis of OM barrier function and the genetic factors required for its maintenance remain incompletely understood. Here, we employed transposon insertion sequencing to identify genes required for Vibrio cholerae resistance to vancomycin and bacitracin, antibiotics that are thought to be too large to efficiently penetrate the OM. The screen yielded several genes whose protein products are predicted to participate in processes important for OM barrier functions and for biofilm formation. In addition, we identified a novel factor, designated vigA (for vancomycin inhibits growth), that has not previously been characterized or linked to outer membrane function. The vigA open reading frame (ORF) codes for an inner membrane protein, and in its absence, cells became highly sensitive to glycopeptide antibiotics (vancomycin and ramoplanin) and bacitracin but not to other large antibiotics or detergents. In contrast to wild-type (WT) cells, the vigA mutant was stained with fluorescent vancomycin. These observations suggest that VigA specifically prevents the periplasmic accumulation of certain large antibiotics without exerting a general role in the maintenance of OM integrity. We also observed marked interspecies variability in the susceptibilities of Gram-negative pathogens to glycopeptides and bacitracin. Collectively, our findings suggest that the OM barrier is not absolute but rather depends on specific OM-antibiotic interactions.
T
he outer membrane (OM) of Gram-negative bacteria is a formidable barrier to many antibiotics (1). The OM is an asymmetric lipid bilayer with an inner leaflet composed of phospholipids (phosphatidylethanolamine, phosphatidylglycerol, and cardiolipin) and an outer leaflet composed of lipopolysaccharide (LPS) (2). LPS is a highly complex molecule consisting of the hexa-acylated diglucosamine lipid A at its base, followed by the core oligosaccharide (an oligomer of glucose, galactose, and heptose in most species) and the O-antigen, a structure with a high level of interspecific variability (3, 4). Due to lipid A’s 6 acyl chains, the lipid bilayer portion of the OM is highly hydrophobic. In addition, lipid A as well as core oligosaccharide contain anionic phosphate and carboxylic acid groups. These anionic residues can engage in electrostatic interactions with cations (e.g., Mg2⫹ and Ca2⫹) across neighboring LPS molecules, resulting in ion bridges that stabilize and stiffen the OM (5). The complex, amphiphilic composition and stiffness of the outer membrane render it largely impermeable to large molecules, such as a variety of antibiotics (5). Smaller molecules (water and nutrients) gain access to the periplasm via OM porins, largely nonspecific transmembrane channels that allow the diffusion of solutes smaller than ⬃600 Da (1). Porins also serve as the entryways for small, hydrophilic beta-lactam antibiotics (6, 7), such as the penicillins and cephalosporins. In contrast, slow cross-membrane diffusion is apparently required for the entry of larger, lipophilic compounds such as rifampin and novobiocin (8), which in part explains the comparatively high concentrations required for their antibacterial action. Some antibiotics seem entirely unable to cross the OM, as Gram-negative bacteria are generally resistant to them, even though the antibiotics’ targets are present. These antibiotics include the very large (⬎1.4-kDa) glycopeptides vancomycin and ramoplanin as well as the polypeptide antibiotic bacitracin. Several conditions and mutations are known to result in an increase in OM permeability. For example, EDTA treatment
August 2016 Volume 60 Number 8
causes the leakage of periplasmic proteins into the surrounding medium and increases OM permeability to lysozyme and large antibiotics (9). EDTA is thought to destabilize ionic interactions between LPS molecules by removing the divalent cations that stabilize the anionic groups of LPS. LPS instability leads to the loss of LPS into the surrounding medium and patchy replacement of the typical asymmetric phospholipid-LPS OM with a symmetric phospholipid bilayer. Phospholipid bilayer patches are more permeable to hydrophobic compounds and susceptible to disruption by detergents (9, 10). The increase in OM permeability observed in many cell envelope mutants, e.g., so-called “rough” mutants defective to various degrees in O-antigen and core oligosaccharide production, is likewise thought to result from the accumulation of symmetric phospholipid bilayer patches associated with LPS instability (5, 9). Some bacteria, including Vibrio cholerae, appear to have naturally occurring symmetric bilayer patches and are thus intrinsically more susceptible to some hydrophobic compounds (such as rifampin and novobiocin), but the genetic basis for the
Received 13 March 2016 Returned for modification 27 April 2016 Accepted 19 May 2016 Accepted manuscript posted online 23 May 2016 Citation Dörr T, Delgado F, Umans BD, Gerding MA, Davis BM, Waldor MK. 2016. A transposon screen identifies genetic determinants of Vibrio cholerae resistance to high-molecular-weight antibiotics. Antimicrob Agents Chemother 60:4757–4763. doi:10.1128/AAC.00576-16. Address correspondence to Tobias Dörr, [email protected], or Matthew K. Waldor, [email protected]. * Present address: Tobias Dörr, Department of Microbiology/Weill Institute for Cell and Molecular Biology, Cornell University, Ithaca, NY, USA. Supplemental material for this article may be found at http://dx.doi.org/10.1128 /AAC.00576-16. Copyright © 2016, American Society for Microbiology. All Rights Reserved.
Antimicrobial Agents and Chemotherapy
aac.asm.org
4757
Dörr et al.
formation of these patches is not known (11). Apart from treatments and mutations that cause drastic changes in OM properties, the genetic factors that contribute to the maintenance of the OM barrier to large antibiotics are only incompletely understood. Here, we investigated the genetic determinants of the OM permeability barrier in V. cholerae, the cholera pathogen. Using a transposon insertion sequencing-based screen, we identified factors required for the resistance of V. cholerae to vancomycin and bacitracin. In addition to factors known or predicted to mediate OM integrity and/or the extracellular barrier to antibiotic diffusion, we also found that a previously uncharacterized inner membrane protein, which we name VigA, is required for vancomycin resistance. Surprisingly, a ⌬vigA mutant is susceptible to vancomycin and ramoplanin but does not exhibit increased susceptibility to other large antibiotics, hydrophobic compounds, or detergents, suggesting that the integrity of its OM is not generally compromised. Finally, we observed that there is a large range in the MICs of glycopeptides and bacitracin (antibiotics generally thought to have a minimal capacity to penetrate the OM) among various Gram-negative pathogens. Collectively, our observations suggest that the OM permeability barrier to large antibiotics is not absolute. Instead, we speculate that this barrier depends on specific attributes of both the OM and the antibiotic in question. MATERIALS AND METHODS Media and growth conditions. All strains were grown in 3 ml LB broth in 10-ml glass tubes at 37°C with shaking (225 rpm). Antibiotics (all from Sigma-Aldrich), where applicable, were used at the following concentrations: 200 g/ml streptomycin, 50 g/ml carbenicillin, 200 g/ml vancomycin, 50 g/ml kanamycin, and 5 g/ml chloramphenicol. Strains and plasmids. All V. cholerae strains are derivatives of El Tor C6706 unless otherwise indicated. Defined transposon mutants were taken from an ordered transposon insertion library (12). The defined vigA deletion strain (replacement of the open reading frame [ORF] with the sequence 5=-T TATCATGCGGCCGCACTCGAGTAATGATAA-3=) was constructed by allele replacement via homologous recombination using suicide plasmid pCVD442 as described previously (13). Regions of homology were amplified by using primers TDP1494 (5=-AGGTATATGTGATGGGTTAAAAAGGAT CGATCCTGACCCCACACTAAACTATCAACAGTTT-3=)/TDP1495 (5=TTATCATGCGGCCGCACTCGAGTAATGATAATCAGCGAAT CAGTTCAGAAATCATCAAG-3=) and TDP1496 (5=-TTATCATTAC TCGAGTGCGGCCGCATGATAAGGCGCTCTCAGAAGAGATA GG-3=)/TDP1497 (5=-CCGGGAGAGCTCGATATCGCATGCGG TACCTCTAGTCGTGAAGTCTTCCATATCGCCCT-3=) and cloned into XbaI-digested pCVD442 by isothermal assembly (14) (boldface type represents homology overhangs for isothermal assembly cloning). For complementation, the vigA ORF was placed under the control of an isopropyl-D-thiogalactopyranoside (IPTG)-inducible promoter followed by integration of the whole construct into the neutral chromosomal lacZ locus. Transposon insertion sequencing. Under each condition (2 with no antibiotic control, 2 with vancomycin, and 2 with bacitracin), 1 ml of a culture of V. cholerae grown overnight was mixed with 1 ml of a culture of donor strain SM10 carrying pSC189 (15) (transposon donor plasmid) grown overnight. The mixture was pelleted by centrifugation (5,000 ⫻ g for 5 min) and resuspended in 1 ml LB broth. This mixture was aliquoted (100 l each) onto 0.45-m filter disks on prewarmed (37°C) LB agar plates. After 3 h of incubation, cells were scraped off the filter disks using a pipette tip into LB medium containing streptomycin, vortexed, and plated directly onto LB agar plates containing streptomycin (to kill the donor strain) and either no antibiotic (control library), 200 g/ml vancomycin, or 200 g/ml bacitracin. After overnight incubation (resulting in ⬃200,000 colonies under each mating condition), cells were scraped off the plates and pooled. Libraries for Illumina sequencing were prepared as
4758
aac.asm.org
described previously (16). Briefly, following genomic DNA extraction, DNA was sheared by using sonication, and overhanging ends were blunted by using a blunting enzyme kit (NEB). This was followed by Illumina adaptor ligation and amplification of genomic DNA-transposon junctions. Sequencing reactions were conducted on an Illumina MiSeq benchtop sequencer (Illumina, San Diego, CA). Data analysis was conducted as described previously (16, 17). Cell fractionation. Cell fractionation was conducted by differential centrifugation essentially as described previously (18), with modifications. Cultures grown overnight were diluted 1,000-fold into 50 ml of fresh LB broth in baffled flasks and grown for 1 h with shaking at 37°C. Where applicable, 100 M IPTG was added to induce the expression of epitope-tagged VigA, followed by incubation for an additional 2 h. Cells were then harvested by centrifugation (15 min at 15,000 ⫻ g at 4°C), washed once in cold phosphate-buffered saline (PBS) (pH 7.4), and resuspended in 1.5 ml of the same buffer supplemented with lysozyme (5 mg/ml), Halt protease inhibitor tablets (Fisher Scientific), and 1 mM phenylmethylsulfonyl fluoride (PMSF). Cells were lysed by sonication (3 times for 10 s) on ice, followed by centrifugation (15 min at 12,000 ⫻ g at 4°C) to remove debris. The supernatant (envelope fraction plus cytoplasmic fraction) was then subjected to ultracentrifugation (200,000 ⫻ g at 4°C) to obtain membranes. To separate the outer from the inner membrane, the latter was solubilized by resuspending the membrane pellet in 1 ml of PBS containing 0.4% reduced Triton X-100 (Sigma), followed by incubation on a rotator for 3 h at 4°C. Subsequent ultracentrifugation (200,000 ⫻ g at 4°C) was then used to separate the inner (solubilized) and outer (pellet) membranes. Fluorescent vancomycin staining. Cells were grown to exponential phase, washed once with PBS, and then resuspended in the same medium to ⬃1 ⫻ 109 CFU/ml. BODIPY FL vancomycin (Thermo Scientific) was added at 100 g/ml for 10 min (ambient temperature). Cells were imaged after washing once in PBS. MICs. MICs were assessed by using Clinical and Laboratory Standards Institute (CLSI) guidelines; however, instead of Mueller-Hinton broth, we used LB broth due to the poor growth of V. cholerae in the former medium. Protein techniques. Western blotting was conducted by using standard protocols. Cell lysates were normalized by the protein content (measured by Qubit; Thermo Fisher Scientific).
RESULTS
Similar to other Gram-negative bacteria, V. cholerae is resistant to large antibiotics like vancomycin (1.5 kDa) and bacitracin (1.4 kDa). Even though the targets for vancomycin (nascent cell wall/ cell wall precursors) and bacitracin (lipid II, a precursor for cell wall synthesis) are widely conserved in Gram-negative bacteria (19), the OM permeability barrier in these organisms renders them resistant to these agents. We used a transposon insertion sequencing-based strategy to identify loci that were required for V. cholerae resistance to vancomycin and/or bacitracin and, thus, presumably for maintenance of the OM permeability barrier. We created Himar1 transposon insertion libraries in an El Tor clinical isolate of V. cholerae and compared the relative frequencies of transposon insertions at each locus (a proxy for mutant fitness) when the library was grown in the absence versus the presence of vancomycin or bacitracin. We identified eight loci for which insertions were significantly underrepresented after growth on vancomycin plates and one locus for which insertions were underrepresented after growth on bacitracin (Fig. 1A and B). The single gene identified in the bacitracin screen, vc1423, was also one of the eight loci detected in the vancomycin screen. The 8 genes that our screens identified constitute candidate determinants of outer membrane barrier function. Two of these genes—galE, coding for UDP-glucose 4-epimerase (20, 21), and
Antimicrobial Agents and Chemotherapy
August 2016 Volume 60 Number 8
Genetic Basis of V. cholerae Glycopeptide Resistance
FIG 1 Identification of genes conferring enhanced susceptibility to vancomycin and bacitracin using transposon insertion sequencing. (A and B) Volcano plots
depicting the ratio of read counts mapped to individual genes in transposon libraries plated onto vancomycin (200 g/ml) (A) or bacitracin (200 g/ml) (B) compared to control libraries plated onto LB agar without antibiotics (mean fold change, x axis) versus mean P values (y axis) (see Materials and Methods for details). Genes shown in red and blue are considered significantly underrepresented. (C) Validation of plating defects on vancomycin of selected mutants identified in the transposon screen. Mutants were plated onto either LB agar containing vancomycin (200 g/ml) or LB agar alone (⫺ vancomycin). (D and E) Sensitivity of a V. cholerae vigA mutant (⌬vc1423) to vancomycin and bacitracin. Strains (including the ⌬vc1423 strain ectopically expressing vc1423 under the control of an IPTG-inducible promoter) were spot plated onto agar containing either vancomycin (200 g/ml) (D) or bacitracin (200 g/ml) (E) and IPTG (100 M). (F) Vancomycin MICs of V. cholerae WT and ⌬vc1423 strains.
cls, coding for the synthase of cardiolipin, an anionic membrane phospholipid (22)— have known roles in cell envelope permeability in other Gram-negative bacteria (cls and galE) and in biofilm formation in V. cholerae (galE only) (20). Biofilms constitute a diffusion barrier for large antibiotics in other bacteria (23). Components of both operons required for the production of exopolysaccharide (EPS), a capsule-like extracellular polymer with a crucial role in biofilm formation (24), were also identified. Finally, both screens identified vc1423, a gene that encodes a hypothetical protein of unknown function and is discussed in more detail below. We validated the results of our transposon screen by spotting defined mutants corresponding to each category on LB agar plates
August 2016 Volume 60 Number 8
with and without vancomycin. Growth of wild-type (WT) V. cholerae was not impaired by 200 g/ml vancomycin, but all mutants displayed at least a modest (5- to 10-fold) plating deficiency, and the ⌬vc1423 mutant exhibited a drastic (10,000-fold) plating deficiency when spotted onto vancomycin plates (Fig. 1C and D). Interestingly, the four mutants that exhibited modest plating deficiencies had no reduction in vancomycin MICs compared to the WT (data not shown), while the ⌬vc1423 mutant exhibited a 4-fold decrease in the MIC against vancomycin (Fig. 1F). Thus, the vancomycin susceptibility of some strains is associated with particular attributes of their environments, e.g., growth on surfaces versus growth in liquid, or the unchanged MIC reports the
Antimicrobial Agents and Chemotherapy
aac.asm.org
4759
Dörr et al.
FIG 2 VigA is an inner membrane protein. (A) Cells carrying either ectopically expressed, His-tagged VigA or chromosomally encoded, His-tagged PBP1A (control protein for inner membrane localization) or LpoA (control protein for outer membrane localization) were lysed and separated into fractions via differential centrifugation. Proteins were visualized by immunoblotting with an antibody recognizing the His epitope. OM, outer membrane fractions; IM, inner membrane fractions. (B) Ectopically expressed N- or C-terminal fusions of VigA with beta-lactamase were plated onto agar containing an inducer (IPTG, 100 M) and carbenicillin. The control is an empty plasmid. bla, beta-lactamase. (C) Model of VigA topology based on the experiments described above. C, C terminus.
growth of a phenotypically resistant subpopulation of cells. The ⌬vc1423 mutant also exhibited a marked plating deficiency when spotted onto plates containing bacitracin (Fig. 1E) or the 2.5-kDa glycopeptide ramoplanin (see Fig. S1 in the supplemental material). Expression of vc1423 in trans complemented the plating defect for all three antibiotics (Fig. 1D and E; see also Fig. S1 in the supplemental material). Thus, vc1423 (renamed vigA for vancomycin inhibits growth) is crucial for V. cholerae to maintain its resistance to these 3 large antibiotics. VigA is an inner membrane protein with an N-in, C-out topology. VigA is predicted to be a mostly helical protein with a short N-terminal cytoplasmic segment, followed by a transmembrane domain and a large periplasmic domain. We carried out cell fractionation experiments to verify that VigA localizes to the cell membrane. Lysates of cells ectopically expressing His-tagged VigA were separated into inner membrane and outer membrane fractions, using differential centrifugation. Immunoblotting was then used to detect epitope-tagged VigA as well as His-tagged versions of known inner and outer membrane proteins: penicillin binding protein 1A (PBP1A) and LpoA, respectively (25, 26). VigA was present in the inner membrane fraction but absent from the outer membrane fraction, consistent with its predicted localization (Fig. 2A). PBP1A and LpoA were highly enriched in the inner membrane and outer membrane samples, respectively, validating the data from the fractionation assay. To validate VigA’s predicted N-in, C-out membrane topology,
4760
aac.asm.org
FIG 3 VigA mutant cells appear more permeable to vancomycin than do WT cells. (A) Cells were grown in LB medium, washed once in PBS, stained for 10 min with fluorescent vancomycin (vancomycin FL), washed twice with PBS, and then imaged. The red arrowheads point to lysed WT cells. (B) Prior to imaging, cells were grown to exponential phase (optical density at 600 nm of 0.5) in LB medium and then exposed to the indicated concentrations of vancomycin for 3 h.
we constructed N-terminal and C-terminal fusions of VigA to beta-lactamase. This protein hydrolyzes beta-lactam antibiotics, conferring resistance to these agents, only when it is exported to the periplasm. Both N-terminal and C-terminal VigA– beta-lactamase fusions were stably expressed, and both fusions were able to complement the plating deficiency on vancomycin of a vigA mutant, indicating that both fusion proteins are functional (see Fig. S2 in the supplemental material). However, only the C-terminal fusion allowed growth in the presence of carbenicillin (Fig. 2B), suggesting that the C-terminal portion of VigA is in the periplasm, while the N terminus localizes to the cytoplasm, as predicted (Fig. 2C). Vancomycin accumulates in the periplasm of a vigA mutant and causes loss of rod shape. The vigA mutant’s pronounced growth deficiency (relative to WT cells) on plates containing vancomycin suggested that this glycopeptide antibiotic was able to penetrate the OM of the vigA mutant and bind to its target (pentapeptide side chains of peptidoglycan (PG) [27]). To test this possibility, we briefly stained WT and vigA mutant cells with fluorescently labeled vancomycin (Van-FL). In the WT culture, Van-FL stained only lysed cells (in which the vancomycin target is exposed due to disruption of the OM) (Fig. 3A, arrowheads), while intact cells did not accumulate any periplasmic fluorescence. In contrast, vigA mutant cells rapidly accumulated peripheral fluorescence, consistent with the idea that Van-FL can enter and/or remain in the periplasm. When vancomycin binds to its target, pentapeptide residues in nascent PG, it sterically prevents transpeptidation by the major cell wall synthases, the penicillin binding proteins, and thus effectively inhibits cell wall synthesis (27). To gauge whether vancomy-
Antimicrobial Agents and Chemotherapy
August 2016 Volume 60 Number 8
Genetic Basis of V. cholerae Glycopeptide Resistance
FIG 4 The general integrity of the OM in the vigA mutant is not compromised. (A) Comparison of MICs of membrane stressors and large antibiotics in V. cholerae WT and ⌬vigA strains. (B) Growth curves of V. cholerae WT and ⌬vigA strains in the presence or absence of 1% bile. (C) Endogenous RpoE levels in WT V. cholerae versus the ⌬vigA mutant. Lysates of exponentially growing cells were subjected to SDS-PAGE followed by immunoblotting with an antibody raised against RpoE. Lysates were normalized by both the optical density at 600 nm (OD600) and protein content before loading. The RNA polymerase subunit RpoB is the internal control.
cin’s ability to inhibit the vigA mutant’s growth was associated with inhibition of cell wall synthesis, we treated growing cultures with various concentrations of vancomycin and observed the effect on cell morphology. Treatment of vigA cells with ⬃0.5⫻ or ⬃1⫻ the vancomycin MIC for the mutant cells resulted in a loss of rod shape and sphere formation (Fig. 3B), as observed when V. cholerae is treated with beta-lactams and other inhibitors of cell wall synthesis (28) whose action results in a loss of the cell wall. Surprisingly, wild-type cells treated with 200 g/ml of vancomycin (approximately the MIC for the vigA mutant and 0.2⫻ the WT MIC), also exhibited changes in morphology but did not form spheres (Fig. 3B). Thus, although WT V. cholerae is resistant to vancomycin, its OM likely does not completely block the entry of this antibiotic and the associated inhibition of cell wall synthesis. However, our data suggest that vancomycin has a more profound effect on cell wall synthesis in the vigA mutant, consistent with this strain’s lower MIC and its periplasmic accumulation of vancomycin. OM integrity is not generally compromised in the vigA mutant. To investigate whether vancomycin sensitivity in the vigA mutant was indicative of a general OM integrity defect, we measured the mutant’s sensitivity to SDS and the high-molecularweight antibiotics rifampin and novobiocin. Surprisingly, MICs of all three agents were the same for the vigA mutant and the wild-type strain (Fig. 4A). MICs of bacitracin and ramoplanin also did not differ significantly between the WT strain and the vigA mutant (Fig. 4A), despite the mutant’s marked plating defect with these agents, providing further evidence that disruption of vigA
August 2016 Volume 60 Number 8
does not result in a general OM defect. Furthermore, we found that the efficiencies of plating of WT and ⌬vigA strains on plates containing SDS and increasing concentrations of EDTA (a standard assay for OM defects [10]) were very similar (see Fig. S3 in the supplemental material), and the vigA mutant grew as well as the WT in the presence of 1% bile, a membrane-disrupting agent (Fig. 4B). Finally, the abundance of the alternate sigma factor RpoE, which increases in response to envelope stress, was not elevated in the vigA mutant (Fig. 4C), nor were there marked differences in inner or outer membrane protein contents (e.g., an increase in the abundance of outer membrane porins) in the mutant compared to the WT (see Fig. S4 in the supplemental material). Collectively, these observations suggest that the increased sensitivity of the vigA mutant to vancomycin does not result from a generalized OM defect in this strain; instead, the vigA deletion results in sensitivity that appears to be specific to certain antibiotics and is also modulated by the bacterial growth environment (i.e., liquid versus solid medium). Variability of glycopeptide/bacitracin resistance in Gramnegative pathogens. Finally, we explored whether several other Gram-negative pathogens exhibit some degree of sensitivity to glycopeptides and bacitracin. While WT strains of V. cholerae (including a clinical isolate from the current outbreak in Haiti) exhibited very similar MIC values against the three compounds tested, there was considerable interspecies variability in MIC values (Table 1). Bacitracin inhibited Acinetobacter baumannii growth at 62 g/ml, while all other bacteria were resistant even at 1,000 g/ml (limit of detection). However, the susceptibility of A.
Antimicrobial Agents and Chemotherapy
aac.asm.org
4761
Dörr et al.
TABLE 1 Variations in susceptibility to glycopeptides and bacitracin among Gram-negative pathogensa MIC (g/ml)
Antibiotic
V. cholerae V. cholerae MW C6706 H1 EHEC
Bacitracin 1,422 >1,000 Vancomycin 1,500 1,000 Ramoplanin 2,500 125
>1,000 1,000 125
Pseudomonas aeruginosa A. baumannii PA14 Lac-4
>1,000 >1,000 500 >1,000 500 >1,000
62 500 125
a Shown are MICs of glycopeptides and bacitracin against selected Gram-negative pathogens. MW, molecular weight. V. cholerae clinical isolate H1 was obtained from the recent cholera outbreak in Haiti. Boldface type indicates MIC values above the limit of detection.
baumannii to vancomycin was comparable to that of the other three pathogens tested. Conversely, enterohemorrhagic Escherichia coli (EHEC) had a significantly higher MIC for ramoplanin (but not vancomycin or bacitracin) than the other pathogens. Interestingly, almost all organisms (with the exception of A. baumannii) were more susceptible to the largest compound (ramoplanin [2.5 kDa]) than to the smallest (bacitracin [1.4 kDa]), suggesting that the size of the antibiotic alone is not a good predictor of antimicrobial efficacy. Taken together, the variability in sensitivity to the two large glycopeptide antibiotics and bacitracin among these 4 Gram-negative pathogens suggests that the OM permeability barrier is not absolute but depends on species-specific biophysical or biochemical interactions between the OM and particular antibiotics. DISCUSSION
Mutations that confer severe and general OM defects have been well described (9, 29), but there is less knowledge of specific genetic requirements for OM barrier function in Gram-negative bacteria. Here, we took advantage of the power of transposon insertion sequencing to identify genetic factors that contribute to the generation or maintenance of the V. cholerae OM permeability barrier against vancomycin and bacitracin. One gene of unknown function, vc1423, here renamed vigA, was identified in both screens. In the absence of VigA, an inner membrane protein, cells became highly sensitive to glycopeptide antibiotics and bacitracin when grown on solid medium. In liquid cultures, exposure to vancomycin resulted in a loss of rod shape for the vigA mutant, and the ⌬vigA strain could be stained with fluorescent vancomycin. The vigA mutant was not sensitive to membrane-damaging agents, either on solid medium or in liquid culture. Thus, VigA appears to specifically prevent the periplasmic accumulation of certain large antibiotics without exerting a general role in the maintenance of OM integrity. Besides vigA, the screen identified Vibrio exopolysaccharide (VPS) biosynthesis genes and galE as requirements for vancomycin resistance. Both VPS and galE are required for biofilm formation (20, 30). While biofilms in general are thought to constitute a significant diffusion barrier to vancomycin, e.g., in Gram-positive Staphylococcus spp. (23, 31), VPS or similar secreted exopolysaccharides have not, to our knowledge, been linked to OM permeability or to limiting the diffusion of large antibiotics into vibrios or other Gram-negative organisms. Given its sensitivity to high-molecular-weight glycopeptides, the vigA mutant’s lack of a general OM defect was unexpected. Most mutations that result in an increase in OM permeability to
4762
aac.asm.org
large compounds also tend to cause a fairly severe impairment of overall OM integrity and an increased permeability for other, unrelated compounds (32, 33). The lack of SDS and bile sensitivity in the ⌬vigA mutant suggests that the mutant’s defect is not attributable to elevations in the amount of phospholipid bilayer patches in its OM, and overall, VigA does not appear to be involved in crucial OM maintenance functions. There are several potential explanations for the vigA mutant’s sensitivity to glycopeptides and bacitracin. One possibility is that the ⌬vigA mutation results in subtle changes in the biophysical properties of LPS, enabling increased diffusion of glycopeptides and bacitracin. Such a subtle OM alteration is also thought to underlie the glycopeptide sensitivity of an unidentified mutation isolated in E. coli (34). Alternatively, VigA might be important for cell wall repair processes, rather than OM homeostasis per se, or for drug efflux. Cell wall repair or modification processes could compensate for damage caused by vancomycin or affect the abundance of its target, while an efflux pump could limit antibiotic accumulation in the cell. While we cannot entirely exclude any explanation, detection of fluorescent vancomycin staining in the vigA mutant suggests increased accumulation (due to enhanced influx or reduced efflux) rather than a cell wall repair defect. VigA has strong homologues only within the family Vibrionaceae. However, VigA is a member of the more widely conserved but poorly characterized HemX superfamily. Future studies will address whether HemX homologues in other Gram-negative bacteria also play a role in glycopeptide resistance. If so, the development of agents that inactivate the activity of VigA-like proteins could enable the use of glycopeptide antibiotics in emerging Gram-negative pathogens such as Acinetobacter baumannii. The Gram-negative OM is often described as an absolute diffusion barrier to large antibiotics. However, exposure to subMICs of vancomycin resulted in shape defects in V. cholerae, suggesting that this organism’s OM is not entirely impermeable to this compound even at concentrations that do not inhibit population growth. Furthermore, there was not a simple correlation between antibiotic molecular weight and MIC against a panel of Gram-negative pathogens, suggesting that the OM diffusion barrier for these compounds is not based simply on size. Instead, our observations suggest that the efficacy of large antibiotics depends on specific interactions of antibiotics with specific OMs. Studies of structure-permeability relationships coupled with medicinal chemistry are an avenue of investigation that could potentially enable the use of our current arsenal of Gram-positive-specific drugs on Gram-negative pathogens. ACKNOWLEDGMENTS We thank the Mekalanos laboratory (in particular William Robins) for the gift of the OmpU antibody and Carol Gross for the RpoE antisera.
FUNDING INFORMATION This work, including the efforts of Matthew K. Waldor, was funded by HHS | National Institutes of Health (NIH) (R37 AI-042347). This work, including the efforts of Matthew K. Waldor, was funded by Howard Hughes Medical Institute (HHMI). This work, including the efforts of Fernanda Delgado, was funded by Howard Hughes Medical Institute (HHMI).
REFERENCES 1. Delcour AH. 2009. Outer membrane permeability and antibiotic resistance. Biochim Biophys Acta 1794:808 – 816. http://dx.doi.org/10.1016/j .bbapap.2008.11.005.
Antimicrobial Agents and Chemotherapy
August 2016 Volume 60 Number 8
Genetic Basis of V. cholerae Glycopeptide Resistance
2. Silhavy TJ, Kahne D, Walker S. 2010. The bacterial cell envelope. Cold Spring Harb Perspect Biol 2:a000414. http://dx.doi.org/10.1101 /cshperspect.a000414. 3. Chatterjee SN, Chaudhuri K. 2003. Lipopolysaccharides of Vibrio cholerae. I. Physical and chemical characterization. Biochim Biophys Acta 1639:65–79. http://dx.doi.org/10.1016/j.bbadis.2003.08.004. 4. Whitfield C, Trent MS. 2014. Biosynthesis and export of bacterial lipopolysaccharides. Annu Rev Biochem 83:99 –128. http://dx.doi.org/10 .1146/annurev-biochem-060713-035600. 5. Nikaido H. 2003. Molecular basis of bacterial outer membrane permeability revisited. Microbiol Mol Biol Rev 67:593– 656. http://dx.doi.org/10 .1128/MMBR.67.4.593-656.2003. 6. Kojima S, Nikaido H. 2013. Permeation rates of penicillins indicate that Escherichia coli porins function principally as nonspecific channels. Proc Natl Acad Sci U S A 110:E2629 –E2634. http://dx.doi.org/10.1073/pnas .1310333110. 7. Yoshimura F, Nikaido H. 1985. Diffusion of beta-lactam antibiotics through the porin channels of Escherichia coli K-12. Antimicrob Agents Chemother 27:84 –92. http://dx.doi.org/10.1128/AAC.27.1.84. 8. Nikaido H. 1976. Outer membrane of Salmonella typhimurium. Transmembrane diffusion of some hydrophobic substances. Biochim Biophys Acta 433:118 –132. 9. Hancock RE. 1984. Alterations in outer membrane permeability. Annu Rev Microbiol 38:237–264. http://dx.doi.org/10.1146/annurev.mi.38 .100184.001321. 10. Malinverni JC, Silhavy TJ. 2009. An ABC transport system that maintains lipid asymmetry in the gram-negative outer membrane. Proc Natl Acad Sci U S A 106:8009 – 8014. http://dx.doi.org/10.1073/pnas.0903229106. 11. Paul S, Chaudhuri K, Chatterjee AN, Das J. 1992. Presence of exposed phospholipids in the outer membrane of Vibrio cholerae. J Gen Microbiol 138:755–761. http://dx.doi.org/10.1099/00221287-138-4-755. 12. Cameron DE, Urbach JM, Mekalanos JJ. 2008. A defined transposon mutant library and its use in identifying motility genes in Vibrio cholerae. Proc Natl Acad Sci U S A 105:8736 – 8741. http://dx.doi.org/10.1073/pnas .0803281105. 13. Donnenberg MS, Kaper JB. 1991. Construction of an eae deletion mutant of enteropathogenic Escherichia coli by using a positive-selection suicide vector. Infect Immun 59:4310 – 4317. 14. Gibson DG, Young L, Chuang RY, Venter JC, Hutchison CA, III, Smith HO. 2009. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods 6:343–345. http://dx.doi.org/10.1038/nmeth .1318. 15. Chiang SL, Rubin EJ. 2002. Construction of a mariner-based transposon for epitope-tagging and genomic targeting. Gene 296:179 –185. http://dx .doi.org/10.1016/S0378-1119(02)00856-9. 16. Chao MC, Pritchard JR, Zhang YJ, Rubin EJ, Livny J, Davis BM, Waldor MK. 2013. High-resolution definition of the Vibrio cholerae essential gene set with hidden Markov model-based analyses of transposoninsertion sequencing data. Nucleic Acids Res 41:9033–9048. http://dx.doi .org/10.1093/nar/gkt654. 17. Pritchard JR, Chao MC, Abel S, Davis BM, Baranowski C, Zhang YJ, Rubin EJ, Waldor MK. 2014. ARTIST: high-resolution genome-wide assessment of fitness using transposon-insertion sequencing. PLoS Genet 10:e1004782. http://dx.doi.org/10.1371/journal.pgen.1004782. 18. Thein M, Sauer G, Paramasivam N, Grin I, Linke D. 2010. Efficient
August 2016 Volume 60 Number 8
19. 20.
21. 22. 23.
24.
25.
26. 27. 28. 29. 30.
31. 32. 33.
34.
subfractionation of gram-negative bacteria for proteomics studies. J Proteome Res 9:6135– 6147. http://dx.doi.org/10.1021/pr1002438. Scheffers DJ, Tol MB. 2015. LipidII: just another brick in the wall? PLoS Pathog 11:e1005213. http://dx.doi.org/10.1371/journal.ppat.1005213. Nesper J, Lauriano CM, Klose KE, Kapfhammer D, Kraiss A, Reidl J. 2001. Characterization of Vibrio cholerae O1 El Tor galU and galE mutants: influence on lipopolysaccharide structure, colonization, and biofilm formation. Infect Immun 69:435– 445. http://dx.doi.org/10.1128/IAI.69.1 .435-445.2001. Robertson BD, Frosch M, van Putten JP. 1993. The role of galE in the biosynthesis and function of gonococcal lipopolysaccharide. Mol Microbiol 8:891–901. http://dx.doi.org/10.1111/j.1365-2958.1993.tb01635.x. Bishop RE. 2014. Emerging roles for anionic non-bilayer phospholipids in fortifying the outer membrane permeability barrier. J Bacteriol 196: 3209 –3213. http://dx.doi.org/10.1128/JB.02043-14. Darouiche RO, Dhir A, Miller AJ, Landon GC, Raad II, Musher DM. 1994. Vancomycin penetration into biofilm covering infected prostheses and effect on bacteria. J Infect Dis 170:720 –723. http://dx.doi.org/10.1093 /infdis/170.3.720. Yildiz FH, Schoolnik GK. 1999. Vibrio cholerae O1 El Tor: identification of a gene cluster required for the rugose colony type, exopolysaccharide production, chlorine resistance, and biofilm formation. Proc Natl Acad Sci U S A 96:4028 – 4033. http://dx.doi.org/10.1073/pnas.96.7.4028. Dorr T, Lam H, Alvarez L, Cava F, Davis BM, Waldor MK. 2014. A novel peptidoglycan binding protein crucial for PBP1A-mediated cell wall biogenesis in Vibrio cholerae. PLoS Genet 10:e1004433. http://dx.doi.org /10.1371/journal.pgen.1004433. Typas A, Banzhaf M, Gross CA, Vollmer W. 2012. From the regulation of peptidoglycan synthesis to bacterial growth and morphology. Nat Rev Microbiol 10:123–136. http://dx.doi.org/10.1038/nrmicro2677. Perkins HR, Nieto M. 1974. The chemical basis for the action of the vancomycin group of antibiotics. Ann N Y Acad Sci 235:348 –363. http: //dx.doi.org/10.1111/j.1749-6632.1974.tb43276.x. Dorr T, Davis BM, Waldor MK. 2015. Endopeptidase-mediated beta lactam tolerance. PLoS Pathog 11:e1004850. http://dx.doi.org/10.1371 /journal.ppat.1004850. Sampson BA, Misra R, Benson SA. 1989. Identification and characterization of a new gene of Escherichia coli K-12 involved in outer membrane permeability. Genetics 122:491–501. Fong JC, Syed KA, Klose KE, Yildiz FH. 2010. Role of Vibrio polysaccharide (vps) genes in VPS production, biofilm formation and Vibrio cholerae pathogenesis. Microbiology 156:2757–2769. http://dx.doi.org/10 .1099/mic.0.040196-0. Choudhry O, Gathwala G, Singh J. 2010. Vancomycin resistant enterococci in neonatal ICU—a rising menace. Indian J Pediatr 77:1446 –1447. http://dx.doi.org/10.1007/s12098-010-0243-6. Vaara M. 1993. Antibiotic-supersusceptible mutants of Escherichia coli and Salmonella typhimurium. Antimicrob Agents Chemother 37:2255– 2260. http://dx.doi.org/10.1128/AAC.37.11.2255. Vuorio R, Vaara M. 1992. The lipid A biosynthesis mutation lpxA2 of Escherichia coli results in drastic antibiotic supersusceptibility. Antimicrob Agents Chemother 36:826 – 829. http://dx.doi.org/10.1128/AAC.36 .4.826. Shlaes DM, Shlaes JH, Davies J, Williamson R. 1989. Escherichia coli susceptible to glycopeptide antibiotics. Antimicrob Agents Chemother 33:192–197. http://dx.doi.org/10.1128/AAC.33.2.192.
Antimicrobial Agents and Chemotherapy
aac.asm.org
4763
A Transposon Screen Identifies Genetic Determinants of Vibrio cholerae Resistance to High-Molecular-Weight Antibiotics Tobias Dörr,* Fernanda Delgado, Benjamin D. Umans, Matthew A. Gerding, Brigid M. Davis, Matthew K. Waldor Howard Hughes Medical Institute, Brigham and Women’s Hospital Division of Infectious Diseases and Harvard Medical School Department of Microbiology and Immunobiology, Boston, Massachusetts, USA
Gram-negative bacteria are notoriously resistant to a variety of high-molecular-weight antibiotics due to the limited permeability of their outer membrane (OM). The basis of OM barrier function and the genetic factors required for its maintenance remain incompletely understood. Here, we employed transposon insertion sequencing to identify genes required for Vibrio cholerae resistance to vancomycin and bacitracin, antibiotics that are thought to be too large to efficiently penetrate the OM. The screen yielded several genes whose protein products are predicted to participate in processes important for OM barrier functions and for biofilm formation. In addition, we identified a novel factor, designated vigA (for vancomycin inhibits growth), that has not previously been characterized or linked to outer membrane function. The vigA open reading frame (ORF) codes for an inner membrane protein, and in its absence, cells became highly sensitive to glycopeptide antibiotics (vancomycin and ramoplanin) and bacitracin but not to other large antibiotics or detergents. In contrast to wild-type (WT) cells, the vigA mutant was stained with fluorescent vancomycin. These observations suggest that VigA specifically prevents the periplasmic accumulation of certain large antibiotics without exerting a general role in the maintenance of OM integrity. We also observed marked interspecies variability in the susceptibilities of Gram-negative pathogens to glycopeptides and bacitracin. Collectively, our findings suggest that the OM barrier is not absolute but rather depends on specific OM-antibiotic interactions.
T
he outer membrane (OM) of Gram-negative bacteria is a formidable barrier to many antibiotics (1). The OM is an asymmetric lipid bilayer with an inner leaflet composed of phospholipids (phosphatidylethanolamine, phosphatidylglycerol, and cardiolipin) and an outer leaflet composed of lipopolysaccharide (LPS) (2). LPS is a highly complex molecule consisting of the hexa-acylated diglucosamine lipid A at its base, followed by the core oligosaccharide (an oligomer of glucose, galactose, and heptose in most species) and the O-antigen, a structure with a high level of interspecific variability (3, 4). Due to lipid A’s 6 acyl chains, the lipid bilayer portion of the OM is highly hydrophobic. In addition, lipid A as well as core oligosaccharide contain anionic phosphate and carboxylic acid groups. These anionic residues can engage in electrostatic interactions with cations (e.g., Mg2⫹ and Ca2⫹) across neighboring LPS molecules, resulting in ion bridges that stabilize and stiffen the OM (5). The complex, amphiphilic composition and stiffness of the outer membrane render it largely impermeable to large molecules, such as a variety of antibiotics (5). Smaller molecules (water and nutrients) gain access to the periplasm via OM porins, largely nonspecific transmembrane channels that allow the diffusion of solutes smaller than ⬃600 Da (1). Porins also serve as the entryways for small, hydrophilic beta-lactam antibiotics (6, 7), such as the penicillins and cephalosporins. In contrast, slow cross-membrane diffusion is apparently required for the entry of larger, lipophilic compounds such as rifampin and novobiocin (8), which in part explains the comparatively high concentrations required for their antibacterial action. Some antibiotics seem entirely unable to cross the OM, as Gram-negative bacteria are generally resistant to them, even though the antibiotics’ targets are present. These antibiotics include the very large (⬎1.4-kDa) glycopeptides vancomycin and ramoplanin as well as the polypeptide antibiotic bacitracin. Several conditions and mutations are known to result in an increase in OM permeability. For example, EDTA treatment
August 2016 Volume 60 Number 8
causes the leakage of periplasmic proteins into the surrounding medium and increases OM permeability to lysozyme and large antibiotics (9). EDTA is thought to destabilize ionic interactions between LPS molecules by removing the divalent cations that stabilize the anionic groups of LPS. LPS instability leads to the loss of LPS into the surrounding medium and patchy replacement of the typical asymmetric phospholipid-LPS OM with a symmetric phospholipid bilayer. Phospholipid bilayer patches are more permeable to hydrophobic compounds and susceptible to disruption by detergents (9, 10). The increase in OM permeability observed in many cell envelope mutants, e.g., so-called “rough” mutants defective to various degrees in O-antigen and core oligosaccharide production, is likewise thought to result from the accumulation of symmetric phospholipid bilayer patches associated with LPS instability (5, 9). Some bacteria, including Vibrio cholerae, appear to have naturally occurring symmetric bilayer patches and are thus intrinsically more susceptible to some hydrophobic compounds (such as rifampin and novobiocin), but the genetic basis for the
Received 13 March 2016 Returned for modification 27 April 2016 Accepted 19 May 2016 Accepted manuscript posted online 23 May 2016 Citation Dörr T, Delgado F, Umans BD, Gerding MA, Davis BM, Waldor MK. 2016. A transposon screen identifies genetic determinants of Vibrio cholerae resistance to high-molecular-weight antibiotics. Antimicrob Agents Chemother 60:4757–4763. doi:10.1128/AAC.00576-16. Address correspondence to Tobias Dörr, [email protected], or Matthew K. Waldor, [email protected]. * Present address: Tobias Dörr, Department of Microbiology/Weill Institute for Cell and Molecular Biology, Cornell University, Ithaca, NY, USA. Supplemental material for this article may be found at http://dx.doi.org/10.1128 /AAC.00576-16. Copyright © 2016, American Society for Microbiology. All Rights Reserved.
Antimicrobial Agents and Chemotherapy
aac.asm.org
4757
Dörr et al.
formation of these patches is not known (11). Apart from treatments and mutations that cause drastic changes in OM properties, the genetic factors that contribute to the maintenance of the OM barrier to large antibiotics are only incompletely understood. Here, we investigated the genetic determinants of the OM permeability barrier in V. cholerae, the cholera pathogen. Using a transposon insertion sequencing-based screen, we identified factors required for the resistance of V. cholerae to vancomycin and bacitracin. In addition to factors known or predicted to mediate OM integrity and/or the extracellular barrier to antibiotic diffusion, we also found that a previously uncharacterized inner membrane protein, which we name VigA, is required for vancomycin resistance. Surprisingly, a ⌬vigA mutant is susceptible to vancomycin and ramoplanin but does not exhibit increased susceptibility to other large antibiotics, hydrophobic compounds, or detergents, suggesting that the integrity of its OM is not generally compromised. Finally, we observed that there is a large range in the MICs of glycopeptides and bacitracin (antibiotics generally thought to have a minimal capacity to penetrate the OM) among various Gram-negative pathogens. Collectively, our observations suggest that the OM permeability barrier to large antibiotics is not absolute. Instead, we speculate that this barrier depends on specific attributes of both the OM and the antibiotic in question. MATERIALS AND METHODS Media and growth conditions. All strains were grown in 3 ml LB broth in 10-ml glass tubes at 37°C with shaking (225 rpm). Antibiotics (all from Sigma-Aldrich), where applicable, were used at the following concentrations: 200 g/ml streptomycin, 50 g/ml carbenicillin, 200 g/ml vancomycin, 50 g/ml kanamycin, and 5 g/ml chloramphenicol. Strains and plasmids. All V. cholerae strains are derivatives of El Tor C6706 unless otherwise indicated. Defined transposon mutants were taken from an ordered transposon insertion library (12). The defined vigA deletion strain (replacement of the open reading frame [ORF] with the sequence 5=-T TATCATGCGGCCGCACTCGAGTAATGATAA-3=) was constructed by allele replacement via homologous recombination using suicide plasmid pCVD442 as described previously (13). Regions of homology were amplified by using primers TDP1494 (5=-AGGTATATGTGATGGGTTAAAAAGGAT CGATCCTGACCCCACACTAAACTATCAACAGTTT-3=)/TDP1495 (5=TTATCATGCGGCCGCACTCGAGTAATGATAATCAGCGAAT CAGTTCAGAAATCATCAAG-3=) and TDP1496 (5=-TTATCATTAC TCGAGTGCGGCCGCATGATAAGGCGCTCTCAGAAGAGATA GG-3=)/TDP1497 (5=-CCGGGAGAGCTCGATATCGCATGCGG TACCTCTAGTCGTGAAGTCTTCCATATCGCCCT-3=) and cloned into XbaI-digested pCVD442 by isothermal assembly (14) (boldface type represents homology overhangs for isothermal assembly cloning). For complementation, the vigA ORF was placed under the control of an isopropyl-D-thiogalactopyranoside (IPTG)-inducible promoter followed by integration of the whole construct into the neutral chromosomal lacZ locus. Transposon insertion sequencing. Under each condition (2 with no antibiotic control, 2 with vancomycin, and 2 with bacitracin), 1 ml of a culture of V. cholerae grown overnight was mixed with 1 ml of a culture of donor strain SM10 carrying pSC189 (15) (transposon donor plasmid) grown overnight. The mixture was pelleted by centrifugation (5,000 ⫻ g for 5 min) and resuspended in 1 ml LB broth. This mixture was aliquoted (100 l each) onto 0.45-m filter disks on prewarmed (37°C) LB agar plates. After 3 h of incubation, cells were scraped off the filter disks using a pipette tip into LB medium containing streptomycin, vortexed, and plated directly onto LB agar plates containing streptomycin (to kill the donor strain) and either no antibiotic (control library), 200 g/ml vancomycin, or 200 g/ml bacitracin. After overnight incubation (resulting in ⬃200,000 colonies under each mating condition), cells were scraped off the plates and pooled. Libraries for Illumina sequencing were prepared as
4758
aac.asm.org
described previously (16). Briefly, following genomic DNA extraction, DNA was sheared by using sonication, and overhanging ends were blunted by using a blunting enzyme kit (NEB). This was followed by Illumina adaptor ligation and amplification of genomic DNA-transposon junctions. Sequencing reactions were conducted on an Illumina MiSeq benchtop sequencer (Illumina, San Diego, CA). Data analysis was conducted as described previously (16, 17). Cell fractionation. Cell fractionation was conducted by differential centrifugation essentially as described previously (18), with modifications. Cultures grown overnight were diluted 1,000-fold into 50 ml of fresh LB broth in baffled flasks and grown for 1 h with shaking at 37°C. Where applicable, 100 M IPTG was added to induce the expression of epitope-tagged VigA, followed by incubation for an additional 2 h. Cells were then harvested by centrifugation (15 min at 15,000 ⫻ g at 4°C), washed once in cold phosphate-buffered saline (PBS) (pH 7.4), and resuspended in 1.5 ml of the same buffer supplemented with lysozyme (5 mg/ml), Halt protease inhibitor tablets (Fisher Scientific), and 1 mM phenylmethylsulfonyl fluoride (PMSF). Cells were lysed by sonication (3 times for 10 s) on ice, followed by centrifugation (15 min at 12,000 ⫻ g at 4°C) to remove debris. The supernatant (envelope fraction plus cytoplasmic fraction) was then subjected to ultracentrifugation (200,000 ⫻ g at 4°C) to obtain membranes. To separate the outer from the inner membrane, the latter was solubilized by resuspending the membrane pellet in 1 ml of PBS containing 0.4% reduced Triton X-100 (Sigma), followed by incubation on a rotator for 3 h at 4°C. Subsequent ultracentrifugation (200,000 ⫻ g at 4°C) was then used to separate the inner (solubilized) and outer (pellet) membranes. Fluorescent vancomycin staining. Cells were grown to exponential phase, washed once with PBS, and then resuspended in the same medium to ⬃1 ⫻ 109 CFU/ml. BODIPY FL vancomycin (Thermo Scientific) was added at 100 g/ml for 10 min (ambient temperature). Cells were imaged after washing once in PBS. MICs. MICs were assessed by using Clinical and Laboratory Standards Institute (CLSI) guidelines; however, instead of Mueller-Hinton broth, we used LB broth due to the poor growth of V. cholerae in the former medium. Protein techniques. Western blotting was conducted by using standard protocols. Cell lysates were normalized by the protein content (measured by Qubit; Thermo Fisher Scientific).
RESULTS
Similar to other Gram-negative bacteria, V. cholerae is resistant to large antibiotics like vancomycin (1.5 kDa) and bacitracin (1.4 kDa). Even though the targets for vancomycin (nascent cell wall/ cell wall precursors) and bacitracin (lipid II, a precursor for cell wall synthesis) are widely conserved in Gram-negative bacteria (19), the OM permeability barrier in these organisms renders them resistant to these agents. We used a transposon insertion sequencing-based strategy to identify loci that were required for V. cholerae resistance to vancomycin and/or bacitracin and, thus, presumably for maintenance of the OM permeability barrier. We created Himar1 transposon insertion libraries in an El Tor clinical isolate of V. cholerae and compared the relative frequencies of transposon insertions at each locus (a proxy for mutant fitness) when the library was grown in the absence versus the presence of vancomycin or bacitracin. We identified eight loci for which insertions were significantly underrepresented after growth on vancomycin plates and one locus for which insertions were underrepresented after growth on bacitracin (Fig. 1A and B). The single gene identified in the bacitracin screen, vc1423, was also one of the eight loci detected in the vancomycin screen. The 8 genes that our screens identified constitute candidate determinants of outer membrane barrier function. Two of these genes—galE, coding for UDP-glucose 4-epimerase (20, 21), and
Antimicrobial Agents and Chemotherapy
August 2016 Volume 60 Number 8
Genetic Basis of V. cholerae Glycopeptide Resistance
FIG 1 Identification of genes conferring enhanced susceptibility to vancomycin and bacitracin using transposon insertion sequencing. (A and B) Volcano plots
depicting the ratio of read counts mapped to individual genes in transposon libraries plated onto vancomycin (200 g/ml) (A) or bacitracin (200 g/ml) (B) compared to control libraries plated onto LB agar without antibiotics (mean fold change, x axis) versus mean P values (y axis) (see Materials and Methods for details). Genes shown in red and blue are considered significantly underrepresented. (C) Validation of plating defects on vancomycin of selected mutants identified in the transposon screen. Mutants were plated onto either LB agar containing vancomycin (200 g/ml) or LB agar alone (⫺ vancomycin). (D and E) Sensitivity of a V. cholerae vigA mutant (⌬vc1423) to vancomycin and bacitracin. Strains (including the ⌬vc1423 strain ectopically expressing vc1423 under the control of an IPTG-inducible promoter) were spot plated onto agar containing either vancomycin (200 g/ml) (D) or bacitracin (200 g/ml) (E) and IPTG (100 M). (F) Vancomycin MICs of V. cholerae WT and ⌬vc1423 strains.
cls, coding for the synthase of cardiolipin, an anionic membrane phospholipid (22)— have known roles in cell envelope permeability in other Gram-negative bacteria (cls and galE) and in biofilm formation in V. cholerae (galE only) (20). Biofilms constitute a diffusion barrier for large antibiotics in other bacteria (23). Components of both operons required for the production of exopolysaccharide (EPS), a capsule-like extracellular polymer with a crucial role in biofilm formation (24), were also identified. Finally, both screens identified vc1423, a gene that encodes a hypothetical protein of unknown function and is discussed in more detail below. We validated the results of our transposon screen by spotting defined mutants corresponding to each category on LB agar plates
August 2016 Volume 60 Number 8
with and without vancomycin. Growth of wild-type (WT) V. cholerae was not impaired by 200 g/ml vancomycin, but all mutants displayed at least a modest (5- to 10-fold) plating deficiency, and the ⌬vc1423 mutant exhibited a drastic (10,000-fold) plating deficiency when spotted onto vancomycin plates (Fig. 1C and D). Interestingly, the four mutants that exhibited modest plating deficiencies had no reduction in vancomycin MICs compared to the WT (data not shown), while the ⌬vc1423 mutant exhibited a 4-fold decrease in the MIC against vancomycin (Fig. 1F). Thus, the vancomycin susceptibility of some strains is associated with particular attributes of their environments, e.g., growth on surfaces versus growth in liquid, or the unchanged MIC reports the
Antimicrobial Agents and Chemotherapy
aac.asm.org
4759
Dörr et al.
FIG 2 VigA is an inner membrane protein. (A) Cells carrying either ectopically expressed, His-tagged VigA or chromosomally encoded, His-tagged PBP1A (control protein for inner membrane localization) or LpoA (control protein for outer membrane localization) were lysed and separated into fractions via differential centrifugation. Proteins were visualized by immunoblotting with an antibody recognizing the His epitope. OM, outer membrane fractions; IM, inner membrane fractions. (B) Ectopically expressed N- or C-terminal fusions of VigA with beta-lactamase were plated onto agar containing an inducer (IPTG, 100 M) and carbenicillin. The control is an empty plasmid. bla, beta-lactamase. (C) Model of VigA topology based on the experiments described above. C, C terminus.
growth of a phenotypically resistant subpopulation of cells. The ⌬vc1423 mutant also exhibited a marked plating deficiency when spotted onto plates containing bacitracin (Fig. 1E) or the 2.5-kDa glycopeptide ramoplanin (see Fig. S1 in the supplemental material). Expression of vc1423 in trans complemented the plating defect for all three antibiotics (Fig. 1D and E; see also Fig. S1 in the supplemental material). Thus, vc1423 (renamed vigA for vancomycin inhibits growth) is crucial for V. cholerae to maintain its resistance to these 3 large antibiotics. VigA is an inner membrane protein with an N-in, C-out topology. VigA is predicted to be a mostly helical protein with a short N-terminal cytoplasmic segment, followed by a transmembrane domain and a large periplasmic domain. We carried out cell fractionation experiments to verify that VigA localizes to the cell membrane. Lysates of cells ectopically expressing His-tagged VigA were separated into inner membrane and outer membrane fractions, using differential centrifugation. Immunoblotting was then used to detect epitope-tagged VigA as well as His-tagged versions of known inner and outer membrane proteins: penicillin binding protein 1A (PBP1A) and LpoA, respectively (25, 26). VigA was present in the inner membrane fraction but absent from the outer membrane fraction, consistent with its predicted localization (Fig. 2A). PBP1A and LpoA were highly enriched in the inner membrane and outer membrane samples, respectively, validating the data from the fractionation assay. To validate VigA’s predicted N-in, C-out membrane topology,
4760
aac.asm.org
FIG 3 VigA mutant cells appear more permeable to vancomycin than do WT cells. (A) Cells were grown in LB medium, washed once in PBS, stained for 10 min with fluorescent vancomycin (vancomycin FL), washed twice with PBS, and then imaged. The red arrowheads point to lysed WT cells. (B) Prior to imaging, cells were grown to exponential phase (optical density at 600 nm of 0.5) in LB medium and then exposed to the indicated concentrations of vancomycin for 3 h.
we constructed N-terminal and C-terminal fusions of VigA to beta-lactamase. This protein hydrolyzes beta-lactam antibiotics, conferring resistance to these agents, only when it is exported to the periplasm. Both N-terminal and C-terminal VigA– beta-lactamase fusions were stably expressed, and both fusions were able to complement the plating deficiency on vancomycin of a vigA mutant, indicating that both fusion proteins are functional (see Fig. S2 in the supplemental material). However, only the C-terminal fusion allowed growth in the presence of carbenicillin (Fig. 2B), suggesting that the C-terminal portion of VigA is in the periplasm, while the N terminus localizes to the cytoplasm, as predicted (Fig. 2C). Vancomycin accumulates in the periplasm of a vigA mutant and causes loss of rod shape. The vigA mutant’s pronounced growth deficiency (relative to WT cells) on plates containing vancomycin suggested that this glycopeptide antibiotic was able to penetrate the OM of the vigA mutant and bind to its target (pentapeptide side chains of peptidoglycan (PG) [27]). To test this possibility, we briefly stained WT and vigA mutant cells with fluorescently labeled vancomycin (Van-FL). In the WT culture, Van-FL stained only lysed cells (in which the vancomycin target is exposed due to disruption of the OM) (Fig. 3A, arrowheads), while intact cells did not accumulate any periplasmic fluorescence. In contrast, vigA mutant cells rapidly accumulated peripheral fluorescence, consistent with the idea that Van-FL can enter and/or remain in the periplasm. When vancomycin binds to its target, pentapeptide residues in nascent PG, it sterically prevents transpeptidation by the major cell wall synthases, the penicillin binding proteins, and thus effectively inhibits cell wall synthesis (27). To gauge whether vancomy-
Antimicrobial Agents and Chemotherapy
August 2016 Volume 60 Number 8
Genetic Basis of V. cholerae Glycopeptide Resistance
FIG 4 The general integrity of the OM in the vigA mutant is not compromised. (A) Comparison of MICs of membrane stressors and large antibiotics in V. cholerae WT and ⌬vigA strains. (B) Growth curves of V. cholerae WT and ⌬vigA strains in the presence or absence of 1% bile. (C) Endogenous RpoE levels in WT V. cholerae versus the ⌬vigA mutant. Lysates of exponentially growing cells were subjected to SDS-PAGE followed by immunoblotting with an antibody raised against RpoE. Lysates were normalized by both the optical density at 600 nm (OD600) and protein content before loading. The RNA polymerase subunit RpoB is the internal control.
cin’s ability to inhibit the vigA mutant’s growth was associated with inhibition of cell wall synthesis, we treated growing cultures with various concentrations of vancomycin and observed the effect on cell morphology. Treatment of vigA cells with ⬃0.5⫻ or ⬃1⫻ the vancomycin MIC for the mutant cells resulted in a loss of rod shape and sphere formation (Fig. 3B), as observed when V. cholerae is treated with beta-lactams and other inhibitors of cell wall synthesis (28) whose action results in a loss of the cell wall. Surprisingly, wild-type cells treated with 200 g/ml of vancomycin (approximately the MIC for the vigA mutant and 0.2⫻ the WT MIC), also exhibited changes in morphology but did not form spheres (Fig. 3B). Thus, although WT V. cholerae is resistant to vancomycin, its OM likely does not completely block the entry of this antibiotic and the associated inhibition of cell wall synthesis. However, our data suggest that vancomycin has a more profound effect on cell wall synthesis in the vigA mutant, consistent with this strain’s lower MIC and its periplasmic accumulation of vancomycin. OM integrity is not generally compromised in the vigA mutant. To investigate whether vancomycin sensitivity in the vigA mutant was indicative of a general OM integrity defect, we measured the mutant’s sensitivity to SDS and the high-molecularweight antibiotics rifampin and novobiocin. Surprisingly, MICs of all three agents were the same for the vigA mutant and the wild-type strain (Fig. 4A). MICs of bacitracin and ramoplanin also did not differ significantly between the WT strain and the vigA mutant (Fig. 4A), despite the mutant’s marked plating defect with these agents, providing further evidence that disruption of vigA
August 2016 Volume 60 Number 8
does not result in a general OM defect. Furthermore, we found that the efficiencies of plating of WT and ⌬vigA strains on plates containing SDS and increasing concentrations of EDTA (a standard assay for OM defects [10]) were very similar (see Fig. S3 in the supplemental material), and the vigA mutant grew as well as the WT in the presence of 1% bile, a membrane-disrupting agent (Fig. 4B). Finally, the abundance of the alternate sigma factor RpoE, which increases in response to envelope stress, was not elevated in the vigA mutant (Fig. 4C), nor were there marked differences in inner or outer membrane protein contents (e.g., an increase in the abundance of outer membrane porins) in the mutant compared to the WT (see Fig. S4 in the supplemental material). Collectively, these observations suggest that the increased sensitivity of the vigA mutant to vancomycin does not result from a generalized OM defect in this strain; instead, the vigA deletion results in sensitivity that appears to be specific to certain antibiotics and is also modulated by the bacterial growth environment (i.e., liquid versus solid medium). Variability of glycopeptide/bacitracin resistance in Gramnegative pathogens. Finally, we explored whether several other Gram-negative pathogens exhibit some degree of sensitivity to glycopeptides and bacitracin. While WT strains of V. cholerae (including a clinical isolate from the current outbreak in Haiti) exhibited very similar MIC values against the three compounds tested, there was considerable interspecies variability in MIC values (Table 1). Bacitracin inhibited Acinetobacter baumannii growth at 62 g/ml, while all other bacteria were resistant even at 1,000 g/ml (limit of detection). However, the susceptibility of A.
Antimicrobial Agents and Chemotherapy
aac.asm.org
4761
Dörr et al.
TABLE 1 Variations in susceptibility to glycopeptides and bacitracin among Gram-negative pathogensa MIC (g/ml)
Antibiotic
V. cholerae V. cholerae MW C6706 H1 EHEC
Bacitracin 1,422 >1,000 Vancomycin 1,500 1,000 Ramoplanin 2,500 125
>1,000 1,000 125
Pseudomonas aeruginosa A. baumannii PA14 Lac-4
>1,000 >1,000 500 >1,000 500 >1,000
62 500 125
a Shown are MICs of glycopeptides and bacitracin against selected Gram-negative pathogens. MW, molecular weight. V. cholerae clinical isolate H1 was obtained from the recent cholera outbreak in Haiti. Boldface type indicates MIC values above the limit of detection.
baumannii to vancomycin was comparable to that of the other three pathogens tested. Conversely, enterohemorrhagic Escherichia coli (EHEC) had a significantly higher MIC for ramoplanin (but not vancomycin or bacitracin) than the other pathogens. Interestingly, almost all organisms (with the exception of A. baumannii) were more susceptible to the largest compound (ramoplanin [2.5 kDa]) than to the smallest (bacitracin [1.4 kDa]), suggesting that the size of the antibiotic alone is not a good predictor of antimicrobial efficacy. Taken together, the variability in sensitivity to the two large glycopeptide antibiotics and bacitracin among these 4 Gram-negative pathogens suggests that the OM permeability barrier is not absolute but depends on species-specific biophysical or biochemical interactions between the OM and particular antibiotics. DISCUSSION
Mutations that confer severe and general OM defects have been well described (9, 29), but there is less knowledge of specific genetic requirements for OM barrier function in Gram-negative bacteria. Here, we took advantage of the power of transposon insertion sequencing to identify genetic factors that contribute to the generation or maintenance of the V. cholerae OM permeability barrier against vancomycin and bacitracin. One gene of unknown function, vc1423, here renamed vigA, was identified in both screens. In the absence of VigA, an inner membrane protein, cells became highly sensitive to glycopeptide antibiotics and bacitracin when grown on solid medium. In liquid cultures, exposure to vancomycin resulted in a loss of rod shape for the vigA mutant, and the ⌬vigA strain could be stained with fluorescent vancomycin. The vigA mutant was not sensitive to membrane-damaging agents, either on solid medium or in liquid culture. Thus, VigA appears to specifically prevent the periplasmic accumulation of certain large antibiotics without exerting a general role in the maintenance of OM integrity. Besides vigA, the screen identified Vibrio exopolysaccharide (VPS) biosynthesis genes and galE as requirements for vancomycin resistance. Both VPS and galE are required for biofilm formation (20, 30). While biofilms in general are thought to constitute a significant diffusion barrier to vancomycin, e.g., in Gram-positive Staphylococcus spp. (23, 31), VPS or similar secreted exopolysaccharides have not, to our knowledge, been linked to OM permeability or to limiting the diffusion of large antibiotics into vibrios or other Gram-negative organisms. Given its sensitivity to high-molecular-weight glycopeptides, the vigA mutant’s lack of a general OM defect was unexpected. Most mutations that result in an increase in OM permeability to
4762
aac.asm.org
large compounds also tend to cause a fairly severe impairment of overall OM integrity and an increased permeability for other, unrelated compounds (32, 33). The lack of SDS and bile sensitivity in the ⌬vigA mutant suggests that the mutant’s defect is not attributable to elevations in the amount of phospholipid bilayer patches in its OM, and overall, VigA does not appear to be involved in crucial OM maintenance functions. There are several potential explanations for the vigA mutant’s sensitivity to glycopeptides and bacitracin. One possibility is that the ⌬vigA mutation results in subtle changes in the biophysical properties of LPS, enabling increased diffusion of glycopeptides and bacitracin. Such a subtle OM alteration is also thought to underlie the glycopeptide sensitivity of an unidentified mutation isolated in E. coli (34). Alternatively, VigA might be important for cell wall repair processes, rather than OM homeostasis per se, or for drug efflux. Cell wall repair or modification processes could compensate for damage caused by vancomycin or affect the abundance of its target, while an efflux pump could limit antibiotic accumulation in the cell. While we cannot entirely exclude any explanation, detection of fluorescent vancomycin staining in the vigA mutant suggests increased accumulation (due to enhanced influx or reduced efflux) rather than a cell wall repair defect. VigA has strong homologues only within the family Vibrionaceae. However, VigA is a member of the more widely conserved but poorly characterized HemX superfamily. Future studies will address whether HemX homologues in other Gram-negative bacteria also play a role in glycopeptide resistance. If so, the development of agents that inactivate the activity of VigA-like proteins could enable the use of glycopeptide antibiotics in emerging Gram-negative pathogens such as Acinetobacter baumannii. The Gram-negative OM is often described as an absolute diffusion barrier to large antibiotics. However, exposure to subMICs of vancomycin resulted in shape defects in V. cholerae, suggesting that this organism’s OM is not entirely impermeable to this compound even at concentrations that do not inhibit population growth. Furthermore, there was not a simple correlation between antibiotic molecular weight and MIC against a panel of Gram-negative pathogens, suggesting that the OM diffusion barrier for these compounds is not based simply on size. Instead, our observations suggest that the efficacy of large antibiotics depends on specific interactions of antibiotics with specific OMs. Studies of structure-permeability relationships coupled with medicinal chemistry are an avenue of investigation that could potentially enable the use of our current arsenal of Gram-positive-specific drugs on Gram-negative pathogens. ACKNOWLEDGMENTS We thank the Mekalanos laboratory (in particular William Robins) for the gift of the OmpU antibody and Carol Gross for the RpoE antisera.
FUNDING INFORMATION This work, including the efforts of Matthew K. Waldor, was funded by HHS | National Institutes of Health (NIH) (R37 AI-042347). This work, including the efforts of Matthew K. Waldor, was funded by Howard Hughes Medical Institute (HHMI). This work, including the efforts of Fernanda Delgado, was funded by Howard Hughes Medical Institute (HHMI).
REFERENCES 1. Delcour AH. 2009. Outer membrane permeability and antibiotic resistance. Biochim Biophys Acta 1794:808 – 816. http://dx.doi.org/10.1016/j .bbapap.2008.11.005.
Antimicrobial Agents and Chemotherapy
August 2016 Volume 60 Number 8
Genetic Basis of V. cholerae Glycopeptide Resistance
2. Silhavy TJ, Kahne D, Walker S. 2010. The bacterial cell envelope. Cold Spring Harb Perspect Biol 2:a000414. http://dx.doi.org/10.1101 /cshperspect.a000414. 3. Chatterjee SN, Chaudhuri K. 2003. Lipopolysaccharides of Vibrio cholerae. I. Physical and chemical characterization. Biochim Biophys Acta 1639:65–79. http://dx.doi.org/10.1016/j.bbadis.2003.08.004. 4. Whitfield C, Trent MS. 2014. Biosynthesis and export of bacterial lipopolysaccharides. Annu Rev Biochem 83:99 –128. http://dx.doi.org/10 .1146/annurev-biochem-060713-035600. 5. Nikaido H. 2003. Molecular basis of bacterial outer membrane permeability revisited. Microbiol Mol Biol Rev 67:593– 656. http://dx.doi.org/10 .1128/MMBR.67.4.593-656.2003. 6. Kojima S, Nikaido H. 2013. Permeation rates of penicillins indicate that Escherichia coli porins function principally as nonspecific channels. Proc Natl Acad Sci U S A 110:E2629 –E2634. http://dx.doi.org/10.1073/pnas .1310333110. 7. Yoshimura F, Nikaido H. 1985. Diffusion of beta-lactam antibiotics through the porin channels of Escherichia coli K-12. Antimicrob Agents Chemother 27:84 –92. http://dx.doi.org/10.1128/AAC.27.1.84. 8. Nikaido H. 1976. Outer membrane of Salmonella typhimurium. Transmembrane diffusion of some hydrophobic substances. Biochim Biophys Acta 433:118 –132. 9. Hancock RE. 1984. Alterations in outer membrane permeability. Annu Rev Microbiol 38:237–264. http://dx.doi.org/10.1146/annurev.mi.38 .100184.001321. 10. Malinverni JC, Silhavy TJ. 2009. An ABC transport system that maintains lipid asymmetry in the gram-negative outer membrane. Proc Natl Acad Sci U S A 106:8009 – 8014. http://dx.doi.org/10.1073/pnas.0903229106. 11. Paul S, Chaudhuri K, Chatterjee AN, Das J. 1992. Presence of exposed phospholipids in the outer membrane of Vibrio cholerae. J Gen Microbiol 138:755–761. http://dx.doi.org/10.1099/00221287-138-4-755. 12. Cameron DE, Urbach JM, Mekalanos JJ. 2008. A defined transposon mutant library and its use in identifying motility genes in Vibrio cholerae. Proc Natl Acad Sci U S A 105:8736 – 8741. http://dx.doi.org/10.1073/pnas .0803281105. 13. Donnenberg MS, Kaper JB. 1991. Construction of an eae deletion mutant of enteropathogenic Escherichia coli by using a positive-selection suicide vector. Infect Immun 59:4310 – 4317. 14. Gibson DG, Young L, Chuang RY, Venter JC, Hutchison CA, III, Smith HO. 2009. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods 6:343–345. http://dx.doi.org/10.1038/nmeth .1318. 15. Chiang SL, Rubin EJ. 2002. Construction of a mariner-based transposon for epitope-tagging and genomic targeting. Gene 296:179 –185. http://dx .doi.org/10.1016/S0378-1119(02)00856-9. 16. Chao MC, Pritchard JR, Zhang YJ, Rubin EJ, Livny J, Davis BM, Waldor MK. 2013. High-resolution definition of the Vibrio cholerae essential gene set with hidden Markov model-based analyses of transposoninsertion sequencing data. Nucleic Acids Res 41:9033–9048. http://dx.doi .org/10.1093/nar/gkt654. 17. Pritchard JR, Chao MC, Abel S, Davis BM, Baranowski C, Zhang YJ, Rubin EJ, Waldor MK. 2014. ARTIST: high-resolution genome-wide assessment of fitness using transposon-insertion sequencing. PLoS Genet 10:e1004782. http://dx.doi.org/10.1371/journal.pgen.1004782. 18. Thein M, Sauer G, Paramasivam N, Grin I, Linke D. 2010. Efficient
August 2016 Volume 60 Number 8
19. 20.
21. 22. 23.
24.
25.
26. 27. 28. 29. 30.
31. 32. 33.
34.
subfractionation of gram-negative bacteria for proteomics studies. J Proteome Res 9:6135– 6147. http://dx.doi.org/10.1021/pr1002438. Scheffers DJ, Tol MB. 2015. LipidII: just another brick in the wall? PLoS Pathog 11:e1005213. http://dx.doi.org/10.1371/journal.ppat.1005213. Nesper J, Lauriano CM, Klose KE, Kapfhammer D, Kraiss A, Reidl J. 2001. Characterization of Vibrio cholerae O1 El Tor galU and galE mutants: influence on lipopolysaccharide structure, colonization, and biofilm formation. Infect Immun 69:435– 445. http://dx.doi.org/10.1128/IAI.69.1 .435-445.2001. Robertson BD, Frosch M, van Putten JP. 1993. The role of galE in the biosynthesis and function of gonococcal lipopolysaccharide. Mol Microbiol 8:891–901. http://dx.doi.org/10.1111/j.1365-2958.1993.tb01635.x. Bishop RE. 2014. Emerging roles for anionic non-bilayer phospholipids in fortifying the outer membrane permeability barrier. J Bacteriol 196: 3209 –3213. http://dx.doi.org/10.1128/JB.02043-14. Darouiche RO, Dhir A, Miller AJ, Landon GC, Raad II, Musher DM. 1994. Vancomycin penetration into biofilm covering infected prostheses and effect on bacteria. J Infect Dis 170:720 –723. http://dx.doi.org/10.1093 /infdis/170.3.720. Yildiz FH, Schoolnik GK. 1999. Vibrio cholerae O1 El Tor: identification of a gene cluster required for the rugose colony type, exopolysaccharide production, chlorine resistance, and biofilm formation. Proc Natl Acad Sci U S A 96:4028 – 4033. http://dx.doi.org/10.1073/pnas.96.7.4028. Dorr T, Lam H, Alvarez L, Cava F, Davis BM, Waldor MK. 2014. A novel peptidoglycan binding protein crucial for PBP1A-mediated cell wall biogenesis in Vibrio cholerae. PLoS Genet 10:e1004433. http://dx.doi.org /10.1371/journal.pgen.1004433. Typas A, Banzhaf M, Gross CA, Vollmer W. 2012. From the regulation of peptidoglycan synthesis to bacterial growth and morphology. Nat Rev Microbiol 10:123–136. http://dx.doi.org/10.1038/nrmicro2677. Perkins HR, Nieto M. 1974. The chemical basis for the action of the vancomycin group of antibiotics. Ann N Y Acad Sci 235:348 –363. http: //dx.doi.org/10.1111/j.1749-6632.1974.tb43276.x. Dorr T, Davis BM, Waldor MK. 2015. Endopeptidase-mediated beta lactam tolerance. PLoS Pathog 11:e1004850. http://dx.doi.org/10.1371 /journal.ppat.1004850. Sampson BA, Misra R, Benson SA. 1989. Identification and characterization of a new gene of Escherichia coli K-12 involved in outer membrane permeability. Genetics 122:491–501. Fong JC, Syed KA, Klose KE, Yildiz FH. 2010. Role of Vibrio polysaccharide (vps) genes in VPS production, biofilm formation and Vibrio cholerae pathogenesis. Microbiology 156:2757–2769. http://dx.doi.org/10 .1099/mic.0.040196-0. Choudhry O, Gathwala G, Singh J. 2010. Vancomycin resistant enterococci in neonatal ICU—a rising menace. Indian J Pediatr 77:1446 –1447. http://dx.doi.org/10.1007/s12098-010-0243-6. Vaara M. 1993. Antibiotic-supersusceptible mutants of Escherichia coli and Salmonella typhimurium. Antimicrob Agents Chemother 37:2255– 2260. http://dx.doi.org/10.1128/AAC.37.11.2255. Vuorio R, Vaara M. 1992. The lipid A biosynthesis mutation lpxA2 of Escherichia coli results in drastic antibiotic supersusceptibility. Antimicrob Agents Chemother 36:826 – 829. http://dx.doi.org/10.1128/AAC.36 .4.826. Shlaes DM, Shlaes JH, Davies J, Williamson R. 1989. Escherichia coli susceptible to glycopeptide antibiotics. Antimicrob Agents Chemother 33:192–197. http://dx.doi.org/10.1128/AAC.33.2.192.
Antimicrobial Agents and Chemotherapy
aac.asm.org
4763
