Genetic Determinants of Salmonella Resistance to the Biofilm-Inhibitory Effects of a Synthetic 4-Oxazolidinone Analog

Biofilms are resistant to killing by disinfectants and antimicrobials. S. enterica biofilms facilitate long-term host colonization and persistence in food processing environments. Synthetic analogs of 4-oxazolidinone natural products show promise as antibiofilm agents. Here, we show that a synthetic 4-oxazolidinone analog inhibits Salmonella biofilm through effects on both motility and biofilm matrix gene expression. Furthermore, we identify three genes that promote Salmonella resistance to the antibiofilm effects of the compound. This work provides insight into the mechanism of antibiofilm effects of a synthetic 4-oxazolidinone analog in Gram-negative bacteria and demonstrates new mechanisms of intrinsic antimicrobial resistance in Salmonella biofilms.

ABSTRACT

Biofilms formed by Salmonella enterica are a frequent source of food supply contamination. Since biofilms are inherently resistant to disinfection, new agents capable of preventing biofilm formation are needed. Synthetic analogs of 4-oxazolidinone containing natural products have shown promise as antibiofilm compounds against Gram-positive bacteria. The purpose of our study was 2-fold: to establish the antibiofilm effects and mechanism of action of a synthetic 4-oxazolidinone analog (JJM-ox-3-70) and to establish mechanisms of resistance to this compound in Salmonella enterica serovar Typhimurium (S. Typhimurium). JJM-ox-3-70 inhibited biofilm formation but had no effect on cell growth. The antibiofilm effects were linked to disruption of curli fimbriae and flagellar gene expression and alteration in swimming motility, suggesting an effect on multiple cellular processes. Using a 2-step screening approach of defined multigene and single-gene deletion mutant libraries, we identified 3 mutants that produced less biofilm in the presence of JJM-ox-3-70 than the isogenic WT, with phenotypes reversed by complementation in trans. Genes responsible for S. Typhimurium resistance to the compound included acrB, a component of the major drug efflux pump AcrAB-TolC, and two genes of unknown function (STM0437 and STM1292). The results of this study suggest that JJM-ox-3-70 inhibits biofilm formation by indirect inhibition of extracellular matrix production that may be linked to disruption of flagellar motility. Further work is needed to establish the role of the newly characterized genes as potential mechanisms of biofilm intrinsic antimicrobial resistance.

IMPORTANCE Biofilms are resistant to killing by disinfectants and antimicrobials. S. enterica biofilms facilitate long-term host colonization and persistence in food processing environments. Synthetic analogs of 4-oxazolidinone natural products show promise as antibiofilm agents. Here, we show that a synthetic 4-oxazolidinone analog inhibits Salmonella biofilm through effects on both motility and biofilm matrix gene expression. Furthermore, we identify three genes that promote Salmonella resistance to the antibiofilm effects of the compound. This work provides insight into the mechanism of antibiofilm effects of a synthetic 4-oxazolidinone analog in Gram-negative bacteria and demonstrates new mechanisms of intrinsic antimicrobial resistance in Salmonella biofilms.

INTRODUCTION

Biofilms are one mechanism of bacterial resistance to external stressors. Biofilms are heterogeneous communities of adherent bacteria encased within an extracellular matrix. The biofilm extracellular matrix and three-dimensional structure alter the transport of nutrients, establish a stratified oxygen environment, and create a physical barrier to damaging compounds (1). In addition, there is a heterogeneity of cellular metabolism, gene expression, and growth rates of the bacteria within the biofilm (2, 3). The combination of a resistant extracellular matrix and metabolic heterogeneity of cells within a biofilm make biofilms intrinsically resistant to antimicrobials and disinfectants (4, 5). Therefore, new strategies are needed to prevent biofilm formation in host and nonhost environments.

Salmonella enterica is a leading cause of bacterial foodborne gastroenteritis that is transmitted predominantly through consumption of contaminated food or water. Salmonella can persist in diverse environments by forming biofilms on biotic and abiotic surfaces (6). Salmonella biofilm is composed of an extracellular matrix, including proteins (curli fimbriae, flagella, and secreted proteins), exopolysaccharides (cellulose, colanic acid, O-antigen capsule [nontyphoidal serotypes], Vi capsule [serotypes Typhi, Paratyphi A, and some Dublin]), and extracellular DNA (reviewed in references 7 and 8). The development of mature biofilms on gallstones contributes to chronic colonization with S. enterica serotype Typhi, providing one route for transmission between individuals (6). In addition to their contribution to chronic typhoid colonization, biofilms are an important mechanism for Salmonella transmission during food processing (9). Single or multispecies biofilms can form on stainless steel, polystyrene, and tile surfaces used in food processing where abrasions and fissures make biofilm attachment easier, leading to unsuccessful disinfection (10–12). Therefore, controlling biofilms both within the host and on abiotic surfaces can help to control Salmonella transmission.

Synthetic analogs of 4-oxazolidinone containing natural products have the potential to be used as antibiofilm compounds. The 4-oxazolidinone core is not contained in commercially available antimicrobials and exhibits a wide range of biological activity (13–15). We created a panel of synthetic 4-oxazolidinone analogs based on the natural product lipoxazolidinone A and found that many of the analogs are potent inhibitors of drug-resistant Gram-positive organisms (16, 17). While the mechanism of action remains unclear, one analog acts in a dose-dependent manner to inhibit cell wall and protein synthesis (16). In spite of their good antimicrobial efficacy against methicillin-resistant strains of Staphylococcus aureus, these compounds fail to inhibit growth of Gram-negative organisms (16, 17). In addition to their antimicrobial effects, some analogs have antibiofilm properties at sub-MICs (18) or can be utilized as adjuvants with existing antimicrobials (19). Understanding how the 4-oxazolidinone analogs may alter biofilm formation will allow us to create analogs with improved efficacy.

The purpose of our study was to determine the mechanism by which the 4-oxazolidinone analog, JJM-ox-3-70 (Fig. 1A), inhibits biofilm in S. enterica serovar Typhimurium (S. Typhimurium) and to establish S. Typhimurium mechanisms of resistance against the antibiofilm effects of the compound. We hypothesized that JJM-ox-3-70 would reduce biofilm by inhibiting extracellular matrix production and that S. Typhimurium’s resistance to its effects would be multifactorial. Our data provide insight into both the mechanism of antibiofilm activity and mechanisms of intrinsic resistance to the synthetic 4-oxazolidinone analog JJM-ox-3-70.

FIG 1.

The small molecule JJM-ox-3-70 inhibits S. Typhimurium biofilm without altering planktonic growth. (A) Chemical structure of the synthetic 4-oxazolidinone, JJM-ox-3-70 (analog 28 from reference 16). (B) Representative image of crystal violet-stained biofilm in indicated growth conditions. (C) Biofilm production in static cultures grown in LB broth with JJM-ox-3-70 or vehicle at room temperature for 96 h as measured by CV-staining. *, significant difference from LB control; **, significant difference between JJM-ox-3-70 and DMSO vehicle control; #, significant difference between JJM-ox-3-70 concentrations as determined by Student's t test with P < 0.05. Bars represent mean ± standard deviation from 3 independent experiments. (D and E) Growth curves in LB broth containing the indicated concentrations of JJM-ox-3-70 (D) or vehicle control (DMSO) (E) in triplicate wells and incubated standing at 25°C. Cell density was measured using optical density at 600 nm. Significant difference between 0 μg/ml and 64 μg/ml from 16 to 24 h and 128 μg/ml from 22 to 24 h by Student's t test with P < 0.05. Data points represent mean ± standard deviation from average of triplicate wells from three independent experiments.

RESULTS

Establishment of JJM-ox-3-70 efficacy.

The synthetic 4-oxazolidinones are potent growth inhibitors for Gram-positive organisms, but the Gram-negative organism Acinetobacter baumannii is resistant to many of the analogs (16). It is likely that both lipopolysaccharide and TolC-mediated drug efflux contribute to Gram-negative resistance to this class of compounds (16, 17). In an effort to understand the effect of the synthetic 4-oxazolidinones on planktonic growth and biofilm production, we expanded efficacy studies to the model Gram-negative pathogen Salmonella Typhimurium. We hypothesized that S. Typhimurium would be resistant to the compound in a TolC-dependent fashion. We found that the wild-type (WT) organism is resistant to antibacterial effects of the compound (MIC > 128 μg/ml), but the growth of the ΔtolC mutant was inhibited at a concentration of 2 μg/ml.

Next, we assayed the effect of various concentrations of JJM-3-ox-70 on biofilm production by crystal violet staining. The vehicle (dimethyl sulfoxide [DMSO]) significantly stimulated biofilm formation at all concentrations compared with the untreated bacteria (Fig. 1B and C). In addition, we found that there was a dose-dependent inhibition of biofilm up to a concentration of 64 μg/ml compared with the equivalent volume of the vehicle control (Fig. 1C) with no effect on culture cell density (data not shown). To establish whether the change in measured biofilm was due to an effect of the compound on the kinetics of planktonic growth, we generated growth curves with various concentrations of the compound or equivalent volume of vehicle. We found a slight reduction in the stationary-phase cell density at compound concentrations ≥ 64 μg/ml (Fig. 1D) but no effect of the vehicle on cellular growth (Fig. 1E).

Biofilm and motility gene expression.

Based on our observations that JJM-ox-3-70 inhibits biofilm formation, we hypothesized that the compound would prevent the transition from planktonic to sessile growth by increasing expression of flagella and reducing expression of curli fimbriae, the main component of the S. Typhimurium extracellular matrix. Using promoter fusions to green fluorescent protein (GFP), we investigated the influence of JJM-ox-3-70 on the activities of promoters for curli (P_csgBA) and the flagellar filament (P_fliC). We observed that JJM-ox-3-70 significantly reduced the activities of both P_csgBA and P_fliC compared to the DMSO control (Fig. 2A and B). Consistent with our findings that DMSO increases biofilm formation, we found that DMSO increased the activity of the csgBA promoter and reduced the activity of the fliC promoter compared to cells in media alone (Fig. 2A and B). To rule out the possibility that JJM-ox-3-70 alters the fluorescence of GFP, we used a reporter plasmid bearing GFP under the control of a tetracycline-inducible promoter (20). There was no significant effect of JJM-ox-3-70 on induced GFP fluorescence, ruling out a direct effect of the compound on GFP fluorescence (Fig. S1 in the supplemental material). Furthermore, JJM-ox-3-70 significantly inhibited swimming motility on semisolid agar (Fig. 3A and B). Together, these data show that JJM-ox-3-70 reduces both curli expression and flagellum-mediated motility.

FIG 2.

JJM-ox-3-70 inhibits activities of csgBA and fliC promoters. The activity of P_csgBA (A) and P_fliC (B) in LB (circle), vehicle control (DMSO) (square), and JJM-ox-3-70 (triangle) as measured by GFP fluorescence (RFU). Measurements were performed on static cultures grown at 25°C with fluorescence (excitation, 485 nm; emission, 528 nm) measured every 12 h. Fluorescence was normalized to a plasmid-less strain (panel A) or a strain containing a plasmid with a promoter-less GFP (panel B). *, significant difference from LB; #, significant difference between JJM-ox-3-70 and DMSO vehicle control as determined by Student's t test (P < 0.05). Experiments were performed in duplicate on three independent occasions. Data points represent mean ± standard error of the mean from three independent experiments.

FIG 3.

JJM-ox-3-70 inhibits swimming motility. (A) WT and ΔmotA mutants were grown overnight and spotted onto 0.3% agar plates with no additive, vehicle (DMSO), or JJM-ox-3-70. The diameter of cell spread from the initial spot was measured after incubation at 37°C for 6 h. Bars indicate mean ± standard error of the mean from four technical repeats performed on six independent occasions. (B) Representative images of swimming plates with the WT (left side of the plate) and ΔmotA mutant (right side of the plate).

Two-step screen for mutants with altered biofilm in response to JJM-ox-3-70.

Next, we took a genetic approach to identify potential targets of the antibiofilm compound JJM-ox-3-70. We performed a hierarchical, two-step screen using two defined kanamycin-marked mutant S. Typhimurium libraries in high throughput (21). In step 1, we tested 151 multigene deletion (MGD) mutants for altered biofilm in the presence of JJM-ox-3-70, vehicle control, or LB alone. Of the 151 MGD mutants tested, 39 mutants produced as much biofilm as the WT (Table S1). Twenty-five mutants had altered cell density in media alone and were excluded from further analysis (Table S2). Of the remaining 87 mutants, there were 59 mutants that produced altered biofilm in media, vehicle, or both and were excluded from further analysis (Fig. S2; Table S2).

Ten MGD mutants produced altered biofilm in all three tested conditions (Fig. S2; Table S3). Within these 10 genomic regions, we identified genes required for biofilm formation, including curli fimbriae, long polar fimbriae, colanic acid biosynthesis, and O-antigen capsule as well as genes expressed in Salmonella biofilms (22–29). Identification of numerous MGD mutants that contained genes necessary for biofilm formation validates our screening method to identify biofilm defects. Since five mutants produced a significantly altered biofilm in the compound compared with the WT when using a stringent statistical cutoff of P < 0.05 compared with the WT, we chose to lower the stringency of our statistical cutoff (to P < 0.1) to increase the number of MGD regions prioritized for further study. One mutant had altered cell density in DMSO and was excluded from further study (Table S2). Therefore, a total of 9 MGD mutants, deleted for 140 genes, were prioritized for further study.

In the second step of our genetic screen, we assayed the effect of JJM-ox-3-70 on the biofilm of 98 single-gene deletion (SGD) mutants corresponding to 9 genomic regions under selection from step 1. We normalized the crystal violet staining intensity in the presence of compound to that of the vehicle control for data analysis to control for the effects of vehicle. We identified 13 mutants with altered biofilm in the presence of the compound compared to the WT (Table S4). The 13 mutants mapped to 7 of the 9 genomic regions selected from step 1. Five mutants produced more biofilm in the compound than the WT, while 8 produced less. The 13 gene products identified through our two-step mutant screening strategy represent potential drug targets as well as resistance mechanisms.

Confirmation of altered biofilm in candidate mutants.

To confirm the effect of JJM-3-ox-70 on biofilm formation in our mutants, each mutation was moved into a clean genetic background and biofilm formation assayed in larger volume to eliminate a potential confounding effect of the high-throughput method on compound efficacy. We found that the compound altered biofilm for 3 of the 13 mutants tested (ΔSTM0437, ΔacrB [ΔSTM0475], and ΔSTM1292) (Fig. 4). Biofilm was significantly reduced by the compound compared to the WT for all mutants. To establish whether the observed alterations in biofilm were due to altered growth kinetics, we generated growth curves for each of our mutants in the presence of compound. We found a small reduction in cell density at stationary phase for the ΔSTM0437 mutant that was not dependent on the compound (Fig. S3) but no alteration in the cell density at 96 h (data not shown).

FIG 4.

Three single-gene deletion mutants have reduced biofilm in the presence of JJM-ox-3-70. Biofilms were grown and quantified as described in the legend to Fig. 1. Bars represent mean ± standard deviation from three independent experiments. *, significant difference in the ratio (JJM-ox-3-70:DMSO) of the mutant compared to the WT by Student's t test (P < 0.05).

To definitively link our observed phenotype with the inactivated gene product in each mutant, we tested each mutant complemented in trans for sensitivity to the antibiofilm effects of JJM-ox-3-70. We cloned each of the 3 genes with the native promoter from mutants producing altered biofilm in the presence of JJM-ox-3-70 onto a stable, low-copy-number vector. The WT bearing the plasmid vector was included as a control to ensure the ampicillin marker did not induce off-target interactions with JJM-ox-3-70. Plasmid carriage did not alter the sensitivity to JJM-ox-3-70 in media or media supplemented with carbenicillin to select for plasmid (Fig. S4). Therefore, we evaluated the effect of the compound on our complemented mutants in media containing carbenicillin. Complementation in trans reversed the phenotype for all 3 of our mutants (Fig. 5), linking STM0437, acrB, and STM1292 to S. Typhimurium resistance to the antibiofilm effects of JJM-ox-3-70.

FIG 5.

Complementation in trans definitively links the sensitivity to JJM-ox-3-70 to the deleted genes. The WT and mutants bearing the empty vector (V) and mutants bearing the appropriate complementing plasmid (C) were grown in LB with carbenicillin in the presence of DMSO or JJM-ox-3-70. Bars indicate normalized ratios of JJM-ox-3-70/DMSO as in Fig. 4. *, significant difference compared to the wild type; **, significant difference between complemented and mutant strains by Student's t test with P < 0.05. Bars represent mean ± standard deviation from three independent experiments.

Periplasmic proteins STM0437 and AcrB.

AcrB is the inner membrane component of the trimeric AcrAB-TolC multidrug efflux system that hydrolyzes ATP to drive efflux. TolC is the outer membrane channel that can also complex with numerous other drug efflux pumps (30). We hypothesized that the ΔacrB mutant would be more sensitive to JJM-ox-3-70 than the WT organism. We found that the MIC of the ΔacrB mutant was identical to that of the WT (>128 μg/ml), and this is substantially greater than that observed for the ΔtolC mutant (2 μg/ml).

STM0437 is annotated as a putative periplasmic protein (31). Since the AcrAB-TolC efflux pump can export compounds from the periplasmic space, we hypothesized that STM0437 and AcrB could operate in the same pathway to facilitate resistance to JJM-ox-3-70. To test this hypothesis, we evaluated the effects of JJM-ox-3-70 on a double mutant lacking both STM0437 and acrB. We found no difference in compound sensitivity for the ΔSTM0437 ΔacrB double mutant compared with the single mutants (Fig. 6), demonstrating no additive effects of the two mutations on resistance to the antibiofilm effects of JJM-ox-3-70.

FIG 6.

Deletion of both STM0437 and acrB does not alter sensitivity to JJM-ox-3-70. Biofilms from the WT (JE447) and ΔSTM0437 (JE1806), ΔacrB (JE1347), and ΔSTM0437ΔacrB (JE1830) mutants prepared as in Fig. 4. Bars are mean ± standard deviation from three independent experiments. *, significant difference from the WT by Student's t test with P < 0.05.

DISCUSSION

In this work, we evaluated a synthetic 4-oxazolidione small-molecule, JJM-ox-3-70, as a lead in the prevention of S. Typhimurium biofilm. Our data demonstrate that JJM-ox-3-70 inhibits expression of both curli fimbriae and flagella, suggesting the antibiofilm effects of the compound are not likely to be through direct inhibition of the biofilm paradigm itself. We found that JJM-ox-3-70 inhibits biofilm formation of S. Typhimurium without relevant effect on planktonic growth. The TolC outer membrane protein is necessary for S. Typhimurium resistance to the antimicrobial effects of JJM-ox-3-70, suggesting that drug efflux is a major mechanism of intrinsic resistance to the compound. We found that 3 genes, acrB, STM0437, and STM1292, mediate intrinsic resistance to the antibiofilm effects of JJM-ox-3-70. STM0437 and AcrB may contribute to the same genetic pathway of resistance. Together, these genes represent mechanisms of intrinsic resistance to the antibiofilm effects of the synthetic 4-oxazolidinone compound JJM-ox-3-70.

Synthetic 4-oxazolidinone compounds have great potential as antimicrobial compounds. These compounds are based on the synoxazolidinone and lipoxazolidinone natural products, which were isolated from marine organisms (13–15). A number of analogs have been developed that have potent antimicrobial activity against methicillin-resistant S. aureus strains and other Gram-positive bacteria. However, they fail to inhibit growth of Gram-negative organisms (16). Gram-negative bacteria with lipopolysaccharide deficiency or with tolC deletion are as sensitive as Gram-positive bacteria, suggesting that both lipopolysaccharide and drug efflux mediate resistance to these compounds (17). Our data are consistent with a role for TolC-mediated drug efflux in S. Typhimurium resistance to the compound. In addition to potent growth inhibition, some analogs have antibiofilm effects on S. aureus at sub-MICs as we found for S. Typhimurium (18, 19). The high concentration of JJM-ox-3-70 needed to reduce S. Typhimurium biofilm makes it an unlikely candidate for development as treatment for Gram-negative biofilm-mediated infections. However, a compound such as JJM-ox-3-70 could be applied topically to prevent Salmonella biofilm development on surfaces, such as within a food processing plant. Further work is needed to demonstrate the efficacy and safety of JJM-ox-3-70 or its derivatives in food production. Studies designed to determine the mechanism of action of these compounds are needed in order to improve upon the efficacy of this class of compounds.

We chose to use one synthetic 4-oxazolidinone analog, JJM-ox-3-70, to interrogate the antibiofilm mechanism using S. Typhimurium as a model organism. Like other Gram-negative organisms, JJM-ox-3-70 has minimal effect on the growth of S. Typhimurium at the concentrations tested. However, we found that JJM-ox-3-70 causes a significant reduction in in vitro biofilm formation as assayed by crystal violet staining. JJM-ox-3-70 is poorly soluble in aqueous solvents, and therefore, the organic solvent DMSO was used to dissolve JJM-ox-3-70. At concentrations much greater than those used in our study (4%), DMSO enhances biofilm formation by increasing curli protein and fiber assembly in Escherichia coli (32). We also found a more robust biofilm when S. Typhimurium was grown in the presence of DMSO (0.24% vol/vol), and this is due to an increase in the activity of the csgBA promoter (Fig. 2A). Since DMSO enhanced biofilm formation in S. Typhimurium, we directly compared the effect of our compound to DMSO to establish compound efficacy. We maximized the compound concentration to minimize the volume of DMSO added to cultures (data not shown). In spite of the confounding effects of the DMSO solvent on biofilm formation, JJM-ox-3-70 caused a dose-dependent reduction in biofilm formation, suggesting that the synthetic 4-oxazolidinone can inhibit biofilm even in the presence of an inducing compound. We chose a concentration of JJM-ox-3-70 that caused maximal biofilm reduction to interrogate the molecular mechanisms by which JJM-ox-3-70 inhibits the Salmonella biofilm.

JJM-ox-3-70 inhibited the promoter activity of both the major flagellar filament (P_fliC) and curli fimbriae (P_csgBA) and reduced swimming motility on semisolid agar. These data suggest that the effect of JJM-ox-3-70 on biofilm is not due to direct disruption of the biofilm regulatory pathway. During the switch from planktonic to sessile state, there is a temporary reduction in flagellar expression and activation of cyclic-di-GMP, which activates the master regulator of biofilm formation in S. Typhimurium, CsgD (33). In turn, CsgD activates expression of curli fimbriae and the diguanylate cyclase adrA, leading to an increase in cellular cyclic-di-GMP (34). Cyclic-di-GMP induces expression of genes needed for additional extracellular matrix components, including cellulose, facilitating organization of the biofilm community. High levels of cyclic-di-GMP reduce motility by facilitating the interaction of YcgR with FliG to alter rotation of the flagellar motor and through a direct inhibitory effect of cellulose on flagellar motility (35–37). CsgD also reduces motility by competing with the class 1 flagellar regulator, FlhDC, for binding to the fliE-fliFGHIJK spacer region (34). Therefore, if our compound had a direct inhibitory effect on csgD expression, we would have expected the compound to either have no effects or stimulate an increase in motility.

An alternate hypothesis for the antibiofilm effects of JJM-ox-3-70 is that the inhibition of motility disrupts initial attachment and maturation of the biofilm. Flagella contribute to the protein component of the Enterobacteriaceae extracellular matrix, and flagellar genes are expressed at various times during biofilm development in liquid and on surfaces (38–40). Furthermore, flagella may facilitate mechanosensing upon contact with solid surfaces, thereby initiating the regulatory cascade that mediates attachment (reviewed in reference 41). We identified 4 mutants deleted for flagellar machinery (lacking genes STM1910 through STM1941 [ΔSTM1910_STM1941], ΔSTM1948_STM1968, ΔSTM1976_STM1982, and ΔSTM1171_STM1183) as defective for biofilm in LB and/or DMSO. Our data are consistent with published work suggesting a role for flagella in biofilm formation on plastic surfaces (28). Therefore, the inhibitory effects of JJM-ox-3-70 on biofilm could be due to disruption of flagellar motility. Further work is needed to determine the mechanism for these findings.

We used two libraries of S. Typhimurium mutants to first screen multigene deletion (MGD) mutants, deleted for multiple contiguous genes followed by subsequent selection of corresponding single-gene deletion (SGD) mutants to identify mutants with altered sensitivity to our compound (21). These libraries have elucidated S. Typhimurium mutants under selection during chick infection and persistence within the cytosol of invaded epithelial cells (42, 43). Our S. Typhimurium mutant library screen evaluated a pool of 151 MGD mutants, covering nearly 2,000 nonessential genes that accounted for nearly half of the total genes present in the genome. We identified no phenotype for 39 of 151 mutants, suggesting it is unlikely that the kanamycin resistance cassette used to create our mutations played a role in our observed phenotypes. A subsequent minilibrary of 98 single-gene deletion mutants corresponding to the MGD regions with altered biofilm only in response to JJM-ox-3-70 was used to identify single genes responsible for altered biofilm in the presence of the compound. This is the first study to our knowledge to use this mutant library within the context of drug discovery. We confirmed 3 single-gene deletion mutants with altered biofilm in the presence of compound. These mutants originated from 3 of the 9 original genomic regions under selection (Table S4). There are several potential reasons for our inability to identify an effect of compound on biofilm for mutants from each of the 9 MGD regions. First, it is possible that a phenotype could be observed only when other genes in a respective region are also absent. Second, we tested only the mutants that were already present in the SGD collection (21); therefore, it is possible that there was a key gene in the region without a corresponding mutant available. Third, it is possible that a secondary site mutation on the chromosome of one or more of the MGD mutants could be responsible for the observed phenotype. Finally, it is possible that there were false-positive results because we chose to reduce the statistical stringency to increase the number of regions from which SGD mutants would originate.

We found that AcrB is necessary for S. Typhimurium resistance to the antibiofilm effects but is insufficient to alter resistance to the growth inhibition effects of JJM-ox-3-70. AcrAB-TolC belongs to the family of resistance nodulation division (RND) efflux pumps and is composed of a trimeric complex containing an inner membrane ATPase (AcrB), a periplasmic adaptor (AcrA), and outer membrane channel (TolC) (44). The AcrAB-TolC efflux pump is responsible for the excretion of a broad scope of substrates creating intrinsic resistance to many antibiotics (30). None of the components of the AcrAB-TolC pump are required for biofilm formation (45). Our data corroborate this observation, showing that the ΔacrB mutant produces adequate biofilm in media (with or without DMSO) but has a significant reduction in biofilm formation in the presence of JJM-ox-3-70. The ΔacrA single-gene mutant was not found in our collection, so it was not tested against JJM-ox-3-70. We found that the MIC for the ΔtolC mutant, but not the ΔacrB mutant, was much lower than that of the WT. The increased sensitivity of the ΔtolC mutant to the compound is likely due to the fact that S. Typhimurium has nine functional drug efflux pumps, seven of which can complex with TolC, likely explaining the resistance of the ΔacrB mutant to the compound (30, 46). The fact that AcrB alone can mediate resistance to the antibiofilm effects of the compound suggests that the functional redundancy of S. Typhimurium efflux mechanisms for growth inhibition does not translate to resistance to biofilm inhibition.

We also found two genes of unknown function that mediate resistance to the compound, STM1292 and STM0437. STM1292 (yeaC) encodes a 90-amino-acid protein of unknown function. Its role in the resistance of S. Typhimurium to the antibiofilm effects of JJM-ox-3-70 is unknown. STM0437 is annotated as a putative periplasmic protein that has a GC content of 43.7%, lower than the genome average of 52.2%, suggesting that it was acquired through horizonal gene transfer (31). STM0437 is predicted to have an N-terminal signal peptide that facilitates secretion through the Sec system and 5 Sel-1-like motifs (47, 48). PredictProtein (49) and Phyre2 (50) analyses showed that STM0437 shares structural similarity with the Sel1 repeat protein EsiB, a protein identified as a potential protective antigen for extraintestinal pathogenic E. coli infections (51, 52). STM0437 is predicted to be cotranscribed with another Sel-1-like repeat protein, STM0438, with which it shares 31% amino acid identity (53). Sel1-like repeat motifs can mediate protein-protein interaction and can be involved in protein signaling pathways in eukaryotes and prokaryotes, in virulence of bacterial pathogens, and in maintenance of cell wall, but the majority of Sel1-like repeat-containing proteins have unknown functions (reviewed in reference 54). The ΔSTM0437 mutant was more sensitive to the compound than the WT organism. We used a genetic approach to determine whether STM0437 and acrB, both predicted to function in the periplasmic space, could operate in the same genetic pathway to promote resistance to JJM-ox-3-70. We saw no change in sensitivity to the compound in the double ΔSTM0437ΔacrB mutant, suggesting the functions of the two proteins could be related. Both STM0437 and STM1292 represent potential novel mechanisms of intrinsic antimicrobial resistance in the context of biofilm formation; further work is needed to characterize the mechanism by which these proteins protect against the antibiofilm effects of JJM-ox-3-70.

Bacterial persistence in biofilms remains a serious threat to human health. In the work presented here, we studied the antibiofilm properties of a synthetic 4-oxazolidinone analog. We show that the compound inhibits S. Typhimurium biofilm, likely due to dysregulation of flagellar motility. S. Typhimurium resists the antimicrobial effects in part by efflux mediated through the promiscuous outer membrane pore TolC. In an unbiased 2-step screening approach, we interrogated the mechanism by which Salmonella resists the antibiofilm effects of JJM-ox-3-70. Herein, we report 3 mutants (ΔSTM0437, ΔacrB, and ΔSTM1292) that are more sensitive to the compound than the isogenic WT. The genes identified through our screen represent potential mechanisms of intrinsic antimicrobial resistance in biofilms. Through understanding the mechanism by which S. Typhimurium is resistant to the antibiofilm effects of the compound, improved antibiofilm strategies can be used to lead to new potent therapeutic weapons to combat bacterial biofilms.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

All Salmonella strains are derivatives of ATCC 14028s. Multigene deletion (MGD) and single-gene deletion (SGD) mutants were previously described (21). Bacterial strains and plasmids used to confirm phenotypes are listed in Table 1. Bacterial cultures were grown on Luria Bertani (LB)-Miller (10 g/liter NaCl) agar or in LB-Miller broth at 37°C with agitation (250 rpm) unless otherwise specified. When necessary, media was supplemented with kanamycin (50 mg/liter), carbenicillin (100 mg/liter), or a combination thereof.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Genotype	Source or reference no.
Strain
JE447	Salmonella enterica serotype Typhimurium 14028s	ATCC
JE1499	14028s + pWSK29; Ampr	This study
JE1367	JE447 ΔphnR::kan; Kanr	This study
JE1369	JE447 ΔSTM0436A::kan; Kanr	This study
JE1371	JE447 ΔSTM0437::kan; Kanr	This study
JE1477	JE1371 + pWSK29; Ampr	This study
JE1479	JE1371 + pSTM0437; Ampr Kanr	This study
JE1345	JE447 ΔrpmJ_1::kan; Kanr	This study
JE1347	JE447 ΔacrB::kan; Kanr	This study
JE1491	JE1347 + pWSK29; Ampr	This study
JE1493	JE1347 + pSTM0475; Ampr Kanr	This study
JE1349	JE447 ΔSTM0481::kan; Kanr	This study
JE1351	JE447 ΔSTM1292::kan; Kanr	This study
JE1495	JE1351 + pWSK29; Ampr	This study
JE1497	JE1351 + pSTM1292; Ampr Kanr	This study
JE1353	JE447 ΔSTM2201::kan; Kanr	This study
JE1355	JE447 ΔSTM3152::kan; Kanr	This study
JE1365	JE447 ΔSTM3153::kan; Kanr	This study
JE1382	JE447 ΔtolC::kan; Kanr	This study
JE1357	JE447 ΔSTM4038::kan; Kanr	This study
JE1359	JE447 ΔSTM4041::kan; Kanr	This study
JE1361	JE447 ΔSTM4521::kan; Kanr	This study
JE1541	JE447 + pMPM-A3ΔPlac null-gfpLVA; Ampr	This study
JE1543	JE447 + pMPM-A3ΔPlac PfliC-gfpLVA; Ampr	This study
JE1545	JE447 + pCT125; Ampr	This study
JE1107	JE447 ΔmotA::kan; Kanr	61
JE1806	JE447 ΔSTM0437::frt	This study
JE1830	JE1806 ΔacrB::kan; Kanr	This study
JE1715	JE447 + pJC45; Kanr	This study
Plasmid
pWSK29	Cloning vector; Ampr	57
pSTM0437	pWSK29::STM0437; Ampr	This study
pSTM0475	pWSK29::STM0475; Ampr	This study
pSTM1292	pWSK29::STM1292; Ampr	This study
pCT125	PcsgBA expression vector; Ampr	60
pMPMA3ΔPlac-PfliC-GFPLVA	PfliC expression vector; Ampr	59
pMPMA3ΔPlac-null-GFPLVA	Promoter-less GFP vector; Ampr	59
pJC45	Tetracycline-inducible GFP expression vector; Kanr	20

The synthetic small-molecule JJM-ox-3-70 was prepared as previously described (16). In short, (S)-lactamide was converted to a protected intermediate that was reacted with a dioxinone derivative under acidic conditions. The resulting product, JJM-ox-3-70, was purified by silica chromatography and fully characterized and was stored lyophilized at −20°C. The compound was diluted in dimethyl sulfoxide (DMSO) to a final concentration of 100 mM and stored at −20°C. The compound was used at a final concentration of 64 μg/ml unless otherwise specified.

To evaluate the effects of JJM-ox-3-70 on growth, overnight cultures were diluted 1:100 in LB supplemented with either JJM-ox-3-70 or an equivalent volume of DMSO and 200 μl of each suspension aliquoted into 96-well plates. Bacteria were incubated standing at 25°C for 18 to 24 h, and the optical density at 600 nm (OD₆₀₀) of each culture was measured every 1 to 2 h as indicated.

Broth microdilution MIC determination.

MIC was determined by broth microdilution according to CLSI guidelines. Briefly, Mueller-Hinton broth (MHB) (BD Difco) was inoculated with bacteria, and cultures were incubated at 37°C with agitation for 6 h or to an optical density at 600 nm equal to 1. Bacteria were diluted to approximately 5 × 10⁵ CFU/ml in 20 ml MHB. JJM-ox-3-70 or ofloxacin was added to the media at 2-fold serial dilutions to a maximum concentration of 128 μg/ml. An equivalent volume of DMSO served as vehicle control. Bacterial suspensions were added to a 96-well plate and incubated at 37°C for 16 h. Following incubation, the MIC values were recorded as the lowest concentration of compound where no visible growth of bacteria was observed.

Biofilm growth and quantification.

To grow biofilms, overnight cultures were diluted 1:100 into LB-Miller broth or LB-Miller broth supplemented with either JJM-ox-3-70 or an equivalent volume of DMSO. Bacteria were incubated standing at 25°C for 96 h. For library screening, biofilms were grown in 150 μl volume in 96-well plates, and for all other assays, biofilms were grown in 1 ml volume in 5-ml test tubes. To measure cell density, planktonic bacteria were resuspended and transferred to a clean container for measurement of optical density (600 nm). Biofilm was quantified by crystal violet staining (55). Briefly, biofilms were washed twice with deionized H₂O and dried on the benchtop prior to staining with 0.1% crystal violet (Amresco). Crystal violet was solubilized with 30% acetic acid and quantified by measuring optical density at 550 nm.

Mutant library screening and confirmation.

We performed a two-step approach to interrogate the S. Typhimurium genome for mutants with altered sensitivity to JJM-ox-3-70. First, we screened mutants in large genomic regions (MGD mutants) containing a kanamycin resistance cassette replacing approximately 4 to 40 contiguous genes, followed by a screen of SGD mutants corresponding to the regions under selection in MGD mutants as previously described (42). For phenotyping of the MGD library, glycerol stocks of S. Typhimurium MGD mutants in 96-well format stamped onto LB agar were transferred into 200 μl of LB broth in 96-well plates and incubated overnight at 37°C. At least three wells contained the WT organism on each plate. Overnight cultures were diluted 1:100 into appropriate media in 96-well plates and incubated at 25°C for 96 h. Biofilms were quantified as described above. For screening the MGD library, the OD₆₀₀ and OD₅₅₀ for each mutant under each condition were compared to that of triplicate wells of the WT using a two-tailed Student's t test with unequal variance. Significance was set at P < 0.05 for the conditions of LB and DMSO and P < 0.1 for JJM-ox-3-70.

A minilibrary of SGD mutants was created from the MGD regions under selection and assembled into a 96-well format with each plate containing at least three wells of the WT. The SGD mutant minilibrary was screened for biofilm formation using the methodology for MGD mutants. To analyze the effect of JJM-ox-3-70 on SGD mutant cell density and biofilm formation, the ratio of OD₆₀₀ or OD₅₅₀ of compound/vehicle control (JJM-ox-3-70/DMSO) was calculated for each strain to determine relative effect of the compound. A two-tailed Student's t test with unequal variance was used to compare the relative compound effect between a mutant and the WT in triplicate with significance set at P < 0.05. To confirm the phenotypes observed in the library screen 96-well format, mutations were moved into a clean genetic background by P22-mediated bacteriophage transduction (56). Transductants were purified twice and mutations confirmed by PCR using primers external to the mutation of interest.

Construction of complementing plasmids.

Genomic DNA was isolated from an overnight culture of the WT organism (GeneElute; Sigma). PCR products were generated from genomic DNA using Q5 polymerase (New England Biolabs) with an annealing temperature of 72°C and extension time appropriate for product length. Primer sequences for each insert are included in Table 2. Restriction sites for endonucleases HindIII (forward) and KpnI (reverse) were integrated into primer sequence design. The size of each PCR product was confirmed by agarose gel electrophoresis. PCR products were purified (QIAquick PCR purification kit; Qiagen) and digested with HindIII and KpnI (New England Biolabs). Gene inserts were cloned into pWSK29 sequentially digested with the HindIII and KpnI and treated with shrimp alkaline phosphatase (New England Biolabs) following primary digestion (57). Plasmid ligations were performed overnight at 16°C with T4 DNA ligase (New England Biolabs) following the manufacturer’s guidelines. Ligated plasmid constructs were transformed into chemically competent DH5α E. coli by heat shock. Transformants were selected on LB agar supplemented with X-gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside, 40 μg/ml) and carbenicillin. Transformants were purified twice to single colonies. Plasmids were isolated by Miniprep (Qiagen), and constructs were verified by Sanger sequencing (Eton Bioscience). Complementing plasmids were transformed into restriction-negative (R⁻), modification-positive (M⁺) S. Typhimurium LB5000 by electroporation (58). Transformants were purified twice to single colonies and complementing plasmids isolated and transformed into appropriate mutants by electroporation. Each mutant was also transformed with the empty plasmid vector. Mutants harboring plasmids were purified twice to single colonies and stored in glycerol stocks at −80°C.

TABLE 2.

Primer sequences used to generate complementing plasmids

Gene insert	Primera	Endonucleases	Product size (bp)
STM0437	Forward: CGCTAAAGCTTGACGGTCCCTAAACCCCA	HindIII	2,205
	Reverse: ATCGCGGTACCCCTGGCATACCGATGCGA	KpnI
acrB	Forward: CGCTAAAGCTTACAGTGCAGCAGCTGGA	HindIII	3,806
	Reverse: ATCGCGGTACCGCTTGCGCGGCCTTATC	KpnI
STM1292	Forward: CGCTAAAGCTTCAGTCACGATTCGTGAGCC	HindIII	925
	Reverse: ATCGCGGTACCCAATGAAACCGCCGGATG	KpnI

^aUnderlining indicates restriction endonuclease recognition sites.

Promotor activity using GFP reporter plasmids.

The activities of the csgBA and fliC promoters were measured using plasmids bearing a transcriptional fusion of the indicated promoter with the coding sequence of green fluorescent protein (GFP) (pMPMA3ΔPlac null-gfpLVA and pMPMA3ΔPlac PfliC-gfpLVA [plasmid nos. 23338 and 23342, respectively] were generous gifts from Olivia Steele-Mortimer Addgene; pCT125 was a generous gift from Çagla Tükel) (59, 60). Nonfluorescent strains containing either the plasmid with promoter-less GFP or the organism without a plasmid were used as negative controls. Overnight cultures were diluted 1:100 into media alone or media supplemented with either JJM-ox-3-70 or an equivalent volume of DMSO. Cultures were incubated at 25°C for 72 h in a black flat-bottom 96-well plate. Promoter activity of each indicated gene was estimated by GFP fluorescence and expressed as relative fluorescence units (RFU) normalized to the RFU of the nonfluorescent control in a given condition. For determination of effect of JJM-ox-3-70 on GFP fluorescence, a plasmid containing GFP under the control of the tetracycline-inducible promoter was used (20). Cells were grown to mid-log in LB broth; the culture was divided and half induced with anhydrous tetracycline (200 nM) and the other half left untreated. Cultures were divided further and treated with JJM-ox-3-70 or vehicle or left untreated. RFU of induced cultures in a given condition were normalized to those of noninduced cultures. Fluorescence (excitation, 485 nm; emission, 528 nm) was determined using a plate reader (Synergy HTX, BioTek).

Swimming motility.

Motility assays were performed as previously described (61). Briefly, overnight cultures were normalized by OD₆₀₀, mixed with 0.1% (vol/vol) Higgins waterproof black India ink, and spotted onto semisolid agar (0.3% Difco Bacto agar, 25 g/liter LB broth base) or semisolid agar containing JJM-ox-3-70 or vehicle. Cultures were incubated at 37°C for 6 h and the distance traveled measured from the edge of the black dye (marker of the starting spotted culture) to the edge of the motility ring. Experiments were performed in 4 replicates on 6 independent occasions.

Statistical analysis.

All experiments were performed on three independent occasions unless otherwise noted. For phenotype confirmation and complementation analyses, the ratio of compound to vehicle control was compared to the WT using a two-tailed Student's t test. For growth curve analysis, the difference between treatment groups was determined using 2-way analysis of variance (ANOVA) with Tukey’s correction for multiple comparisons. For fluorescence gene expression experiments, the fold change in RFU was calculated by dividing the RFU of the GFP-expressing strain by the RFU of the null variant and differences between strains established by two-tailed Student's t test. Unless otherwise specified, P < 0.05 was considered significant. Library screening data analyses were performed using Microsoft Excel, and all other analyses were performed using GraphPad Prism v8.0.