The ryfA gene is involved in resistance to osmotic stress

We previously generated a transposon (Tn) library in UPEC reference strain CFT073 [28]. The library was subjected to osmotic and oxidative stresses as a selection for potentially decreased fitness in an environment in which the bacteria encounter host-derived stresses including osmolarity in urine and interactions with host immune cells. We screened this Tn library for mutants that exhibited reduced tolerance to osmotic and oxidative stress in order to identify genes that may play a role in UPEC stress tolerance. One mutant that exhibited reduced tolerance to osmotic and oxidative stress carried a transposon insertion adjacent to the ryfA gene. To confirm the direct role of ryfA in phenotypes of decreased tolerance to stress, we generated a site-directed deletion to create a ΔryfA mutant of UPEC strain CFT073. To determine whether ryfA contributes to resistance to osmotic stress, the WT parental strain, ΔryfA, and a complemented ΔryfA strain were subjected independently to osmotic stress (0.6 M Urea) and salt stress (0.6 M NaCl), on modified LB agar. The WT and the complemented mutant were able to grow on LB agar containing NaCl or urea (Fig 1A). By contrast, on LB agar with 0.6 M urea or 0.6 M NaCl, the capacity of the ryfA mutant to grow was significantly reduced (Fig 1A). In LB agar, there was no growth difference between CFT073 and CFT073 ΔryfA (S1A Fig). This result suggests that RyfA could play a role in resistance to osmotic stress caused by NaCl and urea.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Role of RyfA in osmotic and oxidative stress resistance.

(A) Growth under conditions of osmotic stress. Strains were grown shaking in LB medium until mid-log phase (O.D.₆₀₀, 0.6) and plated on LB agar and LB agar with 0.6 M NaCl and 0.6 M urea (see Methods). (B) Growth inhibition zones (mm) of UPEC CFT073, ryfA mutant and complemented strains to oxidative stress generating compounds. Strains were grown on either LB agar or (C) M9- glucose agar. ROI-generating compounds tested were 30% hydrogen peroxide (H₂O₂), 53 mM plumbagin, or 40 mM paraquat. (D) Resistance to hydrogen peroxide in LB broth. WT CFT073 and ryfA mutant strains were examined in LB medium containing 5mM H₂O₂ or an equivalent volume of Milli-Q water (H₂O). Samples were collected every 5 minutes, diluted, and plated on LB agar to determine CFUs. The CFT073 ΔoxyR strain was used as a sensitive control strain. The results represent the means of replicate experiments for a minimum of three samples. Vertical bars represent the standard errors of the means. Statistical significance was calculated by one-way ANOVA (A to D): (*, P < 0.05; **, P < 0.005; ***, P < 0.0001).

https://doi.org/10.1371/journal.ppat.1009617.g001

Loss of ryfA increases sensitivity to ROI-generating compounds

To determine if ryfA influences UPEC resistance to oxidative stress, bacterial survival after exposure to oxidative stress was tested using different reactive oxygen intermediate (ROI)-generating agents. We used H₂O₂ and various superoxide generators such as plumbagin and paraquat, during growth on either rich (LB) or minimal (M9-glucose) medium. Sensitivity to these compounds was determined with wild-type UPEC strains CFT073 and 536 and corresponding ΔryfA mutants. On LB medium, the CFT073 ΔryfA strain was more sensitive to the ROI-generating compounds H₂O₂ (p = 0.0022) and plumbagin (p<0.0001) when compared to parental strain CFT073 (Fig 1B). The UPEC 536 ΔryfA mutant was only more sensitive to H₂O₂ (p = 0.0065) when compared to strain 536 (S1B Fig). On minimal medium, both UPEC ryfA mutants were more sensitive to H₂O₂, plumbagin, and paraquat (Figs 1C and S1B). In addition, in LB broth the CFT073 ΔryfA mutant was killed more readily compared to wild-type CFT073 (p = 0.0057) after exposure to 5 mM hydrogen peroxide (Fig 1D). Complementation of the ΔryfA mutant with a single-copy of ryfA on the chromosome restored resistance to ROI-generating compounds (Figs 1B, 1C and 1D). These results show that loss of the ryfA sRNA reduces bacterial resistance to oxidative stress, and that RyfA may contribute to regulation of genes implicated in the oxidative stress response.

RyfA is produced in logarithmic phase

ryfA was originally identified by in silico-based genomic analyses designed to identify genes encoding previously uncharacterized sRNA molecules [23]. In silico analyses revealed that ryfA is chromosomally located between sseA and sseB (S2A Fig). Using BLAST searches, we identified ryfA homologues in other species of Escherichia, Salmonella, Shigella, Klebsiella, Enterobacter, Serratia and Citrobacter (S2A Fig). ryfA from CFT073 shares 98% identity at the nucleic acid level with the non-pathogenic K-12 strain MG1655 (S2B Fig), and has greater than 90% sequence identity to ryfA sequences from other intestinal and extraintestinal pathogenic E. coli and Shigella spp. (S2A Fig).

To determine under what conditions CFT073 produces RyfA, we performed Northern blotting under laboratory conditions. Bacteria were grown in LB medium at 37°C until the culture O.D. of 0.2 to 2.9. Northern blot analysis showed that RyfA was expressed starting at mid-log phase and reached maximal production in late exponential phase (~ 0.9 to 2.3) (Fig 2A). By using quantitative real-time PCR (qRT-PCR) analysis we confirmed that RyfA is produced by wild-type CFT073 under the same conditions (Fig 2B). We did not detect any signal in the ΔryfA strain, confirming the absence of RyfA in the mutant strain (Fig 2A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. CFT073 produces RyfA in mid-logarithmic phase.

(A) Northern blot analysis using 10 mg of total RNA isolated from WT CFT073 following growth in LB medium at 37°C. Lack of signal in ΔryfA strain confirms absence of RyfA in the ΔryfA strain. (B) Quantitative real-time PCR (qRT-PCR) analysis demonstrating that RyfA is produced by wild-type CFT073 under the conditions tested. (C) Northern blot showing RyfA expression in a strain overexpressing RyfA.

https://doi.org/10.1371/journal.ppat.1009617.g002

ryfA is predicted to be 305 nt in length (S2B Fig). The expression of ryfA was placed under the control of an arabinose inducible promoter and the size of the ryfA transcript was confirmed by radiolabeled ryfA-specific antisense RNA probe. A single band of approximately 300 nt was detected by northern blot in RNA samples obtained from strain CFT073 and using the inducible promoter construct (Fig 2C). The secondary structure of RyfA was predicted using the Vienna RNA websuite program [29], and its predicted secondary structure is comprised of multi-stem loops (S2C Fig).

Transcriptomic analysis of the effect of loss of ryfA on UPEC gene expression

Our in vitro experiments revealed an important role for RyfA in response to stresses. We reasoned that the greater sensitivity of the ryfA mutant could be due to numerous changes in gene expression that occur in the absence of this regulatory RNA. To examine the role of RyfA, we performed RNA-Seq to compare RNA profiles between wild-type E. coli CFT073 and the CFT073 ΔryfA mutant. Bacteria were cultured in triplicate at 37°C in LB with aeration to mid-logarithmic growth (O.D. 0.6), mRNAs were extracted and RNA-Seq was performed. In total, the deletion of ryfA significantly altered the expression of 484 genes (120 were upregulated and 364 were downregulated) with a log2 fold-change (FC) greater than or equal to ± 1.7 (S1 Table). The genes that were most highly affected in response to loss of ryfA are listed in Fig 3. Functional analysis revealed that the genes belonging to the categories “metabolism” and “stress response” were by far the most markedly affected. Indeed, increased sensitivity to stresses is likely a consequence of altered gene expression and defective regulation of the stress response.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Genes significantly affected by deletion of ryfA.

RNA-seq analysis was performed on RNA samples (in triplicate) from strain CFT073 and CFT073 ΔryfA grown in LB at mid-log phase of growth (O.D. 0.6). (A) Heatmap of the entire data set (n  =  6) of CFT073 and ΔryfA mutant. Each row of the heatmap represents the log2 fold values transformed with Z-score of 121 differentially expressed genes (upregulation, red, downregulation, yellow). Hierarchical grouping of differentially expressed genes shows clustering. Z-scores are computed on a gene-by-gene basis by subtracting the mean and then dividing by the standard deviation. Specific genes and changes are presented in Fig 3B. (B) Genes classified in different categories based on log2 fold-change difference. Genes upregulated (yellow rectangles) and downregulated (blue rectangles) by at least 1.7-fold were considered significant with P < 0.05. Dark blue is greater or equal to (≥) a 3-fold decrease. Light blue is a decrease from 3-fold to 1.7-fold. Light yellow is an increase from 1.7-fold to 3-fold. Dark yellow is a greater or equal to (≥) 3-fold increase. Genes are classified in different categories with colored squares: General stress responses in red, metabolism and sugar transport in black, metabolism and iron transport in grey, patho-specific in pink, flagella and motility in purple, fatty acids and respiratory chain in green, amino acid metabolism in white, curli and toxin-antitoxin system in orange, RNA in yellow and unknown function in azure. Specific values fold-changes are presented in S1 Table (see supplementary data).

https://doi.org/10.1371/journal.ppat.1009617.g003

Validation of differentially expressed genes

To validate some of the data obtained from the RNA-Seq results, we performed qRT-PCR using primers specific to a series of genes that were differentially expressed: soxS, ibpA, cspA, marA, bssS, rpoH, cadA malE and treC to compare log2 fold change in gene expression between CFT073 ΔryfA and the CFT073 wild-type strains. Strains were cultured under conditions identical to our RNA-Seq experiment. We observed highly similar differences in gene expression as compared to our RNA-Seq results. Specifically, we observed downregulation of soxS, ibpA, cspA, marA, bssS, rpoH and upregulation of cadA, and treC (S3A Fig).

Inactivation of RyfA reduces production of type 1 fimbriae in vitro

Since the ryfA mutant is more sensitive to oxidative and osmotic stresses and given that type 1 fimbriae play an important role for colonization in the murine UTI model [5], the effect of the ryfA mutation on production of type 1 fimbriae was investigated. Production of type 1 fimbriae was determined by yeast agglutination at the mid-log phase of growth in LB medium (O.D. 0.6) and human urine (O.D. 0.4) and in static conditions. Following mid-log growth with shaking in LB broth, and overnight growth in human urine with shaking, the agglutination titer of CFT073 ΔryfA was significantly reduced (p<0.0001 and p = 0.0062, respectively) as compared to parent strain CFT073 (Fig 4A). When cultured statically, agglutination titers were significantly reduced in the CFT073 ΔryfA mutant compared to the WT strain in LB, human urine (Fig 4A) or minimal M9-glucose broth (S4A Fig). The yeast agglutination titer was also significantly reduced in both UPEC ΔryfA strains compared to wild-type UTI89 and 536 parent strains after mid-log phase in LB and human urine (S4B Fig). Yeast agglutination by the ryfA complemented mutant regained titers similar to that of the WT strain (Figs 4A and S4A and S4B). Western blotting against the type 1 fimbrial major subunit FimA confirmed a sharp reduction of FimA production by the ΔryfA mutant compared to WT strain CFT073 and the ryfA complemented strain (Fig 4B). Surprisingly, although production of type 1 fimbriae was reduced in the ryfA mutant, we did not observe significant differences in the RNA-seq data for fim genes under growth conditions tested. This could be explained by the fact that at O.D. 0.6, only 4% of the cell population expresses type 1 fimbriae, as previously reported [35]. However, by comparing expression of fimA in LB and human urine at mid-log phase and after static growth, by RT-PCR, we observed a significant decrease in fimA RNA levels in the ryfA mutant (Fig 4C). These results indicate that deletion of the ryfA sRNA caused a substantial decrease in the levels of type 1 fimbriae production and fimA gene expression.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Effect of deletion of ryfA on type 1 fimbriae (pili) production and motility.

(A) Yeast agglutination titer demonstrating level of production of type 1 fimbriae in strains cultured to the mid-log phase and after overnight static growth in LB broth and in human urine. The Δfim strain was used as a negative control and showed no agglutination. (B) Western blot of fimbrial extracts of strains cultured to the mid-log phase of growth in LB broth. The CFT073 fim-locked ON strain was used as a positive control. (C) Expression of fimA gene by qRT-PCR in the ΔryfA strain compared to the WT CFT073 strain in LB, human urine and infected bladders. (D) Motility of CFT073, ryfA mutant and complemented strain on 0.25% soft or urine agar. Each box and scatter dot plot (min to max) represents the mean diameter of the motility zone. Results are the mean values and standard deviations for at least three biological experiments. Statistical significance was calculated by the Student t test (A, C, D): *, P < 0.05; **, P < 0.005; ***, P < 0.0001. NS, not significant.

https://doi.org/10.1371/journal.ppat.1009617.g004

Loss of RyfA increases swimming motility

The effect of ryfA deletion on motility was determined, since motility contributes to colonization and persistence in the urinary tract [36]. We compared the motility of the ryfA mutant in 0.25% soft and urine agar. We reasoned that since the deletion of ryfA decreased expression of type 1 fimbriae (Fig 4A), motility of the ryfA mutants may potentially be increased. There was a clear and significant increase in motility of the ΔryfA strain compared to wild-type CFT073 parent strain in soft agar and urine agar plates (Fig 4D). Interestingly, the upregulation of flagellar and motility genes also corresponds to the increased swimming motility that was observed on motility agar (Fig 3).

Deletion of ryfA decreased the expression of P fimbriae and increased expression of F1C fimbriae

It has been observed that deletion of the fim operon results in increased expression of other types of fimbriae [13,35]. Since UPEC strains adhere to bladder epithelial cells and epithelial cell adherence is mediated predominantly by type 1 fimbriae, we therefore hypothesized that the ryfA mutant may demonstrate reduced adherence to bladder epithelial cells. To test this possibility, adhesion assays using 5637 human bladder epithelial cells (ATCC HTB-9) were performed. The ryfA mutant adhered to human bladder cells as well as the WT strain (Fig 5A). As expected, addition of 1.5% α-d-mannopyranose, which inhibits type 1 fimbriae mediated binding, greatly decreased adherence of the WT (mean decrease of 17.2%) and complemented strain (mean decrease of 15.9%) (Fig 5A). By contrast, addition of mannopyranose only slightly reduced adherence of the ΔryfA (mean decrease of 2.8%) and Δfim (mean decrease of 5.4%) mutants, to 5637 bladder cells (Fig 5A). In line with these findings, fimA expression was also down-regulated by 4.8-fold in the ryfA mutant, during adherence to 5637 bladder cells (S4C Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Role of ryfA in the murine model of ascending UTI and in P fimbriae (pili) production.

(A) Adherence of strain CFT073 and its derivatives to human 5637 bladder epithelial cells in the presence or absence of 2.5% α-d-mannopyranose was determined. (B) Electron microscopy of CFT073 and the ΔryfA mutant. Arrows show fimbriae on cell surfaces. (C) Production of Pap fimbriae by hemagglutination assay. Results are the mean values and standard deviations for four biological experiments. (D) Single-strain infections to compare wild-type strain CFT073 to ΔryfA mutant. Results are presented as the log₁₀ CFU g⁻¹. Each data point represents a sample from an individual mouse, and horizontal bars indicate the medians. Two independent series of infections were performed: i (CFT073 WT and CFT073 ΔryfA) and ii (CFT073 WT and the complemented mutant). The ΔryfA mutant was attenuated 146-fold in bladder and 10,000-fold in kidneys compared to the WT parent strain. Data are means ± standard errors of the means of 10 mice. Statistical significance was calculated by one-way ANOVA (A and C) or Mann–Whitney Test (D): *, P < 0.05; **, P < 0.005; ***, P < 0.0001. NS, not significant.

https://doi.org/10.1371/journal.ppat.1009617.g005

These results suggest that other adhesins such as the foc-encoded F1C fimbriae, which were upregulated in the ryfA mutant (Fig 3B), could potentially mediate adherence to the bladder epithelial cells. Further, electron microscopy images demonstrate that the type of fimbriae at the surface of cells of the ryfA mutant differ from those produced by the CFT073 wild-type strain and that fimbriae are also somewhat reduced in numbers compared to the WT parent strain (Fig 5B). As such, adherence of the ryfA mutant to bladder cells is largely independent of type 1 fimbriae and is mediated by other adhesins present at the cell surface. We also investigated production of Pap fimbriae by mannose-resistant hemagglutination (MRHA) of human erythrocytes of strains cultured in LB broth. Interestingly, the MRHA titer of the ryfA mutant was significantly reduced compared to parent strain CFT073 (Fig 5C). This supports the hypothesis that P fimbriae production is also reduced in the ryfA mutant. In line with these results, expression of both papA1 and papA2 was reduced by 3.5-fold and 6.3-fold, respectively (S4C Fig). These results demonstrate that the mannose-resistant adherence of the ryfA mutant to bladder epithelial cells is not attributed to increased production of Pap fimbriae, but could be due to increased production of other adhesins such as F1C fimbriae. It was also previously shown that expression of F1C fimbriae increased considerably in a strain lacking the fim and pap gene clusters [37]. Thereby, we investigated by qRT-PCR and Western blotting whether expression of F1C fimbriae was increased in this background. As shown in S4C and S4D Fig, F1C fimbriae were upregulated 7.9-fold in the ryfA mutant. Therefore, upregulation of F1C fimbriae in the ryfA mutant could potentially contribute to the in vitro adherence of the ryfA mutant to bladder cells.

We analyzed the adherence and intracellular bacterial survival of the cystitis isolate UTI89 by using imaging flow cytometry-based assays. The GFP-tagged UTI89 and its derivative strains were allowed to adhere to monolayers of the bladder epithelial cell line 5637 for 2 h (T0), after which extracellular bacteria were killed by addition of the host cell-impermeable antibiotic gentamicin. Following a 4 h (T4) incubation in the continued presence of gentamicin, infected monolayers were washed and intracellular bacterial survival (T4) was quantitatively determined by the percentage of GFP^high 5637 cells. Representative images that correspond to single GFP⁺LAMP1⁺ 5637 cells at T0 and T4 are shown in S5 Fig. Interestingly, the percentages of total cell-associated bacteria for ΔryfA and Δfim mutants were significantly reduced (5.03% and 4.5%, respectively) when compared to the WT parent strain (24.36%) (p<0.0001). At 4 h post-infection, results showed a significantly higher percentage of UTI89 GFP^high labeled WT bacteria within intracellular compartments compared to either the ΔryfA or Δfim mutants (S5 Fig).

Previous studies showed that the Δfim or ΔfimH UTI89 mutant strains were deficient in binding to 5637 cultured bladder epithelial cells in vitro and were unable to colonize and invade bladder epithelium by 6 h p.i. relative to the isogenic WT UTI89 control strain [38,39]. The UTI89 ΔryfA mutant also produced less type 1 fimbriae compared to the WT parent strain in vitro (S4B Fig). These data suggest that deletion of ryfA affects type 1 production in UTI89 and that the adherence to 5637 cells in the absence of the type 1 fimbriae is not sufficient to trigger bacterial internalization into bladder epithelial cells.

Loss of RyfA reduces virulence and expression of type 1 fimbriae in the murine urinary tract

Based upon increased sensitivity of the ΔryfA mutant to environmental stresses, we hypothesized that the impaired osmotic and oxidative stress response could result in a fitness defect during infection of the urinary tract. In the mouse infection model, the ΔryfA strain showed a marked decrease in bacterial numbers in both bladders (146-fold decrease) (P = 0.0021) and the kidneys (10,000-fold decrease) (P<0.0001) compared to the wild-type strain CFT073 (Fig 5D). The ryfA complemented mutant colonized the mouse urinary tract as well as wild-type strain CFT073 (Fig 5D). Likewise, fimA transcription was downregulated 5.8-fold in the bladder (Fig 4C). This reduction in colonization from inactivation of ryfA could be due in part to the decrease in expression of fimA and reduced production of type 1 fimbriae, which was also observed in vitro (Fig 4A and 4C). However, it is likely that additional factors including decreased resistance to cellular stresses due to loss of ryfA contribute to the strong attenuation and decreased bacterial numbers observed during infection with the ΔryfA mutant, particularly in the kidneys.

In addition, we performed a co-infection of ΔryfA mutant with the virulent CFT073 Δlac strain in the CBA/J murine model. The CFT073 Δlac strain has been shown to colonize the urinary tract as well as the CFT073 wild-type parent and presented no statistical difference in a murine UTI model [40]. Interestingly, the ΔryfA mutant was significantly outcompeted by the CFT073 Δlac strain with a mean of 124-fold in the bladders (p<0.0001) and 13-fold (p<0.0001) in the kidneys (S6A Fig). In contrast to the fitness decrease identified for the ΔryfA strain during UTI, there was no growth difference between CFT073 and CFT073 ΔryfA when grown in vitro in either LB broth or in filtered human urine. The latter represents an ex vivo condition that may reflect nutrient availability and environmental conditions encountered in the bladder (S6B and S6C Fig). The growth of the ΔryfA mutant was altered in LB supplemented with 0.6 M urea or 0.6 M NaCl, but not in human urine. Interestingly, concentration ranges of urea (150 mM to 350 mM) and sodium (40 mM to a maximum of 275 mM) in human urine are considerably lower than the concentrations that were inhibitory to the ryfA mutant [41–43]. Thus, the competitive defect of the ryfA mutant during UTI was not due to a general fitness defect, and that during standard in vitro culture conditions, ryfA was dispensable. Overall, these results confirm an important role for the RyfA sRNA during colonization and infection of the urinary tract by strain CFT073.

RyfA contributes to increased uptake and survival of UPEC in professional phagocytes

During infection, inflammatory cells produce a variety of antimicrobial factors. A primary antimicrobial agent in this repertoire is superoxide, which reacts with other molecules to form ROS. The primary source of bactericidal ROS during infection is provided by phagocyte oxidase. We next sought to assess the potential role of ryfA for protection from the host-generated oxidative environment during UTI by testing bacterial interaction with human macrophages. To evaluate bacterial uptake, survival and proliferation within macrophages, primary human monocyte-derived macrophages (HMDMs) and human cultured macrophage-like THP-1 cells were infected with the wild-type strain and the isogenic mutants using a gentamicin protection assay (at a multiplicity of infection (MOI) of 20). The number of bacteria present at different times was determined by viable counts. We also assessed intramacrophage survival of an oxyR mutant for comparison as a sensitive control. The uptake of the ryfA mutant was significantly reduced in HMDMs when compared with uptake of the wild-type strain (Figs 6A and 7). However, for the THP-1 macrophages, similar bacterial loads for the different strains were observed (Fig 6B). In contrast to wild-type strain CFT073, for both the ryfA and oxyR mutants, viable bacteria decreased at 2 h p.i. in all types of macrophages (Figs 6A, 6B and 7). Furthermore, both mutants decreased in the intracellular compartment of macrophages at 24 h p.i. and were unable to reach the number of intracellular bacterial cells observed for the wild-type parent strain. We also obtained similar results by using multi-parameter and imaging flow cytometry (Figs 6D and 7). We monitored the uptake of mCherry-tagged CFT073 wild-type and ryfA mutant strains in CD14⁺ HMDMs, which represent a classic phenotype for HMDMs. Results showed a major drop in survival of the ryfA mutant in HMDMs compared to the wild-type strain at all-time points (2h, 6 h and 10 h p.i.) (Fig 6C). Furthermore, we confirmed this assessment of the same GFP-tagged CFT073 strains including the ryfA-complemented strain via imaging flow cytometry. Representative images were also generated at 2 h p.i. for GFP^high labeled strains within intracellular compartments (Fig 7). Interestingly, the ΔryfA fim locked-ON mutant showed a very high internalization level in HMDM compared to the WT CFT073 strain (Fig 7B). This strain also did not present any survival defect at 2 h p.i. (Fig 7B). Previous studies have suggested that fimbriae-mediated adhesion to macrophages helps E. coli to avoid clearance by the innate immune system [44,45]. In addition, E. coli overexpressing type 1 fimbriae showed increased intracellular survival in macrophages [46]. Of note, we did not observe a significant difference in the uptake of mCherry-labelled CFT073 and ryfA mutant by RAW264.7 macrophages (Fig 6D). However, the ryfA mutant survived significantly less than the parent wild-type strain at 2 h and 24 h p.i. in these cells (Fig 6D). These observations demonstrate that loss of ryfA can contribute to decreased virulence and clearance during extra-intestinal infections such as UTI, since cells lacking ryfA are more susceptible to host phagocytic cells such as macrophages. Importantly, the complemented mutant regained uptake and survival capacity in macrophages similar to wild-type parental strain CFT073 (Figs 6A, 6B and 7). It is of further interest that loss of ryfA also resulted in a reduced level of bacterial uptake by phagocytes, suggesting that some of the changes in expression that occur on the bacterial cell surface led to a reduced initial level of phagocytosis by macrophages. This defect of uptake by macrophages by the ryfA mutant also is not likely to be linked to increased sensitivity to oxidative stress, since the ROS-sensitive oxyR mutant did not demonstrate any differences in uptake by macrophages (Fig 6A and 6B). These results support an important role for the ryfA regulatory RNA in interactions with the host cellular response by contributing to uptake, survival, and replication of UPEC CFT073 inside human macrophages.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Role of RyfA during interaction with professional phagocytes.

Phagocytic cells were infected with CFT073 wild-type strain, ΔryfA or ΔoxyR isogenic mutants or the ΔryfA-complemented strain. (A) in vitro differentiated human monocyte-derived macrophages (HMDMs) and (B) THP-1 human macrophages were infected with different strains for 1 h, followed by gentamicin treatment. Cells were lysed and intracellular bacterial counts (CFU ml⁻¹) were determined at different times post-infection (pi). Infection assessment of mCherry tagged CFT073 strains via flow cytometry in (C) HMDMs and (D) RAW264.7 cells. (E) Reactive oxygen species (ROS) production in RAW264.7 cells. RAW264.7 cells were infected at an MOI of 20 and intracellular generation of ROS was measured at 4 h and 6 h post-infection by using H₂DCFDA. Cells were stained with 10 μM H₂DCFDA. All assays were conducted in triplicate and repeated independently at least three times. The results are expressed as the means ± sem. Significant differences between mutant, WT and complemented mutant strains were determined using One-way ANOVA. *, P < 0.05; **, P < 0.005; ***, P < 0.0001. NS, not significant.

https://doi.org/10.1371/journal.ppat.1009617.g006

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Role of RyfA during interaction with human macrophage cells.

Imaging flow cytometry analysis of human monocyte-derived macrophages (HMDMs) that were infected with CFT073 wild-type strain, ΔryfA mutant, the ΔryfA complemented strain and the ΔryfA fim L-ON strain at 2 h post-infection. (A) Representative images of HMDM singlets CD14⁺GFP^high at 2h post-infection. (B) Infection assessment for double positive percentages of CD14⁺GFP^high HMDMs. Below, representative flow cytometric histograms for an independent experiment. Results are expressed as the means ± sem of the replicate experiments. Significant differences between mutants, WT, and complemented mutant strain were determined using a paired Student t-test. *, P < 0.05.

https://doi.org/10.1371/journal.ppat.1009617.g007

The ryfA mutant generates increased ROS production in RAW264.7 murine macrophage-like cells

The phagocyte NADPH oxidase and ROS production play a key role in the elimination of bacteria following phagocytosis. To measure intracellular ROS levels, RAW264.7 macrophages infected with CFT073, the ΔryfA, ΔoxyR isogenic mutant, or the ΔryfA complemented strain were incubated for 4 h and 6 h with the membrane-permeant fluorescent probe H₂DCFDA. The cells were then stained with a viability marker (propidium iodide) and H₂DCFDA fluorescence was quantified in viable cells. Four hours later, the level of intracellular ROS after interaction with WT strain CFT073 was 3.9-fold higher when compared to uninfected control cells (p = 0.0026) (Fig 6E). The level of intracellular ROS was 25% (p = 0.0093) and 20% (p = 0.0006) higher in cells infected with the ΔryfA and ΔoxyR mutants respectively than in cells infected with the WT parent strain. ROS levels remained high (4-fold higher above basal level) up to 6 h p.i. in cells infected with UPEC CFT073 (Fig 6E). The ROS levels were reduced by 3-fold for the ryfA mutant and by half for the oxyR mutant after 6 h p.i. (Fig 6E), suggesting that both mutants are killed at a faster rate by ROS activity than the wild-type parent. By contrast, no significant difference was observed between the ryfA complemented mutant and the WT parent strain in ROS generation. These results also support the likelihood that attenuation of the ryfA mutant in the bladder and kidneys during UTI in the mouse model may also be due to a greater sensitivity to killing by host phagocytic cells and an increased generation of ROS within these cells.

Effects due to loss of ryfA in UPEC are independent of a predicted small peptide sequence

The mechanisms of interaction of ryfA and its specific role in the adaptive response of pathogenic E. coli to resist environmental stresses and infect the urinary tract remain to be elucidated. A recent study identified that the ryfA gene from Salmonella enterica serovar Typhimurium can encode a small toxic inner membrane protein, called TimP, upon overexpression [27]. Alignment of the ryfA sequence of UPEC CFT073 also identified an open reading frame (ORF) spanning nt +135 to +270 (S7A Fig). This ORF is predicted to be translated into a 45-amino-acid small protein (S7B Fig). We also identified a putative Sec-dependent secretion signal using two different prediction tools: SignalP-5.0 [47] and Phobius [48] and both programs predicted a Sec translocase signal sequence spanning amino acids 1 to 22 (S7B Fig). In order to determine if such a peptide could potentially contribute to some of the regulatory changes observed due to deletion of ryfA, we introduced different point mutations that would eliminate the potential production of such a peptide. Three different variants of ryfA wherein the E. coli ORF predicted to encode a paralog of TimP was disrupted were generated, since some nucleotide changes could potentially affect the secondary structure of the RNA (S8 Fig). We then introduced these variants in single copy into the chromosome to complement the ryfA mutant of strain CFT073. Complementation of the ΔryfA mutant with any of the 3 variants of RyfA restored production of type 1 fimbriae (S7C Fig), reduced motility (S7D Fig) to wild-type levels, led to a regain in resistance to oxidative stress (S7E Fig) and survival in THP-1 macrophages (S7F Fig). These results suggest that the effects observed due to deletion of ryfA in UPEC do not involve the loss of a peptide that it may encode, but are more likely to be linked to the RyfA sRNA.

Ethics statement

This study was performed in accordance with the ethical standards of the University of Quebec, INRS. A protocol for obtaining biological samples from human donors was reviewed and approved by the ethics committee—Comité d’éthique en recherche of INRS (CER 19–507, approved November 19, 2019). Formal written consent was obtained by the donors.

Bacterial strains, growth conditions and plasmids

E. coli strains and plasmids used in this study are listed in S2 Table. E. coli CFT073 was initially isolated from the blood and urine of a patient with acute pyelonephritis [75]. Bacteria were grown in lysogeny broth (LB) (Alpha Bioscience, Baltimore, MD) at 37°C and in human urine. Urine was collected from healthy female volunteers that were from 20 to 40 years old and who had no history of UTI or antibiotic use in the prior 2 months. Each urine sample was immediately filter sterilized (0.2-μm pore size), pooled, and frozen at −80°C and was used within 2 weeks. Antibiotics and reagents were added as required at the following concentrations: kanamycin, 50 μg/ml; ampicillin, 100 μg/ml; chloramphenicol 30 μg/ml.

Construction of site-directed mutants and complementation of strains

All mutants were generated by the procedure described by Datsenko and Wanner using plasmids pKD3 and pKD4 as the template for chloramphenicol and kanamycin resistance cassettes, respectively [76]. Primers used are listed in S3 Table in the supplemental material. Antibiotic resistance cassettes flanked by FLP recombination target (FRT) sequences were removed by transforming the mutant strains with pCP20 expressing the FLP recombinase [77].

Growth under conditions of osmotic stress

Strains were tested for the capacity to grow under conditions of osmotic stress caused by NaCl or urea. Strains were inoculated 1:100 from an overnight pre-culture grown in LB and grown until mid-log phase with shaking. They were serially diluted and plated on LB agar alone and LB agar supplemented with 0.6 M NaCl or 0.6 M urea. Colonies were counted, and growth under each condition was compared to growth on LB agar.

Sensitivity of E. coli strains to reactive oxygen intermediate (ROI)-generating agents

Sensitivity to oxidative stress generating agents was determined by an agar overlay diffusion method on LB plates (1.5% agar) as described by Sabri et al. [78]. Briefly, overnight-grown cultures were used to inoculate (1/100) fresh LB medium without antibiotics, and the resulting cultures were incubated until the O.D. ₆₀₀ was 0.6. Then, 100 μl of each culture were mixed with 3 ml molten top agar and poured onto an LB agar or M9- glucose plate. Whatman filter disks saturated with 10 μl of hydrogen peroxide (30%), plumbagin (53 mM) or paraquat (40 mM) were spotted onto the disks. The plates were then incubated overnight at 37°C. Following growth, the diameters of inhibition zones were measured.

For sensitivity of bacterial cultures to H₂O₂, bacteria were grown at 37°C in LB broth and approximately 2.5 x 10⁸ CFU/ml of wild-type CFT073 or mutant was inoculated into PBS containing either 5 mM H₂O₂ and or H₂O. The test was carried out under static conditions at 37°C and samples were collected at various time points post-inoculation. Samples were diluted and plated on LB agar to determine bacterial counts.

Experimental UTI in CBA/J mice

Experimental infections were carried out using either competitive coinfection or single-strain infection models as described previously [40]. Prior to inoculation, strains were grown for 16 h at 37°C with shaking (250 rpm) in 55 ml of LB medium. For coinfection, cultures were centrifuged and pellets of a lac-negative derivative of the wild-type (WT) and mutant or complemented strains were mixed 1:1. Six-week-old CBA/J female mice were transurethrally inoculated with 20 μl of the 1:1 mixture containing 2 × 10⁹ CFU of the UPEC CFT073 ΔlacZYA strain (QT1081) and 5 × 10⁸ CFU of either the CFT073 ΔryfA (QT5255) strain or its complemented derivative (QT5309). At 48 h p.i., mice were euthanized; bladders and kidneys were aseptically removed, homogenized, diluted, and plated on MacConkey agar to determine bacterial counts. In the single-strain experimental UTI model, mice were infected as described above but with only a single strain (10⁹ CFU), and 48 h p.i., bacterial counts were determined from the bladders and kidneys. Bladders were dissected; one half was used to determine bacterial counts and the other half was resuspended in TRIzol reagent (Invitrogen) for RNA extraction and subsequent analysis of bacterial gene expression.

RNA extraction and quantification of gene expression

Bacterial cultures were grown in triplicate in LB broth or human urine. RNAprotect (Qiagen, Toronto, ON, Canada) was added to cultures and RNA extracted using the RNeasy mini kit (Qiagen, Toronto, ON, Canada). Ambion Turbo DNase (Thermo Fisher Scientific, St. Laurent, QC, Canada) was used to remove contaminating DNA. RNA integrity was assessed using a Nanodrop (ND-1000). RNA was also extracted from infected bladders at 48 h p.i. with TRIzol reagent (Thermo Fisher Scientific, St. Laurent, QC, Canada), followed by DNase 1 treatment. Total RNAs were then reverse-transcribed to cDNAs using TransScript All-in-One First-Strand cDNA Synthesis SuperMix Kit (TransGen, Haidian District, Beijing, China). The resulting cDNA was used for qRT PCR with EvaGreen according to the manufacturer’s instructions (TransGen, Haidian District, Beijing, China). The rpoD gene was used as housekeeping control. Each qRT-PCR run was done in triplicate, and for each reaction the calculated threshold cycle (C_T) was normalized to the C_T of the rpoD gene amplified from the corresponding sample. The fold change was calculated using the 2^−ΔΔCT method [79]. Genes with a fold-change above or below the defined threshold of 2 were considered as differentially expressed. Primers used for qRT-PCR analysis are listed in S3 Table in the supplemental material.

RNA sequencing, mapping and analyses

Cultures were grown in triplicate as described above, and total RNA was isolated using the Qiagen RNeasy Protect Bacteria Mini kit. RNA-seq was performed at Génome Québec Innovation Centre, McGill University. Total RNA was quantified using a NanoDrop Spectrophotometer ND-1000 (NanoDrop Technologies, Inc.) and its integrity was assessed on a 2100 Bioanalyzer (Agilent Technologies, St. Laurent, QC, Canada). rRNA were depleted from 250 ng of total RNA using Ribo-Zero rRNA removal kit specific for bacterial RNA (Illumina). Residual RNA was cleaned up using the Agencourt RNA Clean XP Kit (Beckman Coulter) and eluted in water. cDNA synthesis was achieved with the NEB Next RNA First Strand Synthesis and NEBNext Ultra Directional RNA Second Strand Synthesis Modules (New England BioLabs, Whitby, ON, Canada). The remaining steps of library preparation were done using and the NEBNext Ultra II DNA Library Prep Kit for Illumina (New England BioLabs, Whitby, ON, Canada). Adapters and PCR primers were purchased from New England BioLabs, Whitby, ON, Canada. Libraries were quantified using the Quant-iT PicoGreen dsDNA Assay Kit (Thermo Fisher Scientific, St. Laurent, QC, Canada) and the Kapa Illumina GA with Revised Primers-SYBR Fast Universal kit (Kapa Biosystems, Wilmington, MA, USA). Average size fragment was determined using a LabChip GX (PerkinElmer, Woodbridge, ON, Canada) instrument (Génome Québec).

Reads were checked for quality using the FASTQC tool (Galaxy Version 0.72+galaxy1), poor quality reads by Trimmomatic (Galaxy Version 0.38.0) were converted to FastQ format, or “groomed” for downstream analysis using the FastQ Groomer tool (Galaxy Version 1.1.5). Bowtie (Galaxy Version 2.3.4.3+galaxy0) was used to align RNA-seq reads to the genome of E. coli CFT073 (GenBank accession no. AE014075.1 with default parameters. Finally, the number of reads mapping to each annotated feature was obtained with HTSeq (Galaxy Version 0.9.1). Differential expression between WT CFT073 and the ryfA mutant was analyzed using Degust v4.1.1 (DOI: 10.5281/zenodo.3258932). Transcripts were considered significant if passing the following cut-offs: adjusted p-value of < 0.05 and log₂ fold change of >1.7 and <-1.7. All processing steps were performed using the Galaxy platform (https://usegalaxy.org/).

RNA-seq data have been deposited in the GEO database, accession number GSE157450 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE157450).

Verification of RNA-seq results (validation of RNA-seq results by qPCR)

RNA samples were obtained and prepared as described above for RNA-seq. qRT-PCR was performed to verify the RNA levels of 8 genes selected for validation (soxS, ibpA, cspA, marA, bssS, rpoH, cadA and treC).

Northern blot analysis

The hot phenol method was used to extract total RNA [80]. A sample of 10 μg of total RNA was loaded on a 5% acrylamide (29:1)/8M urea gel. RNA was electro-transferred to a Hybond-XL membrane (Amersham Biosciences) and UV-crosslinked (1200J). Prehybridization was done in 50% formamide, 5× SSC, 5× Denhardt reagent, 1% SDS, and 100 μg/ml sheared salmon sperm DNA for 4 h at 60°C. The radiolabeled RNA probe was added and incubated overnight. Three 15 minutes washes of the membranes were made with 1× SSC/0.1% SDS followed by a final wash with 0.1× SSC/0.1% SDS at 65°C.

RNA probe radio-labeling

The RNA probe antisense to the RyfA sRNA was synthesized in vitro using the T7 RNA polymerase. Oligos EM4969B and EM4970 were used to generate the transcription template. Transcription was carried out in the T7 transcription buffer (40 mM Tris-HCL at pH 8.0, 6 mM MgCl2, 10 mM dithiothreitol, 2 mM spermidine), 400 μM NTPs (A, C and G), 10 μM UTP, 3 μl of α-32P-UTP (3000 Ci/mmol), 20 U RNA guard, 20 U T7 RNA polymerase and 0.5 μg DNA template. After 2 h of incubation at 37°C, the mixture was treated with 2 U of Turbo DNAse (Ambion) for 15 minutes. The radio-labeled probe was then purified on a G50-Sephadex column before hybridization.

Evaluation of type 1 and P fimbriae production

The level of production of type 1 fimbriae was determined by a yeast agglutination assay [35]. Briefly, the strains were cultured at 37°C / 250 rpm in LB broth or human urine to mid-log phase (conditions which were used for transcriptional analyses), or static in LB broth, M9-glucose or human urine overnight. Following centrifugation, 40 μl of an initial suspension of approximately 2 × 10¹¹ cells ml⁻¹ in PBS was transferred and serially diluted 2-fold in microtiter wells containing equal volumes of a 3% commercial yeast suspension in PBS. After 30 min of incubation on ice, yeast aggregation was monitored visually. The agglutination titer was defined as the most diluted sample showing agglutination.

The mannose-resistant hemagglutination (MRHA) assay was determined using human type O⁺ and A⁺ erythrocytes to determine HA that could be mediated by P fimbriae. Bacterial strains were grown to mid-log phase or passaged six times on LB agar at 37°C, after which a bacterial sample was transferred to a well containing 50 μl of 3% (vol/vol) human red blood cells. To inhibit type 1 fimbriae-dependent hemagglutination, a final concentration of 2.5% α-d-mannopyranose was added to samples.

Preparation of fimbrial extracts and Western blotting

Preparation of fimbrial extracts and Western blotting were performed as described previously [81], with anti-FimA serum from E. coli strain B_AM and F1C fimbriae-specific (anti-F165₂) antiserum [82].

Adherence and gentamicin protection (invasion) assays

5637 human bladder cells (ATCC HTB-9) were grown in RPMI 1640 medium (Wisent Bioproducts) supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 10 mM HEPES, 1 mM sodium pyruvate, 4.5 g/liter glucose, and 1.5 g/liter sodium bicarbonate. 5637 cells were grown to confluency in RPMI 1640 and 2 × 10⁵ cells/well were distributed in 24-well plates. UPEC CFT073 and its derivative strains were grown in LB medium at 37°C to the mid-log phase of growth (O.D. 0.6). Immediately before infection, cultures were washed once with PBS to remove dead cells, returned to the infection medium, and infected at an estimated MOI of 10 CFU per cell. The bacterial cells were centrifuged, washed twice with PBS, resuspended in RPMI 1640 medium (Wisent Biocenter, St-Bruno, Canada) supplemented with 10% fetal bovine serum at 10⁶ CFU ml⁻¹, and added to each well. Bacterium-host cell contact was enhanced by a 5-min centrifugation at 600 × g. After 2 h, cells were washed three times and lysed with PBS–0.1% sodium deoxycholate (DOC), serially diluted, and plated on LB agar plates. Quantification of cell-associated bacteria was performed. To block adherence mediated by type 1 fimbriae, 2.5% α-d-mannopyranose was added to culture medium.

For GFP-tagged UTI89, bacteria were grown at 37°C in LB broth supplemented with ampicillin to O.D. 0.6. Cells were infected with UTI89 using a MOI of 10 bacteria per host cell. After a 2 h incubation at 37°C, samples were washed three times with PBS to remove any non-adherent bacteria. Triplicate samples were used to calculate the total number of the associated bacteria present both intra- and extracellularly. To determine invasion frequencies, after the initial 2 h incubation, additional sets of wells were washed three times with PBS and then incubated for another 2 h in medium containing 100 μg/ml membrane-impermeable bactericidal antibiotic gentamicin to kill any extracellular bacteria. Following additional washes with PBS, fresh medium containing a lower concentration of gentamicin (10 μg/ml) was added, and incubations were continued for another 2 h. This submaximal concentration of gentamicin was used to prevent extracellular growth of UPEC while limiting possible leaching of the antibiotic into the host cells during longer incubations. Monolayers were then washed with PBS and the positivity percentage of GFP-tagged bacteria present within the cells were measured by using ImageStream-based assay.

Electron microscopy

Cells for electron microscopy were grown as described above for agglutination assay experiments. A glow-discharged Formvar-coated copper grid was placed under a drop of bacterial culture for 15 min to allow cells to adsorb. Excess liquid was removed using filter paper, just before a drop of 1% phosphotungstic acid (negative stain) was placed onto the grid for 15 sec. Samples were left to air dry and viewed using a Hitachi H-7100 transmission electron microscope.

Swimming motility assay

Motility assays were done using UPEC strains CFT073 and 536 and its mutant derivatives as previously described [28,36] with modifications. Following overnight growth at 37°C, strains were cultured at 37°C in LB broth to mid-log phase. Soft agar (1% tryptone, 0.5% NaCl, 0.25% agar) or urine agar, were stabbed in the center of the plate using an inoculating needle. Care was taken not to touch the bottom of the plate during inoculation to ensure only swimming motility was assessed. After 16 h (soft agar) or 22 h (urine agar) of incubation at 37°C, the motility diameters were measured for each strain. Wild-type E. coli CFT073 was always included as a reference, and five independent motility experiments for each mutant were performed. E. coli CFT073 fim L-ON was used as a negative control. Results were analyzed using a paired t-test.

Infection of human cultured macrophage-like cell line THP-1

The human monocyte cell line THP-1 (ATCC TIB-202) was maintained in RPMI 1640 (Wisent, Saint-Jean-Baptiste, QC, Canada) containing 10 % (v/v) heat-inactivated FBS (Wisent, Saint-Jean-Baptiste, QC, Canada), 1 mM sodium pyruvate (Wisent, Saint-Jean-Baptiste, QC, Canada) and 1 % modified Eagle’s medium with non-essential amino acids (Wisent). A stock culture was maintained as monocyte-like, non-adherent cells at 37°C in an atmosphere containing 5 % (v/v) CO₂. The bacterial strains were cultured at 37°C in LB broth to mid-log phase. Bacterial cells were pelleted, washed with phosphate buffer saline and added to the cell monolayer at a MOI of 20, and plates were centrifuged for 5 min at 800 ×g to synchronize bacterial uptake. After 60 min of incubation at 37°C, extracellular bacteria were removed by washing cells three times with prewarmed PBS, and the infected monolayers were either lysed with PBS-DOC (T0) or incubated for 2 h or 24 h with medium containing 100 μg gentamicin ml^–1 (Wisent) to kill remaining extracellular bacteria, and then with 12 μg gentamicin ml^–1 for the remainder of the experiment. Surviving bacteria were determined (CFU) by plating on LB agar. Results are expressed as the mean±sem of at least three experiments performed in duplicate. One-way ANOVA was used for statistical analysis.

Human monocyte-derived macrophage cultures

Peripheral blood mononuclear cells (PBMC) were obtained from leukaphereses of healthy donors. All participants of the study received approval from the McGill University Health Centre Ethical Review Board (ethic reference number SL-00.069). At least a purity of 96% of human monocytes was obtained using the Human Monocyte Enrichment Kit (Stem Cell Technologies, Vancouver, BC). Afterwards, purified human monocytes were seeded in RPMI 1640 medium supplemented with 2mM glutamine, antibiotics, and 10% FBS in the presence of 50 ng/ml of Human recombinant Macrophage colony-stimulating factor (M-CSF). After 6 days, human monocyte-derived macrophages (HMDMs) were used for infection assays. Differentiation follow-up during 6 days was assessed by flow cytometry analysis.

Multiparametric and imaging flow cytometry analyses

To assess bacterial infection in macrophages, HMDMs were harvested after adding non-enzymatic cell dissociation solution (Sigma-Aldrich, Oakville, ON, Canada) for 30 mins at room temperature. Cells were washed with PBS and prepared in FACS tubes. For surface staining, we used the following specific monoclonal antibodies provided from BD Biosciences for phenotypic analysis of human HMDMs: anti-CD3 (BV605), anti-CD14 (BV650), anti-CD80 (PE Cy7), anti-CD209 (PE) and anti-CD64 (PE). Finally, the viability marker 7-aminoactinomycin D or 7-AAD (ThermoFisher Scientific, St. Laurent, QC, Canada) was used to exclude dead cells from analyses. Afterwards, cells were fixed with 4% paraformaldehyde solution and then washed. For data analysis, samples were analyzed by flow cytometry with a BD LSRII Fortessa flow cytometer and DIVA software (BD). Viable gated cell singlets were analyzed for each sample and the percentages of mCherry^high CFT073 containing CD14⁺ HMDMs were determined. For Imaging flow cytometry, the same procedures were applied while cells were prepared in Eppendorf tubes. For surface staining, we used CD14-V450 and CD80-APC H7 for phenotypic gating of HMDMs. Samples were acquired using the Image Stream X MKII flow cytometer and analyzed with IDEAS software (Amnis) and representative images of singlets CD14⁺ HMDMs were generated expressing GFP^high CFT073 bacterial cells.

Measurement of ROS production

RAW264.7 cells were infected at an MOI 20:1 in complete DMEM medium in 24-well plates. The same conditions as for infection of THP-1 cells were used. After 60 min of incubation at 37°C, extracellular bacteria were removed by washing cells three times with prewarmed PBS, and the cells were then incubated for 4 h or 6 h with medium containing 100 μg gentamicin ml^–1 to kill extracellular bacteria. The fluorescent probe 2′, 7′-dichlorodihydrofluorescein diacetate (H₂DCF-DA) (ThermoFisher Scientific, St. Laurent, QC, Canada), was solubilized at 5 mM in dimethyl sulfoxide (DMSO) (Sigma). Infected or uninfected cells were incubated with 10 μM H₂DCFDA for 45 min at 37°C, and then washed once with PBS 1X. Cells were then trypsin treated, washed once with PBS 1X and cell pellets were suspended in a 1 μg/ml propidium iodide solution (ThermoFisher Scientific, St. Laurent, QC, Canada). The flow cytometer FACSCalibur (Becton Dickinson) was used to measure the fluorescence of the oxidized product dichlorofluorescein at a λ of 488 nm with excitation at 530 nm. The fluorescence of propidium iodide was measured at a λ of 585 nm with an excitation at 542 nm. A total of 10 000 viable cells were analyzed by sample. For all points, experiments were performed in replicates. Statistical differences between different conditions were determined with a two-tailed Student’s t test on the indicated number (n) of experiments.

Biofilm assay

Biofilm formation on polystyrene surfaces was assessed in 96-well plates. Strains were grown at various temperatures (25°C, 30°C, 37°C, and 42°C) for 48 h under static conditions in LB and M9 supplemented with 0.2% glucose. Wells were washed and stained with 0.1% crystal violet (Millipore Sigma) for 15 min, then 200 μl of ethanol-acetone (80:20) solution was added, followed by measuring at an optical density at OD₅₉₅.

Site-directed mutagenesis to generate ryfA variants containing point mutations

Site-directed mutagenesis was performed using the Q5EvaGreen Site-Directed Mutagenesis kit as specified by the manufacturer (New England Biolabs). pIJ546 was used as a template for the construction of the RyfA variants with specific mutations pIJ588, pIJ589 and pIJ590 at 25 to 50 ng per reaction with 10 pmol of each of the complementary primers. Primers used to generate the point mutations were CMD 2712 and CMD2713 for pIJ588 (variant 1), CMD 2715 and CMD2716 for pIJ589 (variant 2) and CMD2718 and CMD2719 for pIJ590 (variant 3). Following mutagenesis, all constructs were verified by PCR and by sequencing.

Statistical analyses

Statistical analyses were performed using the Prism 7.04 software package (GraphPad Software). Statistically significant differences between two groups were established by unpaired t-test and comparisons among three or more groups was done by one-way analysis of variance (ANOVA). For the independent infections, comparisons of the CFU mL−1 or CFU g−1 distributions were analyzed using the Mann–Whitney test. A Wilcoxon signed-rank test (two-tailed; P ≤ 0.05) was used to determine statistical significance for comparison of bacterial numbers in coinfection experiments.