Increased iron utilization and oxidative stress tolerance in a Vibrio cholerae flrA mutant confers resistance to amoeba predation

ABSTRACT

The flagellar transcriptional regulator flrA initiates the regulatory cascade of flagellum synthesis in Vibrio cholerae. Previously, we observed an increase in sub-populations of V. cholerae carrying mutations in the flrA gene during long-term co-adaptation with the amoeba host Acanthamoeba castellanii. The flrA mutants exhibit increased growth and intracellular survival in the amoeba host but the molecular mechanisms were unknown. Using an in-frame deletion of flrA, here we show that the absence of flrA increases pathogen growth and induces a distinct V. cholerae transcriptomic signature during amoeba predation. Transcriptome analysis of a flrA mutant in A. castellanii revealed that several genes involved in iron acquisition and amino acid biosynthesis are highly up-regulated compared to the wild-type strain. Furthermore, we show that iron availability is crucial for the survival of V. cholerae in amoeba. We also report that V. cholerae KatB and KatG catalases confer an increased tolerance to oxidative stress. We conclude that the increased survival of the flrA mutant in amoeba is due to a combination of factors, including resistance to oxidative stress and an increased capacity to utilize essential nutrients such as iron and amino acids. Together, the results presented here detail how a bacterial pathogen increases resistance to protozoan predation, thereby allowing for increased survival in the environment.

IMPORTANCE

Persistence of V. cholerae in the aquatic environment contributes to the fatal diarrheal disease cholera, which remains a global health burden. In the environment, bacteria face predation pressure by heterotrophic protists such as the free-living amoeba A. castellanii. This study explores how a mutant of V. cholerae adapts to acquire essential nutrients and survive predation. Here, we observed that up-regulation of iron acquisition genes and genes regulating resistance to oxidative stress enhances pathogen fitness. Our data show that V. cholerae can defend predation to overcome nutrient limitation and oxidative stress, resulting in an enhanced survival inside the protozoan hosts.

INTRODUCTION

Cholera is an acute life-threatening diarrheal disease that continues to impact many regions in the world, especially in parts of Asia, Africa, and Latin America (1, 2). The causative agent of cholera, Vibrio cholerae, is adapted to survival in aquatic ecosystems as well as in the human gastrointestinal tract (3). To persist in aquatic environments, this bacterium must survive predation by bacterivorous protists (4). Many bacteria, including V. cholerae, are able to survive predation by some unicellular eukaryotic protist hosts and are released back into the environment (5 – 7). However, our understanding of the mechanisms involved in survival and persistence in protist hosts is limited.

Iron is an essential nutrient required for pathogen survival and growth (8). However, biological sources of iron are scarce due to the insoluble nature of iron at neutral pH and due to the fact that iron is chelated with high-affinity iron-binding proteins in the host (9). Hence, to overcome iron-withholding defenses in the eukaryotic host, V. cholerae possesses multiple iron acquisition systems, including receptors and transporters of heme as well as the siderophore vibriobactin to acquire and utilize host-associated iron (9). Although iron is a crucial element for survival, excess amounts of free iron (Fe²⁺) can cause oxidative cell damage by the generation of reactive oxygen species (ROS) due to Fenton reactions (10, 11). Hence, cells must tightly control the cellular pool of iron to maintain proper homeostasis.

In V. cholerae, the expressions of many of the iron acquisition genes are tightly controlled by the ferric uptake regulator (Fur) (12). Fur–iron (Fe²⁺) complexes repress iron acquisition genes in order to mitigate iron-mediated cellular toxicity caused by ROS, thereby ensuring that these genes are only expressed when iron is limited (12). Interestingly, Fur regulation has also been reported to modulate the expression of genes involved in anti-oxidant defenses (13 – 15). V. cholerae also possesses multiple defense systems to mitigate ROS, including the oxidative stress-related transcriptional activator OxyR, catalases (KatB and KatG), and several peroxidases (16, 17). Phagocytic protists such as amoebae exhibit many similarities to mammalian phagocytic cells and use ROS to destroy internalized bacteria (18, 19). Hence, to survive digestion when phagocytized by amoebae, bacterial cells must counteract ROS-mediated damage.

FlrA is a σ⁵⁴-dependent enhancer-binding protein (EBP) that initiates the regulatory cascade for the synthesis of flagellum in V. cholerae (20 – 22). EBPs are known to regulate diverse functions in bacteria, including motility, biofilm formation, quorum sensing, and virulence gene expression (23, 24). FlrA is inhibited by the binding of cyclic-di-GMP and potentially linked to the expression of catalases (25, 26). In our previous study, we co-adapted V. cholerae O1 El Tor strain A1552 in an amoeba host, Acanthamoeba castellanii, for extended periods and observed increases in the sub-populations of V. cholerae carrying mutations in the flrA gene over time (27). Using genetic screening and molecular assays, we found that increased survival and fitness were associated with inactivation of the flrA gene. This work focuses on the mechanisms of increased survival and fitness of the V. cholerae flrA mutants in A. castellanii. Results show that increased survival and fitness of V. cholerae flrA are related to increased expression of iron acquisition-related genes and enhanced tolerance to oxidative stress.

RESULTS

Increased survival of the flrA mutant is not due to the loss of flagella or motility

The flrA mutation leads to the loss of flagellum and, consequently, motility. To determine if it is the loss of flagella or the loss of motility that is responsible for increased survival, we generated non-motile genetic mutants targeting either the major flagellin subunit FlaA, resulting in non-flagellated bacteria, or the flagellar motor protein MotX, resulting in rotation-deficient but fully flagellated bacteria. The lack of motility of these mutants was verified by swimming motility assays using soft agar as described in Materials and Methods (Fig. S1). In-frame deletion knockouts of non-motile mutants flrA, flaA, and motX were subjected to intracellular survival assays in amoeba and compared to the wild-type mutants. Equal numbers of bacterial cells (approximately 1.0 × 10⁷ CFU mL⁻¹) were used to infect approximately 10⁵ cells of amoeba at an MOI of 100:1. The number of intracellular bacterial cells was quantified at 2 and 4 h post-infection using the intracellular survival assay. The CFU, 2 h after infection, revealed a reduced ingestion of the non-motile mutants by amoeba compared to the wild-type mutants (Fig. 1A). After 2 h of infection, non-motile mutants showed approximately 100-fold reduction in ingestion by amoeba compared to the wild-type strain (P < 0.001). Although all the non-motile mutants showed reduced ingestion, only the ∆flrA mutant showed increased intracellular survival in amoeba (Fig. 1B). The fold change of bacterial CFU, 2 and 4 h post-infection, revealed that only the ∆flrA mutant showed approximately 4.8-fold increased survival (P < 0.001) while the other non-motile mutants ∆motX and ∆flaA did not differ significantly compared to the wild-type mutants. Thus, the results indicate that neither the loss of flagella nor the loss of motility is responsible for the increased survival of the ∆flrA mutant in amoeba.

Fig 1.

Uptake and intracellular survival of wild-type and non-motile mutants in amoeba. (A) Uptake of wild-type and non-motile mutants by amoeba expressed as log₁₀ of intracellular CFU at 2 h. (B) Intracellular survival of wild-type and non-motile mutants of V. cholerae in A. castellanii. Fold change survival was calculated by dividing the intracellular CFU at 4  h by the CFU at 2  h. Data were collected from three independent experiments with three biological replicates and are shown as the mean + standard deviation. Statistical analysis was performed using one-way ANOVA and Dunnett’s multiple comparisons test. Statistical significance is indicated by ∗∗∗, P < 0.001 compared to wild type.

Transcriptome profiling of the ∆flrA mutant revealed up-regulation of iron transport systems

To identify the potential mechanisms contributing to the increased survival of the ∆flrA mutant in amoeba, we utilized RNA sequencing (RNA-seq) to compare the ∆flrA mutant and wild-type strains during amoeba predation. Differentially expressed transcripts were determined using the edgeR pipeline (28). Transcripts with log₂ fold change 1.5 and P < 0.05 were considered to be differentially expressed (Supplementary Data S1). The mapping of the transcripts to genes revealed 555 transcripts that corresponds to the equal number of genes that were differentially expressed in the ∆flrA strain, which represent approximately 15% of the V. cholerae genes. Of these 555 differentially expressed genes, 148 and 407 were found to be up- and down-regulated, respectively, in ∆flrA compared to the wild type indicating that FlrA acts primarily as a transcriptional activator in V. cholerae (Fig. 2). A strong correlation was observed between the RNA-seq for biological replicates of both the wild-type and the ∆flrA samples as depicted by multi-dimensional scaling (MDS) plot (Fig. S2) and clustering heatmap of differentially expressed genes (Fig. S3). The MDS plot also revealed a marked difference in gene expression in the ∆flrA compared to the wild type.

Fig 2.

Differential expression analysis between the ∆flrA mutant and the wild type during amoeba predation. The volcano plot represents the differentially expressed transcripts. The negative log of P-value (base 10) is plotted on the y-axis, and the log of the fold change (base 2) is plotted on the x-axis. Up- and down-regulated genes are represented by square and circles, respectively, while non-significant genes are represented by triangles. Transcripts corresponding to genes predicted to be regulated by Fur and low iron are indicated in red. The list of V. cholerae transcripts predicted to be regulated by the Fur and low iron was adopted from Rivera-Chávez et al. (29).

Gene ontology (GO) enrichment analyses revealed a significant over-representation of GO terms associated with iron transport in the up-regulated genes (Supplementary Data S2). GO iron-related terms related to iron with fold enrichment values more than 5 includes iron import to cell (GO:0033212), siderophore metabolic process (GO:0009237), siderophore transport (GO:0015891), iron coordination entity transport (GO:1901678), and iron ion transport (GO:0006826) (Fig. 3). V. cholerae encodes multiple iron acquisition systems, including several receptors, transporters, and binding proteins. The RNA-seq revealed that a majority of the genes involved in the uptake of heme as well as siderophore and vibriobactin were significantly up-regulated in ∆flrA relative to the wild type during amoeba predation (Fig. 2). Of the 148 up-regulated genes, 12.1% (n = 18) were involved in the uptake of heme and vibriobactin-mediated acquisition of iron. On an average, heme and vibriobactin-associated genes showed approximately 4- and 1.8-fold induction, respectively (Fig. 4A). Expressions of many of the iron acquisition genes are tightly controlled by the ferric uptake regulator (Fur). All the Fur-controlled genes that are known to be induced under iron-limiting conditions were also up-regulated in ∆flrA relative to wild type (Fig. 4B). Up- and down-regulation of 13 randomly selected iron genes were validated by quantitative real-time PCR (qRT-PCR) and the relative expression profiles were compared to fold change value of RNA-seq data that revealed similar levels of expression to the genes tested by qRT-PCR (Fig. S4).

Fig 3.

Gene ontology (GO) enrichment analysis of the differentially expressed genes. The top 20 GO terms annotated in the biological process category for the up- and down-regulated genes are selected based on their fold enrichment values. Fold enrichment values were calculated from the number of genes observed in the list of differentially expressed genes (DEGs) divided by the expected number in the reference list for a particular GO term. Fold enrichment values greater than 1 indicates that the category is over-represented for the particular GO term. Conversely, the category is under-represented if it is less than 1. The size of each dot is proportional to the number of DEGs for the given GO term in the reference list. Color bar representing false discovery rate is the Fisher’s exact test P-value adjusted using the Benjamini-Hochberg procedure.

Fig 4.

The ∆flrA mutant showed up-regulation of multiple iron transport systems during amoeba predation. (A) Average log fold change of up-regulated genes categorized as heme utilization genes, vibriobactin genes, and others (up-regulated genes except heme and vibriobactin). (B) Heatmap of RNA-seq data from mean-centered log₂ transformed expression values measured in RPKM units (reads per kilobase of transcript per million reads mapped) for Fur-regulated genes in ∆flrA mutant compared to the wild type. (C) Growth of wild type and ∆flrA under iron-limited conditions. The wild type and ∆flrA were grown in M9 with the presence of Fe³⁺ and the iron-chelator 2,2′-bipyridyl (BP) as indicated in the y-axis. The OD₆₀₀ was taken after 5 h of growth. Statistical analyses were performed using two-way ANOVA and Sidak’s multiple comparisons test. Statistical significance is indicated by ∗∗∗, P < 0.001; ∗∗, P < 0.01; ∗, P < 0.05; ns, not significant compared to wild type. (D) Intracellular survival of wild type and ∆flrA in A. castellanii with varying concentrations of iron as indicated in the x-axis. Fold change survival was calculated by dividing the intracellular CFU at 4  h by the CFU at 2  h. Data were collected from three independent experiments with three biological replicates and are shown as the mean + standard deviation. Statistical analysis was performed using two-way ANOVA and Sidak’s multiple comparisons test. Statistical significance is indicated by ∗∗∗, P < 0.001; ∗∗, P < 0.01; ∗, P < 0.05; ns, not significant compared to the condition without addition of iron for each group.

Iron availability increases the intracellular survival of V. cholerae in amoeba

Since the majority of the genes involved in iron acquisition were up-regulated in the ∆flrA mutant compared to the wild type, the mutant might have an increased capacity for utilization of iron. In order to test this, we performed growth assay in M9 under iron-enriched (by supplementing Fe³⁺) and iron-deficient [by supplementing an iron chelator 2,2′-bipyridyl (BP)] conditions. No significant differences in growth were observed between ∆flrA and wild type in the presence of M9 medium supplemented with glucose (Fig. 4C). However, under iron-enriched conditions (100, 150, and 200 µM Fe³⁺), the ∆flrA showed significant increases in growth compared to the wild-type strain (Fig. 4C). In contrast, under iron-deficient conditions (100 and 200 µM BP), the ∆flrA mutant showed no differences in growth compared to the wild type (Fig. 4C). The ∆flrA mutant also showed no differences in growth compared to the wild type with addition of BP and iron (100, 150 and 200 µM Fe³⁺ with 100 µM BP). These results indicate that iron-deprived conditions confer no growth advantage to the ∆flrA while the presence of iron confers a growth advantage to the ∆flrA over the wild-type strain under in vitro conditions.

To investigate the role of iron in the intracellular survival in amoeba, we performed an intracellular survival assay in the presence of varying concentrations of iron (Fig. 4D). The chelation of iron with an iron chelator in the amoeba co-incubation leads to cyst formation by the amoeba; hence, the intracellular survival assay was not performed with the addition of an iron chelator. However, with varying concentrations of iron (0–50 mM), the intracellular survival of V. cholerae increased; thus, iron is crucial for intracellular survival. Addition of iron with a starting concentration of 0.1 mM significantly improves the survival of wild type by 15-fold (1.2/0.08) and that of ∆flrA by 12-fold (5.1/0.4) compared to the condition without iron. The optimum concentration of iron for increased survival was 1 mM (the concentration that was used in 2M medium), which leads to 16-fold (1.3/0.08) increase in survival in the wild type and 18-fold (7.3/0.4) increase in ∆flrA compared to the condition without iron. Those with higher than the optimum concentrations (2.5 mM) lead to a decline in survival in both wild type and ∆flrA. The high concentration (2.5–50 mM) of iron reduces intracellular survival of the wild type and ∆flrA while the wild type is more affected compared to the ∆flrA. Fenton reaction mediated by iron toxicity due to the generation of free radicals might be the potential driver for the reduction in intracellular survival observed with the high concentration of iron. It is also important to note that the ∆flrA showed approximately 5-, 4-, 5-, 6-, 3-, and 4-fold increases in survival compared to the wild type under conditions supplemented with 0, 0.1, 1, 2.5, 5, and 50 mM iron, respectively. The above results indicate that external supplementation of iron under in vivo conditions enhances the survival of both wild type and ∆flrA; hence, iron utilization is a crucial mechanism by which V. cholerae confers resistance to predation.

Catalases promote increased survival of ∆flrA in amoeba

In addition to iron acquisition genes, two catalase-encoding genes, katB (VC1585) and katG (VC1560), showed opposite patterns of transcriptome expression. The katB gene was significantly up-regulated (1.6-fold) in the flrA mutant while katG was significantly down-regulated (−3.8-fold) during co-incubation with amoeba. Up- and down-regulation of katB, katG, and other oxidative stress-related genes (oxyR and sodC) were validated by qRT-PCR and the relative expression was compared to the fold change value of RNA-seq data (Fig. S4). Since the katB gene was significantly up-regulated in the flrA mutant, we hypothesized that the increased katB activity might contribute to the increased survival in amoeba. To validate this, we generated katB and ∆flrAkatB mutants and tested them for the survival in amoeba using the intracellular survival assay. The results indicate that deletion of katB in ∆flrA attenuates survival of the ∆flrA mutant in amoeba by approximately 1.5-fold (6.3/4.2) (P < 0.01) (Fig. 5A).

Fig 5.

Increased intracellular survival, catalase activity, and oxidative stress resistance of catalase mutants. (A) Intracellular survival of wild-type and catalase mutants of V. cholerae in A. castellanii. Fold change survival was calculated by dividing the intracellular CFUs at 4  h divided by the CFUs at 2  h. (B) Catalase activity of the wild-type and mutant strains extrapolated from the standard curve (Fig. S5). (C) Oxidative stress resistance of wild-type and catalase mutants of V. cholerae. Strains were grown in LB to mid-log phase. The cells were resuspended in PBS and exposed to a final concentration of 2 mM H₂O₂ for 60 min. Surviving bacteria are expressed as a percentage of the initial inoculum. Data were collected from three independent experiments with three biological replicates and are shown as the mean + standard deviation. Statistical analyses were performed using one-way ANOVA and Tukey’s multiple comparisons test. Statistical significance is indicated by ∗∗∗, P < 0.001; ∗∗, P < 0.01; ∗, P < 0.05; ns, not significant compared to wild type.

Since KatB showed partial activity, we tested the prominent catalase KatG. We have tested ΔflrAΔkatG and ΔflrAΔkatBΔkatG and found that the intracellular survival of these two mutant strains drops significantly compared to the wild type. No significant differences were found between the wild-type mutant and the katB mutant; however, ΔkatG and ΔkatBΔkatG showed significant reductions in intracellular survival compared to the wild type. Complementation of katG to ∆katG and ∆katB∆katG strains and induction with arabinose restore increased survival of the mutants to wild-type levels (Fig. S5). The catalase activity of the constructed mutant strains was evaluated by extrapolating from a standard catalase activity curve and comparing with wild-type strain (Fig. S6). Quantification of catalase activity revealed that the ∆flrA showed approximately 3-fold (76/23) increased catalase activity compared to wild type (P < 0.001) (Fig. 5B). The ∆flrAΔkatB showed 1.4-fold (76/54) reduced catalase activity compared to the ∆flrA (P < 0.001). The ΔflrAΔkatG showed 5.4-fold (76/14) and the ΔflrAΔkatBΔkatG showed 10.8-fold (76/7) reduction in catalase activity compared to the ∆flrA (P < 0.001).

To further confirm that the increased catalase activity of the ∆flrA leads to increased oxidative stress resistance and contributes to the enhanced survival of the ∆flrA mutant in amoeba, we performed oxidative stress resistance assays using sub-lethal doses of H₂O₂. The ∆flrA showed approximately 3.6-fold (21/5.8) increased resistance to H₂O₂ compared to the wild type (Fig. 5C). The ∆flrAΔkatB also showed a 1.6-fold (21/12) attenuated survival in response to H₂O₂ compared to the ∆flrA (P < 0.001). The ΔflrAΔkatG and ΔflrAΔkatBΔkatG were approximately 26-fold (21/0.8) attenuated compared to the ∆flrA (P < 0.001). To further confirm that the ROS response is involved in enhanced survival of V. cholerae, we added a ROS scavenger N-acetyl cysteine (NAC) to the intracellular survival assay. The results showed that addition of 100 µM NAC improves the survival of the wild-type strain by 4.5-fold compared to the control (without addition of NAC) (Fig. 6).

Fig 6.

Effect of amino acids and ROS scavenger supplementation in intracellular survival. Fold change survival was calculated by dividing the intracellular CFU at 4  h by the CFU at 2  h. Data were collected from three independent experiments with three biological replicates and are shown as the mean + standard deviation. Statistical analysis was performed using one-way ANOVA and Dunnett’s multiple comparisons test. Statistical significance is indicated by ∗∗∗, P < 0.001; ∗∗, P < 0.01; ∗, P < 0.05 compared to wild type.

Multiple amino acid biosynthetic genes and pathways are also enriched in ∆flrA transcriptome

In addition to the iron transport systems, several GO terms associated with amino acid biosynthesis showed fold-increased enrichment values in the analysis. Most notably, the arginine biosynthetic process via ornithine (GO:0042450) showed the highest fold enrichment (26.2) followed by arginine biosynthetic process (16.7) (Fig. 3). GO terms with enrichment greater than 5 that are associated with other amino acids include threonine, methionine, histidine, glutamine, homoserine, aspartate, and sulfur amino acids. To determine their potential role in intracellular survival in amoeba, we selected the three top-most amino acids based on the GO analysis (arginine, threonine, and methionine) and we supplemented them in intracellular survival assays. The results revealed that supplementation of all these three amino acids significantly enhanced the survival of the wild type (Fig. 6). In contrast, GO enrichment analyses also revealed a significant over-representation of GO terms associated with cell motility and chemotaxis in the down-regulated genes (Fig. 3). In addition to motility and chemotaxis, GO terms associated with lipid/fatty acid catabolic pathways are significantly enriched in the down-regulated genes (Fig. 3). The KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway abundance analysis correlated with the GO enrichment analysis and revealed that, in addition to iron acquisition pathways, several amino acid metabolic pathways were significantly enriched in up-regulated genes including arginine, histidine, alanine, aspartate, glutamate, cysteine, methionine, glycine, serine, and threonine metabolic pathways (Fig. S7). In contrast, pathways involved in flagellar assembly, chemotaxis, fatty acid degradation, and two-component systems were mostly down-regulated in ∆flrA (Fig. S7). The highest proportion of genes involved in chemotaxis pathways (0.6, n = 43) was down-regulated in ∆flrA followed by genes involved in flagellar assembly (0.56, n = 28), fatty acid degradation (0.54, n = 6), benzoate degradation (0.42, n = 3), butanoate (0.41, n = 12), two-component systems (0.36, n = 61), and pyruvate (0.25, n = 12) metabolism (Fig. S7).

DISCUSSION

The results presented here highlight mechanisms that allow bacteria to adapt to predators in the environment. Analysis of the transcriptome of the ∆flrA strain that exhibited increased intracellular survival during amoeba predation revealed strategies that allow for the utilization of essential nutrients, including iron, and amino acids as well as protection against ROS. The transcriptome also revealed metabolic changes associated with increased intracellular survival and revealed putative mechanisms of selection of a particular clone under predation pressure.

The expressions of transcripts related to genes involved in iron acquisition in mammalian hosts have been linked to enhanced fitness and virulence in V. cholerae (29). Here, a similar pattern was observed where there was up-regulation of iron acquisition genes leading to increased fitness and survival inside the unicellular eukaryotic host A. castellanii. The addition of iron significantly improves the survival of V. cholerae and the increased expression of genes involved in iron acquisition reflects the competitive environment for iron inside the amoeba host. Iron acquisition genes are under tight control by the Fur–Fe²⁺ complex (12) and the increased expression of these genes in a ∆flrA background indicates potential regulatory interactions between Fur and FlrA.

The Fur–Fe²⁺ complex has been reported to regulate genes involved in anti-oxidant defense, both positively and negatively (14, 30). It is possible that all the genes involved in anti-oxidant defense are differentially regulated by Fur–Fe²⁺ in V. cholerae as was observed in the transcriptomic analysis. A previous study indicates that KatG is a more potent catalase and is more efficient in detoxifying ROS compared to KatB in V. cholerae wild-type background under different conditions (16). Here, we also observed that KatG is the prominent catalase in V. cholerae that is involved in an enhanced survival and predation defense in amoebal hosts. The differential expression of the two catalases KatB and KatG indicated that the use of only one system may be metabolically advantageous under the conditions tested here. In this study, we have observed down-regulation of KatG in ΔflrA background. Expression of katG in a few cells in a population might be sufficient to mediate the anti-oxidant defense in amoeba, which might not been captured in bulk RNA-seq data, but the use of single-cell RNA-seq might be useful to predict their precise expression and function in the future study. KatB is regarded as a heme catalase that uses heme as a co-factor for the enzymatic conversion of H₂O₂ to H₂O and it has been reported that iron content does not impact catalase expression but affects catalase activity in V. cholerae at the post-translational level (12, 31). Thus, increased iron/heme availability in the ∆flrA mutant may lead to the up-regulation of katB gene while down-regulating the katG. In addition, flrA may have a regulatory role in expression of catalases. In fact, the signaling molecule cyclic-di-GMP increases KatB expression but not KatG (26). In V. cholerae, cyclic-di-GMP modulates gene expression through three transcriptional regulators, VpsT, VpsR, and FlrA. The expression of KatB mediated by cyclic-di-GMP is dependent on the transcription factors VpsT and VpsR (26). However, the role of FlrA in the expression of catalases cannot be ruled out as cyclic-di-GMP also acts as a modulator of FlrA (25). Binding of cyclic-di-GMP inactivates the transcriptional factor FlrA and inhibits the downstream signaling cascade of flagellar synthesis (25). Hence, the absence of FlrA might lead to the co-ordinate regulation of catalases through combined effects of other transcriptional regulators VpsT and VpsR.

The ∆flrA strain showed up-regulation of genes involved in the metabolism of many amino acids and the precise role of the up-regulation of these amino acid genes and their metabolites in predation resistance needed to be elucidated in future studies. The highest fold enrichment was observed for genes involved in arginine biosynthesis, which is considered to be a metabolically flexible amino acid. We have found that supplementation with arginine improves the survival of V. cholerae in amoeba, which can act as a carbon source and is interconvertible with a range of other amino acids and metabolites such as proline, glutamate, nitric oxide, creatine, polyamines, agmatine, and metabolites of the urea cycle (32). These metabolites play a significant role in both bacteria and their eukaryotic hosts. Up-regulation of cysteine and methionine reflects additional anti-oxidant mechanisms of the flrA mutant in addition to catalase as both of the amino acids are involved in anti-oxidant defenses (33). V. cholerae is a facultative anaerobic pathogen and is capable of fermenting diverse carbohydrates including glucose (34). Low levels of oxygen inside amoeba might induce switching to anaerobic metabolism indicated by the down-regulation of carbon and pyruvate metabolism. This down-regulation is indicative of an anaerobic metabolic lifestyle, producing less energy through cycling of carbon through the TCA cycle while favoring synthesis of essential amino acids. The fermentative pathway enables bacteria to produce low levels of energy by altering the flux of pyruvate to produce organic acids and other molecules like amino acids. The flux of pyruvate to the production of the neutral molecule acetoin benefits bacteria at low pH (35). Thus, by reducing carbon–pyruvate flux to energy production, the ∆flrA produces molecules/amino acids that might enhance their ability to survive ROS and low pH encountered inside the amoeba.

Taken together, the results presented here highlight how V. cholerae changes gene expression in response to protozoan predation. Oxidative stress resistance and utilization of essential nutrients confer resistance to protozoan predation.

MATERIALS AND METHODS

Organisms and growth conditions

Organisms, oligonucleotide primers, and plasmids used in this study are listed in Table S1. Unless otherwise stated, V. cholerae O1 El Tor strain A1552 and its derived mutant strains were grown in lysogeny broth at 37°C with shaking at 200  rpm. A. castellanii was routinely maintained axenically in peptone yeast glucose (PYG) medium (20 g protease peptone, 5 g yeast extract, and 50  mL 2 M glucose L⁻¹) at room temperature. To measure the optical density (OD) in iron-enriched and iron-limited conditions, we added Fe³⁺ salts and an iron chelator 2,2′-bipyridyl (BP), respectively, in M9 medium at final concentrations indicated in the results section. The medium was also supplemented with 0.2% glucose as a carbon source. Swimming motility assays of the mutants were performed on LB plates containing 0.3% (w/v) agar. The colonies of the respective mutants were inoculated by stabbing and the plates were photographed after incubation at 30°C for 24  h.

Intracellular survival and uptake assay in amoeba

The intracellular survival assay of V. cholerae in amoeba was determined according to the method described in Hoque et al. (27). Briefly, 3 d old A. castellanii cultures were washed to remove PYG medium and seeded at a concentration of 1 × 10⁵ cells mL⁻¹ in 24-well plates in (2M) marine minimal medium (1 M MOPS, pH 8.2; 132 mM K₂HPO₄; 952 mM NH₄Cl; 0.4 M tricine and 1 mM FeSO₄.7H₂O, pH 7.8 in artificial seawater) supplemented with 0.01% glucose (36). Amoebae were allowed to adhere to the bottom of the wells for 1 h before addition of bacterial cells. V. cholerae strains grown overnight were washed and diluted to OD₆₀₀ = 1.0 in 2M medium and 10⁷ diluted cells were used to infect triplicate wells of previously seeded A. castellanii at a final amoeba–bacteria ratio of 1:100 and incubated at room temperature. The intracellular bacteria were recovered at different time points by lysis of the amoeba cells with 1% Triton-X. Extracellular bacteria were removed by adding 300 µg mL⁻¹ of gentamicin (Sigma-Aldrich, USA) followed by washing the wells with 2M medium 30 min before lysis. The number of V. cholerae cells present in the lysates was determined by enumeration on LB agar plates. The percentage of intracellular survival was calculated using the formula: number of surviving bacteria at 4 h/number of surviving bacteria at 2 h. The uptake of the test strains was determined by the surviving number of bacteria at 2 h measured by CFU on LB agar plates.

Generation of mutants and complementation

The V. cholerae in-frame deletion mutants used in this study were constructed as previously described (27). The V. mutants were constructed by splicing overlap extension PCR (SOE PCR) followed by natural transformation (37). Briefly, two sets of primers were used to amplify upstream and downstream regions of the target gene and fused with the chloramphenicol acetyltransferase (cat) gene amplified from pKD3 using SOE PCR. The resulting construct was transformed into the wild-type strain using chitin-mediated natural transformation (38). The transformants were selected on LB agar plates supplemented with 5 µg mL⁻¹ of chloramphenicol and the cat gene was removed using the TransFLP method to get an in-frame deletion mutant of the target gene (39). The absence of the target genes was confirmed by PCR. For complementation, the katG gene was amplified by PCR from V. cholerae A1552 strain and cloned into pBAD24 using Gibson assembly. The resultant construct was transformed into Escherichia coli using chemical transformation method. Recombinant transformants were selected on LB agar plates containing 100 µg mL⁻¹ of carbenicillin. The recombinant plasmids were transformed into the recipient V. cholerae strains using chitin-mediated transformation.

RNA extraction and sequencing

V. cholerae wild-type and ∆flrA mutants were grown with A. castellanii as described in the previous sections. After overnight co-incubation, cells were washed three times in 2M medium, and amoeba was lysed with 1% Triton-X to release intracellular bacteria. RNA protect (Qiagen, Hilden, Germany) was added immediately to the cell lysate and total RNA was extracted using the RNeasy plus mini kit (Qiagen, Hilden, Germany). A combination of lysozyme and proteinase K was used to ensure complete lysis. RNeasy plus mini kit’s gDNA shearer column as well as RNase free DNase (Qiagen, Hilden, Germany) on column DNA digest were used to ensure sufficient degradation of the contaminant DNA in the sample. The quantity and purity ratios of samples were determined with Nanodrop One. A Clean & Concentrator kit was used to purify the extracted RNA (Zymo Research). RNA samples were quantified fluorometrically using a Qubit fluorometer (Thermo Fisher) and quality was evaluated using TapeStation (Agilent Technologies). Only samples with a RIN number higher than 8 were used for analysis. RNA-seq libraries were prepared using Illumina TruSeq Stranded mRNA kit according to the manufacturer’s protocol (Illumina, San Diego, CA, USA). The libraries were sequenced on the Illumina HiSeq2500 platform (V2 100 × 100 bp) at the Singapore Centre for Environmental Life Sciences Engineering, Nanyang Technological University, Singapore.

Transcriptomic analysis

The RNA-seq reads were filtered to remove adapter contamination and low-quality bases (≤Q20) using TrimGalore and were checked using FastQC before analysis (40). Clean reads were aligned to the reference genome of V. cholerae O1 El Tor strain N16961 (RefSeq accession numbers NC_002505 and NC_002506 for chromosome I and II, respectively) using Subread aligner (41). The mapped reads per gene count were quantified using featureCount (42). The count matrices were uploaded to R version 4.1.0 and differential expression and MDS analysis were performed using edgeR (28). Heatmaps were generated using mean-centered log₂ transformed expression values measured in RPKM (reads per kilobase of transcript per million reads mapped) unit and visualized by pheatmap package in R (43). The GO enrichment analysis of the differentially expressed genes was performed in Gene Ontology Web server (http://geneontology.org/) that uses panther databases (http://www.pantherdb.org) for classification of the genes (44, 45). KEGG pathway analyses were conducted using the kegga function in R package limma. Other visualization was generated with ggplot2 package in R (46).

Quantitative real-time PCR

RNA for qRT-PCR was isolated as described in the earlier section. The reverse transcription reaction and real-time PCR were carried out using the iTaq Universal SYBR Green One-Step Kit (Bio-Rad) according to the manufacturer’s instructions. Briefly, 100 ng of RNA was mixed with iTaq universal SYBR green reaction mix and iScript reverse transcriptase in a 20 µL final volume. The assay was carried out in 96-well plates in a QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) with specific primer pairs for target genes. The relA gene was used as a housekeeping reference gene and the relative gene expression was determined by the 2^-ΔΔCt method (47). All the primers used for qRT-PCR are listed in Table S1.

Oxidative stress sensitivity assay

To assess oxidative stress sensitivity, we measured the viability of V. cholerae strains after exposure to hydrogen peroxide (H₂O₂). Exponential-phase cultures were normalized to an OD₆₀₀ of 0.6 and treated with 2 mM H₂O₂ for 60 min. The viability of the cells was assessed by enumerating CFU after overnight growth on LB agar. Survival was determined by normalizing CFU to the H₂O₂ non-treated group. The assay was performed in triplicate in three independent experiments.

Catalase activity assay

Catalase activity was determined as previously described (48). Briefly, V. cholerae strains were grown in LB at 37°C with shaking at 200 rpm. Cells were washed and adjusted to OD₆₀₀ of 1.5 in phosphate-buffered saline (PBS). Equal volumes of normalized cells were mixed with catalase reaction buffer (1% Triton X-100 and 3% hydrogen peroxide in PBS) in tubes. The tubes were incubated at room temperature and the height of the bubbles was measured when bubbling subsided (approximately after 10 min). A standard curve was generated using purified bovine catalase (Sigma) mixed with catalase reaction buffer for 10 min (Fig. S5). Catalase activities of the samples were determined by linear regression equation calculated by “geom_smooth” function of ggplot2 package in R using the standard curve.

Statistical analyses

The R software package and GraphPad Prism software version 9 (La Jolla, CA, USA; www.graphpad.com) were used for statistical analyses. Statistical analyses for experiments with multiple samples were performed using either two-way ANOVA and Sidak’s multiple comparisons test or one-way ANOVA and Dunnett’s multiple comparisons test.

DATA AVAILABILITY

The RNA-seq data and count matrix have been deposited in the Gene Expression Omnibus under the accession number GSE215287.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.01095-23.

Supplemental file 1. aem.01095-23-s0001.docx.

Fig. S1 to S7 and Table S1

Supplemental file 2. aem.01095-23-s0002.xlsx.

Supplementary Data S1

Supplemental file 3. aem.01095-23-s0003.xlsx.

Supplementary Data S2

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.