Bacterial strains and plasmids used in this study are listed in Supplementary Table S1. In general, bacterial strains were grown at 37°C with aeration and shaking at 250 rpm in Luria-Bertani (LB) (Becton Dickinson and Company, Franklin Lakes, NJ, United States) supplemented with appropriate antibiotics (ampicillin (Ap) at 200 μg/mL, kanamycin (Km) at 50 μg/mL, chloramphenicol (Cm) at 30 μg/mL, and tetracycline (Tc) at 5 μg/mL) when necessary. For microarray analyses, bacterial cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) (Thermo Fisher Scientific, Waltham, MA, United States) under gastrointestinal tract (GIT)-mimicking condition (Donnenberg and Kaper, 1992).

Porcine alveolar macrophage cell line 3D4/31 was maintained in RPMI1640 medium supplemented with 10% fetal bovine serum (FBS), 1% penicillin–streptomycin, and non-essential amino acid (Thermo Fisher Scientific, Waltham, MA, United States) on a 100 mm × 20 mm culture dish (Corning, NY, United States) at 37°C under 5% CO₂ atmosphere.

To investigate the role of ppGpp in EPEC pathogenesis, relA and spoT mutations were introduced into E. coli serotype O127: H6 E2348/69. Since a spoT deletion mutation is lethal to E. coli with a relA⁺ background (Xiao et al., 1991), the double mutant EPEC strain was constructed by consecutive allelic exchange (Miller, 1992) with combination of two conjugative plasmids, pCVD442/relA::aphA-3 and pCVD442/spoT::cat in E. coli strain S17-1 λpir. Initially a strain carrying a single relA deletion mutation (∆relA::aphA-3) was constructed. Then spoT deletion mutation was introduced afterward (∆relA::aphA-3 ∆spoT::cat). Briefly, the recombinant DNA construct pCVD442/relA::aphA-3 was transferred to E2348/69 via conjugation and transconjugants were selected with LB agar plates containing Ap. Several Ap- and Km-resistant transconjugants were grown overnight at 30°C in the presence of 6% sucrose to eliminate the plasmid backbone by homologous recombination. Resultant exconjugants were tested for Ap-susceptible and Km-resistant phenotypes. They were further confirmed by PCR and DNA nucleotide sequencing. To create a double relA- and spoT-deleted mutant, another integration plasmid pCVD442/spoT::cat was transformed into the relA mutant by conjugation as described above, followed by selection with antibiotics. The resulting Ap-susceptible and Cm-resistant exoconjugants were further confirmed by PCR and DNA sequencing and designated a ΔrelAΔspoT double mutant.

For relA and spoT complementation, both multi- and low-copy plasmids expressing intact spoT, or both relA and spoT of EPEC were constructed. Briefly, intact relA and spoT genes were PCR amplified using gene-specific primers with indicated restriction sites. Digested PCR fragments were ligated into corresponding restriction sites in plasmid pUC19 or pACYC184 and transformed into E. coli DH5α. Resulting plasmids were confirmed by DNA sequencing. Mutation and complementation of relA/spoT in EPEC strains were confirmed by growing transformants on minimal glucose (MG) agar plates.

Wildtype, ΔrelAΔspoT mutant, relA- and spoT-complemented strains were grown overnight at 37°C on LB agar plates. Cells were then collected, washed, resuspended in 2.5% glutaraldehyde in phosphate-buffered saline (PBS), and kept on ice. Samples were sent to the Korea Research Institute of Bioscience and Biotechnology (KRIBB, Daejeon, Korea) and observed under transmission electron microscopy (TEM) (JEM-1200EX, JEOL Ltd.). At least two independent sections were prepared and analyzed.

Samples for 2D proteomic analysis were prepared as follows. Bacterial cultures were grown in LB at 37°C for 18 h and diluted 1:100 with 30 ml of pre-warmed LB. They were further incubated at 37°C with aeration and shaking at 250 rpm until optical density at 600 nm (OD₆₀₀) of 0.85. Cells were then placed on ice for 20 min and harvested at 4°C. Bacterial pellets were resuspended in a sample buffer consisting of 7 M urea, 2 M thiourea, 4% (w/v) CHAPS, 1% (w/v) DTT, 2% (v/v) pharmalyte (pH 3.5–10, GE Healthcare, Chicago, IL, United States), and 1 mM benzamidine. These samples were sent to Genomine Inc (Pohang, Korea) for 2D electrophoresis, normalization, and protein identification.

Bacterial cultures grown in LB at 37°C for 18 h were diluted 1:100 with 30 ml of pre-warmed LB and further incubated at 37°C with aeration and shaking at 250 rpm for 6 or 19 h. Whole cell lysates were separated on 10% acrylamide gel with sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Separated proteins were transferred onto Immobilon-P polyvinylidene difluoride (PVDF) membranes (MilliporeSigma, Burlington, MA, United States). Membranes were then blocked with 5% (w/v) non-fat dried milk in PBS plus 0.5% (v/v) Tween-20, incubated with antibodies targeting specific proteins, including affinity-purified rabbit α1 bundlin antiserum (a kind gift from M. Donnenberg at School of Medicine, University of Maryland, United States), anti-GadA (MyBioSource, San Diego, CA, United States), and anti-DnaK (Enzo Life Sciences Inc., Farmingdale, NY, United States) at a dilution of 1:20,000 in PBS containing 1% (w/v) non-fat dried milk and 0.05% (v/v) Tween-20, and reacted with horseradish peroxidase-conjugated anti-rabbit or mouse IgG antibody (Bethyl Laboratories, Montgomery, TX, United States) at a dilution of 1:10,000. Blots were developed with an Enhanced Peroxidase Detection kit (AbClon, Seoul, Korea) following the manufacturer’s instructions.

Glutamate-dependent acid resistance assay was performed as previously described (Lee et al., 2018). Briefly, bacterial cultures grown in LB with 0.4% glucose (to repress stationary sigma factor) at 37°C for 19 h were diluted 1:100 with 40 ml of pre-warmed E minimal media containing 0.4% glucose (EG) medium and further incubated at 37°C with aeration and shaking (250 rpm) for 2 h. Bacterial survival rates were determined by colony forming units (CFU) before and after acid shock. All data are shown as means ± standard error of the mean (SEM) from three independent experiments. Statistical significance was analyzed by Student’s t-test using SPSS 24 software (SPSS Inc., Chicago, IL, United States).

Bacterial cultures grown in LB at 37°C for 18 h were diluted 1:100 with 5 ml of pre-warmed LB and further incubated at 37°C with aeration and shaking (250 rpm) until OD₆₀₀ reached 0.85. Bacterial cultures were kept on ice for 1 min and centrifuged at 13,000 rpm for 5 min. The supernatant was filtered through a 0.45 μM syringe filter to remove any bacterial cells or debris. The filtrate was concentrated with a Vivaspin-500 (MWCO 10 kDa; Sartorius, Goettingen, Germany) by centrifugation at 15,000 rpm for 20 min at 4°C. Protein concentration was determined using a BCA protein assay kit (Thermo Fisher Scientific, Waltham, MA, United States). Concentrated samples were subjected to 10% SDS–PAGE and stained with Coomassie Brilliant Blue.

RNA extraction was performed as follows. For bacterial RNA extraction, overnight LB cultures were diluted 1:100 in pre-warmed LB containing appropriate antibiotics and incubated at 37°C for 6 h with shaking (250 rpm). Bacterial total RNAs were extracted using a HiYield™ Total RNA Mini Kit (Real Biotech Corporation, Taipei, Taiwan). For 3D4/31 RNA extraction, total RNA was extracted from 3D4/31 cells resuspended in Trizol reagent using a RNeasy mini kit (QIAGEN, Hilden, Germany) in accordance with the manufacturer’s instructions. Residual genomic DNA in crude total RNA was eliminated by DNase I (Ambion, Austin, TX, United States) treatment.

For cDNA synthesis, 3 μg of total RNA was reverse transcribed using 25 μg/mL oligo dT or random primers, 0.5 mM deoxynucleotide triphosphate (Promega, Fitchburg, WI, United States), RNase inhibitor, and PrimeScript™ reverse transcriptase (TaKaRa BIO, Kusatsu, Shiga, Japan) in accordance with the manufacturer’s instructions. For qRT-PCR reaction, each well contained 10 μL 2 × FastStart DNA Master SYBR Green I (Roche, Basel, Switzerland), 2 pM primers, 0.5 μL cDNA, and distilled water. Primers used in this study are listed in Supplementary Table S2. A housekeeping gene TBP1 was used an internal reference gene for 3D4/31 qRT-PCR (Nygard et al., 2007) and the rpoB gene was used as internal control for bacterial qRT-PCR (Leverton and Kaper, 2005; Crane et al., 2007). PCR was performed using a LightCycler® 96 system (Roche, Basel, Switzerland). Thermal conditions consisted of a pre-incubation at 95°C for 10 min, followed by 45 cycles with denaturing at 95°C for 10 s, annealing at 60°C for 10 s, and elongation at 72°C for 10 s. To confirm a single PCR product, melting curves were generated after final elongation by increasing temperature from 65°C to 95°C. Gene expression levels were analyzed using the comparative delta–delta Ct (2^−ΔΔCt) method (Livak and Schmittgen, 2001). All data are shown as means ± SEM from three independent experiments. Statistical significance was analyzed by Student’s t-test using SPSS 24 software.

Germ line-defective and temperature-sensitive Caenorhabditis elegans strain SS104 was obtained from the Caenorhabditis Genetic Center at the University of Minnesota. C. elegans were maintained at 15°C on nematode growth media (NGM) (0.3% NaCl, 0.25% peptone, 5 mg/ml cholesterol, 1 M CaCl₂, 1 M MgSO₄, 1 M potassium phosphate buffer, 2.5% agar) plates and propagated on a non-pathogenic feeder strain E. coli OP50. In brief, the C. elegans strain N2 was propagated on NGM plates at 15°C and synchronized. Synchronized worms were fed on a lawn of the non-pathogenic feeder strain E. coli OP50 for 72 h until they reached Larva 4 stage known to be hypersensitive to pathogens. These nematodes were transferred onto NGM plates containing 10 μl of EPEC O127:H6 or E. coli OP50 strains pre-cultured in LB media at 37°C overnight with shaking at 150 rpm and then incubated at 25°C. Live worms were scored every 24 h. Worms were considered dead if they did not respond to tapping with a platinum wire pick. Three independent experiments were performed. Data were recorded as the mean survival percentage ± standard deviation. Biological significance between wild-type and mutant E. coli strains in killing C. elegans was statistically determined by repeated measures Analysis of Variance using SPSS 24 software.

For microarray analyses, bacterial cultures grown in LB at 37°C for 18 h were diluted 1:100 with 30 ml of pre-warmed DMEM and further incubated at 37°C with shaking (250 rpm) until OD₆₀₀ reached 0.85. Cultures were immediately mixed with five volumes of RNAlater (Thermo Fisher Scientific, Waltham, MA, United States) to stabilize RNA and stored at 4°C until further use. Total RNA was extracted using RiboPure™-Bacteria Kit (Thermo Fisher Scientific, Waltham, MA, United States) according to the manufacturer’s instructions. The quality of each RNA sample was confirmed by electrophoresis. Two independent sets of RNA samples were sent to Macrogen (Daejeon, Korea) for microarray analysis using a 35-mer Genomic Microarray for E. coli O127:H6 strain E2348/69 (CombiMatrix Prokaryotic 12 K Platform; CombiMatrix, San Francisco, CA, United States).

For macrophage transcriptome analyses, 1 × 10⁶ CFU of pre-cultured bacteria were inoculated into a pre-warmed LB broth and incubated at 37°C with vigorous shaking for 19 h. Cultured bacterial cells were pelleted, washed, and resuspended in RPMI1640 medium (Thermo Fisher Scientific, Waltham, MA, United States). 3D4/31 cells were cultured in a 100 mm × 20 mm dishes until reaching 80% confluency. They were then subcultured in 6-well cell culture plates (SPL Life Sciences, Pocheon, Korea) at density of 1 × 10⁶ cells per well and incubated at 37°C under 5% CO₂ atmosphere for 16 h. EPEC strains were then inoculated into each well with multiplicity of infection (MOI) of 50. Cells were then incubated at 37°C under 5% CO₂ atmosphere for 3 h. Culture medium containing uninfected bacterial cells in each well was aspirated. Attached macrophage cells were washed, treated with 1 ml TRIzol® (Ambion, Austin, TX, United States), and stored at −80°C until further use. Total RNA from 3D4/31 was isolated using a RNeasy mini kit (QIAGEN, Hilden, Germany) in accordance with the manufacturer’s instructions. Potential genomic DNA contaminations in total RNA samples were eliminated by treatment with DNase I (Ambion, Austin, TX, United States). RNA-sequencing (RNA-seq) library construction, sequencing, normalization, and functional analysis were conducted by Macrogen Inc (Daejeon, Korea).

All experimental and animal care procedures were approved by Kangwon National University Institutional Animal Care and Use Committees (KW-180227-1) and performed in compliance with the standard guidelines. Female C57BL/6 N mice at 8 weeks old were purchased from Koatech Inc (Pyeongtaek, Korea), acclimated for 5 days, and challenged intraperitoneally with 1.0 × 10⁸ CFU of one of the three EPEC strains: wildtype, ΔrelAΔspoT mutant, or relA- and spoT-complemented mutant strain. At 24 and 48 h post infection, blood samples were collected from retro-orbital plexus. After euthanization, peritoneal lavage was performed, and fluid samples were collected with 4 ml of PBS.

To calculate the number of EPEC survived in peritoneal cavity, samples were serially diluted, plated onto LB agar supplemented with Tc, and incubated at 37°C for 19 h. The number of viable colonies was counted and statistical significance between groups was obtained with Student’s t-test using SPSS 24 software.

Cytokine samples from 3D4/31 cells were prepared as follows. Macrophages were infected with EPEC strains as aforementioned. Culture supernatants were transferred to 1.5 mL tubes, centrifuged to remove bacterial pellets and debris, and stored at −80°C until further use. Levels of porcine interleukin 6 (IL-6) and interleukin 8 (IL-8) in cell culture supernatants were quantified with pig IL-6 and IL-8 Quantikine ELISA kit (R&D systems, Minneapolis, MN, United States) using instructions provided by the manufacturer.

Cytokine samples from murine challenge experiments were prepared as follows. Blood samples collected from retro-orbital plexus were centrifuged at 6,000 rpm for 15 min. The upper layer containing the serum was transferred to a 1.5-ml tube and stored at −80°C until further use. Concentrations of murine IL-6 and monocyte chemoattractant protein-1 (MCP-1) in the serum and peritoneal lavage fluid were determined with mouse IL-6 and MCP-1 ELISA kits (Invitrogen, Carlsbad, CA, United States) following instructions provided by the manufacturer.

At least three independent experiments were performed in this study. All quantitative data are presented as mean ± SEM. Statistical analysis was performed via Student’s t-test using SPSS 24 software.

In this study, both relA and spoT genes on the chromosome of EPEC E2348/69 strain were mutated by a sequential conjugation. Because a single spoT deletion is known to be lethal, a single ΔrelA mutation was initially introduced, followed by ΔspoT mutation to create a ΔrelAΔspoT double mutant strain (Figure 1A). Functional inactivation was confirmed by assessing growth defect on a MG agar medium as a previous report showed that relA and spoT were essential for EPEC growth on MG medium (Jin et al., 2012). As expected, ΔrelAΔspoT mutant strain was unable to grow on MG medium at 37°C (Figure 1B). Complementation of spoT with plasmid pUC/spoT restored its growth, confirming functional inactivation of relA and spoT. Since the ΔrelAΔspoT double mutant could not acquire a coccoid form (Magnusson et al., 2007), we further validated ΔrelAΔspoT mutation by observing morphological changes using TEM. As a result, ΔrelAΔspoT mutant strain showed a more elongated morphology than its wildtype strain and a spoT-complemented mutant strain (Figure 1C). Collectively, a ΔrelAΔspoT double mutant was constructed and verified by matching previously reported phenotypes.

We then performed a 2D proteomic analysis using whole bacterial proteins to identify EPEC proteins differentially expressed in the ΔrelAΔspoT double mutant grown in a rich LB medium. Our experimental analyses from two independent trials revealed that 153 out of 1,600 protein spots were consistently up- or down-regulated in the ΔrelAΔspoT double mutant than in its parental strain (Figure 2). Top 20 protein spots with significant differences in their fold changes were then subjected to matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) analysis for identification (Tables 1, 2). Notably, the expression level of BfpF, the eighth gene in the bfp cluster which encoded BFP, was significantly increased (>726-fold) in the ΔrelAΔspoT double mutant. Repression of three proteins, a universal stress protein UspA and gad operon proteins GadA and GadB, isoforms of glutamate decarboxylase (Gad), was also observed (0.6, 0.2, and 0.1, respectively). Overall, our 2D proteomic analysis showed that relA and spoT depletion affected expression patterns of various proteins including virulence factors in EPEC (Supplementary Figure S1).

As we identified that one of the bfp operon genes was induced in the ΔrelAΔspoT mutant, we speculated that other genes in the bfp operon might also be induced. To confirm this, we examined protein level of BfpA in the ΔrelAΔspoT mutant by western blotting. When grown in LB, the level of BfpA in ΔrelAΔspoT mutant was significantly increased, whereas it was not detectable in the parent strain (Figure 3A). Interestingly, the BfpA level in the spoT-complemented mutant strain was similar to that in the ΔrelAΔspoT mutant. However, it was not detectable in relA- and spoT-complemented mutant strain, coinciding with that in the wildtype strain. This observation suggests that ppGpp can negatively regulate BFP expression in EPEC and that partial restoration of ppGpp synthesis is insufficient to regulate BFP expression. We also assessed transcriptional levels of bfp operon genes by qRT-PCR. Notably, bfpA and perA genes were significantly up-regulated (approximately 10-fold) in the ΔrelAΔspoT mutant than in the wildtype (Figure 3B). This suggests that intracellular ppGpp can negatively affect the expression of both PerA (an activator of the bfp operon) and BfpA.

As seen in our 2D proteomic analysis, Gad proteins in EPEC were significantly downregulated by relA and spoT depletion. This finding was also confirmed by western blotting using anti-GadAB antibody (Figure 4A). GadAB proteins play a vital role in GDAR. They also contribute to the extremely low infectious dose of pathogenic E. coli, including EHEC O157:H7 (Lin et al., 1996). Decreased expression of GadAB could result in disruption of GDAR responses and increased susceptibility of the bacterium to acid insults. To confirm this, we performed acid resistance assay using EG medium (pH 3.0; 1.5 mM glutamate). As a result, ΔrelAΔspoT mutant was more susceptible to acid stress than the wildtype or the complemented mutant strains (Figure 4B). These results imply that ppGpp-mediated signaling is required for proper Gad expression to enhance bacterial survival in an acidic environment via GDAR.

As aforementioned, ppGpp can regulate various virulence factors in pathogenic bacteria. Since some virulence factors are readily secreted, we surmised that ppGpp might affect the overall EPEC protein secretion. To confirm this, we compared secreteomic profiles of wildtype, ΔrelAΔspoT mutant, and spoT-complemented mutant strain using SDS-PAGE. As shown in Figure 5A, general secretion abilities of the ΔrelAΔspoT mutant were significantly reduced compared to those of other strains. Interestingly, sizes of several hypo-secreted proteins in the ΔrelAΔspoT mutant corresponded to those of T3SS components such as EspA, EspB, and EspC (25, 35 and 110 kDa, respectively). T3SS encoded in the LEE pathogenicity island (LPI) plays a central role in the virulence of EPEC (Sekiya et al., 2001). We then performed qRT-PCR to further examine transcriptional patterns of LPI virulence factors, including ler (encoded in LEE1), sepD (LEE2), escC (LEE2), escV (LEE3), espA (LEE4), espB (LEE4), eae (LEE5), and tir (LEE5). Transcriptional levels of all T3SS genes were significantly decreased in the ΔrelAΔspoT mutant in comparison with those in its parental strain (Figure 5B), coinciding with secretome results. These results indicate that ppGpp can up-regulate the expression of LPI genes.

Our work demonstrated that loss of ppGpp-mediated signaling altered protein expression not normally observed when growing in a rich medium. Previous EPEC studies used DMEM to mimic a host environment. To address the effect of ΔrelAΔspoT mutation on transcriptional activities in EPEC under a host-like condition, we performed two independent microarray analyses using wildtype and ΔrelAΔspoT mutant grown in DMEM. A total of 1,564 genes were differentially transcribed in the ΔrelAΔspoT mutant compared to the wildtype, including 824 upregulated genes and 740 downregulated genes (Supplementary Table S3). Consistent with results of our experiments, bfp genes (such as bfpA, bfpC, bfpF, and bfpI) and their regulator perABC were activated in the ΔrelAΔspoT mutant, whereas genes encoding T3SS structural proteins (escC, escD, escR, escS, and escV) and GDAR genes (gadA and gadBC) were repressed (Table 3). Collectively, our findings address the importance of ppGpp-mediated signaling in the regulation of major virulence genes (e.g., LEE, BFP, and GAD) in EPEC.

To further assess the effect of ppGpp depletion on EPEC virulence, we designed an in vivo infection assay using Caenorhabditis elegans. Although there are no definitive animal models available for EPEC infection, a previous article has used C. elegans as an infection model for pathogenic bacteria (Couillault and Ewbank, 2002). Because other numerous bacterial infection studies have also used nematode as an infection host (Tan et al., 1999; Labrousse et al., 2000), we selected this organism as our host system. C. elegans were fed with E. coli OP50, a non-pathogenic strain known as a food source for nematodes, wildtype, and ΔrelAΔspoT mutant, respectively. Their survival ratios were then monitored up to 10 days. Approximately 50% of C. elegans fed with the wildtype survived on the fourth day. However, all died on the last day of observation. On the other hand, the death rate was significantly delayed in the group fed with the ΔrelAΔspoT mutant, similar to that in the group fed with OP50 (Figure 6). This result suggests that loss of ppGpp signal can significantly attenuate EPEC pathogenicity in vivo.

After constructing the ΔrelAΔspoT mutant and assessing its effect on EPEC pathogenicity, we then aimed to identify host cell genes differentially expressed upon exposure to the ΔrelAΔspoT mutant. A porcine alveolar macrophage cell line 3D4/31 was infected with wildtype or ΔrelAΔspoT mutant and subjected to RNA-seq to analyze differentially expressed genes. Genes with their fold-change more than 2.0-fold and their normalized FPKM values over 30 were considered differentially expressed. The results revealed that 82 genes were differentially expressed in 3D4/31 infected with ΔrelAΔspoT mutant as compared to those in 3D4/31 infected with the wildtype. Among these genes, 77 genes were up-regulated, and five genes were down-regulated. Differentially expressed genes were categorized according to their functions. Notably, transcription factors and cytokine related genes accounted for a large proportion of up-regulated genes (Supplementary Table S4). To validate our transcriptome results, 11 genes were randomly selected and subjected to qRT-PCR. Expression patterns of 10 genes matched RNA-seq results (value of p <0.05; Supplementary Table S5), indicating that transcriptomic analyses derived from this work were highly valid. The datasets presented in this study can be found in the NCBI Sequence Read Archive¹ under accession number PRJNA917101. Raw reads of the RNA-seq from two samples were deposited in the NCBI SRA² under accession numbers SRR23329177 and SRR23329178.

RNA-seq analyses revealed increased expression levels of five proinflammatory cytokine genes (IL-6, IL-8, GM-CSF, MCP-1, and MIP2-Α) in 3D4/31 exposed to ΔrelAΔspoT mutant as compared to those in 3D4/31 exposed to the wildtype (Supplementary Table S6). Notably, IL-6 and IL-8 showed the highest fold increases (2.8- and 5.4-fold, respectively). Their expression levels were further assessed using qRT-PCR and ELISA (Table 4). Results showed that IL-6 and IL-8 expression at both transcription and protein levels were significantly increased in 3D4/31 cells infected with the ΔrelAΔspoT mutant when compared with those in 3D4/31 cells infected with the wild-type strain or the relA- and spoT-complemented mutant strain (Table 4).

We speculated that infection with the ΔrelAΔspoT mutant could also induce cytokine expression in vivo. To confirm this, we analyzed cytokine expression levels after EPEC infection using a murine peritoneal infection model. A single dose of 1 × 10⁸ CFU of wildtype, ΔrelAΔspoT mutant, or relA- and spoT-complemented mutant was injected intraperitoneally into each C57BL/6 N mouse. Blood and peritoneal lavage fluid samples were then taken at 24- and 48-h post infection and subjected to ELISA analysis using either IL-6 or MCP-1 antibody. At 24 h post infection, mice infected with the ΔrelAΔspoT mutant showed increased levels of IL-6 in both serum (Figure 7A) and lavage fluid (Figure 7B) samples in comparison with those infected with wildtype or complemented strain. At 48 h post infection, all groups had low levels of IL-6. The same expression patterns were observed for MCP-1 known to be controlled by IL-6 (Romano et al., 1997; Biswas et al., 1998) (Figures 7C,D). Collectively, these results suggest that infection with ΔrelAΔspoT EPEC can induce IL-6 expression both in vitro and in vivo.

Cytokines play a vital role in early innate immune response. IL-6 and MCP-1 are known to contribute to bacterial clearance by inducing chemotaxis of neutrophils and macrophages (Itakura et al., 2017). We thus speculated that the number of ΔrelAΔspoT mutant survived in peritoneal lavage fluid would be significantly lower than that of wildtype or complemented strain in vivo due to increased levels of cytokines. To confirm this, we compared bacterial CFU in peritoneal lavage fluid samples by plating diluents on LB agar plates. At the 24 h timepoint, all EPEC strains showed similar numbers of survived cells. At the 48 h timepoint, however, the number of ΔrelAΔspoT mutant cells (3.2 ± 0.6 × 10³) was significantly reduced as compared to that of the wildtype (2.1 ± 0.7 × 10⁴) or the complemented strain (9.7 ± 2.1 × 10³) (Figure 8). This clearly indicates that bacterial clearance is favored in a host system that is exposed to ΔrelAΔspoT EPEC.