sci1 expression is heterogeneous within clonal population of EAEC

Expression of the EAEC sci1 T6SS gene cluster is repressed by Fur under iron-rich conditions, such as in LB medium, and induced under iron depletion, such as in M9/glycerol minimal medium (sci1-inducing medium, SIM [7]). Using a translational reporter strain expressing tssB-sfGFP to visualize T6SS sheath dynamics, we observed heterogeneous fluorescence in iron-limiting SIM medium (Fig 1A). Three distinct subpopulations were observed: cells that do not produce TssB-sfGFP (no fluorescence), cells that produce TssB-sfGFP (diffuse fluorescence) and cells that produce TssB-sfGFP and assemble sheaths (Figs 1A and S1). These observations suggest that heterogeneity occurs at both the transcriptional/translational level (expression/production versus no expression/production) and the post-translational level (assembly versus no assembly).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Phenotypic heterogeneity of T6SS expression and assembly reaches a stable equilibrium over generations.

(a) Analyses of TssB-GFP cells. Representative confocal fields (top, merged bright and fluorescence channels; scale bar, 5 μm) and distribution of T6SS sheath (bottom, max/mean fluorescence intensity ratio) in TssB-GFP cells population. The y-axis represents the probability distribution of single-cell fluorescence intensities. (b) Analyses of TssC-GFP-TssK cells. Representative confocal fields (top, merged bright and fluorescence channels; scale bar, 5 μm) and distribution of fluorescence (bottom, mean fluorescence intensity ratio) in TssC-GFP-TssK cells population. The y-axis represents the probability distribution of single-cell fluorescence intensities. (c) Coefficient of variation (CV) of TssC-GFP-TssK populations and T6SS⁺ cell frequency (f(T6SS+)) in TssB-GFP populations as a function of generation number. T6SS⁺ cell frequency was calculated as the proportion of cells with a max/mean fluorescence ratio > 2. Approximately 2,000 cells were analyzed per generation from ≥3 independent replicates. The data represent means (circles) and standard deviations (SD, error bars) from ~2,000 cells per generation from ≥3 independent replicates. (d) Representative confocal fields of the 3 types of TssC-GFP-TssK (upper panels) or of TssB-GFP (lower panels) microcolonies (homogeneous ON, heterogeneous, homogeneous OFF). Scale bars, 10 μm. (e) Distribution of CV in TssC-GFP-TssK (left) and f(T6SS⁺) in TssB-GFP (right) across microcolonies from single-cell experiments. The dotted lines indicate the mean values. TssC-GFP-TssK: 8,500 cells analyzed (55 microcolonies from 4 independent replicates); TssB-GFP: 5,900 cells analyzed (43 microcolonies from 3 independent replicates). (f) Dynamics of f(T6SS⁺⁾ in sorted subpopulations over time. The data represent means (circles) ± standard deviations (SD, error bars) from at least three biological replicates. Statistical significance (one-tailed Wilcoxon’s t test; ns, nonsignificant; *p < 0.1; ***p < 0.001) is indicated. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003838.g001

To gain further information on transcriptional heterogeneity, we engineered a reporter strain in which sfGFP was inserted between the tssC and tssK genes (TssC-GFP-TssK). This construct revealed a bimodal distribution of fluorescence (Fig 1B), with approximately 40% of cells in a low-fluorescence OFF state and 60% in a high-fluorescence ON state (Figs 1B and S1). The partial overlap between peaks (Fig 1B) suggests the presence of intermediate or transient activation states. Flow cytometric sorting of ON and OFF subpopulations confirmed the coexistence of these states (S4A Fig, see below).

The ON/OFF equilibrium of sci1 T6SS expression is stable across generations

To quantify heterogeneity, we calculated the coefficient of variation (CV) of single-cell fluorescence, i.e., the ratio between the standard deviation of the fluorescence represented by values from each cell with the mean fluorescence of the population (CV = SD/µ), over generations. With this parameter, higher CV values indicate greater heterogeneity. Both the CV and the proportion of T6SS⁺ cells remained stable over multiple generations in SIM, indicating that the ON/OFF equilibrium is maintained (Figs 1C and S2).

Heterogeneous expression is stable and heritable across generations

To investigate how heterogeneity arises and propagates, we performed single-cell time-lapse microscopy using the TssB-GFP and TssC-GFP-TssK reporter strains. Cells grown to equilibrium were plated on a microfluidic chip at single-cell density. Individual founder cells gave rise to predominantly OFF, ON, or heterogeneous microcolonies after 7–10 generations (Figs 1D and S3). ON-enriched and heterogeneous colonies were more frequent than OFF-enriched colonies, resulting in a population distribution similar to that observed in liquid culture (CV > 0.20; Fig 1E). The overall frequency of T6SS⁺ cells was reduced in microfluidics, likely reflecting slower growth associated with lower oxygen availability compared with liquid cultures (Fig 1E). The presence of mostly homogeneous colonies indicates that the ON or OFF state can be transmitted from mother to progeny, with ON states appearing more frequently. It is, however, noteworthy that following transcriptional repression, visualization of the OFF state by fluorescence microscopy depends on the half-life of sfGFP.

To further assess the stability of the T6SS ON and OFF states, subpopulations were isolated by flow cytometry. The ratio of OFF and ON cells was comparable to that observed by fluorescence microscopy (S4A and S4B Fig). Sorted cells largely retained their activation state in the short term: ON cells predominantly formed ON colonies, and OFF cells predominantly formed OFF colonies (Figs 1F, S4C and S4D). However, both subpopulations reverted to the heterogeneous ON/OFF equilibrium within 15–20 generations in SIM (Fig 1F).

Fur binding and GATC methylation govern heterogeneous Psci1 expression

The ability of ON and OFF subpopulations to revert to equilibrium suggested that sci1 heterogeneity is genetically controlled. The sci1 promoter (Psci1) is comprised of multiple regulatory elements, including two Fur boxes (proximal F1 and distal F2) and three GATC sites, targeted by the Dam methyltransferase (proximal G1 to distal G3) (Fig 2A) [7]. G1 lies within F1, overlapping the −10 element, G2 is located between the −10 and −35 elements, while F2 and G3 are upstream of the −35 (Fig 2A). Fur binding onto F1 and F2 represses sci1 expression and prevents G1 methylation, whereas G1 methylation antagonizes Fur binding, thereby promoting transcription [7].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Fur and Dam binding sites F2 and G3 control heterogeneous T6SS expression.

(a) Schematic representation of the sci1 T6SS promoter (Psci1). Psci1 contains two Fur binding boxes (F1 and F2) and three Dam methylation GATC sites (G1, G2, and G3). The F1 box includes the G1 site and overlaps with the transcriptional −10 element [7]. (b) TssB-GFP f(T6SS⁺) (bar plots) and TssC-GFP-TssK CV (line points) for WT and promoter variants. The data represent means (vertical bars) ± SD (error bars) from at least three biological replicates. Statistical significance compared to the WT strain (one-tailed Wilcoxon’s t test; ns, nonsignificant; *p < 0.1; **p < 0.01; ***p < 0.001) is indicated. (c) Competition assay between EAEC WT or promoter variants (attackers) and sensitive E. coli W3110 (recipient) in SIM (top) or LB (bottom) medium. Killing activity was visualized by the LAGA method [40], where color intensity reflects attacker activity (from yellow (no killing) to purple (intense killing), see heatmap on right). Images are representative results of biological triplicates. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003838.g002

To dissect the contributions of these regulatory elements to heterogeneity, we engineered chromosomal point mutations disrupting individual Fur boxes or GATC sites (S6A Fig). Preventing adenine methylation at GATC sites reduced Psci1 activity, as shown by decreased T6SS sheath formation in G1, G3, and triple G123 mutants (Fig 2B). Conversely, disrupting the F2 Fur box relieved repression as the F2 mutant exhibited increased T6SS sheath numbers (Figs 2B and S5). When assessed in the TssC-GFP-TssK reporter strain, Fur box mutants displayed an increased ON state, whereas G3 and G123 mutants were homogeneously OFF (Figs 2B, S6B and S7). G1 and G2 mutants remained heterogeneous but with ON/OFF ratios distinct from wild-type (Figs 2B, S6 and S7). We thus concluded that the promoter architecture, via competing Fur binding and Dam methylation, is essential for establishing heterogeneous sci1 expression.

We next assessed whether these mutations affected T6SS-mediated killing of E. coli W3110 target cells using the LAGA assay [40]. Figure 2C shows that the T6SS activity roughly correlated with the proportion of T6SS⁺ cells: G2 mutant cells exhibited reduced killing, while G1 and G3 cells were significantly impaired in killing and G123 mutant cells were inactive under inducing conditions (SIM, Fig 2C). In contrast, the F2 mutant showed increased T6SS activity compared to the WT and killed target cells even under iron-rich, repressive conditions (LB, Fig 2C). These findings support the model that ON cells deploy T6SS for antagonism, whereas OFF cells remain inactive.

T6SS heterogeneity acts as an attenuation strategy against retaliatory T6SS attacks

Because the F2 and G123 mutants mimic homogeneous ON (T6SS⁺) and OFF (T6SS⁻) subpopulations, respectively, we used these strains to explore the functional significance of heterogeneity, particularly the role of OFF cells. One long-standing hypothesis is that T6SS imposes a substantial energetic burden, favoring heterogeneous expression as a cost-sharing strategy. However, consistent with previous findings [36], neither overexpression (F2) nor complete inhibition (G123) affected growth in liquid culture (S8A Fig) even in the presence of bile salts to mimic the EAEC ecological environment, indicating that T6SS assembly and activity are not energetically costly in EAEC and that the OFF subpopulation does not serve to counteract the growth defects associated to the production, assembly or activity of the T6SS.

As the T6SS was recently shown to serve as a foraging strategy to acquire nutrients by lysing neighboring cells [41–43], we next considered a suicide model in which OFF cells act as a nutritional reservoir for ON cells. In this case, OFF cells will not produce the Tli1 immunity protein and therefore can be killed by ON cells to support their growth. To test this hypothesis, we conducted competition experiments using the SGK assay [40]. The SGK assay is a growth recovery method that measures the time necessary for surviving bacteria to reach mid-exponential phase, which is inversely proportional to the initial number of surviving cells. We did not detect kin killing in EAEC: contrarily to Δsci1 mutant cells (i.e., deleted of the whole T6SS gene cluster) that are susceptible to WT attacks, G123 cells were not (S8B Fig). In addition, competition with the ON-homogeneous F2 strain did not increase mortality of WT or of G123 OFF cells (S8B Fig). This result suggests that basal expression of the immunity gene tli1 from the internal promoter P₄₅₃₂ [44] is sufficient to confer protection against kin ON cells.

A third possibility is that T6SS heterogeneity is an attenuation strategy into the host. During infection, the immune system recognizes virulence factors to neutralize pathogens. An attenuation strategy, such as LPS modification in Burkholderia pseudomallei [45] or antigenic variation in Borrelia burgdorferi [46], allows to escape the immune system. However, we observed that homogeneous ON populations exhibited the highest infective success in Galleria mellonella, whereas OFF cells alone failed to efficiently colonize this wax moth host model (S8C Fig), arguing against this hypothesis.

Finally, we hypothesized that heterogeneity mitigates antagonism in polymicrobial environments where defensive T6SS⁺ species deploy tit-for-tat retaliation [47]. Heterogeneous expression could thus serve as an attenuation strategy to reduce defence intensity and to improve survival. Consistent with this hypothesis, SGK assays showed that OFF cells (G123) alone failed to kill E. coli W3110, whereas increasing the ON/OFF ratio using synthetic mixes between the F2 and G123 populations proportionally enhanced antibacterial activity (Fig 3A). However, when competing with the model retaliatory T6SS-positive competitor Pseudomonas aeruginosa PAO1ΔretS (i.e., a strain that constitutively expresses the H1-T6SS but only assembles it and fires upon attacks from competitors [48]), increasing the ON/OFF ratio led to higher EAEC mortality (Fig 3B). These findings indicate that ON cells provide antagonistic capacity but simultaneously increase susceptibility to retaliatory T6SS attacks, whereas heterogeneity dampens this cost. Thus, T6SS heterogeneity functions as an attenuation strategy, balancing offensive capacity with survival.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Phenotypic heterogeneity modulates T6SS strength and mediates a trade-off between killing efficiency and survival against T6SS⁺ competitors.

(a) Competition assays between WT EAEC or mixed populations with increasing ON/OFF ratios (0%–100% ON) and E. coli W3110 recipients. F2 and G123 mutants were used as ON and OFF populations, respectively, and combined at different ratios to generate distinct heterogeneity states. Recipient mortality was quantified by the SGK method [40]. The data represent the three independent values (gray circles) and means (horizontal bars) from three biological replicates. Statistical significance between strains (Student t test; ns, nonsignificant; *p < 0.1; **p < 0.01) is indicated. (b) Competition assays between WT EAEC or ON/OFF mixtures and the T6SS⁺ defensive bacterium Pseudomonas aeruginosa PAO1ΔretS. EAEC mortality was quantified by the SGK method [40]. Data are means of 3 replicates. The data represent the three independent values (gray circles) and means (horizontal bars) from three biological replicates. Statistical significance between strains (Student t test; ns, nonsignificant; *p < 0.1) is indicated. (c) Three-species competition assays between E. coli K-12 W3110 (gray), EAEC strain 17-2 (WT or Δsci1, blue), and P. aeruginosa PAO1ΔretS or PAO1ΔretSΔH1-T6SS (orange) at a 1:1:1 ratio. Mortality of each bacterial species was quantified by the SGK method [40]. (d) Three-species competition assays between E. coli K-12 W3110 (gray), EAEC strain 17-2 (WT or increasing ON/OFF ratios (0%–100%), blue), and P. aeruginosa PAO1ΔretS or PAO1ΔretSΔH1-T6SS (orange) at a 1:1:1 ratio. Mortality of each bacterial species was quantified by the SGK method [40]. The data represent means (vertical bars) ± SD (error bars) from nine replicates (technical triplicates of three biological replicates). Statistical significance compared to the WT strain (one-tailed Wilcoxon’s t test; ns, nonsignificant; *p < 0.1; **p < 0.01; ***p < 0.001) is indicated with the strain color code. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003838.g003

To further evaluate the ecological relevance of EAEC T6SS sci1 heterogeneity, we performed three-species competition assays with EAEC, the T6SS-negative and susceptible E. coli W3110 strain, and P. aeruginosa PAO1ΔretS mixed in a 1:1:1 ratio. Figure 3C shows that in this three-species competition assay, EAEC and P. aeruginosa engaged in reciprocal T6SS-mediated killing, leading to high mortality in both populations. Under these conditions, the strong reduction in EAEC abundance limited its ability to eliminate the susceptible W3110 strain. In contrast, when the H1-T6SS of P. aeruginosa was inactivated, EAEC efficiently eliminated both competitors while experiencing reduced mortality. Inactivation of both the EAEC Sci1 T6SS and P. aeruginosa H1-T6SS led to a stable equilibrium of the three strains during co-incubation.

To quantitatively assess the impact of phenotypic heterogeneity, we varied the proportion of T6SS-active (ON) cells within the EAEC population (Fig 3D). Increasing the fraction of ON cells enhanced killing of P. aeruginosa, but also resulted in an increase in EAEC mortality due to retaliatory attacks (Fig 3D). As a consequence, W3110 survival increased as EAEC populations became depleted through reciprocal killing.

Bacterial strains and culture conditions

Bacteria and plasmids used in this study are listed in S1 Table. E. coli strains were routinely grown in LB medium (iron-rich conditions) or in sci1 inducing minimal medium (SIM, M9 minimal medium supplemented with glycerol 0.2%, vitamin B1 200 μg/mL, casaminoacids 40 μg/mL, MgSO₄ 1 mM and CaCl₂ 0.1 mM; iron-starved conditions). When required, media were supplemented with antibiotics (chloramphenicol 40 µg/mL, kanamycin 50 µg/mL, ampicillin 100 µg/mL, gentamicin 15–20 µg/mL) or isopropyl-β-D-galactopyranoside (IPTG, 0.1 mM). Cultures were grown at 37 °C with shaking (180 rpm). Chloro-Phenol-Red-Galactopyranoside (CPRG, 2 mM) was prepared in distilled water.

Plasmid construction

Plasmid and oligos used in this study are listed in S2 and S3 Tables, respectively. PCRs were performed using a Biometra thermocycler and Q5 polymerase (New England Biolabs). The pKO3-Psci1 vector was generated by restriction-ligation. A 1-kb DNA fragment containing the Psci1 promoter (from −500 pb to + 500 bp relative to the −10 element) was amplified from EAEC 17−2 genomic DNA with oligonucleotides carrying BamHI and SalI restriction sites at 5′ and 3′, respectively. The purified fragment and pKO3 plasmid were digested by BamHI and SalI (New England Biolabs) for 3 h at 37 °C and ligated with T4 DNA ligase (New England Biolabs) for 1 h at room temperature. Promoter point mutations were introduced by site-directed mutagenesis. Complementary oligonucleotide pairs bearing the desired substitution were used to amplify the entire pKO3-Psci1 plasmid. PCR products were then digested with DpnI to remove template plasmids and transformed into DH5α competent cells. All pKO3-Psci1 variants were verified by PCR and DNA sequencing (Eurofins Genomics).

Strain construction

The fluorescence transcriptional reporter TssC-GFP-TssK was constructed by λ-Red recombination [65] using plasmid pKOBEG [66]. Electrocompetent cells harboring pKOBEG were transformed with a PCR product amplified from pKD4-Nt-sfGFP, which contained the sfGFP gene and a kanamycin cassette flanked by 50-bp extensions homologous to the upstream and downstream regions of the insertion site. Transformants were selected on kanamycin plates and verified by colony PCR before and sequencing of the site of insertion. The kanamycin cassette was then excised using pCP20 [65]. Directed mutagenesis was performed by allelic exchange using the pKO3 vector, which carries the sacB gene whose expression is toxic in the presence of sucrose [67]. Cells were transformed with pKO3-Psci1 variants and incubated at 30 °C for plasmid replication. Clones were restreaked at 37 °C to select for chromosomal integration. Second crossover events were obtained by plating on LB agar plates supplemented with 5% sucrose at 37 °C. Colonies were restreaked on LB agar plates and randomly sequenced (Eurofins Genomics) to confirm the desired mutation.

Competition assay

Intra-species competition assays were performed with EAEC 17-2 (or derivatives) as attacker strains and E. coli W3110 Kanᴿ or EAEC 17-2 (or derivatives) bearing a kanamycin-resistance cassette inserted into the lacZ gene [7] as recipients. Overnight cultures were diluted into fresh SIM medium and grown to an optical density at λ = 600 nm (OD₆₀₀) of 0.8. For the lysis-associated β-galactosidase assay (LAGA), attacker and recipient strains were mixed at a 4:1 ratio; for the survivor growth kinetics assay (SGK), the ratio was 1:4 [40]. Mixtures (10 µL) were spotted on SIM agar and incubated for 4 h at 37 °C. In the qualitative LAGA assay, β-galactosidase released from lysed W3110 cells cleaves the membrane-impermeant β-galactosidase chromogenic substrate chlorophenol-red β-D-galactopyranoside (CPRG, yellow) into chlorophenol red (CPR, purple). 10 μL of CPRG 2 mM was then added on top of the spots. The color intensity correlates with killing efficiency. For quantitative SGK assays, spots were resuspended in LB supplemented with 50 μg/mL of kanamycin, serially 10-fold diluted in LB supplemented with kanamycin and inoculated in 96-well microplates. Recipient survivor growth was monitored using a TECAN microplate reader. Growth parameters were extracted with the R package growthcurver [68]. Because the time to resume growth depends on the initial cell concentration (i.e., the time required for the recipient population to resume exponential growth is inversely correlated with the initial number of surviving cells), the mid-exponential time point (Tmid) was used as a proxy for recipient mortality. For inter-species competition assays, Pseudomonas aeruginosa PAO1ΔretS was used as the attacker and EAEC wild-type strain 17-2 and its derivatives carrying the pBBR5-Gm^R vector as recipients. PAO1 was grown overnight in LB, diluted 25-fold, and cultured to OD₆₀₀ = 2. One mL was pelleted and resuspended to OD₆₀₀ = 10. EAEC strains were grown overnight in LB supplemented with 15 µg/mL gentamicin, diluted into SIM, and harvested at OD₆₀₀ = 0.8. Three-species competition assays were performed with EAEC 17-2 (or its Δsci1 derivative) with a kanamycin cassette inserted in the lacZ gene, PAO1 ΔretS (or its ΔH1-T6SS derivative) carrying the pBBR-Gm^R vector, and E. coli W3110 carrying the Amp^R pmCherry vector. Cells were grown as described above, mixed in a 1:1:1 ratio, and 10-μL mixtures were spotted on SIM agar plates. After 4h of incubation at 37 °C, spots were resuspended in LB and serially diluted in LB supplemented with 50 μg/mL of kanamycin, 20 μg/mL of gentamycin or 100 μg/mL of ampicillin for measuring the survival of EAEC, PAO1, and W3110, respectively. Mixtures containing defined F2(ON)/G123(OFF) ratios (0%–100%) were prepared and mixed with competitor cells, spotted on SIM agar, and incubated for 4 h at 37 °C and processed as above for quantitative measurements.

Fluorescence microscopy

Cells were grown as described, concentrated 10-fold, and 2 µL were spotted onto a 2% agarose pad mounted in a Gene Frame. Slides were sealed with a coverslip and imaged on a Nikon Eclipse Ti2 microscope equipped with a 100 × NA 1.45 Ph3 objective, an ORCA-Fusiondigital camera (Hamamatsu) and a Perfect Focus System. sfGFP was excited at 488 nm. Images were analyzed with MicrobeJ after segmentation with MiSiC [69] to extract single-cell fluorescence. Distributions were plotted using ggplot2 and geom_density within the R package [70]. The coefficient of variation (CV = SD/mean) was used as a heterogeneity metric for TssC-GFP-TssK cell fluorescence data. Because the T6SS sheath gives the brighter pixel in a cell compared to a diffuse fluorescence, a ratio of maximum/mean pixel fluorescence > 2 was considered positive for sheath detection for strain TssB-GFP.

Flow cytometry and cell sorting

Flow cytometry and fluorescence-activated cell sorting (FACS) were performed on an Influx v7 Cell Sorter (Becton) equipped with a stream-in-air 70 µm nozzle and a blue 488-nm Sapphire OPS laser (400 mW). The 488-nm laser was used for the analysis of the FSC (488/10 nm band pass filter, PMT1), which provides information related on cell size, and of the SSC (trigger signal, 488/10 nm band pass filter, PMT2), which provides information on cell density and the GFP fluorescence (530/40 nm band pass filter, PMT3). Light was detected with PMTs (Hamamatsu R3896 PMTs in C6270 sockets). FACSFlow buffer (Becton, Dickinson and Company) ran at 33 psi with an event rate of ~3,000 events s⁻¹. The instrument was calibrated daily with 1-μm blue and 2-μm yellow-green FluoSphere beads (Thermo Fisher Scientific) in the linear range. For calibration in the logarithmic range and for spiking into the sample, 0.5-μm and 1-µm yellow-green beads (Thermo Fisher Scientific) were used. Acquisition was stopped after 50,000 events were recorded in the FSC against the SSC cell gate. FACS was performed as described [71] with the most accurate sorting mode ’1.0 drop pure’ and at a maximum sorting speed of 3,000 cells s⁻¹. A total of 2 × 10⁶ cells per gate were collected for microscopy, mixed with glycerol (15% final concentration), and stored at −80 °C. 2D-plots were created using FlowJo V10 (Becton, Dickinson and Company). In addition, proportions of cells were analyzed in 2D-plots FSC versus green fluorescence in the GFP cell gate equal to 100% with respect to the 3 gates assembler, producer (ON) and nonproducer (OFF) for the 4 strains EAEC 17-2, TssB-GFP, TssC-GFP-TssK, and TssC-GFP-TssK F2, respectively. For the calculation, the beads in the GFP cell gate were subtracted.

Single-cell experiments

Single-cell assays were performed in an Ibidi sticky-Slide VI 0.4. A 2% BD agar layer was poured into the channel, then cut to create a flow channel. The cell suspension was spotted on the agar, overlaid with a coverslip, and the input well was filled with 80 µL of culture medium. The device was incubated at 37 °C in a custom thermostable chamber on the microscope. Microcolonies growth was imaged after 4 h using the settings described above.

Wax moth infection assay

Galleria mellonella larvae were used as an infection model. Larvae were starved for 48 h at 28 °C in the dark. EAEC 17-2 WT, F2 (ON) and G123 (OFF) strains bearing the kanamycin cassette inserted into the lacZ gene were grown to OD₆₀₀ = 0.8. Starved larvae were infected by ingestion of 2 × 10⁵ CFU/mL and incubated for 24 h at 28 °C in the dark. After 24 h, larvae were washed for 10 s in sterile water, sacrificed, and their intern organs were collected and homogenized in 500 µL of PBS with 6 glass beads using a FastPrep-24 5G (MP biomedicals) for 30 s at power 4.5. Homogenates were serially diluted and plated on LB agar plate supplemented with 50 µg/mL of kanamycin for EAEC recovery. The assay was performed in triplicate.

Statistical analysis

All statistical analyses were performed using R. Pairwise comparisons were conducted using unpaired two-tailed Student’s t-tests unless otherwise specified. For nonparametric datasets, statistical significance was assessed using one-tailed Wilcoxon tests. When comparing multiple groups, one-way ANOVA followed by Tukey’s post-hoc test was applied. All experiments were performed with at least three independent biological replicates.