An acidic environment is essential for LF82 intracellular replication within macrophage phagolysosomes (Bringer et al., 2006). To monitor pH fluctuations in macrophages, we employed LysoSensor, which emits blue and yellow fluorescence (yellow fluorescence indicates pH<7, stronger yellow fluorescence reflects a lower pH level) to compare pH variations between uninfected and LF82-infected macrophages. The result showed that infected macrophages exhibited a yellow hue in comparison to uninfected cells ( Figure 1A ), indicating a rapid decline in pH levels upon LF82 infection. This finding is consistent with previous reports suggesting that intracellular acidification results from LF82 replication within macrophages(Bringer et al., 2012).

NH₄Cl and chloroquine diphosphate (CQ) have been demonstrated to effectively neutralize the acidic pH in macrophages (Bringer et al., 2006; Bringer et al., 2012). A replication assay demonstrated that neutralization of acidic environment in macrophages with 30 mM NH₄Cl or 10 μM CQ resulted in a reduction of LF82 replication by 3.3-fold and 4.2-fold, respectively ( Figure 1B ). Meanwhile, confocal scanning revealed a significant decrease in bacterial load within macrophages treated with NH₄Cl and CQ at 24 h post-infection (p.i.) compared to the untreated control group ( Figure 1C ). An in vitro growth curve demonstrated that 30 mM NH₄Cl or 10 µM CQ had no impact on LF82 proliferation in LB or RPMI 1640 medium, indicating that the reduction of LF82 replication in macrophages treated with either NH₄Cl or CQ was not due to the growth defect of LF82 in LB and 1640 medium ( Figures 1D, E ). These results indicate that the acidic microenvironment within macrophages facilitates LF82 replication.

To investigate the mechanism underlying LF82 replication in an acid environment, we simulated the pH of LF82 in acidic macrophages (Bringer et al., 2005; Bringer et al., 2007), and the transcriptome of LF82 cultured in a pH 5.8 acid environment for 30 min was analyzed. After filtering low-quality reads, a total of 24,815,427 and 18,614,200 reads were obtained from the acid-treated and control group, respectively. Approximately 97.22% of the total reads for the acid treated group and 98.31% of those for the control group were uniquely mapped to the reference genome. Subsequently, gene expression profiles were compared. The results showed that a total of 1,997 genes were differentially expressed in the acid-treated group compared to those in the control group. Of these, 995 and 1002 genes were categorized as upregulated and downregulated, respectively (fold change > 2 and P value < 0.05), show as Table S1 and the volcano plot of DEGs reported in Figure 2A . To verify the transcriptome results, 10 differentially expressed genes (DEGs) were randomly selected for qRT-PCR analysis. The results showed that the expression patterns of these DEGs were consistent with those of RNA-Seq data, indicating the robustness and validity of our RNA-Seq approach ( Figure 2B ).

Subsequently, DEGs in the acid treated and control group were classified using the NCBI COG functional categories annotation system. The COG categories significantly enriched in the group of upregulated genes were primarily associated with post-translational modification, protein turnover, chaperones, mobilome, translation, ribosomal structure, and biogenesis. The COG categories significantly enriched in the list of downregulated genes included carbohydrate transport and metabolism, amino acid transport and metabolism, energy production, and conversion ( Figure 2C ). These results indicate that the impact of acid conditions on LF82 gene expression is bidirectional, which provides a basis for us to study the replication mechanism of LF82 in acid macrophages from the transcriptional level.

Based on RNA-Seq results, the expression of asr (acid shock gene) was significantly upregulated, by 622-fold, in the acid-treated group compared to the control group ( Figures S2A , 3A ). RT-qPCR results further verified that asr was 10.63-fold upregulated in the acid-treated group compared to the control group ( Figure 3B ). Meanwhile, the expression of asr was significantly upregulated, by 128.69-fold, in intra-macrophage LF82 at 1 h p.i. and maintained a high level of expression at 6 h (14.85 fold) and 24 h (3.63 fold) p.i., respectively ( Figure 3C ). We further generated asr mutant (Δasr) to evaluate whether asr affects the intracellular replication of LF82. The results showed that Δasr led to a 3.85-fold decrease in LF82 replication ability within macrophages, these differences could be restored to wild-type levels when complementary plasmids pSWK129-asr was introduced into the Δasr mutant ( Table S3 ; Figure 3H ), indicating that asr promotes the replication of LF82 in macrophages. However, there was no difference in the replication abilities between WT and Δasr in macrophages which neutralized the acid condition with NH₄Cl or CQ ( Table S3 , Figures 3I, J ). These results indicated that asr plays an important role in survival and replication of LF82 in macrophages under acidic conditions.

In addition, the volcano map and heat map results showed that the expression of several acid fitness genes was also upregulated in the LF82 acid-treated group compared to the control group ( Figures S2A , 3A ). Further qRT-PCR results confirmed the upregulation of gadA, hdeA, and hdeB ( Figures 3D, F ). To investigate the expression of gadA in LF82-infected macrophages at 1 h, 6 h, and 24 h p.i., qRT-PCR was performed. The results showed that gadA was upregulated by 4.77-fold, 6.03-fold, and 1.93-fold in LF82 infected macrophages at 1 h, 6 h, and 24 h p.i., respectively ( Figure 3E ). Additionally, the expression of hdeA in LF82-infected macrophages was upregulated by 7.02-fold, 3.96-fold, and 2.53-fold at 1 h, 6 h, and 24 h p.i., respectively ( Figure 3G ). Furthermore, hdeB exhibited a significant increase in expression levels, 9.20-fold at 1 h p.i., followed by increases of 2.66-fold and 2.05-fold at 6 h and 24 h p.i. ( Figure 3G ). These results suggest that gadA, hdeA, and hdeB play an important role after LF82 infects macrophages. To identify the effect of gadA, hdeA, and hdeB in LF82 intracellular replication, we generated ΔgadA, ΔhdeA, and ΔhdeB. As shown in Table S3 and Figure 3H , the replication ability of ΔgadA in macrophages was reduced by 3.57-fold compared to that of WT, complementary strains of ΔgadA+pgadA could recover the replication ability of WT, indicating that gadA is required for the replication of LF82 in macrophages. When the acidic pH of macrophages was neutralized by NH₄Cl and CQ, the replication ability of WT in macrophages was significantly reduced, while there was no significant difference in the replication ability between WT and ΔgadA in untreated and acid-neutralized macrophages ( Figures 3I, J ). In addition, ΔhdeA and ΔhdeB shared silimiar trends with ΔgadA ( Table S3 ; Figures 3H–J ). These findings suggest that acid regulatory genes play crucial roles in the survival and replication of LF82 within macrophages.

In addition, the results of the growth curve showed that there was no difference between gene deletion mutants (Δasr, ΔgadA, ΔhdeA and ΔhdeB) and WT in control 1640 medium. But there was significant difference between gene deletion mutants and WT when growth in 1640 medium (pH=5.8) ( Figures S3A, S3B ). These results further indicated that asr, gadA, hdeA and hdeB plays an important role in survival and replication of LF82 in macrophages under acidic conditions.

Transcriptome analysis showed that several nitrate-metabolism-related genes were upregulated, including narGHIJ gene cluster, narK, nitrate sensing two-component system narX/narL, and hmp, as shown in the volcano plot and heat map ( Figures S2A , 4A ). The upregulation of genes involved in nitrogen metabolism, namely, narK, narL, narX, narP, and hmp, was confirmed by qRT-PCR analysis following acid treatment, with fold changes of 2.13-, 3.67-, 3.77-, 2.59- and 3.39-fold, respectively ( Figure 4B ). As replication within host macrophages of LF82 is essential for chronic inflammation, the impact of nitrate utilization on LF82 replication in macrophages was assessed. The addition of 0.3 mM Na₂NO₃ to the 1640 medium resulted in a significant increase (3.25-fold) in LF82 replication within macrophages ( Figure 4C ). Meanwhile, confocal scanning results demonstrated that Na₂NO₃ treatment significantly increases bacterial quantity in macrophages compared to the untreated group at 24 h p.i. ( Figure 4D ), indicating a potential role of nitrate metabolism in promoting LF82 replication within macrophages.

Furthermore, we generated mutants lacking narK (ΔnarK), narXL (ΔnarXL), narP (ΔnarP), and hmp (Δhmp), which are genes related to nitrate metabolism. ΔnarK, ΔnarXL, ΔnarP, and Δhmp resulted in a significant reduction in LF82 replication within macrophages ( Table S4 ; Figures 4E ). In addition, the gene deletion strains ΔnarK, ΔnarXL, ΔnarP and Δhmp lost the ability to enhance nitrate metabolism and then enhance replication ability in macrophages ( Table S4 ; Figures 4F ). These differences could be restored to wild-type levels by ΔnarK+pnarK, ΔnarXL+pnarXL, ΔnarP+pnarP, and Δhmp+phmp complementary strains ( Table S4 ; Figures 4E, F ), indicating that narK, narXL, narP and hmp affected the replication of LF82 in macrophages by influencing the metabolism of nitrate.

In addition, the results of the growth curve showed that there was no difference between gene deletion mutants (ΔnarK, ΔnarXL, ΔnarP and Δhmp) and WT in control 1640 medium and 1640 medium (pH=5.8) ( Figures S4A, S4B ). These results further indicated that narK, narXL, narP and hmp did not affect the growth of LF82 in macrophages under acidic conditions.

Overexpression of flagellar genes have been found to attenuate the virulence of Salmonella, while suppressing flagellar expression in acidic environments facilitates intracellular replication within macrophages (Yang et al., 2012). The analysis of transcriptome data revealed a significant downregulation of flagellar genes in the acid-treated group ( Figure 5A ), which was further confirmed by qRT-PCR ( Figure 5B ). The motility of LF82 in semisolid media was significantly impaired at pH levels below 6, with a reduction in swimming ability observed particularly at pH values lower than 5.5, where the diameter of swimming decreased to approximately 0.4-fold compared to that seen under neutral conditions ( Figures 5C, D ). These results suggest that acidic environments may limit flagellar formation and movement.

We also compared the expression levels of flagellar genes of LF82 cultured in 1640 medium and infected macrophages using qRT-PCR. At 1 h p.i., fliA, flhD, and fliD were downregulated by approximately 10-fold, while fliC was downregulated by approximately 5-fold ( Figure 5E ). These results suggest that the decreased expression of LF82 flagellar genes in macrophages is associated with the acidic environment within host cells.

We obtained ΔfliC and ΔfliD mutants; these displayed a reduced swimming ability of 0.7-fold and 0.7-fold, respectively, compared to WT ( Figures 5F, G ). The adhesion abilities of ΔfliC and ΔfliD to epithelial cells were reduced to 39.1% and 18.6% compared to WT, respectively ( Figures 5H ). The invasion abilities of ΔfliC and ΔfliD to epithelial cells were reduced to 20.2% and 15.2% compared to WT, respectively ( Figure 5I ). However, there was no significant difference of intracellular replication abilities within macrophages between ΔfliC, ΔfliD, and WT ( Figure 5J ). These findings suggest that the regulation of flagellar expression by fliC and fliD may contribute to the motility, adhesion, and invasion ability of LF82 in epithelial cells. However, deletion of these genes did not significantly affect LF82 replication in macrophages.

Shown as Figures 5K, L , there was no difference between fliC, fliD and WT in LB medium and DMEM medium. These results indicated that fliC and fliD reduced swimming ability of LF82 and reduced the adhesion and invasion ability to Hela cells, not because the growth was affected by the deletion of fliC and fliD genes.

In addition, the results of the growth curve showed that there was no difference between gene deletion mutants (ΔfliC and Δ fliD) and WT in control 1640 medium and 1640 medium (pH=5.8) ( Figures S5A, S5B ). These results further indicated that narK, narXL, narP and Δhmp did not affect the growth of LF82 in macrophages under acidic conditions.

The bacterial strains and plasmids used in this study are listed in Table S2 . E. coli LF82 O83:H7 was used as the WT strain. Mutant strains were generated using the λ Red recombinase system of pSim6 and primers carrying the 39~45 bp homologous regions flanking the start and stop codons of the gene to be deleted, as previously described (Liu et al., 2019). Plasmid pUC57 carrying red fluorescent protein (mCherry) was used for confocal microscope observation and analysis. Plasmid vectors pSWK129 were used for complementation.Bacteria were cultured overnight at 37°C in LB liquid or LB agar plate. When necessary, appropriate antibiotics were added: ampicillin, 50 μg/ml; kanamycin, 50 μg/ml; chloramphenicol, 25 μg/ml; gentamicin, 20 μg/mL or 100 μg/mL.

The LF82 glycerin tube frozen in the refrigerator at -80°C was frozen on ice, inoculated in LB liquid medium at a ratio of 1:1000 for overnight culture (about 16 hours), then inoculated in LB liquid medium at a ratio of 1: 100 for logarithmic phase and centrifuged at 5,500 rpm for 5 min. The supernatant was discarded to collect the bacteria and washed with 1640 (containing 10% FBS (fetal bovine serum)) medium 3 times. The precipitate of the bacteria was suspended in 1640 (containing 10% FBS, pH=5.8, the pH was adjusted to 5.8 with HCl) medium as acid treated group and 1640 (containing 10% FBS, pH=7.5) as control group, cultured in a shaking table at 37°C for 180 rpm for 30 min, and centrifuged at 5500 r for 5 min to collect the precipitate.

To determine the growth curve of each strain, overnight cultures were washed with PBS 3 times and diluted (1:1000) in LB without antibiotics. A 200 μL aliquot was added to a 96-well flat-bottom microplate, and 200 μL LB medium was used as negative control and incubated at 37 °C with shaking at 180 rpm for 24 h, as previously described (Liu et al., 2021). The absorbance at 600 nm was recorded. Experiments were independently performed three times.

Raw 264.7, a murine macrophage cell line, 5×10⁵/well in 12-well cell culture plate was cultured with RPMI 1640 (containing 10% FBS and 1% penicillin and streptomycin (P/S) when necessary) medium at 37°C in 5% CO₂. Bacterial uptake, survival, and replication were measured by gentamicin protection test. Before infection, the bacteria were washed with PBS and resuspended in RPMI 1640 and incubated at 37°C with shaking at 180 rpm for 30 min, infected 5×10⁷/well with multiplicity of infection (MOI) of 100∶1, centrifuged at 1,000× g for 10 min, incubated with 5% CO₂ at 37°C for 10 min, and then extracellular bacteria were killed with 100 μg/ml gentamicin for 40 min (defined as T1). Ammonium chloride (NH₄Cl, Sigma) and chloroquine diphosphate (CQ, APExBIO) were used for acid neutralization to study the effect of pH on the replication of LF82 in macrophages.; Na₂NO₃ was added to RPMI 1640 to study the effect of nitrate on the replication of LF82 in macrophages. In order to determine the number of bacteria in cells, 1% triton X-100 was added to each well for 5 min to lyse eukaryotic cells. Triton X-100 at this concentration has no effect on bacterial viability for at least 30 min. Samples were diluted and spread on LB agar plates to determine the amount of CFU recovered from the cracked monolayer. T24/T1 is the replication multiple of gentamicin treatment for 24 hours compared with gentamicin treatment for 1 hour.

The complementary DNA (cDNA) was generated from 1 μg of total RNA using the Primescript 1st strand cDNA synthesis kit (Takara, Shiga, Japan). The cDNA sample was diluted three-fold prior to performing downstream experiments. qRT-PCR was performed using the Applied Biosystems 7500 real-time PCR system and SYBR green PCR master mix (Applied Biosystems, Waltham, MA, USA). All the data were normalized to levels of housekeeping gene 16S (Tasara and Stephan, 2007). The relative expression level of each gene was calculated using the cycle threshold method (2^−ΔΔCt) (Livak and Schmittgen, 2001). At least three biological replicates were carried out for each experiment. All the oligonucleotides used for qRT-PCR are listed in Table S3 .

Overnight bacteria were subcultured in DMEM (containing 10% FBS) medium at 37°C until OD₆₀₀ ≈ 0.6-0.8. HeLa cells were washed with PBS three times before infection. The cell culture medium was replaced with fresh DMEM without antibiotics or FBS. Then, the cells were infected with bacteria cultured in DMEM with MOI of 100∶1. After incubation with HeLa cells for 3 hours (Duan et al., 2013), unattached bacteria were removed by washing with PBS 6 times. Then, HeLa cells were lysed with 0.1% SDS (H₂O as solvent). The lysate was spread on an LB agar plate to count the number of surviving adherent bacteria. In order to determine the number of intracellular bacteria, fresh cell culture medium containing 100 g/mL gentamicin (Sigma) was added for 1 hour to kill extracellular bacteria. Cells were then lysed with 1% Triton X-100, and bacteria were quantified as described above.

Bacteria were cultured overnight in LB liquid medium at 37°C and 180 rpm, and 2 μL of culture was inoculated on 0.3% LB agar plate. The plate was cultured at 37°C for 9 h, and the motility was quantitatively evaluated by examining the circular swimming movement of the growing active bacterial cells.

After infection with bacteria containing mCherry plasmid, cells were washed with PBS to eliminate non-adsorbed bacteria or extracellular bacteria, and fixed with 3% paraformaldehyde for 10 min. Subsequently, the cells were washed with PBS, washed with PBS, and permeabilized with 0.1% Triton X-100 for 20 min. After washing with PBS, the slides were incubated with PBS-0.2% gelatin twice for 10 min each time. Then, the monolayer was washed with PBS and distilled water, and the stationary solution was fixed on the glass slide. The coverslips were mounted on slides and cell images were acquired using a confocal laser scanning microscope (Zeiss LSM800) and analyzed with ZEN 2.3 (blue edition) and format design by Adobe Illustrator CC 2018 softwares. Data is based on the clearest results preserved from five fields of three slides.

Statistical analysis was conducted using GraphPad Prism software (v9.3.0; GraphPad Software, San Diego, CA, USA). The mean ± SD from three independent experiments is shown in the figures. Differences between two mean values were evaluated using a two-tailed Student’s t-test. Statistical significance was set at P < 0.05.