sRNA centered signaling activates nitrate respiration and enhances Cronobacter sakazakii virulence in host environments

Abstract

Cronobacter sakazakii is an important neonatal pathogen frequently associated with powdered infant formula. However, the mechanisms by which C. sakazakii adapts to the host environment and establishes systemic dissemination remain poorly understood. Here, we reveal a signal transduction pathway centered on a novel sRNA, CsrN, which facilitates C. sakazakii in utilizing nitrate respiration in response to oxygen-limited environments within the host, thereby enhancing its virulence in vivo. C. sakazakii infection triggers an inflammatory response, leading to the accumulation of host-derived nitrate, a key alternative electron acceptor. The expression of CsrN is induced under anaerobic conditions via the ArcAB two-component regulation system. CsrN subsequently enhances the expression of the narGHJI operon, which encodes a nitrate reductase complex. This promotes the colonization of C. sakazakii in the gastrointestinal tract and benefits its survival within macrophages, ultimately leading to increased systemic bacterial dissemination and virulence in the host. We show that administration of tungstate, a specific inhibitor of nitrate respiration, significantly attenuates C. sakazakii virulence in animal experiments. This work provides novel insights into the survival and pathogenicity mechanisms employed by C. sakazakii in host environments and suggests nitrate respiration as a potential therapeutic target for combating C. sakazakii infections.

This study reveals an sRNA-centered signaling pathway that activates nitrate respiration and enhances Cronobacter sakazakii virulence in oxygen-limited host environments, offering important insights into therapeutic strategies.

Introduction

Cronobacter spp, are Gram-negative, facultatively anaerobic bacteria recognized as opportunistic pathogens that pose a serious threat to neonates and immunocompromised infants^1,2. Among them, Cronobacter sakazakii is the most clinically significant species, frequently associated with both outbreaks and sporadic cases worldwide³. Infection caused by C. sakazakii can result in life-threatening conditions, such as bacteremia, sepsis, necrotizing enterocolitis, meningitis, and brain abscesses, with a reported mortality rate reaching up to 80% based on epidemiological data^4,5. Survivors often suffer from long-term neurological complications, including quadriplegia, hydrocephalus, and delayed neural development⁶. Although C. sakazakii is ubiquitous in the environment, its primary transmission source is reconstituted powdered infant formula (PIF)⁷. This is largely attributed to the bacterium’s remarkable resistance to desiccation, osmotic stress, and high temperature^1,2,6,8. Furthermore, the emergence of multi-drug resistant C. sakazakii renders the effective treatment more challenging, which has been regarded as an important concern in public health⁹.

Beyond reducing contamination risks during PIF production, a comprehensive understanding of the virulence mechanisms of C. sakazakii is essential for developing effective preventive and therapeutic strategies. The infection process of C. sakazakii typically begins with the colonization of the gastrointestinal tract (GIT)². Several virulence factors contribute to the ability of C. sakazakii to adhere to intestinal mucosal surfaces, among which OmpA and OmpX play particularly prominent roles⁷. Following colonization, C. sakazakii can translocate into the bloodstream, where it exhibits systemic persistence by evading immune responses and replicating within macrophages^7,10,11. Survival within macrophages is an essential step for C. sakazakii to establish systemic dissemination, as the pathogen can evade the host immune defense, penetrate the tissue barrier, and translocate to the bloodstream, liver, and spleen through macrophages^2,7,12. Furthermore, C. sakazakii is capable of penetrating the blood-brain barrier (BBB) by invading human brain microvascular endothelial cells, ultimately leading to meningitis^7,13.

Adenosine triphosphate (ATP) is essential for bacterial proliferation, serving as the primary energy source for biosynthetic processes. Bacterial cells generate ATP through a series of redox reactions, whereby electrons are transferred from electron donors, such as glucose, to electron acceptors, with oxygen (O₂) being the most efficient due to its high redox potential. This process releases substantial free energy that fuels cellular activities¹⁴. However, within host environments, bacteria frequently encounters oxygen-limited conditions, which necessitate the use of alternative electron acceptors, such as tetrathionate, nitrate, and fumarate^15,16. The ability to efficiently exploit these alternative respiratory pathways enable pathogens to maximize energy production, thereby facilitating rapid proliferation and competitive advantage over commensal microbes¹⁴. However, the mechanisms by which C. sakazakii adapts to the gut environment and establishes systemic dissemination remain largely unclear.

The process of C. sakazakii infection is shaped by a complex array of host environment factors. In bacteria, small RNAs (sRNAs) play crucial roles in regulating gene expression in response to changing environmental conditions¹⁷. A critical regulatory mechanism employed by many pathogens involves post-transcriptional modulation via sRNAs, which function by base-pairing with specific mRNA targets to influence processes such as growth, metabolism, cell signaling, and immune evasion^17–19. Despite their recognized importance in other bacterial pathogens, the specific roles of sRNA in C. sakazakii virulence remain largely unexplored.

In this study, we identified a signal transduction pathway that activates nitrate respiration, thereby enhancing the pathogenicity of C. sakazakii through an sRNA, CsrN. We found that the expression of CsrN is induced by the ArcAB two-component systems under anaerobic conditions. CsrN increases the expression of narGHJI, which encodes the nitrate reductase, by directly binding to the 5′ untranslated region (UTR) of narGHJI mRNA. Activation of nitrate respiration significantly promotes the survival of C. sakazakii in both the gut lumen and macrophages, ultimately facilitating its systemic dissemination within the host. Furthermore, administration of tungsten to neonatal rats significantly attenuated gastrointestinal colonization and systemic dissemination of C. sakazakii by inhibiting the molybdenum-cofactor-dependent nitrate respiration pathway. Our results suggest that targeting nitrate respiration may represent a promising therapeutic strategy for combating C. sakazakii infection.

Results

CsrN is a small RNA in C. sakazakii that is induced during infection

Using publicly available RNA-seq data from C. sakazakii ATCC 29544 during its infection of RAW264.7 cells²⁰, we identified a previously uncharacterized sRNA located in the intergenic region between the genes yhcB (inner membrane protein) and degQ (HtrA family serine endoproteases), and named it CsrN (Fig. 1A). Notably, the expression of CsrN was found to be upregulated during infection²⁰, indicating a potential role in the virulence of C. sakazakii (Fig. 1B).

Fig. 1. Identification and characterization of the sRNA CsrN in C. sakazakii.

A The position of the CsrN sequence in the genome of C. sakazakii ATCC 29544. B qRT-PCR analysis of C. sakazakii CsrN expression in RAW264.7 at 1, 4, 24 h post-infection (hpi) compared with that in LB medium. n = 3 independent experiments. C Northern blot analysis of CsrN expression in C. sakazakii WT, ΔCsrN, and ΔCsrN complemented (ΔCsrN+pCsrN) strains. 5S rRNA (ribosomal RNA) served as a loading control. n = 3 independent experiments. D Secondary structures of CsrN sRNA were predicted using RNAfold (http://rna.tbi.univie.ac.at/). In (B), the data are presented as the mean ± SD. Two-tailed Student’s t test (B) was applied. Source data are provided as a Source Data file.

To determine whether CsrN functions as an sRNA in C. sakazakii ATCC 29544, we conducted Northern blotting using a CsrN-specific probe. A distinct RNA band was detected in both the wild-type (WT) and the ΔCsrN complemented strain (ΔCsrN+pCsrN), but was absent in the ΔCsrN strain, indicating that CsrN is transcribed from the reverse strand of the C. sakazakii ATCC 29544 genome (Fig. 1C). To precisely map the ends of CsrN, we employed 5′ and 3′ rapid amplification of cDNA ends (RACE), which confirmed that CsrN was 181 nt in length, spanning coordinates 608,538 to 608,718 in the genome (Supplementary Fig. 1A). Secondary structure prediction using RNAfold²¹ revealed that CsrN consists of six hairpin loops, five interior loops and one multiloop (Fig. 1D).

Given that the RNA chaperone Hfq is often involved in sRNA stability and function²², we investigated whether CsrN requires Hfq. Northern blotting analysis revealed no significant difference in CsrN expression between Δhfq and WT strains (Supplementary Fig. 1B). Additionally, a rifampicin transcription inhibition assay²² showed that the degradation rates of CsrN exhibited no significant difference between Δhfq and WT strains, indicating that Hfq does not influence the stability of CsrN (Supplementary Fig. 1C). These results suggest that CsrN is an Hfq-independent sRNA.

A sequence homology search showed that CsrN is present only in Cronobacter spp. In order to explore the distribution of CsrN in C. sakazakii, we performed a comparative genomics analysis using 188 publicly available complete C. sakazakii genomes (Supplementary Data 1). The results showed that CsrN was present in 97.3% C. sakazakii strains (183/188), suggesting the prevalence and importance of CsrN in C. sakazakii.

CsrN enhances C. sakazakii virulence in vitro and in vivo

Given the replication within macrophages enables pathogens to establish systemic disease, we infected RAW264.7 cells with C. sakazakii to assess whether CsrN contributes to the survival of C. sakazakii in macrophages. Growth curves analysis showed that WT and ΔCsrN exhibited similar growth rates in monoculture in vitro (Supplementary Fig. 2A, B). However, the survival of ΔCsrN within RAW264.7 cells was significantly decreased compared to that of WT and ΔCsrN complemented strains at 1, 4, 24 h post infection (hpi) (Fig. 2A). These results indicate that CsrN promotes the intracellular survival of C. sakazakii in macrophages. Meanwhile, we also detected the cytotoxicity in macrophages infected with WT, ΔCsrN, or ΔCsrN complemented strain, as measured by LDH release at 1, 4, 24 hpi. The result showed that the ΔCsrN strain was significantly less cytotoxic to macrophages compared to the WT strain (Supplementary Fig. 2C). These results indicate that deletion of CsrN in C. sakazakii reduces bacterial intracellular survival and cytotoxicity in macrophages.

Fig. 2. CsrN enhances C. sakazakii virulence in vitro and in vivo.

A Bacterial burdens of C. sakazakii WT, ΔCsrN, or ΔCsrN+pCsrN strain in RAW264.7 at 1, 4, and 24 hpi. n = 3 independent experiments. B Bacterial counts recovered from the blood, liver, spleen, small intestine, and colon of 2- to 5-day-old rats orally infected with 1 × 10⁸ CFU of C. sakazakii WT, ΔCsrN or ΔCsrN+pCsrN strain at 48 hpi. n = 8 rats per group. C Survival curves of rats orally infected with 1 × 10⁷ CFU of the C. sakazakii WT, ΔCsrN or ΔCsrN+pCsrN strain. n = 12 rats per group. D In vivo bioluminescent imaging of rats orally infected with 1 × 10⁸ CFU of the WT-Lux, ΔCsrN-Lux or ΔCsrN+pCsrN-Lux strain at 48 hpi. n = 6 rats per group. Box plot bounds represent the interquartile range between the 25th and 75th percentiles with the horizontal bar representing the median. Whiskers extend to the full range of data. Representation images (E) and histological score (F) of rat colon orally infected with 1 × 10⁸ CFU of C. sakazakii WT, ΔCsrN or ΔCsrN+pCsrN strain at 48 hpi. n = 5 rats per group. ns, nonsignificant. In (A and F), the data are presented as the mean ± SD. One-way ANOVA (A), Two-tailed Mann-Whitney U test (B, D and F), and the log-rank (Mantel-Cox) test (C) were applied. Source data are provided as a Source Data file.

To investigate the role of CsrN in bacterial pathogenicity in vivo, we orally infected 2- to 5-day-old rats with WT, ΔCsrN, and ΔCsrN complemented strains. Specimens from the small intestine, colon, blood, liver, and spleen were collected at 48 hpi and plated on the LB agar plate to assess GIT colonization and systemic dissemination. The results showed that rats infected with ΔCsrN exhibited significant lower bacterial burdens in the small intestine, colon, blood, liver, and spleen compared to those infected with WT and ΔCsrN complemented strains, suggesting that CsrN contributes to both GIT colonization and systemic dissemination of C. sakazakii within the host (Fig. 2B). Furthermore, rats infected with ΔCsrN exhibited prolonged survival times compared to those infected with WT and ΔCsrN complemented strains (Fig. 2C). To further validate these findings, we constructed a stable luciferase reporter plasmid using the frr promotor and luxCDABE operon, following the method described by Kevin Howe²³, and transformed it into WT, ΔCsrN, and ΔCsrN complemented strains, respectively. This allowed us to monitor infection level in rats using live imaging. Bioluminescence imaging (BLI) showed that rats infected with ΔCsrN-Lux exhibited significantly reduced dissemination of bacteria into systemic organs compared with those infected with WT-Lux or ΔCsrN complemented-Lux strain (Fig. 2D). We further investigated the inflammatory factor production and histological damage in the intestines of neonatal rats infected with C. sakazakii. The results showed that the levels of IL-6 and TNF-α, as well as histopathology scores, were significantly higher in WT-infected rats than those in ΔCsrN-infected rats (Fig. 2E, F, and Supplementary Fig. 2D), indicating that the severity of disease caused by ΔCsrN infection in neonatal rats was significantly decreased compared to that caused by WT infection. Collectively, these results indicate that CsrN plays a critical role in the virulence of C. sakazakii in vivo.

CsrN enhances narGHJI expression by stabilizing its mRNA

To investigate the molecular mechanism by which CsrN regulates the virulence of C. sakazakii, we compared the transcriptional profiles of WT and ΔCsrN using RNA-seq. A total of 124 differentially expressed genes between the two strains were identified, including 29 genes that were significantly down-regulated and 95 genes that were up-regulated in the absence of CsrN (Fig. 3A, Supplementary Data 2).

Fig. 3. CsrN binds to the 5′ UTR of narHJI to enhance narGHJI expression by stabilizing its mRNA.

A The volcano plot of differentially expressed genes (DEGs) in C. sakazakii WT and ∆CsrN strains in LB medium. B KEGG Chord plot of 10 related KEGG pathways. Chords indicated a detailed relationship between expression levels of DEGs (left semicircle perimeter) and their enriched KEGG pathways (right semicircle perimeter). C The position of narGHJI cluster in the genome of C. sakazakii ATCC 29544. qRT-PCR analysis of narGHJI expression in the C. sakazakii WT and ∆CsrN strains in LB medium (D), RAW264.7 at 24 hpi (E), or small intestine of rats infected with WT or ∆CsrN at 48 hpi (F). n = 3 independent experiments. G Predicted CsrN-5′ UTR of narGHJI RNA base pairing. Point mutations to generate the disrupted alleles. REMSA of CsrN and 5′ UTR of narGHJI (H), CsrN^mut and 5′ UTR of narGHJI (I), CsrN and 5′ UTR^mut of narGHJI (J), and CsrN^mut and 5′ UTR^mut of narGHJI (K). L, M Expression of the 5′ UTR or 5′ UTR^mut of narGHJI-lux fusion when CsrN or CsrN^mut was overexpressed as indicated. n = 3 independent experiments. qRT-PCR analysis of narH (N) and narJ (O) expression in rifampicin-treated C. sakazakii WT, ΔCsrN and ΔCsrN+pCsrN strains. n = 3 independent experiments. ns, nonsignificant. In (D–F and L–O), the data are presented as the mean ± SD. For differentially expressed genes (DEGs) analysis, raw read counts were normalized by the TMM method, and DEGs were performed using the edgeR package with a significance threshold of |log2(fold change)| > 1 and false discovery rate (FDR)-adjusted p-values < 0.005 (Benjamini-Hochberg correction). A Two-tailed Student’s t test (D–F), One-way ANOVA (L, M), Two-way ANOVA (N, O) were applied. Source data are provided as a Source Data file.

KEGG pathway enrichment analysis revealed distinct regulatory patterns in the ΔCsrN strain, with the top five significantly up- and down-regulated pathways further visualized through chord diagrams (Fig. 3B). Based on RNA-seq results, we found that the expression of narGHJI operon (Fig. 3C), which encodes a membrane-bound anaerobic nitrate reductase complex²⁴ related to nitrogen metabolism in KEGG pathways, was significantly down-regulated in ΔCsrN (Fig. 3A, B).

The nitrate reductase complex functions to reduce nitrate to nitrite, a key process in anaerobic nitrate respiration¹⁴. To confirm that narGHJI is regulated by CsrN, we performed qRT-PCR analysis on WT and ΔCsrN grown in LB medium. The results showed a significant reduction in narGHJI expression in ΔCsrN compared to WT strain (Fig. 3D), consistent with our RNA-seq data (Fig. 3A, B). Additionally, during macrophage infection, the expression of narGHJI in ΔCsrN was also significantly reduced at 1, 4, and 24 hpi (Fig. 3E and Supplementary Fig. 3A, B). A similar down-regulation trend was also observed in the small intestine and colon of rats infected with ΔCsrN compared with those infected by WT (Fig. 3F and Supplementary Fig. 3C). These results indicated that CsrN positively regulates the expression of narGHJI both in vitro and in vivo.

To investigate whether CsrN directly regulates the expression of narGHJI, we analyzed potential base-pairing interactions between CsrN and narGHJI using IntaRNA²⁵. The analysis revealed a 14-nucleotide complementary region between the 5′ UTR of narGHJI mRNA (GCGUUACCGUCCCG) and the CsrN sRNA (CGGGACGGCAGCGC), with a 3-nucleotide motif in the CsrN hairpin loop (CGG) that may facilitate the binding (Fig. 3G). Subsequently, we conducted RNA electrophoretic mobility shift assay (REMSA) to test whether there is a direct interaction in vitro, using CsrN sRNA and 5′ UTR of narGHJI mRNA obtained through in vitro-transcription. The results demonstrated that CsrN sRNA binds to the 5′ UTR of narGHJI mRNA, as well as to its complementary strand (CsrN+, positive control), in a dose-dependent manner (Fig. 3H and Supplementary Fig. 3D). In contrast, no interaction was observed between the mutant CsrN sRNA (CsrN^mut, with the binding site CGG mutated to AGC) and 5′ UTR of narGHJI mRNA (Fig. 3I), or between CsrN sRNA and a mutant 5′ UTR of narGHJI mRNA (5′ UTR^mut of narGHJI, with the binding site CCG mutated to GCU) (Fig. 3J). However, a direct interaction was observed between the CsrN^mut sRNA and 5′ UTR^mut of narGHJI mRNA (Fig. 3K). These results indicate that CsrN increases the expression of narGHJI by directly binding to the 5′ UTR of narGHJI mRNA through a specific binding motif in the hairpin loop.

To further confirm the regulatory role of CsrN, we constructed a luciferase reporter plasmid by fusing the promoter and 5′ UTR of narGHJI or its mutant (5′ UTR^mut of narGHJI, with CCG mutated to GCU) with luxCDABE. These plasmids were transformed into ΔCsrN to generate strains ΔCsrN narGHJI-lux and ΔCsrN narGHJI^mut-lux, respectively. Overexpression of CsrN in ΔCsrN narGHJI-lux strain significantly increased luminescence intensity, while overexpression of CsrN^mut had no effect (Fig. 3L). In contrast, the luminescence intensity of ΔCsrN narGHJI^mut-lux exhibited no significant change upon overexpression of CsrN, but exhibited a significant increase when CsrN^mut was overexpressed (Fig. 3M), confirming that CsrN increases the expression of narGHJI by directly binding to the 5′ UTR of narGHJI mRNA.

To investigate whether CsrN increases the expression of narGHJI by influencing its mRNA stability, we performed a rifampicin transcription inhibition assay. The results showed that narGHJI mRNA degraded more rapidly in ΔCsrN compared to WT, as evidenced by reduced expression of narH and narJ in ΔCsrN (Fig. 3N, O), indicating that CsrN enhances the stability of narGHJI mRNA. Additionally, the expression of narH and narJ in Δhfq strain showed no significant difference compared to the WT strain (Supplementary Fig. 3E, F). These results demonstrate that CsrN directly binds to the 5′ UTR of narGHJI mRNA in an Hfq-independent manner, promoting its expression by stabilizing the mRNA.

narGHJI promotes the pathogenicity of C. sakazakii

To investigate the role of narGHJI in C. sakazakii virulence, we constructed a ΔnarGHJI mutant strain and a double-knockout strain ΔCsrNΔnarGHJI. Growth curve analysis revealed no significant difference in growth rates among WT, ΔnarGHJI and ΔCsrNΔnarGHJI strains when cultured in vitro (Supplementary Fig. 4A, B). However, the survival of ΔnarGHJI in macrophages was significantly reduced compared with that of WT (Fig. 4A). This survival deficiency was restored to WT levels through plasmid complementation (Fig. 4A). It indicates that narGHJI benefits the survival of C. sakazakii in macrophages. In addition, the survival of ΔCsrNΔnarGHJI exhibited no significant difference compared with that of ΔnarGHJI or ΔCsrNΔnarGHJI complemented with CsrN (ΔCsrNΔnarGHJI + pCsrN) strain, indicating that CsrN deletion did not affect the bacterial survival in macrophages in a ΔnarGHJI background (Fig. 4A and Supplementary Fig. 4C). The analysis of cytotoxicity in macrophages showed that the ΔnarGHJI strain was significantly less cytotoxic to macrophages compared to the WT strain (Supplementary Fig. 4D). These results suggest that deletion of narGHJI in C. sakazakii reduces bacterial intracellular survival and cytotoxicity in macrophages.

Fig. 4. narGHJI promotes the pathogenicity of C. sakazakii.

A Bacterial burdens of C. sakazakii WT, ΔnarGHJI, ΔCsrNΔnarGHJI, or ΔnarGHJI complemented (ΔnarGHJI+pnarGHJI) strains in RAW264.7 at 1, 4, and 24 hpi. n = 3 independent experiments. B Bacterial counts recovered from the blood, liver, spleen, small intestine, and colon of 2- to 5-day-old rats orally infected with 1 × 10⁸ CFU of C. sakazakii WT, ΔnarGHJI, ΔCsrNΔnarGHJI or ΔnarGHJI+pnarGHJI strain at 48 hpi. n = 8 rats per group. C Survival curves of rats orally infected with 1 × 10⁷ CFU of the C. sakazakii WT, ΔnarGHJI, ΔCsrNΔnarGHJI, or ΔnarGHJI+pnarGHJI strain. n = 12 rats per group. D In vivo bioluminescent imaging of rats orally infected with 1 × 10⁸ CFU of the WT-Lux, ΔnarGHJI-Lux, or ΔnarGHJI+pnarGHJI-Lux strain at 48 hpi. n = 6 rats per group. Box plot bounds represent the interquartile range between the 25th and 75th percentiles, with the horizontal bar representing the median. Whiskers extend to the full range of data. Representation images (E) and histological score (F) of rat colon orally infected with 1 × 10⁸ CFU of C. sakazakii WT, ΔnarGHJI, or ΔnarGHJI+pnarGHJI at 48 hpi. n = 5 rats per group. ns, nonsignificant. In (A, F), the data are presented as the mean ± SD. One-way ANOVA (A), Two-tailed Mann-Whitney U test (B, D and F), the log-rank (Mantel-Cox) test (C) were applied. Source data are provided as a Source Data file.

To investigate whether narGHJI contributes to systemic dissemination in vivo, neonatal rats were orally infected with the WT, ΔnarGHJI, ΔCsrNΔnarGHJI, and ΔnarGHJI complemented strains (ΔnarGHJI+pnarGHJI), respectively. The results showed that rats infected with ΔnarGHJI exhibited significantly lower bacterial burdens in the blood, liver, and spleen compared with those infected with WT or ΔnarGHJI complemented strain (Fig. 4B). We also evaluated the colonization of ΔnarGHJI in the small intestine and colon, an environment with limited oxygen¹⁴. The results revealed that narGHJI was essential for the colonization of C. sakazakii in the GIT, as rats infected with ΔnarGHJI exhibited a significantly decreased bacterial burdens in the small intestine and colon compared with those infected with WT or ΔnarGHJI complemented strain (Fig. 4B). Moreover, rats infected with ΔnarGHJI exhibited prolonged survival times compared to those infected with WT or ΔnarGHJI complemented strain (Fig. 4C). BLI analysis also showed that ΔnarGHJI-Lux exhibited reduced dissemination into systemic organs compared with WT-Lux or ΔnarGHJI complemented-Lux strain (Fig. 4D). These data indicate that narGHJI contributes to the virulence of C. sakazakii in vivo. Furthermore, no significant differences in the survival rates were observed among rats infected with ΔCsrNΔnarGHJI, ΔnarGHJI or ΔCsrNΔnarGHJI + pCsrN strain (Fig. 4C and Supplementary Fig. 4E). The bacterial burdens in the blood, liver, spleen, small intestine and colon of rats infected with ΔCsrNΔnarGHJI exhibited no significant difference from those in rats infected with ΔnarGHJI or ΔCsrNΔnarGHJI + pCsrN strain, confirming that CsrN’s impact on C. sakazakii pathogenicity is medicated through narGHJI (Fig. 4B and Supplementary Fig. 4F). In addition, we found that the levels of IL-6 and TNF-α, as well as the histopathology scores, in the intestines of neonatal rats infected with WT were significantly higher than those of rats infected with ΔnarGHJI (Fig. 4E, F and Supplementary Fig. 4G). Collectively, these results showed that narGHJI plays a critical role in the pathogenicity of C. sakazakii within the host.

Moreover, we showed that the 5′ UTR^mut-nar strain (mutated the binding motif of narGHJI 5′ UTR (CCG to GCU) on the chromosome of WT) and CsrN^mut strain (mutated the binding motif of CsrN (CGG to AGC) on the chromosome of WT) exhibited significantly reduced survival in macrophage at 1, 4, 24 hpi, as well as markedly attenuated systemic dissemination levels in neonatal rats compared to the WT strain (Supplementary Fig. 4H, I). In contrast, the 5′ UTR^mut-narCsrN^mut strain, in which the binding motifs in 5′ UTR of narGHJI (CCG to GCU) and CsrN (CGG to AGC) were simultaneously mutated on the chromosome of WT, exhibited restored bacterial survival in macrophage and in vivo bacterial burdens in neonatal rats comparable to WT level (Supplementary Fig. 4H, I). These results further confirm that CsrN positively regulates narGHJI expression through a specific RNA-RNA interaction mediated by its binding motif, thereby promoting C. sakazakii survival in macrophages and facilitating systemic dissemination in the host.

Nitrate respiration is essential for the full virulence of C. sakazakii

Previous studies have demonstrated that intestinal inflammation can create a nutrient-rich niche that promotes bacterial replication within the host²⁶. For instance, Salmonella exploits host-derived nitrate and tetrathionate generated during inflammation as alternative electron acceptors to support anaerobic respiration, providing a metabolic advantage over other gut microbes^16,27. To investigate whether the attenuated virulence observed in ΔnarGHJI is associated with impaired nitrate respiration, we first measured nitrate levels in the small intestine and colon of C. sakazakii-infected rats, as well as in the supernatant of infected macrophage cultures at 1, 4, and 24 hpi in vitro. Compared to uninfected controls, nitrate levels were significantly elevated in both intestinal tissues and macrophage supernatants following infection (Fig. 5A, B). Then we performed qRT-PCR to analyze the expression of host proinflammatory markers (Tnf, Nos2) in the small intestine and colon of C. sakazakii-infected rats and in C. sakazakii-infected macrophages at 1, 4, and 24 hpi in vitro. Notably, Nos2 encodes inducible nitric oxide synthases (iNOS), which catalyzes the production of nitric oxide (NO). NO is subsequently oxidized or reacts with reactive oxygen species to form nitrate²⁸. We found that the expression of these host proinflammatory genes was significantly increased in infected samples in vivo and in vitro (Fig. 5C, D). These results indicate that C. sakazakii infection induces an inflammatory response, leading to elevated Nos2 expression and nitrate accumulation within hosts, which may facilitate the growth of C. sakazakii through nitrate respiration.

Fig. 5. Nitrate respiration is essential for the full virulence of C. sakazakii.

A Nitrate levels in the small intestine and colon of 2- to 5-day-old rats orally infected with 1 × 10⁸ CFU of C. sakazakii WT at 48 hpi versus those uninfected. n = 6 rats per group. B Nitrate levels in RAW264.7 uninfected or infected with C. sakazakii WT at 0, 1, 4, and 24 hpi. n = 3 independent experiments. qRT-PCR analysis of Nos2 and Tnf expression in the small intestine and colon of rats at 48 hpi (C) or RAW264.7 at 1, 4 and 24 hpi (D) infected with C. sakazakii WT strain versus those uninfected. n = 3 independent experiments. E qRT-PCR analysis of narGHJI and narK expression in C. sakazakii WT strain in RAW264.7 at 24 hpi compared with that grown in LB medium. n = 3 independent experiments. F Bacterial burdens of C. sakazakii WT and ΔnarK strains in RAW264.7 at 1, 4, and 24 hpi. n = 3 independent experiments. G Bacterial counts recovered form in blood, liver, spleen, small intestine, and colon of 2- to 5-day-old rats orally infected with 1 × 10⁸ CFU of C. sakazakii WT or ΔnarK strain at 48 hpi. n = 8 rats per group. H Replication of C. sakazakii WT, ΔnarGHJI, and ΔnarK strains in LB medium with or without NaNO₃ in anaerobic environments. Relative replication ratio was the fold change in growth at 1, 4, and 24 hours/0 hours. n = 3 independent experiments. I Replication of C. sakazakii WT, ΔnarGHJI and ΔnarK strains in RAW264.7 with or without NaNO₃. Relative replication ratio was the fold change in bacterial burdens at 4 and 24 hpi/1 hpi. n = 3 independent experiments. ns, nonsignificant. In (A–F and H, I), the data are presented as the mean ± SD. Two-tailed Student’s t test (A and C–F), Two-tailed Mann-Whitney U test (G), Two-way ANOVA (B and H, I) were applied. Source data are provided as a Source Data file.

NarK, a nitrate/nitrite transporter, is always highly expressed under anaerobic conditions in the presence of nitrate and plays a central role in bacterial nitrate uptake²⁹. To further confirm the role of nitrate respiration on C. sakazakii pathogenicity, we examined the expression of narGHJI and narK by qRT-PCR in infected macrophages. The results showed that both the expression of narGHJI and narK in C. sakazakii was significantly upregulated in infected macrophages at 1, 4 and 24 hpi, compared with that of bacteria grown in LB medium (Fig. 5E and Supplementary Fig. 5A, B). Then we constructed a ΔnarK mutant and showed that its survival in macrophages was significantly decreased compared to that of WT strain (Fig. 5F). Additionally, rats infected with ΔnarK exhibited significantly lower bacterial burdens in the small intestine, colon, blood, liver, and spleen compared with those of WT-infected rats (Fig. 5G). These findings further underscore the essential role of nitrate respiration in C. sakazakii survival and systemic dissemination during host infection.

Due to its high redox potential, nitrate serves as a preferred electron acceptor under anaerobic respiration³⁰. To investigate its role in C. sakazakii growth, we performed bacterial growth experiments under both aerobic and anaerobic conditions, with or without nitrate supplementation. The results showed that the growth of ΔnarGHJI, ΔnarK, and WT exhibited no significant difference in aerobic environments, regardless of nitrate supplementation (Supplementary Fig. 5C). However, under anaerobic conditions, the WT strain exhibited significantly enhanced replication compared to the ΔnarGHJI and ΔnarK at 1, 4, 24 hours (Fig. 5H), and showed slightly increased growth compared to the ΔCsrN strain, although this difference was not statistically significant (Supplementary Fig. 5D). Moreover, nitrate supplementation significantly promoted the anaerobic growth of both WT and ΔCsrN strains, but had no significant effect on the growth of ΔnarGHJI and ΔnarK strains (Fig. 5H and Supplementary Fig. 5D). Notably, under nitrate-supplemented conditions, the WT strain showed significantly greater growth than ΔCsrN (Supplementary Fig. 5D). Infection assays in macrophages with or without nitrate supplementation demonstrated that nitrate supplementation significantly enhances the intracellular survival of WT and ΔCsrN strains (Fig. 5I and Supplementary Fig. 5E), but had no effect on the survival of either ΔnarGHJI or ΔnarK (Fig. 5I). These results indicate that C. sakazakii is capable of utilize nitrate as an electron acceptor to support anaerobic respiration, which in turn promotes its survival and proliferation.

ArcAB senses anaerobic condition to enhance narGHJI expression via CsrN

To investigate the mechanism through which the expression of CsrN and narGHJI is induced within host, we performed a DNA pull-down assay to identify potential regulators binding to the promoter of CsrN. The mass spectrometry result showed that ArcA is a potential regulator that binds to the promoter of CsrN (Supplementary Data 3). ArcA, together with ArcB, forms a two-component regulation system (TCS) that sense changes in oxygen availability and regulate the expression of downstream genes³¹. We showed that the expression of CsrN was significantly decreased in ΔarcA mutant compared to WT when grown in LB medium under anaerobic conditions, but there was no significant difference under aerobic conditions (Fig. 6A and Supplementary Fig. 6A). Additionally, the expression of both CsrN and narGHJI was significantly decreased in ∆arcA mutant compared to the WT strain in infected macrophages which represent an oxygen-limited condition³² (Fig. 6B and Supplementary Fig. 6B, C), indicating that ArcA positively regulates CsrN and narGHJI expression under anaerobic conditions. Moreover, no significant difference in narGHJI expression was observed between ΔCsrN and ΔCsrNΔarcA strains under either aerobic or anaerobic conditions (Fig. 6C and Supplementary Fig. 6D), suggesting that ArcA regulate the expression of narGHJI through CsrN. Electrophoretic mobility shift assay (EMSA) showed that phosphorylated ArcA (but not unphosphorylated ArcA) binds to the CsrN promoter in a dose-dependent manner (Fig. 6D). In contrast, ArcA showed no binding to a DNA fragment derived from the kana gene (negative control) (Supplementary Fig. 6E). Chromatin immunoprecipitation (ChIP) qPCR demonstrated that the CsrN promoter (P_CsrN) was enriched 2.78-fold in the ArcA-ChIP samples compared to the Mock-ChIP samples (Fig. 6E). As a control, no significant enrichment was observed for rpoS. These results indicate that ArcA directly activates the expression of CsrN by binding to its promoter.

Fig. 6. ArcAB senses anaerobic condition to enhance narGHJI expression via CsrN.

A qRT-PCR analysis of CsrN expression in the C. sakazakii WT and ΔarcA in LB medium. n = 3 independent experiments. B qRT-PCR analysis of CsrN and narGHJI expression in WT- or ΔarcA-infected RAW264.7 at 24 hpi. n = 3 independent experiments. C qRT-PCR analysis of narGHJI expression in ΔCsrN and ΔCsrNΔarcA strains in LB medium. n = 3 independent experiments. D EMSA of the CsrN promoter (CsrN-pro) with purified ArcA protein. E The fold enrichment of the promoters for CsrN (P_CsrN) and the negative control (rpoS) in ArcA-chromatin immunoprecipitation (ChIP) samples. n = 3 independent experiments. Northern blot analysis (F, H) and qRT-PCR analysis (G, I) of CsrN and narGHJI expression in the C. sakazakii WT (F, G) and ΔarcA (H, I) under aerobic and anaerobic conditions. n = 3 independent experiments. J Bacterial counts recovered from blood, liver, spleen, small intestine, and colon of 2- to 5-day-old rats orally infected with 1 × 10⁸ CFU of C. sakazakii WT, ΔarcA, or ΔarcA+parcA strain at 48 hpi. n = 8 rats per group. ns, nonsignificant. In (A–C and E–I), the data are presented as the mean ± SD. Two-tailed Student’s t-test (A–C and E–I), Two-tailed Mann-Whitney U test (J) were applied. Source data are provided as a Source Data file.

ArcA, a key regulator of anaerobic responses, enables bacteria to adapt to anaerobic environments³³. qRT-PCR assays and northern blot analysis were performed to further analyze CsrN expression in C. sakazakii cultured aerobically or anaerobically in minimal media. The results showed a significant upregulation of CsrN expression in WT under anaerobic conditions compared to aerobic conditions (Fig. 6F, G). Consistently, the expression of the narGHJI, the downstream target of CsrN, was also significantly elevated under anaerobic conditions compared to aerobic conditions (Fig. 6G). In contrast, no significant difference in the expression of CsrN and narGHJI was observed in ΔarcA between aerobic and anaerobic conditions (Fig. 6H, I). These results further confirmed that ArcA promotes the expression of CsrN and narGHJI under anaerobic conditions. Furthermore, qRT-PCR assays showed that the CsrN and narGHJI expression of C. sakazakii in WT-infected rats, was significantly upregulated in the small intestine and colon, in which oxygen availability is limited, compared to that grown in LB medium (Supplementary Fig. 6F, G), in agreement with in vitro anaerobic culture data (Fig. 6F, G). However, the expression of CsrN and narGHJI in ΔarcA in the small intestine and colon of infected rats exhibited no significant difference compared to that grown in LB medium (Supplementary Fig. 6H, I). Previous study reported that the intracellular environment is also oxygen-limited³², consistent with our observation of elevated expression of CsrN and narGHJI in macrophages (Figs. 1B, 5E, and Supplementary Fig. 5A, B). In contrast, when we infected macrophages with ΔarcA, the expression of CsrN and narGHJI at 1, 4, 24 hpi exhibited no significant difference compared to that of ΔarcA grown in LB medium (Supplementary Fig. 6J–L). Moreover, we found that the survival of ΔarcA in macrophages and its system dissemination in vivo were significantly reduced compared to those of the WT or ΔarcA complemented strain (Fig. 6J and Supplementary Fig. 6M). Collectively, these results demonstrate that ArcA functions as an oxygen sensor within the host, activating CsrN expression under anaerobic conditions. This, in turn, leads to the induction of narGHJI and promotes nitrate respiration, thereby contributing to the pathogenicity of C. sakazakii.

Targeting nitrate respiration to control C. sakazakii infection

Bacteria exhibit remarkable metabolic versatility, enabling them to adapt to varying oxygen levels and electron acceptor availability³⁴. To evaluate the role of aerobic respiration, anaerobic respiration, and fermentation on C. sakazakii pathogenicity, we constructed mutants in several key metabolic pathways: cydAB (encoding cytochrome bd oxidase, a terminal oxidase facilitating aerobic respiration under low-oxygen conditions), frdABCD (encoding fumarate reductase, which reduces fumarate as an alternative terminal electron acceptor in anaerobic respiration), dmsABC (encoding DMSO reductases involved in DMSO respiration under anaerobic conditions), adhE (involved in ethanol fermentation), and ldhA (involved in lactate fermentation) strains. Growth curve analysis showed no significant difference in growth between these mutants and WT under both aerobic (Supplementary Fig. 7A, B) and anaerobic conditions (Fig. 7A). Intracellular survival assays showed that the deletion of cydAB, frdABCD, dmsABC, adhE or ldhA did not influence the survival of C. sakazakii in macrophages (Fig. 7B). In addition, infection assays in rats revealed no significant difference in bacterial burdens in the intestine or systemic organs between the mutants and WT (Fig. 7C). These findings suggest that none of the aerobic or anaerobic metabolic pathways, including those governed by cydAB, frdABCD, dmsABC, adhE, and ldhA, are essential for the virulence of C. sakazakii. Instead, these results imply that nitrate respiration is the primary metabolic pathways utilized by C. sakazakii to thrive within the host.

Fig. 7. Nitrate respiration is a crucial energy metabolic pathway in C. sakazakii.

A Replication of C. sakazakii WT, ΔcydAB, ΔfrdABCD, ΔdmsABC, ΔadhE or ΔldhA strains in LB medium under anaerobic environments. Relative replication ratio was the fold change in growth at 1, 4, and 24 hours/0 hours. n = 3 independent experiments. B Bacterial burdens of C. sakazakii WT, ΔcydAB, ΔfrdABCD, ΔdmsABC, ΔadhE, or ΔldhA strains in RAW264.7 at 1, 4, and 24 hpi. n = 3 independent experiments. C Bacterial counts recovered from the blood, liver, spleen, small intestine, and colon of 2- to 5-day-old rats orally infected with 1 × 10⁸ CFU of C. sakazakii WT, ΔcydAB, ΔfrdABCD, ΔdmsABC, ΔadhE, or ΔldhA strains at 48 hpi. n = 8 rats per group. Bacterial counts recovered from the blood, liver, spleen, small intestine, and colon of tungstate-treated or untreated rats orally infected with 1 × 10⁸ CFU of C. sakazakii WT strain (D), or ΔnarGHJI strain (E) at 48 hpi. n = 8 rats per group. ns, nonsignificant. In (A, B), the data are presented as the mean ± SD. Two-way ANOVA (A), One-way ANOVA (B), Two-tailed Mann-Whitney U test (C–E) were applied. Source data are provided as a Source Data file.

Our above results showed that nitrate respiration plays a vital role in the colonization of the GIT and systemic dissemination by C. sakazakii within the host. Tungstate has been shown to disrupt nitrate respiration in Enterobacteriaceae by substituting molybdenum in the molybdopterin cofactor, an essential component of enzymes driving this metabolic pathway^35,36. We found that the addition of nitrate significantly promoted WT growth under anaerobic conditions. However, this growth-promoting effect was markedly inhibited by co-supplementation with tungstate (Supplementary Fig. 7C). Notably, in the presence of both nitrate and tungstate, the addition of increasing concentrations of molybdate progressively restored the anaerobic growth of the WT (Supplementary Fig. 7C). In contrast, supplementation with nitrate, tungstate or molybdate had no significant effect on the growth of ΔnarGHJI under anaerobic conditions (Supplementary Fig. 7D). Additionally, the growth of C. sakazakii WT and ΔnarGHJI showed no significant difference in the presence of tungstate or molybdate under aerobic conditions (Supplementary Fig. 7E). These results indicate that supplementation with tungstate disrupts nitrate respiration of C. sakazakii, thereby inhibiting the bacterial growth under anaerobic conditions.

To assess whether tungstate treatment would inhibit the infection of C. sakazakii, neonatal rats were orally infected with C. sakazakii and then treated with tungstate. The results revealed that tungstate treatment significantly reduced C. sakazakii burdens in the small intestine, colon, blood, liver, and spleen at 48 hpi compared to untreated rats (Fig. 7D). In contrast, tungstate treatment had no effect on the virulence of ΔnarGHJI in vivo (Fig. 7E). Meanwhile, we also performed 16S rRNA sequencing on bacterial DNA from the cecal content of tungstate-treated and control rats. The Chao index, Shannon index, and the PCoA based on Bray-Curtis distances showed that tungstate treatment did not significantly alter intestinal bacterial diversity (Supplementary Fig. 7F, G). Moreover, we assessed the relative abundance of the gut microbiota in each group. Consistently, variations at the genus level also showed that tungstate treatment did not alter the overall community structure (Supplementary Fig. 7H, I). The results indicate that tungstate effectively inhibits C. sakazakii colonization in the GIT and its systemic dissemination within the host, without significantly altering the gut microbiota, highlighting targeting nitrate respiration may serve as a promising therapeutic intervention for C. sakazakii infections.

Discussion

Nitrate is a preferred electron acceptor due to its high redox potential³⁷. Under normal physiological conditions, dietary intake serves as the major source of nitrate. However, during systemic inflammatory responses, such as sepsis or severe gastroenteritis, Nos2 is highly expressed, leading to elevated production of NO, which is eventually converted to nitrate^16,38. In the mucosa of the small and large intestine, the lumen-apposed epithelia limit the access of oxygen to the intestinal lumen, resulting in oxygen availability is limited in the small intestine (~1-6% O₂) and colon (<1% O₂)^14,38,39. Beyond the gut, intracellular oxygen levels in human macrophages are typically around 1%, even under ambient oxygen conditions³². Additionally, during phagocytosis, macrophages produce reactive nitrogen species, such as NO and peroxynitrite (ONOO⁻), which further inhibit the aerobic respiration of engulfed bacteria^40,41. In this study, we showed that the anaerobic environment within the host intestine, along with host-driven production of nitrate induced by C. sakazakii infection, promotes the pathogen to use nitrate-dependent anaerobic respiration, through an ArcA-CsrN-narGHJI regulatory pathway (Fig. 8). This metabolic adaptation not only allows C. sakazakii to compete with fermentative gut microbes, but also enhances its survival within macrophages, highlighting a critical strategy for persistence during infection.

Fig. 8. Model of nitrate respiration regulation by CsrN in C. sakazakii.

CsrN-Mediated regulation of nitrate respiration enhances C. sakazakii survival under anaerobic conditions in the gut.

C. sakazakii has been found in various foods, including dairy products, meats, drinking water, and dried fish products, highlighting its infectious nature^42,43. A recent study found that the LysR-type transcriptional regulator, YeiE, promotes the expression of sulfite reductase CysJI, which mediates the rapid decrease in sulfite and generation of H₂S in C. sakazakii, helpi ng the bacterium counteract ROS produced by neutrophils¹². However, the energy metabolism of C. sakazakii within the host remains poorly studied. While aerobic respiration has been reported as essential for the virulence of pathogens, such as Salmonella enterica, Streptococcus agalactiae, and Staphylococcus aureus, in systemic disease models^44–46, our study found that deletion of the cydAB, which encodes cytochrome bd oxidase, had no significant effect on the virulence of C. sakazakii both in vivo and in vitro (Fig. 7A–C). Previous studies have also demonstrated the crucial roles of anaerobic respiration and associated metabolic pathways in bacterial survival and virulence. For instance, frdABCD, which encodes fumarate reductase, facilitates anaerobic respiration and enhances bacterial motility, aiding the survival of bacteria in viscous environments⁴⁷. Similarly, adhE is involved in metabolic adaptation under anaerobic conditions and contribute to host cell adhesion in E. coli⁴⁸. In addition, the use of alternative electron acceptors such as DMSO and TMAO has been well-characterized in E. coli under anoxic conditions⁴⁹. However, C. sakazakii lacks TMAO reductase gene in its genome, and blocking other anaerobic metabolism pathways have no effect on its virulence, as demonstrate in our study (Fig. 7B, C). Therefore, we conclude that nitrate respiration is the dominant metabolic pathway utilized by C. sakazakii for survival and pathogenicity within the host.

C. sakazakii, an opportunistic foodborne pathogen, has shown a troubling rise in global outbreak frequency in recent years⁵⁰. Although antimicrobial therapy has long served as the primary therapeutic approach against C. sakazakii infection⁵¹, this strategy is being challenged by the emergence of multidrug-resistant strains isolated from both food and clinical samples worldwide^52–54. Therefore, it is important to develop safe and effective alternative therapeutic strategies for the treatment of severe C. sakazakii infection. Our findings reveal that targeted inhibition of nitrate respiration pathways substantially impairs bacterial survival and virulence, suggesting this metabolic pathway as a promising therapeutic target.

C. sakazakii primarily infects newborns, causing life-threatening meningitis and sepsis. While conventional antibiotics can combat bacterial infections, they often indiscriminately disrupt the gut microbiota, eliminating beneficial bacteria that are essential for gut health⁵⁵. This disruption has a profound impact on children’s health and their subsequent growth and development⁵⁶. In contrast, tungsten-based treatment specifically targets C. sakazakii populations that rely on nitrate respiration under disease conditions. In light of both our work (Supplementary Fig. 7F–I) and prior evidence⁵⁷, tungsten appears to have a minimal effect on commensal microbiota, though further validation is warranted. These results suggest that inhibition of bacterial nitrate respiration may offer a precision therapeutic alternative to conventional broad-spectrum antibiotics against the infection of C. sakazakii.

Methods

Ethics statement

All animal experiments were conducted in compliance with the guidelines outlined in the Guide for the Care and Use of Laboratory Animals. All animal studies were conducted according to protocols approved by the Institutional Animal Care Committee of Nankai University (Tianjin, China) and were performed under protocol no. IACUC 2020030501.

Bacterial strains, primers, plasmids, and growth conditions

Bacterial strains, plasmids, and primers used in this study are provided in Supplementary Table 1 and 2. C. sakazakii strain ATCC 29544, which spontaneously developed resistance to nalidixic acid, was used as the wild-type (WT) strain throughout this study. Mutant strains of C. sakazakii were generated using the λ-red recombination system by pKD3 or pKD4 plasmid with chloramphenicol or kanamycin resistance genes⁵⁸. A helper plasmid named pCP20 encoding the FLP recombinases was used to eliminate the antibiotic resistance gene, when required. Complementary plasmids were constructed using pACYC184 and electroporated into the corresponding mutant strains. For purification of ArcA-6×His, arcA was fused with the pET28a expression vector and transferred into E. coli BL21 (DE3). To generate the frr-lux reporter fusion, the amplification products of frr promoter region were digested and cloned into the Xho I–BamH I site of the plasmid pMS402, which carries a promoterless luxCDABE reporter gene cluster. For ChIP DNA purification, arcA was tagged with 3×Flag, and fused with pTRC99A and electroporated into C. sakazakii ATCC 29544. All the strains constructed in the study were verified by sequencing. Bacteria were inoculated into fresh Luria-Bertani (LB) or M9 minimal medium (1×M9 salts, 2 mM MgSO₄, 0.1 mM CaCl₂, 0.4% glucose) and grown at 37 °C under aerobic (250 rpm) or anaerobic (90 % N₂, 5 % CO₂, 5 % H₂; Sheldon Manufacturing) conditions. When needed, isopropyl β-D-thiogalactoside (IPTG) and antibiotics were added to the medium at the following final concentrations: 0.1 mM IPTG, 50 μg/ml nalidixic acid, 25 μg/ml chloramphenicol (Cm), 50 μg/ml kanamycin (Kan), and 10 μg/ml tetracycline (TE).

Northern blot analysis

Total RNA (5–10 µg) was normalized, and samples were mixed with 2 × RNA loading buffer (Beyotime, China). RNA was separated on 6% acrylamide gels containing 8 M urea after heating 65 °C for 15 min. Bands were transferred to nylon membranes and crosslinked by UV. Biotinylated oligonucleotides were used as probes, and hybridizations were performed in the BeyoHyb™ Quick Hybridization Solution (Beyotime, China) at 42 °C overnight. The membranes were washed twice with low-stringency wash solution at room temperature for 5 min each time, and washed once at 42 °C for 2 min. The subsequences were subjected to the Biotin Northern Blot Kit (for Small RNA) Kit (Beyotime, China). Northern blots were visualized using Amersham Imager 600, 5S ribosomal RNA (rRNA) was used as an internal control.

Race assays

Simultaneous determination of 5′ and 3′ ends of sRNA was performed using rapid amplification of complementary DNA ends (RACE) kits (Invitrogen, USA) essentially as the manufacturer’s instructions. The oligonucleotides used for RACE are listed in Supplementary Table 2. In each case, the generated PCR products by specific reverse primer of the correct size were directly cloned into the pEASY-T1 vector (TransGen, China). Plasmids with sRNA RACE inserts were sequenced using M13 forward and reverse oligonucleotides.

Cell culture and C. sakazakii infection of macrophages

Murine macrophage cell line RAW264.7 (ATCC TIB-71) was obtained from the Shanghai Institute of Biochemistry and cultured in RPMI 1640 supplemented with 10% (v/v) FBS (Hyclone, USA) at 37 °C under a 5% CO₂ atmosphere.

RAW264.7 cells were infected with WT or the indicated mutant or complemented strains that grown to the logarithmic phase at a multiplicity of infection (MOI) of 100. The cell monolayers were incubated for 45 minutes following infection, rinsed three times in PBS, and incubated for another 1 hour in RPMI 1640 media containing gentamicin (100 μg/ml) to kill the extracellular bacteria. Then each well containing bacterial cells was rinsed three times in PBS and replenished with RPMI 1640 containing gentamicin (10 μg/ml) supplemented with 1 mM NaNO₃ when required and incubated at 37 °C with 5% CO₂. From this point, the hour post-infection (hpi) was deemed as 0. At 1, 4, and 24 hpi, the infected cells were washed three times with PBS and lysed with 0.05% Triton X-100 (Solarbio, China). The intracellular bacteria were enumerated on LB agar plates for CFU analysis or collected for qRT-PCR analysis. The fold intracellular replication of the indicated strains was expressed as the CFUs recovered at 4 or 24 hpi relative to those at 1 hpi.

RNA isolation and Quantitative RT–PCR (qRT–PCR)

Overnight culture of C. sakazakii WT strain and mutant strains was 1:100 subcultured in 20 ml fresh LB medium at 37 °C with shaking at 180 rpm until the exponential growth phase was reached (OD₆₀₀ = 0.6–0.8). For C. sakazakii collected from the cells, RAW264.7 cells were infected by C. sakazakii WT or mutant strains at an MOI of 100:1 and collected at 1, 4, and 24 hpi. For C. sakazakii collected from the small intestine and colon, 2- to 5-day-old Sprague-Dawley rat pups were orally infected by 1 × 10⁸ CFU of C. sakazakii WT or mutant strains, and collected following the contents at 48 hpi. For C. sakazakii treated with rifampicin, C. sakazakii WT strain and mutant strains were incubated with 100 μg/ml of rifampicin and collected at 0, 2, 4, 8, 16, 32 min. For the effect of oxygen levels on C. sakazakii, mid-logarithmic phase C. sakazakii WT or mutant strains incubated in LB medium were collected and reincubated in M9 minimal medium supplemented under both aerobic and anaerobic conditions for 30 minutes.

Total RNA was extracted and purified using TRIzol reagent (Invitrogen, USA). RNA was treated with DNase I at 37°C for 30 min and successful removal of DNA contamination was verified by PCR. Then the RNA was quantified and qualified by measuring the A₂₆₀/A₂₈₀ absorbance ratio using a NanoDrop spectrophotometer (Thermo Fisher, USA).

cDNA synthesis was performed by StarScript III RT MasterMix (GenStar, China). qRT-PCR was performed using 2×RealStar Power SYBR qPCR Mix (Low ROX) (GenStar, China) on QuantStudio 5 (Applied Biosystems, USA). The expression levels of C. sakazakii transcripts were normalized to that of the rpoA gene⁵⁹, whereas those of the small intestine and colon transcripts were normalized to GAPDH expression. The fold change in the expression of each target gene was estimated using the 2^−ΔΔCt method.

RNA sequencing

Overnight-cultured bacteria of C. sakazakii WT and ΔCsrN strains were inoculated into fresh LB medium at a ratio of 1:100, further cultured till the OD₆₀₀ value reached approximately 0.6, and then collected by centrifugation for RNA extraction with three biological replicates per sample. The RNA quality was assessed with an Agilent 2100 Bioanalyzer (Santa Clara, CA) and agarose gel electrophoresis (AGE). Illumina library preparation, sequencing, and data analysis were conducted by Novogene (Beijing, China). Sequenced reads were checked for base quality, trimmed where 50% of the bases were below quality score 20, and filtered to exclude adapters.

The clean reads were mapped to the C. sakazakii strain ATCC 29544 reference genome (CP011047.1) by using the short-sequence alignment software Bowtie 2. Gene expression levels were normalized and reported as FPKM. For differentially expressed genes (DEGs) analysis, raw read counts were normalized by the TMM method, and DEGs were identified using the edgeR package with a significance threshold of |log2(fold change)| > 1 and false discovery rate (FDR)-adjusted p-values < 0.005 (Benjamini-Hochberg correction).

Growth assays

For aerobic growth, overnight cultures were diluted 1:1000 into fresh LB, RPMI 1640 medium, or M9 minimal medium. Where indicated, the media were supplemented with 40 mM nitrate, 0.5 mM tungstate, or 1 mM molybdate. 200 μl aliquot was added to a 96-well microplate and incubated at 37 °C at 180 rpm for 24 h aerobically. The optical density at 600 nm was measured every 20 min. The experiment was performed in triplicate for each strain. For anaerobic growth, cultures under exponential growth phase (OD₆₀₀ = 0.6) at a final concentration of 1 × 10⁸ CFU/ml, supplemented as follows: (1) 40 mM NaNO₃; (2) 0.5 mM tungstate; (3) 1 mM molybdate; (4) 40 mM NaNO₃ and 0.5 mM tungstate; and (5) 40 mM NaNO₃ with both tungstate and molybdate at the following molar ratios, 1:2, 10:1, and 100:1, and incubated under anaerobic (90 % N₂, 5 % CO₂, 5 % H₂; Sheldon Manufacturing)⁶⁰ growth conditions at 37 °C. The bacterial suspension was diluted with PBS and spread on agar plates for enumeration of bacterial CFUs at 1, 4, and 24 hours, respectively.

Protein expression, purification, and electrophoretic mobility shift assays (EMSAs)

The constructed pET28a-ArcA plasmid was transformed into E. coli BL21 (DE3). pET-ArcA-containing strain was cultured overnight in LB and 1:100 subcultured in fresh LB to OD₆₀₀ = 0.6–0.8 at 37 °C. Then IPTG was added to induce the expression of ArcA and growth at 16 °C for 20 h. The bacteria were lysed by using sonication and centrifuged at 4 °C (12,000 × g, 30 min). The supernatant containing ArcA was subsequently purified using HiTrap Ni²⁺-chelating column. Protein concentration was determined according to the Bradford method⁶¹.

CsrN fragments of the promoter regions and point mutants were amplified by PCR and purified using the StarPrep DNA Gel Extraction Kit (GenStar, China), as well as the DNA fragments of kana DNA (negative control). The DNA fragment (40 ng) was incubated at 37 °C for 40 min with different concentrations of 6 × His-tagged ArcA protein (0-1.75 μM) in 20 μl binding buffer (100 mM Tris-HCl [pH 7.4], 100 mM KCl, 10 mM MgCl, 10% glycerol, and 2 mM dithiothreitol). When needed, 30 mM acetyl phosphate was added to the binding system. Samples were then separated on 6% polyacrylamide gel in 0.5 × Tris-Borate-EDTA (TBE) for electrophoresis. DNA fragments were stained with GelRed (Life-iLab, China).

mRNA stability assay

Overnight culture of C. sakazakii WT strain and mutant strains were 1:100 subcultured in 20 ml fresh LB medium at 37 °C with shaking at 180 rpm until the exponential growth phase was reached (OD₆₀₀ = 0.6). The medium was then centrifuged, and the bacteria were re-suspended with M9, to which rifampicin (100 μg/ml) was added. At each indicated time point after rifampicin treatment, the same volume of bacteria was collected, and the corresponding RNA was extracted immediately for qRT-PCR.

REMSAs

The target DNA templates were transcribed into CsrN, CsrN+ (positive control), CsrN^mut sRNA, 5′ UTR-narGHJI, 5′ UTR-narGHJI^mut mRNA using T7 High Efficiency Transcription Kit (TransGen, China) and purified by MagicPure® RNA Beads (TransGen, China). REMSA was performed with CsrN, CsrN^mut sRNA (0, 10, 20, and 30 μM) and 5′ UTR-narGHJI, 5′ UTR-narGHJI^mut mRNA or CsrN+ (10 μM) in REMSA binding buffer (10 mM HEPES [pH 7.3], 20 mM KCl, 2.4 mM MgCl₂, and 2.4 mM DTT). The reaction mixtures were incubated for 3 min at 85 °C and then at 37 °C for 45 min. RNA was separated by 6% native polyacrylamide gel electrophoresis (PAGE) using a Native-PAGE Preparation kit (Sangon, China) and stained with GelRed (Life-iLab, China) for 10 min, and imaged using a gel imaging system (Tanon).

DNA affinity pull-down assay

DNA pull-down assays were conducted as previously described⁶², with modifications to optimize protein-DNA interaction capture. A biotin-labeled DNA fragment encompassing the CsrN promoter region was PCR-amplified from C. sakazakii and immobilized on streptavidin-coated Dynabeads (Invitrogen, USA). The DNA-bound beads were incubated with crude protein extracts from C. sakazakii wild-type strains under optimized binding conditions. Following rigorous washing to remove nonspecifically bound proteins, the retained proteins were eluted using a high-salt buffer (500 mM NaCl), resolved by SDS-PAGE, and visualized through Coomassie Brilliant Blue staining. Protein bands of interest were excised from the gel, subjected to in-gel tryptic digestion, and analyzed by matrix-assisted laser desorption ionization-time-of-flight tandem mass spectrometry (MALDI-TOF-MS/MS). Peptide sequences and mass spectra were compared against the NCBI non-redundant database for protein identification.

Animal model

2- to 5-day-old Sprague-Dawley rats were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (licensed by Charles River), and housed in specific-pathogen-free (SPF) mouse facilities. For colonization experiments, rats were infected orally with 10⁸ CFU of C. sakazakii WT strain and mutant strains in the logarithmic phase of growth⁶³. 48 hours later, blood specimens were collected for bacterial cultures (CFU). The small intestine and colon were collected, and the contents were removed. Liver, spleen, small intestine, and colon were weighed and homogenized with PBS, and spread on agar plates for enumeration of bacterial CFUs. For survival assays, 2- to 5-day-old Sprague-Dawley rat pups were infected orally with 10⁷ CFU of C. sakazakii WT strain and mutant strains in the logarithmic phase of growth⁶⁴. The number of animals alive was recorded every 8 hours.

In vivo bioluminescence imaging (BLI)

The DNA fragment of frr promotor were amplified, purified, and digested with BamH I/Xho I. The products were ligated to pMS402 plasmid to construct of lux fusion plasmids²³, which was introduced into C. sakazakii WT, ΔCsrN, ΔnarGHJI and ΔCsrNΔnarGHJI strains, respectively. The plasmids and insertions were verified by Sanger sequencing. 2- to 5-day-old Sprague-Dawley rats were infected by oral gavage with the strains harboring lux fusions (10⁸ CFU). At 48 hours, rats were imaged and followed by in vivo BLI using the in vivo imaging system Lumina II (Caliper Life Sciences, USA). Data acquisition and analysis were performed using Living Image software (Caliper Life Sciences, USA).

Lux reporter assay

For lux reporter fusion, the 5′ UTR or 5′ UTR^mut of narGHJI were cloned into the pMS402 plasmid through BamH I/Xho I, respectively. For sRNA overexpression plasmids, the DNA fragments of CsrN and CsrN^mut were cloned into the pNM12 plasmid through Msc I/EcoR I, respectively. The empty pNM12 vector was labeled as control. The lux reporters and sRNA overexpression plasmids were then transformed into the ΔCsrN strain. The plasmids and insertions were verified by Sanger sequencing. The expression of lux-based reporters from ΔCsrN strains growing in liquid culture with 0.1% arabinose (inducing the expression of sRNA) was measured using the Spark multimode microplate reader (Tecan, Switzerland).

Measurement of nitrate and nitrite concentration

The small intestine and colon of normal rat pups or those infected with C. sakazakii WT were collected, the contents were removed, and homogenized in PBS containing protease inhibitors of choice, then centrifuged at 10,000 × g for 20 minutes at 4 °C. Transfer the supernatant to another tube for tissue samples. The cell supernatant of RAW264.7 infected with C. sakazakii WT at 1, 4, and 24 hpi, respectively, and the uninfected cell supernatant were collected as cell samples. The following procedures were carried out according to the specification of the Nitrate/Nitrite Colorimetric Assay Kit (Cayman, USA). Firstly, to determine the concentration of total nitrate and nitrite, 80 μl samples were added to 96-wells, then 10 μl nitrate reductase cofactors preparation and 10 μl nitrate reductase enzyme preparation were added. After incubation for 2–3 hours, 50 µl of Griess reagent 1 and Griess reagent 2 were added, and absorbance was measured at 540 nm about 10 minutes later. Secondly, to determine the concentration of nitrite, 100 μl samples were added to 96-wells, then 50 µl of Griess reagent 1 and Griess reagent 2 were added, and absorbance was measured at 540 nm about 10 minutes later. The nitrate concentration is equal to the total nitrate and nitrite concentration calculated using the nitrate standard curve minus the nitrite concentration calculated using the nitrite standard curve.

Cytotoxicity assay

A lactate dehydrogenase (LDH) release assay was performed to assess cytotoxicity during bacterial infection. Following the infection of RAW264.7 cells with C. sakazakii WT, the indicated mutant or complemented strain, culture supernatants were collected at 1, 4, and 24 hpi. The supernatants from uninfected cells were used as a negative control. Cytotoxicity was quantified using the Cytotox 96 non-radioactive cytotoxicity assay (Promega), and the percentage of cytotoxicity was calculated according to the manufacturer’s instructions.

16S rRNA gene sequencing and analysis

Cecal content samples were collected from 2- to 5-day-old rat pups, which were allocated into an untreated control group (Mock) and a tungstate-treated group (Tungstate; 48 hours of treatment), with 8 animals per group, and the total DNA was extracted. The V3-V4 hypervariable region of the bacterial 16S rRNA gene was amplified by PCR using the primer pair 338 F (5’-ACTCCTACGGGAGGCAGCA-3’) and 806 R (5’-GGACTACHVGGGTWTCTAAT-3’). The resulting PCR products were purified, and sequencing libraries were constructed. Paired-end sequencing was subsequently performed on a NextSeq 2000 system (Illumina, USA). All the analysis was carried out on the Majorbio platform on the website at https://cloud.majorbio.com. α-diversity was calculated by the software of Mothur (http://www.mothur.org/wiki/Calculators, version v1.30.2), and Principal Co-ordinates Analysis (PCoA) was conducted for the analysis of β-diversity. The comparisons among the microbiota compositions on the genus level were conducted using the Vegan v2.5-3 package.

Statistics

For each experiment, the number of independent replicates is indicated in the figure legends, along with the corresponding statistical test used. Statistical significance was analyzed with GraphPad Prism 9.5.0 software (GraphPad Inc). Data were analyzed with either a Student’s t-test to compare two conditions with normal distributions or a Mann–Whitney U-test for two conditions without normal distributions. For assessing statistical significance among three or more groups, a one-way analysis of variance (ANOVA), two-way ANOVA, or Dunnett’s test was utilized as indicated. Survival analysis was used log-rank (Mantel-Cox) test. Differences were considered significant at p < 0.05.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

Conceptualization, B.L., L.W., X.L. and H.S.; methodology, X.L., H.S., X.Y., L.F., Y.N., B.X., J.Q., J.W., Z.L., L.W. and L.F.; visualization, X.L. and H.S.; supervision, B.L.; writing–original draft, B.L., L.W., X.L. and H.S.; writing– review & editing, B.L., L.W., X.L., and H.S.

Peer review

Peer review information

Nature Communications thanks Charlotte Michaux, Haiyan Zeng, and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

RNA-seq data have been deposited in the NCBI SRA database under accession code PRJNA1278892. The proteomics data were submitted to the iProX (integrated proteome resources) (project ID: IPX0012657000). The raw sequencing data of 16S rRNA have been deposited in the Genome Sequence Archive in BIG Data Center, Beijing Institute of Genomics (BIG), Chinese Academy of Sciences, under the accession number: CRA033865. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-70257-x.