C. sakazakii ATCC 29544 strain grown overnight in Luria Bertani (LB) medium was inoculated into fresh LB medium at a ratio of 1:100, and bacterial cells cultured to an OD₆₀₀ of 0.6 were collected for subsequent analysis. E. coli BL21 (DE3) strain and pET-26b(+) expression plasmid was used to express 6×His-tagged recombinant proteins. Gene silencing was performed using plasmid dCas9 (pdCas9, resistant to chloramphenicol) containing dCas9 nuclease and plasmid TargetF (pTargetF, resistant to spectinomycin) containing sgRNA. The oligonucleotide primers used in this study are shown in Table 1.Table 1.

The oligonucleotide primers used in this study.

Construction of phylogenetic trees based on amino acid sequences of GapA and Gap proteins in different organisms. The amino acid sequences of GapA and Gap of C. sakazakii ATCC 29544 were set as reference sequences, respectively, and the BLASTP tool was used to search for the amino acid sequences of GapA or Gap from a total of 39 organisms in five different categories of organisms, including animals, plants, fungi, protists, and bacteria. The amino acid sequences were aligned using the ClustalW tool in MEGA-X software v.10.1.8 with default parameters, followed by construction of phylogenetic trees using MEGA-X and the neighbour-joining (NJ) method with 1000 bootstrap replicates [14]. The visualization of the phylogenetic tree was performed using the Interactive Tree of Life webserver [15].

gapA and gap with the addition of NdeI and XhoI restriction sites were amplified by the polymerase-chain reaction (PCR). pET-26b(+) and amplified DNA were digested by NdeI and XhoI restriction enzymes (Takara, Japan) and ligated by T4 DNA ligase (Thermo Fisher Scientific, USA), followed by verification of insertion sequences in the recombinant plasmids by sequencing and transformation of recombinant plasmids into E. coli BL21 (DE3) chemically competent cells using the heat shock method. The production of recombinant proteins was induced by 0.5 mm Isopropyl β-D-Thiogalactoside for 8 h at 25°C. Bacterial cells were broken using a high-pressure cell crusher (Constant Systems, UK), followed by centrifugation for 15 min at 12,400 × g to collect the supernatant. The supernatant was then applied to a High-Affinity Ni-Charged Resin FF column (Genscript, China), and the recombinant proteins were eluted with different concentrations of imidazole under native conditions. The production efficiency of recombinant proteins was verified by 12% sodium dodecyl sulphate-poly acrylamide gel electrophoresis (SDS-PAGE) gels, and imidazole was removed by dialysis with PB buffer (25.22 mm Na₂HPO₄ and 13.00 mm NaH₂PO₄). The recombinant proteins were concentrated and quantified using the Centrifugal Filter Device (Ultra-15, 10000 NMWL, Millipore, USA) and the Coomassie Blue Fast Staining Solution (Beyotime Biotechnology, China), respectively [16,17].

Polyclonal antibodies were prepared according to the previous study [18]. Rabbits (New Zealand White) were immunized with 2 mg of recombinant GapA (rGapA) or Gap (rGap) proteins each time, followed by purification of polyclonal antibodies from serum using the Protein A Resin column (Genscript, China) following the manufacturer’s instructions.

The transcription of gapA and gap was silenced using the CRISPRi system according to the previous methods [12,13]. The CRISPR spacer sequence for dCas9 nucleases was chosen using Cas-Designer (http://www.rgenome.net/cas-designer/), followed by insertion of the spacer sequence into pTargetF by the inverse PCR. The information of the CRISPRi target site sequences is listed in Table 2. The linearized recombinant pTargetF was ligated by T4 DNA ligase for 16 h, followed by transfection of recombinant pTargetF into C. sakazakii ATCC 29544 chemically competent cells transfected with pdCas9. Gene silencing was performed using 1 μM of anhydrotetracycline to catalyse the co-expression of the dCas9/sgRNA complex (gene-knockdown treat, KT), while the control check (CK) strain was cultured without the addition of anhydrotetracycline.Table 2.

The CRISPRi target site sequences used in this study.

The effect of gene silencing on mRNA expression was verified by a real-time reverse transcription PCR (RT-PCR) assay according to the previous study [4,19] with slight modifications. Bacterial cells were collected by centrifugation for 5 min at 3341 × g, followed by extraction of total RNA using the SteadyPure RNA extraction kit (Agbio, China). Total RNA was quantified using a Colibri Microvolume Spectrometer (Berthold, Germany), and cDNA was synthesized using the Evo M-MLV reverse transcription kit (Agbio, China). RT-PCR assay was performed using a Mastercycler ep realplex (Eppendorf, Germany) and SYBR Green Pro Taq HS qPCR kit (Agbio, China). The fold change in gene expression was calculated by the ΔΔCt method and normalized to the fold change of the 16S rRNA gene.

The effect of gene silencing on protein expression was assessed by SDS-PAGE and Western blot (WB) assays and performed according to the previous study [20] with slight modifications. Bacterial cells were collected by centrifugation for 5 min at 3341 × g and the cell-broken supernatant was collected by centrifugation for 10 min at 13,362 × g. Proteins in the supernatant were separated using 12% SDS-PAGE gels and then transferred onto a 0.22 μM polyvinylidene fluoride membrane (Millipore, USA) using a TRANS-BLOT System (Bio-Rad, USA). The membrane was blocked with Tris-buffered saline containing 0.1% Tween 20 (TBST) and 5% skimmed milk for 1 h at room temperature, followed by incubation with prepared primary antibodies against GapA and Gap (1:10000 dilution) for 2 h, with anti-rabbit IgG-HRP secondary antibody (1:5000 dilution, Solarbio, China) for 1 h. Protein-antibody complexes were imaged using a Chemiluminescence Image Analysis System (Tanon, China), and protein expression was assessed by calculating the pixel intensities of the bands using ImageJ (National Institutes of Mental Health, USA).

The OD₆₀₀ of bacterial cells was recorded at 30 min intervals to construct growth curves. In addition, 100 μL of the bacterial suspension was diluted and dispersed on an LB agar plate, and the colony-forming unit (CFU) was counted to determine the viability of bacterial cells.

Bacterial cells were imaged using a Scanning Electron Microscope (SEM, Hitachi, Japan). Briefly, bacterial cells with the same viability were fixed with 2.5% glutaraldehyde for 12 h at 4°C and then washed with 30%, 50%, 70%, 80%, 90%, and 100% ethanol for 10 min, respectively, followed by drops on a copper mesh and dried in a filter before imaging.

To assess the swimming motility, special LB agar plates containing 1 g/100 mL peptone, 0.5 g/100 mL yeast extract, 1 g/100 mL NaCl, and 0.3% agar powder (wt/vol) were prepared. 10 μL of bacterial cells was dropped in the centre of the plate, followed by incubation for 12 h at 37°C. Swimming motility was assessed based on the diameter of the swimming zone.

Human ileocecal adenocarcinoma cell line (HCT-8) and human intestinal epithelial cell line-6 (HIEC-6) cells were cultured in RPMI 1640 medium (GIBCO, USA) supplemented with 10% foetal bovine serum (FBS), 100 U/mL of streptomycin, and 100 U/mL of penicillin G and incubated in a humidified incubator at 37°C with 5% CO₂.

The cytotoxicity of recombinant proteins on both cells was assessed by the Thiazolyl Blue Tetrazolium Bromide (MTT) method described previously [21] with slight modifications. Cells forming confluent monolayers in 96-well plates were continued to be incubated for 24 and 48 h in RPMI 1640 medium (with 2% FBS) supplemented with 2, 4, 8, 16, and 32 μg of the recombinant proteins, followed by 20 μL of MTT solution (5 μg/mL, SolarBio, China) added to each well for another 3 h. The formazan in each well was completely dissolved with DMSO (SolarBio, China) after removing the culture supernatants, and the absorbance was determined at 490 nm using a Varioskan™ LUX Multimode Microplate Reader Comes Equipped (Thermo Fisher Scientific, USA). The fold change in absorbance was calculated to determine the relative survival rate.

The capability of bacteria to adhere to and invade both cells was assessed according to the previous methods [22,23] with slight modifications. 2 × 10⁴ to 6 × 10⁴ cells were inoculated into 24-well plates and cultured to form confluent monolayers. The bacterial cells washed with PBS were resuspended in fresh RPMI 1640 medium and adjusted to an OD₆₀₀ of 0.2 (~1 × 10⁷ CFU). Bacterial cells with the same viability were inoculated into 24-well plates so that each eukaryotic cell was infected by ~10 bacterial cells, and the total volume of each well was supplemented to 1 mL. To assess adhesion, after 1.5 h of co-culture, the cells were washed three times with PBS to remove non-adherent bacteria cells, followed by lysing with 0.1% Triton X-100 to count CFUs (adherent and intracellular bacteria).

To assess invasion, after 1.5 h of co-culture, the bacteria-containing medium was replaced with fresh RPMI 1640 medium supplemented with 100 μg/mL gentamicin. The cells were continued to be cultured for 1.5 h to kill adherent bacteria, followed by washing, lysing, and counting of CFUs (intracellular bacteria).

An additional 50 μg of inactivated or active recombinant proteins were added to each well to assess the effect on C. sakazakii adhesion to and invasion of cells. The fold change in CFUs was calculated for the KT strain versus the CK strain to determine the relative adhesion and invasion rates.

The expression levels of proteins related to the NF-κB signalling pathway was assessed by WB assay with monoclonal antibodies, including Anti-IκB alpha (ab32518), Anti-IκB alpha (phospho S36) (ab133462), Anti-NF-κB p105/p50 (ab32360), Anti-NF-κB p65 (ab16502) (Abcam, UK), and β-actin (AC004) (ABclonal, China). The proteins from both cells were collected using the Native lysis Buffer (SolarBio, China) followed by SDS-PAGE and WB assays. The fold change in pixel intensity was calculated for the rGapA (or rGap) versus the control and normalized to that of β-actin to determine the relative adhesion and invasion rates.

The secretion levels of inflammatory cytokines such as TNFα, IL-6, and IL-8 in culture supernatants were assessed using the ELISA Kit (NovusBio, USA) according to the manufacturer’s instructions.

The expression levels of inflammatory factors and TJs in intestinal cells were assessed by RT-PCR assay. Cells forming confluent monolayers in 24-well plates were incubated for 24 h in RPMI 1640 medium (with 2% FBS) supplemented with 50 μg of inactivated and active recombinant protein, followed by total RNA extraction and assessment of inflammatory cytokines (including interleukin 6 (IL-6), IL-8, IL-1β, and tumour necrosis factor-alpha (TNFα)) and TJs (including occludin (OCLN), claudin-1 (CLDN-1), CLDN-2, zonula occludens-1 (ZO-1), and ZO-2). The fold change in gene expression was calculated by the ΔΔCt method and normalized to the fold change of GAPDH.

Neonatal rats, obtained from Spyford Biotechnology Co., Ltd., were selected to construct a systemic infection model of C. sakazakii, as described by the previously described method [24] with slight modifications. Twenty-eight 3-day-old neonatal rats were randomly divided into four groups (seven neonatal rats per group, and cages were randomly placed) and did not live with their biological mothers: the negative control group (fed with goat’s milk), the positive control group (fed with goat’s milk containing the CK strain), the gapA silencing group (fed with goat’s milk containing the gapA-KT strain), and the gap silencing group (fed with goat’s milk containing the gap-KT strain). Among them, bacterial cells washed with PBS were resuspended in goat’s milk and adjusted to ~1 × 10⁸ CFU/100 μL. Each group of neonatal rats was first acclimated for 2 days and given 100 μL of goat’s milk every 4 h, and then 100 μL of prepared sample was fed through intra-oesophageal gavage and again 24 h later.

All neonatal rats were euthanized 48 h after the first feeding of samples, and blood, brain, liver, spleen, kidney, and colon were immediately collected and divided into portions under aseptic conditions, followed by counting of C. sakazakii CFUs in blood and homogenized organs (excluding the colon) using the C. sakazakii Chromogenic Medium (Hopebiol, China). Five neonatal rats in each group were randomly selected for counting (n = 5).

The expression levels of inflammatory cytokines such as IL-1β, IL-6, IL-8, TNFα, and interferon-γ (IFN-γ) in blood was determined by the ELISA Kit (MMBIO, China). The blood of the four neonatal rats in each group was randomly selected for determination (n = 4).

Total RNA from brains and colons frozen in liquid nitrogen was extracted to determine the expression levels of inflammatory cytokines (including IL-6, IL-10, IL-1β, TNFα, and IFN-γ) and TJs (including OCLN, CLDN-1, CLDN-2, ZO-1, and ZO-2). The brains and colons of the three neonatal rats in each group were randomly selected for determination (n = 3). The fold change in gene expression was normalized to that of GAPDH.

Brains and colons fixed in 10% formalin solution were embedded with paraffin and stained with haematoxylin-eosin. Brain and colon damage was imaged using an ECLIPSE CI upright microscope and a DS-U3 imaging system (NIKON, Japan). All animal experiments were approved by the Institutional Animal Care and Use Committee of the Tianjin University of Science & Technology (20221112) and performed under the guidance. This study was conducted in compliance with the ARRIVE guidelines to ensure transparent and complete reporting of research involving animals.

The statistical analyses were assessed using GraphPad Prism 8.0.2 (GraphPad Software, USA) and SPSS 27.0 (SPSS Inc., USA) and performed using a one-way analysis of variance (ANOVA) and Student’s t tests as the mean ± standard deviation (SD) of at least three independent experiments. P-values of 0.05 or less were considered statistically significant and were expressed as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001, and ns for not significant.

GapA is widely regarded as a conserved protein, the conservation between the amino acid sequences of both GapA and Gap proteins of C. sakazakii and these two proteins from different organisms was assessed (Figure S1). The GapA sequences are highly conserved in many of the organisms listed, and the GapA sequences of the seven Cronobacter spp. were particularly highly conserved, with up to 100% identity (Figure S1a). In addition, the GapA sequences of several different genera in bacteria are also highly conserved, up to 97.89%. The Gap sequences are also conserved in these organisms, especially in Cronobacter spp. with up to 99.70% identity (Figure S1b). Compared to the GapA, Gap is less conserved than GapA.

Since GapA is indispensable in numerous organisms such as C. sakazakii, the virulence properties of GapA and Gap were investigated by gene silencing, and the silencing efficiency was verified by RT-PCR and WB assays. Gene silencing efficiently inhibited gene transcription, with a 77.68% and 70.37% reduction in the expression levels of gapA and gap, respectively (Figure 1a,b). These two recombinant proteins prepared with molecular weights of ~37.0 and ~35.0 kDa were recognized by the prepared polyclonal antibodies against rGapA and rGap, respectively, and the WB assay showed that the expression levels of GapA and Gap were reduced by 70.73% and 40.10%, respectively, after gene silencing (Figure 1c,d). Figure 1.

Gene silencing reduces the expression of GapA and Gap and reduces the swimming motility of C. sakazakii.

Silencing with normal transcription of genes may affect bacterial growth, including growth rate and viability. The time required for C. sakazakii to grow to the logarithmic phase increased following a decrease in the expression levels of GapA and Gap (Figure S2a). The CK strain grew to an OD₆₀₀ of 0.6 after ~2.5 h of culture compared with ~3.5 h and ~4.0 h for the gapA-KT and gap-KT strains, respectively. Bacterial cell volume was adjusted based on the calculated viable bacteria to ensure that the CK and KT strains with the same viability were used in the experiments. The CFUs of the gapA-KT strain were significantly less than that of the CK strain at OD₆₀₀ at 0.4, while the CFUs of the gap-KT strain were even significantly less than that of the gapA-KT strain at OD₆₀₀ of 0.4, 0.6, 0.8, 1.0, and 1.2 (Figure S2b).

Gene silencing also affected bacterial morphology. Some gapA-KT strains had shorter cell lengths compared with the CK strains, and conversely, some gap-KT strains grew longer and even wrinkled (Figure S2c). These results indicate that the normal expression of both moonlighting proteins is indispensable for maintaining the normal growth of bacterial cells.

Swimming motility is one of the main characteristics of C. sakazakii, and the diameter of the swimming zone of the gapA-KT and gap-KT strains was reduced by 48.05% and 61.76%, respectively, compared to the CK strain (Figure 1e). Thus, silenced expression of both proteins reduced the swimming motility of C. sakazakii.

The cytotoxicity of the two recombinant proteins was determined prior to the assessment of adhesion and invasion. Both recombinant proteins were co-cultured with cells for 24 and 48 h to determine cytotoxicity (Figure S3a-d). The addition of 32 μg of recombinant protein did not result in more than 50% cell death after 24 and 48 h of co-culture, and therefore neither recombinant protein has severe cytotoxicity to either cell.

C. sakazakii is able to colonize intestinal cells, so the virulence properties of GapA and Gap on bacterial adhesion to and invasion of both cells were investigated (Figure 2a-h). The relative adhesion and invasion rates of the gapA-KT and gap-KT strains were reduced by more than 74.21% compared with the CK strain, and the additional active recombinant proteins enhanced the adhesion capability of these two strains, but not the invasion. Surprisingly, both active recombinant proteins were able to enhance the CK strain adhesion to and invasion of both cells. These results indicate that these two proteins play key roles in enhancing C. sakazakii adhesion to and invasion of intestinal cells. Figure 2.

GapA and Gap enhance C. sakazakii adhesion to and invasion of intestinal cells.

C. sakazakii can enhance the phosphorylation of NF-κB in intestinal cells and promote the release of inflammatory cytokines, which contribute to host cell colonization. Therefore, it was assessed whether GapA and Gap have the property of enhancing NF-κB phosphorylation. rGapA and rGap induced increased expression level of p-IκB in HCT-8 cells and at a higher level in HIEC-6 cells (Figure 3a and d). rGapA and rGap induced increased expression levels of NF-κB p50 subunit and p65 subunit in both cells (Figure 3b, c, and e), especially inducing an increase in the expression levels of p65 subunit in HIEC-6 cells by 215.17% and 273.94%, respectively (Figure 3f). The results indicate that the two recombinant proteins play similar roles in eliciting and enhancing phosphorylation and nuclear translocation of the NF-κB signalling pathway. Figure 3.

GapA and Gap induce phosphorylation of NF-κB signalling pathway in intestinal cells.

The secretion levels of inflammatory cytokines induced by rGapA and rGap were further assessed by ELISA method. Both rGapA and rGap induced an increase in the secretion levels of TNFα, IL-6, and IL-8 in both cells, and in particular, the rGap induced a 14.05% and 12.52% increase in the secretion levels of IL-6 and IL-8, respectively, in HIEC-6 cells (Figure 4a and b). Figure 4.

GapA and Gap induce the secretion of TNFα, IL-6, and IL-8 in cell culture supernatants.

In addition, the expression-level changes of four inflammatory cytokines (i.e. IL-6, IL-8, IL-1β, and TNFα) in both cells were determined by RT-PCR assay. rGapA was able to induce increased expression levels of IL-1β and TNFα in HCT-8 cells and IL-8 and TNFα in HIEC-6 cells, respectively, and rGap was able to induce increased expression levels of IL-8, IL-1β, and TNFα in HCT-8 cells and IL-6 in HIEC-6 cells, respectively (Figure S4a and c). Both rGapA and rGap induced a decrease in the expression levels of OCLN and ZO-2 in both cells, and they also led to a significant decrease in the expression level of CLDN-1 in HIEC-6 cells (Figure S4b and d). These results indicate that the two recombinant proteins can induce increased expression levels of several inflammatory cytokines (especially TNFα) in intestinal cells and disrupt the intestinal epithelial barrier integrity.

C. sakazakii is able to colonize the intestinal barrier and cause systemic infection. The invasion capability of the CK and KT strains was assessed by counting the CFUs of C. sakazakii in blood and homogenate tissues, including the brain, spleen, kidney, and liver. Compared with the CK strain, the CFUs of the gapA-KT and gap-KT strains that invaded the blood and brain were significantly reduced, but there was no significant change in the CFUs of bacteria that invaded the spleen, kidney, and liver (Figure 5a-e).Figure 5.

GapA and Gap enhance bacterial infection in neonatal rats and induce inflammatory cytokine expression.

C. sakazakii infection resulted in a significant increase in the secretion levels of five inflammatory cytokines in the blood of neonatal rats, but the silenced expression of both proteins reduced the induction capability (Figure 5f-j). Importantly, the expression levels of TNFα were also highest in the CK, gapA silencing, and gap silencing groups compared to the negative control group.

The expression levels of inflammatory cytokines and TJs in the brain and colon were also determined. Both silenced expression of GapA and Gap decreased the expression levels of IL-6 and IL-1β induced by C. sakazakii in the brain and colon (Figure S5a and c). Silenced expression of GapA weakened the capability of the gapA-KT strain to decrease the expression levels of ZO-2 in the brain and CLDN-2 in the colon, respectively, and silenced expression of Gap weakened the capability of the gap-KT strain to decrease the expression levels of CLDN-1 and ZO-1 in the brain and CLDN-1 and CLDN-2 in the colon, respectively (Figure S5b and d).

Finally, brain and colon damage caused by the CK, gapA-KT, and gap-KT strains were evaluated. When gavaged goat’s milk without C. sakazakii, the brain tissue and leptomeninges of the neonatal rats were intact, no inflammatory cell infiltrating was observed, and no necrosis or neuronal degeneration was observed in the cerebral cortex (Figure 6a). Mild symptoms of meningitis and inflammatory cells predominantly lymphocytes were observed in the leptomeninges when gavaged goat’s milk containing the CK strain, as indicated by the orange arrows (Figure 6b). In contrast, neither gapA-KT nor gap-KT strains caused meningitis symptoms and inflammatory cell infiltrating in the brain (Figure 6c and d). No structural changes or inflammatory cell infiltrating were observed in the colon when gavaged goat’s milk (Figure 6e). In addition, the epithelial cells and crypts in the mucosal layer were neatly arranged and tightly packed, and there was no sign of oedema below the mucosal layer. The CK strain caused moderate damage to the colon structure of neonatal rats, with large numbers of inflammatory cells, predominantly neutrophils, accumulating in the colon, as indicated by the orange arrows (Figure 6f). Neither gapA-KT nor gap-KT strains caused serious damage to the colon (Figure 6g and h). The results indicated that both proteins also play indispensable roles in enhancing the capability of C. sakazakii to invade neonatal rats and in stimulating host inflammatory responses. Figure 6.

GapA and Gap are involved in the pathogenicity of C. sakazakii.