Aeromonas hydrophila strain GD18 was isolated from the diseased grass carp (Ctenopharyngodon idella) in 2017, which led to MAS in Guangdong Province of China (Li et al., 2021). All the bacterial strains, plasmids, and primers used in this study are listed in Tables 1, 2 (Kennedy et al., 1999; Kang et al., 2002; Liu et al., 2016). Aeromonas hydrophila strains GD18 and E. coli were grown in Luria Broth (LB) (pH 7.4) at 28°C and 37°C, respectively. When necessary, the antibiotic ampicillin (Amp; 100 μg/mL), chloramphenicol (Cm; 30 μg/mL), and kanamycin (Kan; 50 μg/mL) were supplemented in the medium. The Deferoxamine (200 μmol/mL) was added to the LB medium to prepare iron-depleted media.

Healthy grass carp weighing 200 ± 20 g were bought from Xiantao Fish Hatchery, Hubei province, China. Wild-type AB strain zebrafish were obtained from the Institute of Hydrobiology of Chinese Academy of Sciences (Wuhan, China). Before intraperitoneal injection, the fish were acclimatized in the laboratory for 2 weeks. The animal experiment protocol was approved by the Laboratory Animal Monitoring Committee of Huazhong Agricultural University. All efforts were made to minimize suffering.

BLASTN and BLASTX (BLAST, basic local alignment search tool) were used to examine DNA sequences. Promoter identification was carried using the prokaryotic promoter prediction program tools, available at.¹ The Fur binding box was searched using the T6SS genes’ 500-bp upstream sequence (from position −500 bp to 0 bp relative to the corresponding gene translation start site).

Mutants were constructed using a previously described unmarked deletion method (Zhou et al., 2020). The primers were constructed using the gene sequence of fur in A. hydrophila GD18. Primer pairs p1/p2 and p3/p4 were used to amplify upstream and downstream flanking fur fragments, respectively. The primer pair p1/p4 was used in the fusion PCR. The PCR products amplified by the primer pair p1/p4 were introduced into the suicide plasmid pRE112, obtaining pRE-fur. Single-crossover mutants screened on Luria Agar (LA) media containing chloramphenicol. Allelic exchange was accomplished by culturing the single-crossover mutants on LA plates containing 7% sucrose. PCR using primers p1/p4 and direct DNA sequencing verified the Δfur. The fur fragment was amplified from the wild type (WT) strain for gene complementation. The PCR product was cloned into pMMB207, resulting in pMMB-CΔfur. The pMMB-CΔfur was transconjugated into Δfur strain to obtain the complemented strain CΔfur. hcp1 and hcp2 double-deletion mutant Δhcp1/2 as a T6SS- control strain was generated by our group (Li et al., 2021).

To detect the A. hydrophila strains growth kinetics, overnight cultures were diluted 1:500 and place in LB medium at 28°C with 200 rpm shaking. Hourly samples were taken, and optical densities were evaluated at 600 nm (OD₆₀₀).

The LDH Cytotoxicity Assay Kit (Promega, USA) was used to quantify LDH release. Ctenopharyngodon idella kidney (CIK) cell line monolayers were infected with A. hydrophila at a multiplicity of infection (MOI) of 5 for 2 h, and supernatants were collected to evaluate LDH release. The cytotoxicity % was estimated using the manufacturer’s recommendations. LDH release from uninfected cells into the culture supernatant was represented by OD₄₉₀ spontaneous, while LDH release acquired by lysis of uninfected cells was represented by OD₄₉₀ maximal release.

The adhesion experiments were carried out in triplicate as described (Dou et al., 2018). The monolayers of CIK were grown for 24 h in a 12-well cell culture plate, and then infected with A. hydrophila at a MOI of 5. Non-adherent bacteria were eliminated by washing three times with PBS after two hours of incubation at 28°C and 5% CO₂, and adhering bacteria were enumerated by plating on LA medium. Un-washed wells bacteria were enumerated as the total CFU after incubation. The adhesion rate was calculated by dividing the adhesion bacteria’s colony forming units (CFU) by the total CFU after incubation.

To assess the pathogenicity of the A. hydrophila strains, the median lethal dosage (LD₅₀) was carried out in zebrafish as previously described (Dong et al., 2021). Briefly, each A. hydrophila strain was cultured overnight with LB medium, then overnight cultures were seeded into new medium at a rate of 1:100. After 5 h culture, the mid-logarithmic bacterial culture was washed and resuspended in sterile PBS and serially diluted tenfold from 10² to 10⁷ CFU/ml. Six groups of 10 zebrafish per group were injected intraperitoneally (i.p.) with 10 μl of the bacterial solution in PBS for each A. hydrophila strain. Furthermore, as a negative control group, 10 zebrafish were injected just with PBS. The fish were monitored daily for 14 d. On 14 d after infection, surviving fish were euthanized. LD₅₀ values were calculated according to Karber’s methods.

Bacterial loads in the blood, spleen, kidney, and liver were assessed for grass carp as described previously (Hu et al., 2020), with some modifications. At 24 h post-infection, grass carp i.p. infected with 2.73 × 10³ CFU/fish by A. hydrophila were euthanized and dissected. The target organs were collected, weighed, and homogenized with PBS to obtain supernatant. Using a 10-fold dilution method, blood and supernatant were plated on LB plates for number of bacterial.

Bacterial competition experiments were conducted between A. hydrophila and E. coli in accordance with previously described methods with minor modifications (MacIntyre et al., 2010). Aeromonas hydrophila GD18 and its derived strains were ampicillin resistant, and kanamycin resistance was conferred in E. coli DH5α after it was transformed with pET-28a (Kan^r). The bacterial strains were cultured to OD₆₀₀ of 0.5, and then the fresh attacker (A. hydrophila) and prey (E. coli DH5α) were combined at a 1:5 ratio in LB plates with nitrocellulose membranes for 3 h at 28°C. To select for prey bacteria, bacteria were harvested, diluted, and placed onto LB plates containing kanamycin. The prey strain’s survival rate was then determined. The experiment was carried out in triplicate.

To construct pET-23b-fur, the open reading frame of fur of A. hydrophila GD18 was amplified by PCR with the primer pair fur F1/R1, and the PCR products were ligated into pET-23b (see Table 2). The recombinant plasmid was transformed into E. coli BL21 (DE3). Protein expression in E. coli BL21(DE3) cells was induced with 0.5 mM isopropyl-β-D-thiogalactoside (IPTG) at 37°C for 6 h. Fur protein was affinity-purified using Ni-NTA resin column (BBI, China) as described previously (Cui et al., 2019). The purified protein was measured with Pierce BCA protein assay reagent (Thermo Fisher Scientific).

EMSA was performed using the EMSA kit (Thermo Fisher Scientific, USA) in accordance with the manufacturer’s instructions. Purified Fur protein (5.0 μg) was incubated in a 20 μL reaction with the biotin-labeled and purified oligonucleotide. The DNA/protein complexes were placed onto a 2% agarose gel, electrophoresed, and transported to a nylon membrane. A chemiluminescent substrate was used to detect the biotin-labeled DNA. Only DNA probes and only Fur protein were set as control.

We amplified the vipA gene promoter (from position bp −500 to +100 bp relative to the transcriptional start site) and cloned into the pMMB-207 vector harboring a GFP reporter to generate the transcriptional GFP fusion plasmid. The recombinant plasmid was transformed into the WT and Δfur strains, respectively. Strains transformed with recombinant plasmids were harvested at the OD₆₀₀ of about 1.0. The GFP activity was measured using a SpectraMax i3X (Tecan, Austria). The GFP activity was calculated as light units/OD₄₇₀.

T6SS genes expression levels were examined using RT-qPCR as described in previous research (Wang et al., 2019). A. hydrophila were cultured for 12 h in LB medium or LB medium with Deferoxamine and the bacteria were harvested for RNA-extraction. TRizol reagent (Invitrogen, USA) was used to extract total RNA in accordance with the manufacturer’s instructions. Reverse transcription was conducted using the random primers within M-MLV reverse system (Promega, USA). A RT-qPCR reaction was performed with Fast SYBR Green PCR Master Mix (Bio-Rad, USA). The CFX96 Real-Time System (Bio-Rad, USA) was used to perform all reactions. Table 2 shows all the primers that were utilized. The 2^-ΔΔCT method was used to calculate gene expression. 16sRNA was employed as a reference gene. Relative expression level was shown as fold change relative to the level of the wild-type strain GD18 grown in LB medium. All of the qPCR experiments were conducted for at least three times.

Bacteria were cultured in LB medium or LB medium with Deferoxamine. After 12 h, a 0.22-μm filter (Millipore, USA) was used to collect and filter supernatants of LB bacterial culture. The bacteria were rinsed with PBS buffer and centrifuged at 10000 g for 10 min. Anti-Hcp and anti-GADPH (Abclonal, China), were used as primary antibodies. An HRP-goat anti-mouse or HRP-goat anti-rabbit IgG polyclonal antibody (Abclonal, China) was used as the secondary antibody. Blot images were captured in an Image Quant LAS 4000 (GE Healthcare, USA).

All statistical analyses were conducted using GraphPad Prism v 8.0 (GraphPad Software Inc., USA). Data are presented as mean ± standard deviation. A two tailed unpaired t test was used to compare independent group, and a value of p of <0.05 was considered statistically significant (*p < 0.05, **p < 0.01, ***p < 0.005, ****p < 0.0001).

Iron is one of the essential nutrients consumed by vertebrates as a defense mechanism against bacterial infection. This shortage is an environmental signal sensed by some pathogenic bacteria to modulate the expression of virulence genes. To investigate the response of theT6SS under iron-limiting conditions in A. hydrophila, Hcp expressed in the supernatant, as the indicator of functional T6SS was examined by western blot. The results showed that the secretion of Hcp decreased under iron-depleted conditions (Figure 1A). Consistent with this observation, RT-qPCR results showed varying degrees of down-regulation of T6SS core genes hcp, vgrG, vipA, and vasH (Figure 1B), indicating the depression of T6SS genes of A. hydrophila GD18 under iron-limiting conditions.

Given that Fur has been implicated in the modulation of bacterial genes associated with environmental tolerance and virulence by binding to the Fur box (Porcheron and Dozois, 2015), we constructed the fur mutant and complemented strains to determine whether T6SS genes are potentially regulated by Fur in A. hydrophila GD18. Then, western blot was performed to further detect the production levels of Hcp in different strains. The secretion of Hcp protein in Δfur was totally abolished, while complementation of the fur gene restored Hcp secretion (Figure 2A). The transcripts of the T6SS core genes were also measured by RT-qPCR. Consistent with the western blot results, inactivation of fur led to different degrees of down-regulation of T6SS core genes, including tle1 (2.6-fold), hcp (3.5-fold), vgrG (3-fold), vipA (3.6-fold), and vasH (8.4-fold; Figure 2B). Based on these findings, Fur appears to be a positive regulator of T6SS expression in A. hydrophila.

To investigate the mechanism of Fur regulation of T6SS expression, we then examined the potential Fur binding boxes in the promoter regions of the whole T6SS gene cluster in A. hydrophila strain GD18 (Figure 3A). Three putative Fur-binding sites were predicted to locate upstream of vipA open reading frame. The three putative Fur binding boxes, named Fur1, Fur2, and Fur3, showed 63 to 84% homology to the consensus Fur binding site (Figure 3B). Fur1, Fur2, and Fur3 were present at positions −341 to −322 bp, −239 to −221 bp, and − 64 to −45 bp upstream of the vipA transcriptional start site, respectively (Figure 3A). No Fur binding box was found in other promoters of the T6SS gene cluster.

To validate whether Fur binds these Fur-binding sites upstream of vipA, EMSA was conducted using recombinant protein and the biotin-labeled PCR fragments encompassing the potential Fur boxes in the vipA promoter. Fur protein interacted with the Fur1and Fur2 probe but not with the Fur3 probe (Figure 3C), suggesting that Fur binds straight to the promoter region of vipA in the T6SS gene cluster to activate the transcription of the genes in A. hydrophila.

To further verify this assumption, we constructed a GFP transcriptional reporter fusion plasmid in which GFP is governed by the vipA promoter, and then transferred it into the A. hydrophila strains to examine the regulatory effect of Fur on the promoter activities. The results indicated that the expression of GFP under the control of vipA promoter in Δfur was reduced (56%) and complementation of fur restored expression of GFP to the WT strain level (Figure 3D). Overall, these results indicate that Fur positively regulates the expression of T6SS by binding straight to the vipA promoter in A. hydrophila.

Previous studies by our group and others supported A. hydrophila T6SS contributes to the interbacterial competition (Cui et al., 2019; Li et al., 2021). Thus, we further analyzed whether the antibacterial activity was influenced in the Δfur strain. E. coli DH5α strain was taken as prey and A. hydrophila strains were taken as predators (Figure 4). As expected, after 3 h co-incubation with A. hydrophila WT strain, the surviving E. coli exhibited 7log₁₀ reduction in comparison to E. coli cultured alone. The bactericidal capacity of Δfur strain was markedly reduced (p < 0.01) and complementation of fur restored bactericidal activity to the WT strain level. Δhcp1/2 was set as a T6SS-negative control, suggesting that the deletion of T6SS resulted in abolished bactericidal effect. In conclusion, the activation of Fur promotes the interbacterial competition of A. hydrophila.

The effect of Fur on bacterial adhesion and cytotoxicity to CIK cells was investigated. CIK cells were infected with the WT, Δhcp1/2, Δfur, and the complementation strain CΔfur at 5 MOI. Firstly, the numbers of bacteria adhering to CIK cells were determined. The mutant strain Δfur presented a 3.4-fold decreased adherence ability in comparison with that of the WT strain (p < 0.001; Figure 5A). Adhesion capacity was restored in the complementation strain CΔfur. Interestingly, Δhcp1/2, the T6SS-negative mutant strain exhibited stronger (1.4-fold) adherence capacity compared with Δfur (p < 0.001).

The release of LDH in CIK cells was measured to analyze the cytotoxicity of A. hydrophila strains. Compared with the WT strain, Δfur infection led to a significant decreased (90%) of LDH release after 2 h of infection (Figure 5B) and complementation of CΔfur restored cytotoxicity to the WT level. Similarly, Δhcp1/2 was also strongly outcompeted by the WT strain. The analysis revealed that bacterial adhesion and cytotoxicity of the mutant strain Δfur were significantly decreased in comparison with those of the WT strain GD18.

Fur family members are significant virulence determinants of many bacterial pathogens (Troxell and Hassan, 2013). Therefore, we determined the roles of Fur during systemic infection of A. hydrophila. The LD₅₀ values of the WT strain GD18, Δfur, and Δhcp1/2 strains were 2.73 × 10² CFU/fish, 1.00 × 10⁴ CFU/fish, and 5.62 × 10³ CFU/fish, respectively (Table 3). The Δfur was attenuated in virulence with a 37-fold increase of LD₅₀, while Δhcp1/2 with a 21-fold increase.

To determine whether the reduced virulence was linked with failure in systemic spread, we assessed the bacterial loads of A. hydrophila strains in the blood, spleen, liver, and kidney of grass carp at 24 h post-infection. The results showed bacterial loads in the above organs of the Δfur infected group were decreased by 95, 99, 94, and 93% respectively, compared to those of WT strain infected group (p < 0.05 or p < 0.01; Figure 6). The complementation strain CΔfur displayed a similar capacity of systemic infection with the WT strain. These results indicated that Fur potentiates the systemic infection and virulence of A. hydrophila in vivo.