IFN-γ Manipulates NOD1-Mediated Interaction of Autophagy and Edwardsiella piscicida to Augment Intracellular Clearance in Fish

This article is featured in Top Reads, p.999

Accumulating evidence suggests that autophagy is an evolutionarily conserved defense mechanism to dispose of bacteria through lysosome in hosts (1). In particular, autophagy is significantly induced by intracellular bacteria, such as Salmonella, Mycobacterium, and group A Streptococcus, in epithelial cells (2). In the case of Salmonella, autophagy-mediated engulfment and subsequent colocalization with the lysosome limit this intracellular bacteria’s growth in macrophages (3). Additionally, autophagy-mediated elimination of some intracellular pathogens, such as Streptococcus pyogenes (4), Staphylococcus aureus (5), and Shigella flexneri (6), has also been demonstrated in either human epithelial cells or mouse embryonic fibroblasts. In response to host clearance, some bacterial species have strategies to subvert or evade autophagy (7, 8). For instance, Listeria monocytogenes can escape from autophagic degradation by markedly enhancing the replication rate and reducing its colocalization with microtubule-associated protein L chain 3 (LC3) at the later stage of infection (9, 10). These studies suggest that autophagy plays an important role in the host defense against intracellular bacteria.

Upon the initiation of autophagy, autophagy-related gene (ATG) 16L1 conjugates with ATG5 and ATG12 to form a complex essential to autophagosome formation, and then, the autophagosome encloses cytoplasmic materials and eventually fuses with lysosomes, in which the autophagic cargoes are degraded (11, 12). Accordingly, the role of autophagy in the defense against bacterial infection has attracted more and more attention. In this context, there is increasing evidence that autophagy can be triggered by pathogen-associated molecular patterns, which are recognized by pattern-recognition receptors (PRRs), such as TLRs (13), nucleotide-binding oligomerization domain (NOD) receptors (6), and NOD-like receptors (14). Among these PRRs, NOD1 and NOD2 play a significant role in sensing the bacterial peptidoglycan and recognizing intracellular pathogens (15), thereby emerging as candidates for linking bacterial invasion to the initiation of autophagy.

Edwardsiella piscicida is an intracellular pathogen with typical characteristics of the Enterobacteriaceae family, previously identified as Edwardsiella tarda in fish (16, 17). E. piscicida infection leads to a basic systemic disease called edwardsiellosis that causes enormous economic losses in farmed fish (18). Accordingly, increasing attentions have been directed to the pathogenic and invasion characteristics of E. piscicida in fish phagocytes (19), leukocytes (20), and cell lines (21). Given that E. piscicida can enter, survive, and replicate in fish cells, the mechanism mediating the cellular pathways for E. piscicida entry has attracted the most interest, and these studies suggest the contribution of phagocytosis to the intracellular traffic of E. piscicida (22–24). However, the cellular and molecular mechanisms underlying the interaction between E. piscicida and host cells are still far from understanding.

Because no previous studies report the occurrence of autophagy during E. piscicida infection, whether autophagy plays a role in the detection and elimination of this intracellular bacterium is unclear. In this study, we explored the autophagic machinery during E. piscicida infection in grass carp (Ctenopharyngodon idella, an economically important fish species widely cultured in Asia) monocytes/macrophages, and a carp cell line, Epithelioma papulosum cyprini (EPC). We examined the autophagic responses following E. piscicida challenge and detected cellular colocalization of the bacteria with autophagosomes or lysosomes. Moreover, we elucidated NOD1-mediated molecular mechanism governing the cross-talk between autophagy and E. piscicida and investigated how E. piscicida and its virulence factors regulated NOD1 and ATG16L1 expression to interfere with autophagy. Furthermore, the modulation of IFN-γ as an intrinsic regulator on the colocalization of E. piscicida with NOD1 and autophagosomes was determined, and the autophagic role of IFN-γ–promoted E. piscicida clearance was assessed in monocytes/macrophages. These works unveil the mechanism and significance of the cross-talk between host autophagy and intracellular bacteria in fish species.

Healthy Chinese grass carp, weighing ∼1.0 kg, were purchased from Chengdu Tongwei Aquatic Products, kept at 20°C in a running water system, and fed to satiation daily with commercial carp pellets. The fish were acclimated to this environment for at least 1 wk prior to use. During the procedures of cell preparation and tissue isolation, the fish were sacrificed by anesthesia in 0.05% MS222 (E10505; Merck). All animal experiments complied with the regulation of animal experimentation of Sichuan province, China, and were approved by the ethics committee of the University of Electronic Science and Technology of China. EPC cells were obtained from China Center for Type Culture Collection and kept in our laboratory. EPC cells were cultured at 28°C under 5% CO₂ and saturated humidity with MEM medium (41500034; Thermo Fisher Scientific) supplemented with 10% FBS (10099141; Thermo Fisher Scientific).

Grass carp head kidney monocytes/macrophages were prepared by discontinuous density gradient centrifugation followed by adherent culture as described previously (25). In brief, the head kidney was collected from freshly sacrificed grass carp, washed twice, and gently pressed in DMEM/F12 medium (12400024; Thermo Fisher Scientific). The cell suspension was filtrated through a 200-gauge stainless steel mesh and centrifuged at 500 × g for 25 min in a density gradient column from a fish monocyte/macrophage preparation kit (TBD2011FP; TBDscience). After that, cells at the interface were collected and washed with the PBS (10 mM) (pH 7.4) twice. The cells were resuspended in DMEM/F12 supplemented with 5% FBS, seeded at the density of 3 × 10⁷ cells/ml, and incubated at 28°C under 5% CO₂ and saturated humidity. After incubation for 2 h, nonadherent cells were washed off, and the attached cells were cultured in the complete medium (DMEM/F12 and 10% FBS).

The sequence of red fluorescent protein (RFP) was cloned from pACYCDuet-1-mamW-lac-RFP kept in our laboratory using the primers of RFP-T forward and RFP-T reverse (Table I) and ligated into pGEM-T Easy (A1360; Promega). After sequencing analysis, the RFP-pGEM-T Easy plasmid was transformed into E. piscicida (EIB 202) by electroporation. The RFP-labeled E. piscicida was confirmed by fluorescence microscopy.

Cells were collected and lysed in radioimmunoprecipitation assay lysis buffer (P0013B; Beyotime Biotechnology). Protein samples were separated into 8%, 12%, or 15% SDS-PAGE gel and transferred to 0.22-µm polyvinylidene difluoride membranes (ISEQ00010; Millipore) by a wet electroblotting system (Bio-Rad Laboratories). The membranes were blocked by TBST buffer containing 10% (w/v) defatted dry milk at room temperature for 2 h and then incubated with rabbit anti-human LC3 polyclonal Ab (pAb) (at a dilution of 1:1000; PM036; MBL International), anti–β-actin mAb (at a dilution of 1:5000; 200068-8F10; Zen Bioscience), anti-NOD1 mAb (at a dilution of 1:200; sc-398696; Santa Cruz Biotechnology), anti-ATG16L1 pAb (at a dilution of 1:1000; ET7106-65; HUABIO), or anti-ubiquitin (Ub) pAb (at a dilution of 1:1000; ET1609-21; HUABIO) at 4°C overnight with shaking. Then, HRP-conjugated goat anti-rabbit (for LC3, ATG16L1, and Ub) (at a dilution of 1:5000; ZB-2301; ZSGB-Bio) or anti-mouse (for β-actin and NOD1) (at a dilution of 1:5000; ZB-2305; ZSGB-Bio) secondary Ab was applied for 2 h at room temperature. Finally, signals were detected using an ECL kit (180-501; Tanon) according to the manufacturer’s instructions.

Twenty-one E. piscicida virulence factors and their GenBank accession numbers were as follows: E. piscicida virulent protein (EVP) F (AAS58124), EVPB (AAR83928), ESEG (AAX76916), ESED (AAV69405), EVPG (AAS58125), ESCA (AAV69403), ESCB (AAX76917), ESEB (AAX76903), EVPE (AAS58123), EVPJ (ABW69082), EVPN (ABW69086), EVPC (AAR83929), EVPD (AAR83930), EVPI (ABW69081), EVPM (ABW69085), EVPA (AAR83927), EVPO (ABW69087), EVPK (ABW69083), ESAN (AAX76920), ESEE (AAV69406), and EVPL (ABW69084). Based on the cDNA sequences of grass carp NOD1 (FJ937972.1) and gcATG16L1 (MK402292), as well as those virulence factors in GenBank (http://www.ncbi.nlm.nih.gov/genbank/), the designed primers (Table I) with restriction endonuclease recognition sites were used to clone their open reading frames using Pfu polymerase (M7741; Promega) separately. PCR products were sequenced and ligated into pEGFP-N1, p3xFLAG-CMV-7.1, or pcDNA3.1/myc-His (−) A kept in our laboratory to generate expression plasmids. In particular, gcNOD1, gcATG16L1, ESCB, and EVPC expression constructs were indicated as NOD1-GFP, ATG16L1-GFP, NOD1-FLAG, ATG16L1-FLAG, ATG16L1-HIS, EVPC-FLAG, and ESCB-FLAG. Finally, these constructs were prepared using an endotoxin-free plasmid extraction kit (DP118; Tiangen) for transfection in EPC.

After transfection or drug treatment, the cell culture medium was removed, and cells were lysed with 500 µl of coimmunoprecipitation lysis buffer (P0013; Beyotime Biotechnology). Five micrograms FLAG pAb (14793; Cell Signaling Technology) or rabbit IgG (as a negative control; A7016; Beyotime Biotechnology) was added into 200 µl lysate and incubated at 4°C overnight with gentle rotation. To detect the ubiquitinoylation of NOD1, 5 µg of anti-NOD1 mAb (sc-398696; Santa Cruz Biotechnology) or mouse IgG (as a negative control; A7028; Beyotime Biotechnology) was added into 200 µl of lysate and incubated at 4°C overnight with gentle rotation. Subsequently, 20 µl of prebalanced Protein A magnetic beads (73778; Cell Signaling Technology) was added to each sample and incubated at room temperature for 30 min with gentle rotation. Next, the immunoprecipitated proteins were eluted from the beads with the elution buffer (187.5 mM Tris-HCl [648310-M; Merck] [pH 6.8], 6% SDS [L3771; Merck], 30% glycerol [G5516; Merck], 0.03% bromphenol blue [B0126; Merck], and 125 mM DTT [D0623; Merck]). Finally, immunoprecipitates (IP samples) or cell lysates (Input samples) were separated on 8% or 12% SDS-PAGE gel, and the proteins were detected using anti-HIS Ab (at a dilution of 1:5000; SAB2702219; Merck), anti-FLAG Ab (at a dilution of 1:2000; 14793; Cell Signaling Technology), or anti-Ub Ab (at a dilution of 1:1000; ET1609-21; HUABIO), separately.

Grass carp primary monocytes/macrophages or EPC were cultured in antibiotic-free DMEM/F12 supplemented with 10% FBS in 35-mm dishes (Corning). Alternatively, the cells were pretreated with 500 ng/ml grass carp IFN-γ prepared according to our previous study (25) or 10 mM 3-methyladenine (3-MA) (S2767; Selleckchem) for 8 h. The RFP-labeled E. piscicida at a multiplicity of infection (MOI) of 10 resuspended in the antibiotic-free medium was added to the cells. After incubation for 1 h, the cells were washed with PBS five times, fixed with 4% paraformaldehyde (158127; Merck) for 15 min, permeabilized with 0.3% Triton X-100 (T8787; Merck) in PBS for 10 min, and blocked in 10% goat serum (C0265; Beyotime Biotechnology) for 1 h at room temperature. After that, the cells were incubated with anti-human LC3 pAb (at a dilution of 1:400; PM036; MBL International) overnight followed by anti-rabbit IgG conjugated with Alexa Fluor 488 (AF488) (at a dilution of 1:500; ZF-0511; ZSGB-Bio). To detect the colocalization of NOD1 and E. piscicida, the cells were incubated with anti-NOD1 mAb (at a dilution of 1:30; sc-398696; Santa Cruz Biotechnology) overnight followed by anti-mouse IgG conjugated with AF488 (at a dilution of 1:500; ZF-0512; ZSGB-Bio). Alternatively, Alexa Fluor 594 (AF594) (1:500; ZF-0516, ZSGB-Bio) was used to label the anti-FLAG Ab (at a dilution of 1:300; 14793; Cell Signaling Technology). Meanwhile, the nuclei of the cells and E. piscicida were stained with DAPI (C1002; Beyotime Biotechnology) for 5 min and 24 h, respectively. Images were obtained on a Zeiss LSM 800 confocal microscope (Carl Zeiss) using a ×63 oil immersion lens. To reconstruct the three-dimensional (3D) images, the focus was adjusted to locate the coordinates of the top and bottom of cells on the z-axis, and “slides” were set to “10.” For the quantification of AF488-labeled LC3 puncta formation in monocytes/macrophages, the LC3 puncta per cell were counted in a total of >30 cells from independent triplicates using confocal microscopy. For quantification of LC3⁺ or NOD1⁺ bacteria (bacteria colocalized with AF488-labeled LC3 or NOD1), the numbers of LC3⁺ or NOD1⁺ bacteria per cell were determined by counting a total of >30 cells from independent triplicates under confocal microscopy.

Grass carp primary monocytes/macrophages were cultured in antibiotic-free DMEM/F12 supplemented with 10% FBS in a 35-mm dish (Corning), and then RFP-labeled E. piscicida resuspended in the antibiotic-free medium was added into each dish. After incubation for 1 h, the monocytes/macrophages were washed with PBS five times and then cultured with 75 nM LysoTracker Green DND-26 (Thermo Fisher Scientific, L7526), which was prewarmed at 37°C (L7526; Thermo Fisher Scientific) for 30 min. After that, the probes were removed, and images of cells were acquired immediately on a Zeiss LSM 800 confocal microscope (Carl Zeiss) using a ×63 oil immersion lens.

Grass carp monocytes/macrophages seeded in 24-well plates (Corning) were treated with IFN-γ (500 ng/ml) in the absence or presence of 3-MA (10 mM) or rapamycin (10 µM; S1039; Selleckchem) for 8 h and challenged with E. piscicida (1 × 10⁷ CFU/well) for 1 h. The cells were washed with PBS three times, and the cell culture medium was changed to fresh DMEM/F12 containing 100 µg/ml gentamicin (E003632; Merck) to kill the bacteria, which were not internalized in monocytes/macrophages. After incubation for 2 h, the cells were washed with PBS five times, lysed by 1% Triton X-100, and spread on Luria–Bertani agar plates. After incubation at 37°C for 16 h, the colonies on the plates were counted. Counts were obtained from at least three independent experiments.

Grass carp monocytes/macrophages were seeded in 24-well plates (Corning) and incubated with IFN-γ (500 ng/ml), 3-MA (10 mM), N^G-methyl-l-arginine acetate (L-NMMA; 500 µM) (M7033; Merck), or their combination, as indicated in the figures, for 8 h. After treatment, NO production was measured using the Griess reaction. Briefly, 75 µl of culture medium was collected and incubated with 200 µl of Griess reagent (1% sulfanilamide; S9251; Merck), 0.1% naphthyl ethylenediamine dihydrochloride (and 2.5% H₃PO₄; 222488; Merck) at room temperature for 15 min. The absorbance at 540 nm was measured in an iMark Microplate Absorbance Reader (Bio-Rad Laboratories). NO amounts were determined using the serially diluted sodium nitrite (237213; Merck) as a standard.

Total RNA was extracted from grass carp monocytes/macrophages, EPC, head kidney, spleen, liver, or intestine using Total RNA Extraction Reagent (R401-01; Vazyme Biotech) according to the manufacturer’s protocol. RNA concentration was determined by NanoDrop spectrophotometry (Thermo Fisher Scientific). Approximately 2 µg of total RNA was subjected to reverse transcription using the M-MLV reverse transcriptase (M1705; Promega). The primers for quantitative PCR (qPCR) are listed in Table I. RT-qPCR was performed on the qTOWER³ G Thermocycler (Analytik Jena). To estimate the amplification efficiency, the standard curve for each target gene was generated by 10-fold serial dilutions (from 10⁻¹ to 10⁻⁶ fmol/µl) of a plasmid containing the target gene sequences as the PCR template. In these experiments, EF1A or β-actin was amplified as the reference gene.

Images were acquired by chemiluminescence charge-coupled device camera–based digital imaging instruments (Tanon 5200 imaging system), and the densitometric data of LC3-II, NOD1, ATG16L1, and β-actin protein levels were analyzed with ImageJ software (National Institutes of Health). Statistical analysis was performed using Prism 6.0 software (GraphPad Software, La Jolla, CA). Each experiment was repeated at least three times. Two-tailed Student t test was used to analyze statistical differences.

To assess whether E. piscicida can induce autophagy, grass carp monocytes/macrophages and EPC cells were incubated with E. piscicida at an MOI of 10 for 1 h, and the AF488-immunolabeled LC3 puncta formation was examined by confocal microscopy. Compared with uninfected cells, the LC3 puncta number significantly increased in monocytes/macrophages (Fig. 1A, 1B) and EPC cells (Fig. 1C, 1D). Moreover, the effect of E. piscicida at the same conditions on the protein expression of LC3-II was assessed in these two types of cells, and an 8-h treatment with 10 µM rapamycin was used as positive control. Western blotting (WB) assay showed that both E. piscicida and rapamycin challenge significantly enhanced the protein levels of LC3-II in monocytes/macrophages and EPC cells (Fig. 1E). Furthermore, grass carp monocytes/macrophages were challenged with RFP-labeled E. piscicida for 2 h, and the colocalization of E. piscicida (red) with AF488-immunolabeled LC3 (green) was observed by confocal microscopy (Fig. 1F), and the colocalization was further verified by 3D reconstruction (Fig. 1G, 1H). Moreover, RFP-labeled E. piscicida was found to be colocalized with LysoTracker-labeled lysosomes (green) in monocytes/macrophages (Fig. 1I).

To evaluate whether the NOD1 signaling pathway was involved in E. piscicida–induced autophagy, grass carp monocytes/macrophages were infected with RFP-labeled E. piscicida for 1 h, and the results revealed that AF594-immunolabled NOD1 was colocalized with E. piscicida (Fig. 2A). In addition, ATG16L1-FLAG and NOD1-GFP were overexpressed in EPC cells for 48 h, and then the cells were fixed. To detect FLAG-ATG16L1, AF594 (red) was used to label the anti-FLAG Ab; the results showed that FLAG-ATG16L1 (red) was colocalized with GFP-NOD1 (green) (Fig. 2B), and the colocalization was further verified by 3D reconstruction (Fig. 2C). In addition, we performed coimmunoprecipitation assays, showing that HIS-tagged ATG16L1 and ATG16L1 was coimmunoprecipitated by NOD1 in EPC cells (Fig. 2D) and monocytes/macrophages (Fig. 2E), respectively. To detect the cellular colocalization of the bacteria with NOD1-ATG16L1 complex, EPC cells with overexpression of NOD1-FLAG and ATG16L1-GFP were exposed to E. piscicida for 2 h. In this experiment, the cells transfected with blank vector pCMV-3xFLAG and pEGFP-N1 were used as control groups. As shown in (Fig. 2F (left two panels), FLAG alone was not detected, as there was no stop codon in the blank vector, and GFP alone was diffused in the cytosol of cells and not overlapped with DAPI-stained E. piscicida, which differed from the distribution pattern of ATG16L1-GFP. In parallel, the white arrows indicated E. piscicida were surrounded by ATG16L1-GFP overlapping with AF594-immunolabeled NOD1-FLAG (red) (Fig. 2F, right two panels).

The involvement of NOD1 and ATG16L1 in E. piscicida–induced autophagy prompted us to examine the effect of E. piscicida infection on their expression in grass carp monocytes/macrophages. Results uncovered that an increasing MOI (0.3–30) of E. piscicida resulted in a dose-dependent decrease of NOD1 protein levels (Fig. 3A), whereas E. piscicida with an MOI of 10 decreased the protein expression of NOD1 from 10 to 120 min (Fig. 3B). Furthermore, E. piscicida with an MOI of 10 was revealed to block the mRNA expression of NOD1 and ATG16L1 at 3 and 6 h, respectively, in EPC cells (Fig. 3C). Moreover, we also found that E. piscicida infection reduced the gene expression of estrogen-related receptor-α (ESRRA), a regulator of ATG gene expression (26), in both monocytes/macrophages and EPC cells (Fig. 3D). Subsequently, by using ESRRA inhibitor kaempferol, our results showed that kaempferol inhibited the transcription and protein levels of NOD1 and ATG16L1 in grass carp monocytes/macrophages, consistent with the inhibitory effects observed in the E. piscicida challenge group (Fig. 3E, 3F). These results were confirmed by performing grass carp ESRRA overexpression in EPC cells, showing that overexpression partially reversed E. piscicida–attenuated mRNA (Fig. 3G) and protein (Fig. 3H) levels of NOD1 and ATG16L1.

Immunoprecipitation assays revealed that live E. piscicida was able to induce the ubiquitinoylation of NOD1 in grass carp monocytes/macrophages (Fig. 4A). Moreover, MG-132 (10 µM; S2619; Selleckchem), a Ub pathway inhibitor, significantly reversed E. piscicida–attenuated protein levels of NOD1 in the same cell model (Fig. 4B). Given that live but not heat-inactivated E. piscicida decreased the protein levels of NOD1 (Fig. 4C), we hypothesized that E. piscicida virulence factors contributed to modulate NOD1 expression. In this case, we overexpressed 21 virulence factors in EPC cells by using the primers for vector construction (Table I), finding that only ESCB and EVPC, but not other factors, significantly decreased the protein expression of NOD1 (data not shown). In accordance with this, overexpressing ESCB and EVPC significantly stimulated NOD1 ubiquitinoylation (Fig. 4D). Moreover, the inhibitory effects of two factor overexpression on NOD1 protein expression were reversed by MG-132 (Fig. 4E).

To determine whether the interplay between autophagy and E. piscicida was regulated in host cells upon bacterial infection, we examined the effect of IFN-γ on E. piscicida–induced autophagy in grass carp monocytes/macrophages. Corresponding to its role in immune defenses, IFN-γ transcription level was increased 20-fold and 10-fold in the head kidney and head kidney–derived monocytes/macrophages after E. piscicida challenge, respectively (Fig. 5A). Consistently, we showed that LC3-II protein levels, stimulated by E. piscicida infection, were further enhanced by IFN-γ in an additive manner, which was impaired by the addition of 3-MA (Fig. 5B). Upon bacteria challenge, IFN-γ dramatically increased the colocalization of RFP-labeled E. piscicida with AF488-immunolabeled LC3 in an additive manner (Fig. 5C), whereas the number of LC3⁺ bacteria (Fig. 5D) and the percentage of LC3⁺ bacteria (Fig. 5E) were significantly increased by IFN-γ. However, these stimulatory effects were hindered by 3-MA (Fig. 5D, 5E). Additionally, IFN-γ dramatically induced the colocalization of E. piscicida–RFP with AF488-immunolabeled NOD1 (Fig. 5F). Moreover, IFN-γ treatment significantly enhanced the quantity of NOD1⁺ bacteria (Fig. 5G) and the percentage of NOD1⁺ bacteria (Fig. 5H).

The modification on the interaction between autophagy and E. piscicida by IFN-γ suggested that this cytokine may directly induce autophagy and further modulate E. piscicida survival. To verify this hypothesis, the effect of IFN-γ on autophagic response was assessed in grass carp monocytes/macrophages. Not unexpectedly, an 8-h treatment with IFN-γ (100–500 ng/ml) significantly enhanced LC3-II protein levels (Fig. 6A). Moreover, when cells were treated with IFN-γ (500 ng/ml) in the presence or absence of with 3-MA (10 mM) for 8 h, the quantity of AF488-immunolabeled LC3 puncta was enhanced by IFN-γ, and this enhancement was attenuated by 3-MA (Fig. 6B, 6C). To confirm the role of autophagy in mediating E. piscicida clearance, a canonical autophagy inducer, rapamycin, was used in this study. The results showed that CFU counts of bacteria were significantly reduced (∼40%; p < 0.05) when monocytes/macrophages were treated with 10 µM rapamycin (Fig. 6D). In parallel, IFN-γ (500 ng/ml) significantly impeded E. piscicida survival, and the impediment was completely reversed by 3-MA (Fig. 6E). Given that IFN-γ is well known to induce NO production, we evaluated whether the bactericidal activity of IFN-γ is associated with the induction of NO. As shown in (Fig. 6F, L-NMMA (500 µM), an NO synthase inhibitor, did not alter the effect of an 8-h treatment with IFN-γ (500 ng/ml) in eliminating the CFU count of E. piscicida in monocytes/macrophages. In fact, IFN-γ (500 ng/ml) did not induce NO production within the first 8 h of incubation (Fig. 6G). Nevertheless, it significantly promoted NO production from 24 to 72 h, which was abrogated by L-NMMA (500 µM) (Fig. 6G).

As one of the main pathogenic bacteria in fish species, E. piscicida can survive and multiply in various fish cells (19–21), indicating the need for elucidating the cellular mechanism responsible for this bacterium infection. Currently, it seems that much attention is focused on the evaluation of cellular phagocytosis and its impact on the uptake of E. piscicida in fish cells (22–24). However, the autophagic response upon E. piscicida infection remains unknown. Even in the most recent studies, transcriptome analysis on E. piscicida–infected carp species Labeo catla (27) and mudskipper monocytes/macrophages (24) does not provide information on autophagic response. In this study, we unveiled an E. piscicida–driven autophagy process in which LC3 puncta and bacteria-autophagosome interaction were detected in fish cells. Besides, the colocalization of E. piscicida with LysoTracker-labeled lysosomes was observed in grass carp monocytes/macrophages (Fig. 1), implying an autophagy-driven lysosomal degradation of E. piscicida in host cells.

We hypothesized that NOD1 functions as a cytosolic sensor to interact with E. piscicida in fish. In mammals, NOD1 is involved in the process of autophagy by recruiting the critical autophagy protein ATG16L1 to the plasma membrane at the site of bacterial invasion (6), supporting the notion that NOD1 connects intracellular bacterial sensing with the induction of autophagy (28). This intracellular PRR-autophagy axis is a highly conserved mechanism. As an example, intracellular receptors in fruit fly can recognize peptidoglycan produced by Gram-negative bacteria and trigger autophagy, which in turn constrained L. monocytogenes infection and promoted survival of the host (29–31). Even in plants, R proteins (homologous to NOD proteins) play a role in pathogen-triggered hypersensitivity response in Arabidopsis thaliana through autophagy (32). In line with these findings, our study demonstrated direct interaction of NOD1 and ATG16L1, as well as colocalization of the NOD1-ATG16L1 complex with E. piscicida in the cells (Fig. 2). These results uncover the autophagy machinery targeting E. piscicida in fish.

The close cross-talk of autophagy and E. piscicida posed a question of whether this bacterium suppressed host autophagy. With respect to the role of the NOD1-ATG16L1 axis in E. piscicida–triggered autophagy, this bacterium significantly repressed transcription and protein synthesis of ATG16L1 and NOD1 in grass carp monocytes/macrophages. In contrast, virus, LPS, and polyinosinic-polycytidylic acid treatment stimulated the mRNA expression of NOD1 in grass carp head kidney, trunk kidney, and spleen (33), indicating a possible discrepancy in the regulatory mechanisms on NOD1 mRNA expression under different conditions. In this study, we revealed an ESRRA-dependent mechanism monitoring NOD1 and ATG16L1 expression using an ESRRA inhibitor and overexpressing ESRRA (Fig. 3). In support of our results, ESRRA contains a DNA-binding domain involving transcriptional regulation of ATG genes (26) and can promote innate host defense against microbes through autophagy in macrophages and intestine (34). Furthermore, given that the heat-inactivated E. piscicida did not alter the protein expression of NOD1 (Fig. 4C), the possible regulation of live E. piscicida might be subject to its virulence factors. Among 21 virulence factors tested, only EVPC and ESCB, which belong to the type VI secretion system and chaperone proteins in the bacteria, respectively, displayed inhibitory effects on NOD1 protein expression. Further studies demonstrated that they reduced NOD1 protein levels by Ub-dependent proteasomal degradation (Fig. 4). In general, the downregulation of PRRs is not beneficial to host defense because PRRs are indispensable for detecting pathogens and triggering immune responses. For instance, Ub-dependent proteasomal degradation of NOD2 causes a greater bacterial load in the host (35, 36). At present, ESCB is known as a chaperone that binds to another toxin protein, ESEG, to facilitate their secretion (37), whereas EVPC is suggested to modulate the replication rates of E. piscicida in gourami phagocytes (38). Therefore, our results disclosed novel aspects of ESEB and EVPC to help the bacterium evade immune attack by disturbing autophagy. These findings suggested that cross-talk between autophagy and E. piscicida might be a crucial aspect of the innate immunity in fish.

The above results prompted us to decode the intrinsic factors of host governing the interplay of autophagy and E. piscicida. In the current study, IFN-γ was proved as a natural regulator involving this interplay because the macrophage activation is predominantly induced by IFN-γ, and this cytokine appears to be specifically momentous to vertebrates combating intracellular pathogens (39). In support of our speculation, E. piscicida challenge robustly enhanced IFN-γ transcript expression in head kidney and monocytes/macrophages (Fig. 5A), and IFN-γ was able to amplify E. piscicida–induced autophagic responses in monocytes/macrophages (Fig. 6A–D). Additionally, unlike its upregulation on NOD1 and NOD2 expression in mammals (40), IFN-γ did not alter E. piscicida–inhibited NOD1 expression (data not shown). However, it could promote the colocalization of NOD1 and E. piscicida, as well as the number and percentage of NOD1⁺ bacteria (Fig. 5), supporting the notion that NOD1 was involved in the IFN-γ–manipulated interaction between autophagy and E. piscicida in fish. Along these lines, IFN-γ was proven to trigger the autophagic response, which was effective in inhibiting the survival of the intracellular E. piscicida in the same cell model. Similarly, IFN-γ has been reported to promote the eradication of intracellular bacteria, such as Mycobacterium tuberculosis (41) and Chlamydia (42), and simultaneously induce autophagy (43). In teleost, IFN-γ is effective at protecting olive flounder (44) against E. piscicida infection, but the exact mechanism remains elusive. Generally, IFN-γ–elicited NO production is considered a common mechanism for resisting bacterial infection (45). In this study, we found that the protective effect of IFN-γ against E. piscicida was not subjected to NO induction (Fig. 6), particularly highlighting the potential of using IFN-γ to activate autophagy in host defense against E. piscicida.

In summary, our findings demonstrated a cellular mechanistic link between NOD1-mediated autophagy and E. piscicida infection and suggest that autophagy is a critical defense pathway against E. piscicida pathogenesis. Consequently, enhanced autophagy might be a new strategy for preventing E. piscicida infection.

A pathogenic strain of E. piscicida, EIB 202, was a generous gift from Prof. Qin Liu at East China University of Science and Technology (Shanghai, China).

The authors have no financial conflicts of interest.