Interactome of E. piscicida and grouper liver proteins reveals strategies of bacterial infection and host immune response

The occurrence of infectious diseases is related to heterogeneous protein interactions between a host and a microbe. Therefore, elucidating the host-pathogen interplay is essential. We previously revealed the protein interactome between Edwardsiella piscicida and fish gill cells, and the present study identified the protein interactome between E. piscicida and E. drummondhayi liver cells. E. drummondhayi liver cells and bacterial pull-down approaches were used to identify E. piscicida outer membrane proteins that bind to liver cells and fish liver cell proteins that interact with bacterial cells, respectively. Eight bacterial proteins and 11 fish proteins were characterized. Heterogeneous protein-protein interactions between these bacterial cells and fish liver cells were investigated through far-Western blotting and co-immunoprecipitation. A network was constructed based on 42 heterogeneous protein-protein interactions between seven bacterial proteins and 10 fish proteins. A comparison of the new interactome with the previously reported interactome showed that four bacterial proteins overlapped, whereas all of the identified fish proteins were new, suggesting a difference between bacterial tricks for evading host immunity and the host strategy for combating bacterial infection. Furthermore, these bacterial proteins were found to regulate the expression of host innate immune-related proteins. These findings indicate that the interactome contributes to bacterial infection and host immunity.

annexin max1 and flavin-containing monooxygenase were reduced, while that of acetyl-Coenzyme acyltransferase and cathepsin K was elevated (Fig. 2F).
Response of E. piscicida outer membrane proteins to grouper plasma. We further investigated how E. piscicida outer membrane proteins respond to grouper plasma. Genes encoding these seven proteins were cloned from EIB202, expressed in E. coli and purified for antibody generation. The specificity of the prepared antibodies was validated using Western blotting. Only single stained bands were detected at the predicted molecular masses of the corresponding proteins ( Fig. 4A-D).
To explore how these proteins respond to the stress imposed by host's plasma, E. piscicida cells were exposed to grouper plasma. The abundance of EvpB, ETAE_0245, ETAE_1826 and OmpA was increased, whereas the expression of ETAE_3048, ETAE_2675 and ETAE_2572 was reduced (Fig. 4E). These results indicate that these outer membrane proteins are regulated by host plasma.

Interactions between E. piscicida outer membrane proteins and E. drummondhayi liver proteins.
To validate the interactions between E. piscicida and E. drummondhayi liver proteins, far-Western blotting was applied. Membranes containing E. piscicida outer membrane proteins were incubated with the purified recombinant liver proteins individually, and then detected separately using the antibodies against acetyl-Coenzyme A acyltransferase, glutaryl-CoA dehydrogenase, annexin max1, cathepsin K, hyperosmotic glycine-rich protein, hemoglobin beta chain, ribosomal protein S16, ribosomal protein S26, immunoglobulin mu heavy chain, and flavin-containing monooxygenase. Far-Western blotting analysis showed that the liver proteins were interacting with one or more bacterial outer membrane proteins, e.g. acetyl-Coenzyme A acyltransferase interacted with ETAE_1826; glutaryl-CoA dehydrogenase interacted with ETAE_1826, OmpA and ETAE_0245; annexin max1 interacted with EvpB, ETAE_1826 and OmpA; cathepsin K interacted with all seven proteins; hyperosmotic glycine-rich protein interacted with EvpB, ETAE_1826 and OmpA; hemoglobin beta chain interacted with all of these proteins except for ETAE_0245; ribosomal protein S16 interacted with EvpB, ETAE_3048, OmpA and ETAE_2675; ribosomal protein S26 interacted with ETAE_1826, ETAE_3048, OmpA and ETAE_2572; immunoglobulin mu heavy chain interacted with EvpB, ETAE_3048, OmpA and ETAE_2675; flavin-containing monooxygenase interacted with all of the seven proteins (Fig. 5A). Furthermore, the interactions of cathepsin K to EvpB, ETAE_3048 and ETAE_0245 were confirmed by co-immunoprecipitation (Fig. 5B). Thus, the interactome  Table 1. Identification of protein bands isolated from liver cells that bound to E. piscicida*. *Reliability of these proteins was further validated by Western blot using these bound proteins as antigens in Fig. 2D(b).
between the seven bacterial proteins and ten grouper liver proteins was constructed consisting of 42 heterogeneous protein-protein interactions, as shown in Fig. 5C.
Investigation of zebrafish, D. rerio, innate immune-related proteins in response to outer membrane proteins. To investigate the role of bacterial outer membrane proteins involved in host-pathogen inteaction, D. rerio were randomly divided into groups and acclimatized for one week. D. rerio were separately injected with the recombinant outer membrane proteins ETAE_0245, ETAE_1826, EvpB, ETAE_2572, ETAE_2675, ETAE_3048 and OmpA emulsified with sterile montanide IMS 1312 VG, where the negative control group was injected with montanide IMS 1312 VG only. The body fluids of the D. rerio were drawn 72 hrs post-injection. the abundance of innate immune-related proteins, Ucp2, IL-1β , Bcl 2 and CC-chemokine, were measured by Western blot. ETAE_0245 upregulated the abundance of Ucp2, IL-1β and CC-chemokine; ETAE_1826 and ETAE_2675 upregulated the abundance of Ucp2, IL-1β and CC-chemokine but down-regulated Bcl2; EvpB upregulated IL-1β and CC-chemokine but down-regulated Ucp2 and Bcl2; ETAE_2572 upregulated the level of Ucp2, and CC-chemokine but down-regulated the level of IL-1β and Bcl2; ETAE_3048 upregulated Ucp2 and Bcl2 but down-regulated CC-chemokine; OmpA upregulated the level of Ucp2 and IL-1β but  down-regulated Bcl2 and CC-chemokine (Fig. 6). These results indicate that the outer membrane proteins elicit an innate immune response.

Discussion
Bacterial pathogenesis has been extensively studied at molecular and cellular level. Many molecules have been defined during bacterial infection, e.g. toxins, effectors by type III secretions system. However, very few studies have been done at tissue level, which is rich of information regarding the interplay between host and pathogen at the onset of bacterial invasion. Thus, we considered the pathogen and tissue represents as two independent entity to invesitigate how the interplay occurred. To do that, we developed a pull-down-based 1-DE proteomics approach to characterize the heterogeneous interactome between host proteins and bacterial outer membrane proteins, representing the frontlines of interaction for both of host and pathogen. In a previous study, we investigated the interaction between E. drummondhayi gill cells and E. piscicida outer membrane proteins, and constructed an interactome with five heterogeneous protein-protein interactions 22 . However, bacterial infections involve multiple events that include entering the host, circulating in the host, and colonizing to the target organs to establish local/systematic infections. The interaction of E. piscicida with E. drummondhayi gill only represents one of them. Liver is one of the organs being targeted by E. piscicida. Thus, elucidating the interplay between E. piscicida and liver protein would gain information on the pathogensis of bacteria but anti-infection strategy by the host. To achieve this, the whole bacteria or grouper liver proteins were used as bait protein to capture grouper liver proteins or E. piscicida outer membrane proteins, respectively. Bound proteins were identified through mass   Table 2. Identification of protein bands isolated from E. piscicida outer membrane proteins that bound to liver proteins*. *Reliability of these proteins was further validated by Western blot using these bound proteins as antigens in Fig. 4D.
Scientific RepoRts | 7:39824 | DOI: 10.1038/srep39824 spectrometry. To validate the interactions, far-Western blotting and Co-IP were applied. This is the first report of an interaction network between host liver proteins and bacterial outer membrane proteins. The constructed interactome have two distinct features. First, the protein-protein interactions are not restricted to one-to-one. Some of the proteins have multiple binding targets. Second, the abudance of the interacting proteins were dynamic rather than static, implying the active involvement of these proteins in bacterial infection or host anti-infection, e.g. 1) interactions of increased host proteins with increased outer membrane proteins, such as acetyl-Coenzyme A acyltransferase with EvpB; 2) interactions of increased host proteins with decreased outer membrane proteins, such as cathepsin K with ETAE_2675; 3) interactions of decreased host proteins with increased outer membrane proteins, such as immunoglobulin mu heavy chain with EvpB; 4) interactions of decreased host proteins with decreased outer membrane proteins, such as immunoglobulin mu heavy chain with ETAE_2675. These results indicate that the heterogeneous interaction proteome may contribute to bacterial invasion and host immunity as a result of immunity or infection through these interactions.
Compared to the interactome of E. drummondhayi gill and E. piscicida outer membrane proteins 22 , some of proteins involved in grouper liver proteins and E. piscicida outer membrane proteins were identical. This feature implied that host might mount strategies involving different sets of proteins to eliminate bacterial pathogens in a tissue-specific manner. In the contrast, pathogens like E. piscicida use several key proteins to invade the host. In this study, ETAE_0245, ETAE_1826 (OmpS2), ETAE_2675 and OmpA are the key proteins by E. piscicida to interact with both of the grouper gill and liver. However, Pnp, EvpB, FliC and OmpF2 were only present in E. piscicida-grouper gill interactome, and EvpB, ETAE_2572 and ETAE_3048 were only detected in grouper  The purified recombinant bacterial outer membrane proteins were transferred to NC membranes and used as bait proteins to capture prey proteins from the purified recombinant liver proteins. Then, antibodies were used to recognize the prey proteins. Bands were detected using antibodies against acetyl-Coenzyme A acyltransferase, glutaryl-CoA dehydrogenase, annexin max1, cathepsin K, hyperosmotic glycine-rich protein, hemoglobin beta chain, ribosomal protein S16, ribosomal protein S26, immunoglobulin mu heavy chain and flavin-containing monooxygenase. 1: EvpB 2: ETAE_1826 3: ETAE_3048 4: OmpA 5: ETAE_2675 6: ETAE_2572 7: ETAE_0245 (B) Immunoprecipitation using antiserum against cathepsin K. Bands were detected using anti-EvpB, anti-ETAE_3048 and anti-ETAE_0245. C1, Control, C2, Test group. (C) Network of interactional proteins. Green represents proteins upregulated under external stimuli, while red represents downregulated proteins, and yellow represents proteins for which no change was detected.
Interestingly, several of the outer membrane proteins, OmpA, ETAE_1826 and EvpB, identified in the present study were not directly linking to E. piscicida infection. E. piscicida OmpA was detected in the outer membrane protein fraction using a two-dimensional electrophoresis-based proteomics approach and was shown to present immunogenic potential 24 . ETAE_1826 exhibits high homology to OmpS2, which is essential for a disseminated infection caused by pathogenic E. piscicida 25 . EvpB is a component of the type VI secretion system in E. piscicida and is vital for E. piscicida pathogenesis 26 . On the other hand, several of liver proteins are active in grouper immune response. The immunoglobulin mu chain is an indicator for the physiological maturity of the immune system, and more importantly, is the main immunoglobulin responsible for humoral adaptive immunity in most teleost fish 27,28 . Flavin-containing monooxygenase converts trimethylamine into trimethylamine oxide through an oxidation reaction, which is associated with LPS-induced inflammation 29 . Deficiency of glutaryl-CoA dehydrogenase causes type I glutaric acidemia, usually triggered by childhood infection 30 . The annexins are Ca 2+ -dependent phospholipid-binding proteins involved in many cellular processes, whose overexpression were observed in patients infected with Helicobacter pylori and channel catfish infected with Edwardsiella ictaluri 31,32 . Hyperosmotic glycine-rich protein contributes to the regulation of ion transfer in rainbow trout 33 . Hemoglobin β chain is involved in the stress response to changes in the environment 34 . Serum amyloid A-5 SAA is a common acute-phase protein and is activated in response to various sources of stress in fish 35 . Although the roles of these proteins in fish immunity have been well documented, our study neverthelessly expand those studies by providing the evidence that bacteria could directly interact with these proteins, which may trigger immune response.
The expression of the D. rerio innate immune-related proteins Ucp2, IL-1β , Bcl 2 and CC-chemokine was adjusted in response to the presence of bacterial outer membrane proteins, indicating that these bacterial proteins are important in inducing grouper immune response. Ucp2 is mitochondrial carrier protein, which controls immune cell activation and the production of mitochondrial reactive oxygen species, cytokines and nitric oxide 36,37 . Interleukin-1β is a critical cytokine associated with inflammation 38,39 . Bcl2 regulates outer mitochondrial membrane channel (VDAC) opening, which modulates the mitochondrial membrane potential and, thus, controls the production of reactive oxygen species (ROS) and the release of cytochrome C by mitochondria. Of notice, both of Bcl2 and cytochrome C are the potent inducers of cell apoptosis 40 . CC-chemokine prompts blood cell supplementation and activation when inflammation occurs in the body, which is an important component of fish innate immunity 41 . The alteration of these immune-related molecules strongly suggested that the bacterial outer membrane proteins are the potential antigens triggering immune response, thus being vaccine candidates. Immune response of fish to outer membrane proteins isolated from species of Edwardsiella have been reported 42,43 , where pooled outer membrane proteins were used instead of recombinant proteins. Although 20 μ g and 12.5-100 μ g outer membrane proteins were used in 40.2 g of Labeo robita and 16 g of Ictalurus punctatus, respectively, to achieve optimal immune efficiency, we and others found that 1.5 μ g of recombinant proteins are the optimal dose for 0.3-0.4 g of D. rerio to generate effective immune response 44 . Thus, these results suggest that dose of immunogens is relatd to fish species.
In conclusion, our study of establishing grouper liver protein-E. piscicida interactome has at least two important implications. First, as compared to our previous grouper gill-E. piscicida interactome, different host organs may adopte different strategies against the same pathogen, while bacteria may have several key proteins during infection. Second but not last, several of the outer membranes proteins were recognized by the host would trigger immune response, signifying the importance of these proteins in the initation of immune response by the host. Therefore, these proteins would be tested for their vaccine ability in future studies.

Infection of groupers with E. piscicida.
The bacterial strain used in current study was E. piscicida EIB202, whose complete genome sequence is available 45 . To propagate the bacterium, a single colony was inoculated to 5 mL of TSB medium, followed by shaking at 30 °C for 24 h at a speed of 200 rpm. Bacterial culture was collected through centrifugation at 6,000 g for 5 min at 4 °C, and was washed twice with 25 mL saline solution. The resulting bacterial cells were resuspended with saline solution and adjusted to OD 600 1.0. For bacterial infection, twenty groupers were anesthetized by immersion in 100 ng/mL of tricaine methanesulphonate (MS-222, Sigma, USA), and were infected with EIB202 via intraperitoneal injection (50 μ L 8 × 10 7 cells each), where the same volume of saline solution was used as a control. Fourty-eight hours post infection, E. drummondhayi were anesthetized similarily as described aboved for liver isolation, which follows the previously described protocol 46 . Both of the livers from dying E. drummondhayi and healthy E. drummondhayi were isolated immediately after anesthetia for protein preparation.
Isolation of grouper liver proteins. Liver proteins were prepared as described previously with a few modifications 46 . Briefly, five liver tissues were freshly collected from groupers, E. drummondhayi, and then Scientific RepoRts | 7:39824 | DOI: 10.1038/srep39824 washed in 0.85% sterile saline solution three times. Liquid nitrogen was added, followed by grinding in a Dounce tissue grinder. Next, 1 mL of protein extraction buffer (66 mmol/L Tris/HC1, pH7.2, 3% (v/v) NP40, 0.1 mmol/L PMSF) was added, and the samples were placed on in ice for 30 min. Finally, the supernatants were harvested via centrifugation of 8,000 g at 4 °C for 20 min, and the protein concentration in the supernatants was determined using the Bradford method 47 .

Isolation of bacterial outer membrane proteins through lauryl sarcosinate extraction. Bacterial
outer membrane proteins were prepared as described previously 39 . Briefly, a single colony was propagated in TSB medium at 30 °C for 24 h. The cultures were then diluted 1:100 using fresh TSB medium and grown to an OD600 of 1.0 at 30 °C. The bacterial cells were harvested via centrifugation at 6,000 g for 5 min at 4 °C. The bacterial pellet was washed in sterile saline buffer (0.85% NaCl) three times, then re-suspended in sonication buffer (50 mM Tris/HCl, pH 7.4) and disrupted through intermittent sonic oscillation of the power output 60% for 7 s with intervals of 7 s on ice for a total of 40 min. Unbroken cells and cellular debris were removed via centrifugation at 6,000 g for 15 min at 4 °C. The turbid supernatant was subjected to ultracentrifugation at 100,000 g for 1 h at 4 °C. The pellet was solubilized with 2% (w/v) sodium lauryl sarcosinate (Sigma) at 4 °C and incubated at room temperature for 40 min. Following ultracentrifugation at 100,000 g for 1 h at 4 °C, the resulting pellet was dissolved in sterile pure water (approximately 5 mg/mL) and stored at − 80 °C until use.
E. piscicida pull-down assay for isolation of E. drummondhayi liver proteins. The bacterial pulldown assay was performed as described previously 48 . E. piscicida EIB202 was grown overnight at 30 °C in a shaker bath. A fresh overnight culture seed was then inoculated into TSB medium (1% (w/v) peptone, 0.5% (w/v) yeast extract, 1% (w/v) NaCl, pH 7.4), cultured at 30 °C and grown to an OD600 of 1.0. The cultures were harvested through centrifugation at 6,000 g for 5 min at 4 °C and washed three times via resuspension in 0.85% (w/v) NaCl. These bacterial cells were then suspended in 1% oxymethylene (w/v), inactivated at 80 °C for 90 min and centrifuged at 6,000 g for 5 min at 4 °C. The pellet was resuspended in 0.05 M Tris-HCl (pH 8.0) including 4.5 M urea under gentle rotation for 4.5 h, followed by harvesting through centrifugation at 6,000 g for 5 min at 4 °C. After washing three times via resuspension in 0.85% NaCl, the bacterial pellet was mixed with either 2 mg E. drummondhayi liver proteins, as the treatment group, or 0.85% NaCl, as a negative control. The mixture was incubated for 1 h, followed by harvesting at 6,000 g for 5 min at 4 °C after being washed three times through resuspension in 0.85% NaCl. The pellet was resuspended in 0.05 M Tris-HCl (pH 8.0) containing 4 M urea, under gentle rotation for 1 h, and the proteins were harvested via centrifugation at 8,000 g for 5 min at 4 °C. The pellet was subsequently concentrated in a three-fold volume of acetone for 12 h at − 40 °C, collected via centrifugation at 8,000 g for 10 min at 4 °C and then resuspended in 20 μ L of 0.05 M Tris-HCl. The isolation was repeated twice.
Sepharose 4B-bound E. drummondhayi liver proteins for isolation of E. piscicida outer membrane proteins. Cyanogen bromide Sepharose 4B 0.5 g was sufficiently expanded in 1 mM HCl for 15 min, then added to the column and washed with 50 mL of 1 mM HCl. After washing with 20 mL of binding buffer, liver proteins were added, followed by shaking for 2.5 h. The column was then washed to remove free liver proteins and blocked with 0.1 M Tris-HCl for 3 h. Following balancing using a 10-fold volume of PBS buffer, 1 mL (1 μ g/μ L) outer membrane proteins were added, and the column was incubated for 2 h. A 10-fold volume of PBS buffer and an equal column volume of eluted buffer were used to wash and elute the unbound and bound outer membrane proteins, respectively. The elutes were centrifuged at 8,000 g for 5 min at 4 °C. The resulting bound proteins were concentrated in acetone at − 40 °C for 12 h, then collected through centrifugation at 8,000 g for 10 min at 4 °C and resuspended in 0.05 M Tris-HCl. The isolation was repeated twice.

SDS-PAGE and mass spectrometric (MS) analysis. A discontinuous electrophoresis buffer system
involving Laemmli buffer, 4% stacking gels and 12% resolving gels was used to separate the proteins. All samples were boiled for 5 min after the addition of sample loading buffer and subsequently electrophoresed at a constant voltage of 120 V for the resolving gels, until the tracking dye (bromophenol blue) reached the bottom of the gels. Protein bands were visualized by staining with Coomassie Brilliant Blue R-250. The resultant bands were excised from gels and digested with trypsin through a routine procedure 49 . The sample solution (30-100 ppm), together with equivalent matrix solution, was applied to the MALDI TOF-Target system using HCCA as a MALDI matrix for peptide mapping and was prepared for MALDI-TOF/MS analysis. MALDI-TOF spectra were calibrated using trypsin autolysis peptide signals and matrix ion signals. Proteins with low confidence were further identified using MALDI TOF/TOF. For MS/MS spectra, the five most abundant precursor ions per sample were selected for subsequent fragmentation, and 1000-1200 Da laser shots were accumulated per precursor ion. The criterion for precursor selection was a minimum S/N of 50. All MALDI analyses were performed with a fuzzy logic feedback control system (Reflex III MALDI-TOF system, Bruker) equipped with delayed ion extraction. Both the MS and MS/MS data were interpreted and processed using Cloning of genes, purification of recombinant proteins and preparation of antisera. Standard PCR and molecular biology protocols were used to amplify the ighM, acaA2, anxA4, ctsK, cirbp, rps16, rps26, hbbE2, fmo5, gcdH genes of E. drummondhayi, and ucp2, IL-1, bcl2 and CC-chemokine genes of D. rerio and ETAE_3048, ETAE_0960 genes of E. piscicida. Primers for these genes were designed according to the Epinephelus coioides, or Danio or E. tarda EIB202 ORF sequences released by GenBank (Supplementary Table 1). PCR fragments were detected through agarose electrophoresis and were directionally cloned into the pMAL-c2X plasmid for fish genes and pET-32a for E. piscicida genes. Recombinant plasmids were checked via digestion with restriction endonucleases and transformed into E. coli BL21. Overnight cultures of E. coli BL21 harboring recombinant plasmids were diluted 1:100 (v/v) in fresh LB containing ampicillin (100 μ g/mL), then incubated at 37 °C until the absorbent optical density reached 0.6 at 600 nm (OD600). Protein expression was induced with 1 mM isopropyl-β -D-thiogalactoside (IPTG, from BBI) for 5 h at 37 °C after the optimization of expression conditions, including the culture temperature, IPTG concentrations and IPTG-induced period. Bacterial cells were harvested through centrifugation at 10,000 g for 20 min at 4 °C and washed with 0.85% NaCl, then resuspended in 50 mM sodium phosphate buffer (pH 8.0) containing 8 M urea and incubated for 30 min in ice bath. The cell suspension was disrupted via sonication in an ice bath (350 W, 3 × 10 min), followed by centrifugation at 12,000 g for 20 min at 4 °C. The clarified supernatant was loaded into a column packed with Ni 2+ nitriloaceate, which was charged with 50 mM NiSO 4 , and purified through affinity chromatography on Ni-NTA super-flow resin according to the manufacturer's instructions (Qiagen, Germany). The concentrations of the proteins were determined via the Bradford method. Solutions were stored at − 80 °C until use. Antisera against the purified recombinant proteins were raised separately by immunizing mice with 100 μ g of purified protein emulsified with Freund's complete adjuvant. The first injection was followed by two other injections with Freund's incomplete adjuvant at intervals of two weeks. Sera were collected and stored at − 80 °C until use.
Exposure of E. piscicida to fish plasma. The exposure of bacterial samples to stress caused by fish plasma was carried out as described previously 50 . Briefly, 4 mL of OD600 1.0 E. piscicida EIB202 cells were collected and washed with saline solution. Then, 200 μ L of E. drummondhayi plasma or saline solution was added in the test and control groups, respectively. The mixtures were incubated at 37 °C for 2 h following resuspension, after which the cells were prepared for Western blotting.
Western blotting and far-Western blotting analyses. The two assays were performed as described previously 51,52 . Mouse antisera against the fish immunoglobulin mu heavy chain, acetyl-Coenzyme A acyltransferase, annexin max1, ribosomal protein S16, cathepsin K, hemoglobin beta chain, ribosomal protein S26, glutaryl-CoA dehydrogenase, hyperosmotic glycine-rich protein and flavin-containing monooxygenase and against E. piscicida EvpB, ETAE_1826, OmpA, ETAE_3048 ETAE_0245, ETAE_2675 and ETAE_2572 were used as the primary antibodies, and a horseradish peroxidase (HRP)-conjugated rabbit anti-mouse antibody was used as the secondary antibody. For Western blotting, proteins separated from gels were transferred to 0.22 μ m nitrocellulose (NC) membranes at a constant voltage of 80 V for 1 h at 4 °C, and membranes were stained with Ponceau S to evaluate the transfer efficiency. The membranes were then blocked overnight in 5% non-fat milk in Tris-NaCl-Tween (TNT) buffer at 4 °C. After rinsing three times for 15 min with TNT buffer, the membranes were separately incubated with mouse antibodies for 2 h on a gentle shaker at room temperature. The membranes were subsequently rinsed again and then incubated with rabbit anti-mouse horseradish peroxidase (HRP)-conjugated secondary antibodies for 2 h under the same conditions. The membranes were washed and developed with a dimethylaminoazobenzene (DAB) substrate system until the appearance of maximum color. All of the primary antibodies were diluted 1:100-200 in blocking buffer, and the secondary antibodies were diluted to 1:2000.
For far-Western blotting, the purified recombinant bacterial outer membrane proteins were transferred to NC membranes and used as bait proteins to capture prey proteins (purified recombinant liver proteins). The membranes were then washed with TNT and incubated with the same primary and secondary antibodies used in the Western blotting assay, after which they were developed with the DAB system as described above.
Coimmunoprecipitation (Co-IP). Co-IP was carried out as described previously 53 . Recombinant cathepsin K and E. piscicida outer membrane proteins were incubated at room temperature for 2 h on a gentle shaker. Then, 10 μ L of mouse antiserum against cathepsin K was added to the test group, and pre-immune antiserum was added to the control. After incubation under the same conditions, 20 μ L of nProtein A Sepharose 4 Fast Flow (Amersham Biosciences Corp.) was added, followed by incubation for 12 h at 4 °C on a gentle shaker. The nProtein A Sepharose 4 Fast Flow was collected via centrifugation at 3000 g for 5 min and cleaned six times for 10 min each with pH 7.0 Tris-HCl buffer, followed by incubation in 50 μ L of 1 mM pH 2.4 glycine-HCl buffer for 2 h at room temperature. After centrifugation at 3,000 g for 5 min, the supernatant was employed for Western blotting using anti-EvpB, -ETAE_3048 and -ETAE_0245 as the primary antibodies.
Innate immune response to recombinant outer membrane proteins. Vaccination was carried out as described previously 44 . D. rerio were acclimatized for one week and randomly divided into groups, ten each. These animals were intramuscularly injected injected with the recombinant outer membrane proteins (1.5 μ g per fish) ETAE_0245, ETAE_1826, EvpB, ETAE_2572, ETAE_2675, ETAE_3048 and OmpA emulsified with sterile montanide IMS 1312 VG (Seppic, France), where the negative control group was injected with montanide IMS 1312 VG only. No any toxicity effect on fish was observed. Humoral fluid was collected after 3 days and used for analysis of innate immunity through Western blotting.