Subunit vaccine candidates against Aeromonas salmonicida in rainbow trout Oncorhynchus mykiss

Aeromonas salmonicida subsp. salmonicida is the etiological agent of furunculosis and a major fish health problem in salmonid aquaculture worldwide. Injection vaccination with commercial mineral oil-adjuvanted bacterin vaccines has been partly successful in preventing the disease but in Danish rainbow trout (Oncorhynchus mykiss, Walbaum) aquaculture furunculosis outbreaks still occur. In this study we tested the efficacy of experimental subunit vaccines against A. salmonicida infection in rainbow trout. We utilized in silico screening of the proteome of A. salmonicida subsp. salmonicida strain A449 and identified potential protective protein antigens that were tested by in vivo challenge trial. A total of 14 proteins were recombinantly expressed in Escherichia coli and prepared in 3 different subunit vaccine combinations to immunize 3 groups of rainbow trout by intraperitoneal (i.p.) injection. The fish were exposed to virulent A. salmonicida 7 weeks after immunization. To assess the efficacy of the subunit vaccines we evaluated the immune response in fish after immunization and challenge infection by measuring the antibody levels and monitoring the survival of fish in different groups. The survival of fish at 3 weeks after challenge infection showed that all 3 groups of fish immunized with 3 different protein combinations exhibited significantly lower mortalities (17–30%) compared to the control groups (48% and 56%). The ELISA results revealed significantly elevated antibody levels in fish against several protein antigens, which in some cases were positively correlated to the survival.


Introduction
Aeromonas salmonicida subsp. salmonicida (hereafter AS) is the causative agent of typical furunculosis in aquacultured salmonid fish. Furunculosis causes bacterial septicemia that leads to significant economic losses due to fish morbidity and mortality [1].
Vaccination of salmonids against furunculosis is generally applied with injection vaccines containing formalin-killed AS bacteria combined with mineral oil adjuvant. These vaccines a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 provide efficient protection and induce long-lasting immunity against the bacterium under certain conditions [2]. However the prophylactic effect of the vaccines in Danish rainbow trout is suboptimal under field conditions and the maricultured fish still experience furunculosis outbreaks during the warmer summer months [3,4]. Furthermore, adverse side-effects like intra-abdominal lesions, retarded growth, pigmentation and autoimmunity [2,[5][6][7][8][9] have been associated with vaccine administration. Consequently, research for developing more effective furunculosis vaccines with fewer side-effects is ongoing.
In this study we applied an in silico approach to select potential vaccine candidates for experimental furunculosis vaccines and selected 14 proteins for in vivo trial. The proteins were recombinantly expressed in E. coli and prepared in 3 different vaccine combinations to immunize groups of rainbow trout by intraperitoneal (i.p.) injection. The vaccine efficacy was assessed by infection trial and by measuring the antibody reactivity in immunized fish on grounds that the antibody response has in several studies confirmed to be closely correlated to protection [14,37,38].

Rationale selection
The functionality and domain classification of 14 proteins (Table 1) was conducted by Inter-Pro [39]. The subcellular localization of the individual proteins was predicted by the CELLO and pSORTb predictor [40,41] that provide an output with reliability score for each location of each protein. Commonly, protective B-cell protein antigens are located in the outer membrane and extracellular environment, hence these predicted subcellular locations were targets for selection [42,43]. The conservation study was done by local sequence alignment (Smith-Waterman algorithm) [44] based on the amino-acid sequence of the protein antigens tested for conservation across the public available NCBI chromosome data. The pair-wise alignment values (% similarity and % coverage) were calculated and the most similar protein to a given genome was selected and classified as homologous if the % similarity was >75% on at least 75% of the total protein length (% coverage). Protein conservation correlates with an increased probability of success, due to the ability to elicit protection across different bacterial strains [45]. Therefore, we performed a conservation analysis of the initially selected A. salmonicida proteins across the identified incomplete chromosome genome sequences available in the NCBI database. The primary strain (complete chromosome from NCBI) was strain A449 [46]. The 4 sub-strains (incomplete chromosomes from NCBI) were strain: 01-B526, CBA100, NBRC_13784 and pectinolytica_34mel.

Recombinant construct design
The 14 proteins were expressed in constructs based on conventional in silico analysis such as prediction of signal peptides (SignalP-4.1) [47], transmembrane regions (TmHmm-2.0) [48], non-classical secretion proteins (SecretomeP-2.) [49], functional and structural domains (InterPro [39] and DomCut [50]) ( Table 2). The rationale of expressing the protein in fragments was due to: 1) enhanced probability of expressing the native protein structure, 2) these authors are articulated in the 'author contributions' section. The commercial affiliation (Evaxion Biotech) did not provide financial support in this study. The commercial affiliation (Evaxion Biotech) provided the protein sequences tested in this study but played no role in data collection, analysis or decision to publish.

Competing interests:
The submission is related to a Evaxion Biotech's patent entitled: Vaccines against Aeromonas salmonicida infection. Patent application number: 20825EP00. This does not alter our adherence to PLOS ONE policies on sharing data and materials. expressing the protective part of the protein, 3) establishing a successful recombinant expression in E. coli. Predicted signal peptide or/and transmembrane and intracellular regions were removed. Proteins bigger than 1000aa were split in N-and C-terminal fragments and only the fragment comprising the predicted active functional site domains was expressed. The splitting point of these fragments where decided based on structural domain predictor aiming at keeping the native structure of each structural domain.

Protein expression
All recombinant protein constructs were expressed in E. coli by Creative Biomart (Shirley, NY, USA).

Vaccine preparation
The 14 recombinant protein constructs were allocated in 3 groups as follows: VacA and VacB (both 5 proteins) and vacC (4 proteins) ( Table 2). A total of 25 μg of an individual protein was prepared per fish in the vaccine mixture corresponding to a total of 100-125 μg of mixed proteins per fish. The proteins were allowed to bind to Al(OH) 3 by adding aluminium hydroxide gel adjuvant Alhydrogel (Brenntag, Denmark) to each vial of mixed proteins. A volume of 100 μL Al(OH) 3 was added per 160 μg protein and incubated at room temperature with endover-end rotation for 1 hour. The vials were then centrifuged at 216 G (Heraeus multifuge 3 L-R, Thermo Scientific, Denmark) at room temperature for 2 minutes. The supernatant was removed and checked for the absence of proteins with NanoDrop 2000 (Thermo Scientific, Denmark). The vials containing proteins bound to Al(OH) 3 were washed twice with 0.9% sterile NaCl-each time all vials were centrifuged at 216 G for 2 minutes at room temperature whereafter the supernatant was removed. After the last wash an equal volume of Freund's Incomplete Adjuvant (FIA) (F5506, Sigma-Aldrich, Denmark) was added and the vials were vigorously vortexed at room temperature for 1 hour. The vaccine control containing only adjuvants was made accordingly without the protein constructs. The vaccines were stored at 4˚C until the further use.

Fish
Disinfected eyed rainbow trout eggs originating from Fousing Trout Farm, Jutland, Denmark were translocated and hatched in a pathogen-free rearing facility at Bornholm Salmon Hatchery (AquaBaltic, Denmark). Details of retaining the pathogen-free status have previously been described [51]. Fish were immunized and then reared for 6 weeks (568 degree-days) in the system containing recirculated municipal water at 13-14˚C in 700 L (1 m 3 ) tanks and fed 1% biomass per day with dry pellet feed (BioMar A/S, Denmark). Fish were transported to the fish keeping facility at the University of Copenhagen, Frederiksberg, Denmark and acclimatized by gradually raising the water temperature to 19˚C over the course of the week before the challenge (exposure to live AS) was performed. The study was approved under the license no. 2015-15-0201-00655 issued by the Animal Experiments Inspectorate, Ministry of Environment and Food, Denmark.

Vaccination
A total of 360 rainbow trout (average body weight of 30 g) were randomly divided into 6 groups of 60 fish (Table 3)

Challenge experiment
At 7 weeks post-immunization (wpi) (690 degree-days) the 25 fish (now with an average body weight of 40 g) in all the duplicate tanks were challenged with 5.4×10 8 CFU mL -1 AS strain 090710-1/23 by a challenge method using a multi-puncture device [37] producing 10 perforations in the tail fin with subsequent exposure to bacteria, for 90 seconds. Morbidity was monitored every second hour during 3 weeks post-challenge (wpc). The moribund state was defined as a complete loss of equilibrium, strong discoloration and development of skin hemorrhages. Moribund fish were immediately removed for euthanasia (300 mg L -1 MS222) and recorded as mortalities. Swabs from head kidney of all freshly euthanized fish were plated onto 5% blood agar plates (SSI Diagnostica, Denmark) for bacteriological analysis. The confirmation of AS was done according to Dalsgaard and Madsen [4] and the re-isolated bacteria from the dead fish were confirmed as A. salmonicida subsp. salmonicida. The challenge trial was terminated 3 wpc (24 days).

Sampling
Blood samples for ELISA (10 fish per group) were taken after vaccination (6 wpi) and after challenge (3 wpc) ( Table 3). Blood were collected by caudal vein puncture from euthanized fish (300 mg L -1 MS222), allowed to clot at 4˚C overnight and serum was separated at 4˚C by centrifugation (3000 G) for 10 minutes and stored at -80˚C until further analysis.

ELISA
Enzyme-linked immunosorbent assay (ELISA) was performed according to previously established protocol [52]. In brief, the 96-well microtiter plates (MaxiSorp™, Nunc, Denmark) were coated overnight at 4˚C with either sonicated lysate of AS bacteria strain 090710-1/23 (protein conc 5 μg mL -1 ) or individual recombinant proteins (conc 1 μg mL -1 ) and blocked with 2% Bovine Serum Albumin (BSA, A4503, Sigma-Aldrich, Denmark) for 1 hour at room temperature. A total of 9 proteins (contained in VacB and VacC vaccines) showing best protection were chosen for plate coating for ELISA. Working dilutions were selected considering preliminary results. Dilutions 1:500 and 1:5000 were chosen for testing specific antibodies against sonicated bacteria and dilutions 1:100, 1:1000 and 1:10 000 were chosen for testing specific antibodies against individual proteins. All serum samples were diluted with assay diluent (0.1% BSA in wash buffer) and 100 μL of sample was added to each duplicate ELISA plate well and incubated at 4˚C overnight. The plates were thereafter incubated with 100 μL mouse anti-salmonid Ig (MCA2182 AbD Serotec, Germany, diluted 1:500) and 100 μL HRP-conjugated rabbit antimouse IgG for 1 hour (STAR13B, AbD Serotec, Germany, diluted 1:500). The color reaction was developed with 100 μL tetramethylbenzidine (TMB) PLUS substrate (BUF042A, AbD Serotec, Germany) and stopped after 10 minutes by adding 100 μL 1N HCL. All ELISA plates included 4 wells of pure assay diluent for the background absorbance measurement. The optical density (OD) was measured at 450 nm in an Epoch spectrophotometer (BioTek, USA) in duplicate wells.

Statistical analysis
The relative percentage of survival (RPS) was calculated as proposed by Amend [53]: RPS = [1-(% mortality in vaccinated fish / % mortality in control fish)] × 100. All statistical tests were performed using GraphPad Prism version 4.00 for Windows (GraphPad Software, USA, www. graphpad.com) and P-values < 0.05 were considered statistically significant. Mortality data was analyzed using Kaplan-Meier survival analysis and log-rank test. Duplicate groups were pooled after survival curve comparison showed no significant difference in mortalities. ELISA results were compared using one-way ANOVA followed by Tukey's Multiple Comparison post hoc test. Fish from duplicate tanks were pooled after t-test had confirmed no significant difference between them. Correlation was calculated as Pearson correlation coefficient (r).

Challenge experiment
All fish were challenged at 7 wpi by multi-puncture of the upper caudal fin and local exposure to a known number of AS. The mortality post-challenge (pc) is shown in Fig 1. Mortality started at day 3 pc in all fish groups and subsided by day 6 pc. Thereafter minor mortality still occurred until day 15 pc. The highest cumulative mortality was recorded in the saline control group (56%) followed by the adjuvant only control group (48%). Mortalities in the experimental subunit vaccine groups VacA (30%, RPS 46), VacB (20%, RPS 64) and VacC (17%, RPS 70) were significantly lower than in the saline control group. The lowest cumulative mortality, but ELISA ELISA analyses were conducted to measure the antibody response towards sonicated bacteria and individual recombinant proteins in different groups of immunized fish. Here we highlight only significant differences (p<0.05) in antibody response that occurred when comparing the immunized fish groups (Alpha Ject 1 3000, VacA, VacB, VacC) with the control fish groups (saline control and adjuvant control). Antibodies against sonicated AS. Antibody levels against sonicated bacteria in sera (diluted 1:500) from fish vaccinated with Alpha Ject 1 3000 were significantly higher compared to all other groups at 7 wpi and 3 wpc (Fig 2). A slight but non-significant antibody increase against sonicated bacteria was seen in the groups immunized with the subunit vaccines when compared to the control groups.
Antibodies against 9 recombinant proteins. The fish immunized with the Alpha Ject 1 3000 had significantly higher antibody response against all 9 recombinant proteins when compared to the control groups at 7 wpi and 3 wpc. Fish immunized with experimental vaccines (recombinant proteins) showed diverse antibody responses to these proteins (Figs 3 and 4).

Discussion
This study utilized recombinant proteins encoded by the genome of A. salmonicida subsp. salmonicida A449 for immunization against furunculosis. The in silico approach also known as reverse vaccinology [54] has been used to identify vaccine candidates for bacterial fish diseases in few studies previously [55][56][57] but to the best of our knowledge-this is the first time it has been applied for A. salmonicida subsp. salmonicida. The 14 antigens selected for this study were presumed to be associated with hemoglobin binding, adhesion, iron-acquisition, calcium and siderophore binding, motility, enzymatic and collagenase activity and possess the right characteristics to be surface exposed proteins. The protective properties of these 14 recombinant protein antigens were tested by in vivo challenge experiment. The experimental setup included 3 groups of fish immunized with protein antigens, one group immunized with a commercial vaccine Alpha Ject 1 3000 and 2 control groups injected with either saline or adjuvant only. Antigens administered in experimental subunit vaccines induced protection in the immunized fish. Thus, all the groups of fish immunized with experimental vaccines experienced significantly lower mortalities compared to the saline control group or the group injected with adjuvant (Fig 1). The fish vaccinated with Alpha Ject 1 3000 had significantly higher antibody levels against sonicated bacteria when compared to all the other groups ( Fig  2) and performed best (mortality 8%) throughout the challenge study but not significantly better compared to the vaccine groups VacB and VacC (20% and 17% mortality, respectively). The experimental vaccine group VacA exhibited the highest mortality (30%) among the groups immunized with in silico predicted antigens but the mortality was still significantly lower compared to the control fish injected with saline (56%).
The ELISA analysis measuring antibody response against the individual proteins showed that fish vaccinated with the commercial vaccine had significantly higher antibody levels against all individual proteins included in experimental vaccines VacB and VacC when compared to the saline control group. The fish immunized with the experimental vaccine VacC reacted with elevated antibody levels against ASA_1675-a protein that was included in the VacC vaccine formulation. Antibody reactivity against ASA_1675 was significantly higher both after the immunization and challenge when compared to the control groups and other experimental vaccine groups (Fig 4). This suggests that ASA_1675 is highly immunogenic and may play a role in protection considering VacC vaccine fish exhibited the lowest mortality among the experimental vaccine groups and that the antibody levels against ASA-1675 measured before challenge were positively correlated with the fish survival (Fig 5). However as the experimental vaccines consisted of different proteins, their individual contribution to vaccine induced protection is uncertain and the role of ASA_1675 should be assessed in further studies. The protein ASA_1675 is a hemolysin-type calcium binding repeat-containing protein belonging to a family of putative RTX (repeats-in-toxin) proteins involved in various biological functions like S-layer formation, adhesion, pore formation, hemolysis, cytolysis etc. [58]. The RTX family proteins are secreted by the type I secretion system (TISS) of Gram negative bacteria and have shown to be important virulence factors in various bacteria [58,59] including Aeromonas hydrophila ATCC 7966 T [60] and Vibrio anguillarum [61]. The secretion of the protein in AS seems to occur only in vivo [62]. Other proteins in VacC vaccine formation mentioned previously in relation to AS virulence [46] are ASA_3328 and ASA_3723. The ASA_3328 is a putative heme receptor (hupA) involved in heme utilization [63] and has observed exclusively up-regulated under in vivo and in iron-restricted conditions in vitro [16]. The ASA_3723 is a microbial collagenase orthologue to a putative virulence determinant AHA_0517 in A. hydrophila [60] and among one of the most secreted proteins in the supernatant of virulent AS [62,64]. The ASA_4105 is a M24/M37 family zinc metallopeptidase and has not been previously mentioned as potential virulent factor of AS but was chosen in VacB vaccine due to its outer membrane location and the conservation across A. salmonicida substrains.
The fish that were immunized with the VacB vaccine exhibited significantly higher antibody levels against 4 out of 5 proteins they were immunized with (Fig 4) but the antibody reactivity against these proteins was not as distinguished as the antibody response against ASA_1675 in fish vaccinated with VacC vaccine. The VacB proteins inducing higher antibody response in fish vaccinated with VacB vaccine included ASA_1342, ASA_2532, ASA_4042 and ASA_P5G035. ASA_1342 is a the polar flagella hook-length control protein (FliK) that has previously been mentioned as potential virulent factor of AS having homologues to flagella proteins in A. hydrophila [46]. ASA_2532 is a type IV pilus assembly protein (TapV) shown to contribute to virulence of AS by being engaged in bacterial adhesion [65] and is homologues to FimV in Pseudomonas aeruginosa [66] and similar to pilus assembly proteins in other pathogenic bacteria [67]. The other 3 proteins in VacB vaccine: the hypothetical uncharacterized protein ASA_4042, the conjugal transfer mating pair stabilization protein (TraN) ASA_P5G035 and the tolA protein ASA_0744 have not been previously implicated in AS virulence and were chosen for this study on the bases of conservation, functionality and cellular location. Furthermore, the tolA protein in E. coli has shown to be a major virulence factor [68]. The fish in VacB vaccine group experienced a mortality of 20% suggesting the protective nature (at least some) of the VacB proteins. In addition, the antibody reactivity against ASA_2532 and ASA_P5G035 before the challenge was positively correlated to the survival (Fig 5). Unexpectedly, the fish immunized with VacA vaccine developed significantly elevated antibody levels against few proteins in VacB vaccine-ASA_1342, ASA_P5G035 (after immunization) and ASA_0744 (after challenge) suggesting some unspecific binding of antibodies.
The VacA subunit vaccine included two secreted enzymes: the alpha-amylase (AmyA) ASA_3455 and the endochitinase (chiB) ASA_3320. Both have previously been mentioned as potential virulence factors of AS by Reith [46]. In addition, ASA_3320 has been found abundantly present in supernatant of virulent AS [62,64]. Another previously mentioned putative virulence factor included in the vaccine was ASA_0826-the RTX protein (asx) that belongs to the RTX family associated with virulence in numerous Gram-negative bacteria [58,69]. ASA_0826 is an orthologue to a putative virulence factor AHA_3491 in A. hydrophila [46,60] and have been found abundantly present in AS supernatant [62][63][64]. ASA_3883 -a TonB dependent ferric siderophore receptor-is an outer membrane protein and was chosen as vaccine candidate based on its conservation and functionality. In addition, siderophore receptors have been used as potential vaccine candidates against E. coli in mammalian studies [70][71][72]. ASA_2321 -a cytochrome C-552 (nrfA) was selected due to functionality and conservation among sub-strains. VacA vaccine proteins were not included in the ELISA analysis due to the lower protective effect and the data on antibody response is not available.
Fish vaccinated with the commercial vaccine was expected to develop a high antibody response against both sonicated bacteria and individual proteins as they were injected with whole cell bacterin, exposing them to a high concentration of a wide array of antigens compared to the fish immunized only with 100-125 μg of potentially immunogenic proteins. The latter fish exhibited a lower antibody response compared to the fish vaccinated with the commercial vaccine but still had high survival. This could indicate that the antibody response is not the only protective factor and that the cellular immunity could be involved in protection. The role of these antigens in vaccine protection should be further elucidated in future studies where proteins are evaluated individually in immunization assays.
Some of the candidate proteins have been previously mentioned as potential virulence factors and protective antigens [16,46,62,64,65] but the in vivo protective potential of those antigens has not been assessed before. This study showed that immunization with previously proposed and few new candidate proteins increased the survival of fish and the survival can be related to the antibody producing protective properties that these proteins may have.

Future perspectives
Recombinant AS 449 proteins, when applied in experimental vaccines, induce protection against furunculosis and could therefore be considered as vaccine candidates. However the protective input of each antigen needs to be evaluated by single protein immunization to confirm their individual contribution to protection. The relation between antigen concentration and protection should be determined and suitable adjuvants must be selected.
The side-effects in fish immunized with the subunit vaccines adjuvanted with a mixture of Al(OH) 3 and FIA were not assessed in this study. However, side-effects studies are crucial when testing new vaccine formulations. It is therefore necessary to develop and evaluate the side-effects of suitable adjuvants in future studies.