Development and Characterization of a Novel Live Attenuated Vaccine Against Enteric Septicemia of Catfish

Edwardsiella ictaluri is a Gram-negative intracellular pathogen causing enteric septicemia of channel catfish (ESC). Type six secretion system (T6SS) is a sophisticated nanomachine that delivers effector proteins into eukaryotic host cells as well as other bacteria. In the current work, we in-frame deleted the E. ictaluri evpB gene located in the T6SS operon by allelic exchange. The safety and efficacy of EiΔevpB as well as Aquavac-ESC, a commercial vaccine manufactured by Intervet/Merck Animal Health, were evaluated in channel catfish (Ictalurus punctatus) fingerlings and fry by immersion exposure. Our results showed that the EiΔevpB strain was avirulent and fully protective in catfish fingerlings. The EiΔevpB strain was also safe in catfish fry, and immersion vaccination with EiΔevpB at doses 106 and 107 CFU/ml in water resulted in 34.24 and 80.34% survival after wild-type immersion challenge compared to sham-vaccinated fry (1.79% survival). Catfish fry vaccinated with EiΔevpB at doses 106, 107, and 108 CFU/ml in water exhibited dose-dependent protection. When compared with Aquavac-ESC, EiΔevpB provided significantly higher protection in catfish fingerlings and fry (p < 0.05). Results indicate that the EiΔevpB strain is safe and can be used to protect catfish fingerlings and fry against E. ictaluri.


INTRODUCTION
Channel catfish (Ictalurus punctatus) is the most significant aquaculture commodity in the United States, and Edwardsiella ictaluri causes enteric septicemia of channel catfish (ESC) (Hawke, 1979). The disease occurs as acute enteric septicemia or chronic encephalitis (Shotts et al., 1986). Use of antibiotic-added feed (Smith et al., 1994;Plumb et al., 1995) and feed restriction (Wise et al., 2004) are traditional means used to control ESC. Although these practices could reduce mortalities, feed restriction results in reduced production through lost feeding days. Medicated feed is expensive, useful only in fish that accept feed, and could yield antibiotic-resistant strains.
Vaccination is a vital prophylactic strategy for prevention of bacterial diseases in aquaculture (Shoemaker et al., 2009;Villumsen et al., 2014). Because E. ictaluri species has been shown to be very homogeneous (Plumb and Vinitnantharat, 1989;Bertolini et al., 1990), a vaccine strain could potentially provide wide-range of protection against different E. ictaluri strains in different fish species. The commercial ESC vaccine Aquavac-ESC (RE-33) was developed by serial passage in increasing concentrations of rifampicin . Aquavac-ESC is safe in catfish fry (Klesius and Shoemaker, 1999), but it has not been accepted widely due to marginal economic returns (Bebak and Wagner, 2012). Another live attenuated E. ictaluri vaccine was developed by serial passage on media containing increasing concentrations of rifamycin, which protected catfish fingerlings when added in feed and administered orally (Wise et al., 2015). Under current catfish production practices, catfish fry are transferred from hatchery to nursery ponds when they are 1-2week-old. Therefore, there is an urgent need for an effective ESC vaccine that can be delivered to catfish fry before their release into nursery ponds. Type six secretion system (T6SS) is a virulence factor for many pathogenic bacteria (Rao et al., 2004;Pukatzki et al., 2006). This system is highly conserved and widely distributed in Gramnegative bacteria as one or more copies (Bingle et al., 2008). T6SS delivers protein effectors into the periplasm of the target cells directly upon cell-to-cell contact. Therefore, contributing to different processes ranging from inter-bacterial killing to pathogenesis. The number of genes encoded within T6SS clusters usually varies between 16 and 38 genes (Cascales, 2008;Murdoch et al., 2011), with a minimal set of 13 genes required to assemble a functional T6SS (Boyer et al., 2009;Lin et al., 2013). T6SS is also required to kill other bacterial cells by secreting anti-bacterial proteins (Ho et al., 2014).
Our previous proteomics study showed that EvpB protein is differentially regulated during in vitro iron-restricted conditions (Dumpala et al., 2015). Thus, we hypothesized that EvpB protein could have a crucial role in the T6SS of E. ictaluri. In the current work, we report the construction of an evpB in-frame deletion mutant and its vaccine potential in catfish fingerlings and fry.

Sequence Analysis
The nucleotide sequences of the T6SS operon were obtained from the E. ictaluri strain 93-146 genome (GenBank accession: CP001600) (Williams et al., 2012). The Basic Local Alignment Search Tool (BLAST) was used to determine the sequence of the evpB open reading frame and adjacent sequences.

Construction of Ei evpB In-frame Deletion Mutant
In-frame deletion of the E. ictaluri evpB gene (NT01EI_RS11895) was accomplished by following the procedures described previously (Abdelhamed et al., 2013). Briefly, 1,210 bp upstream and 1,155 bp downstream regions of evpB were amplified from E. ictaluri strain 93-146 genomic DNA with A/B and C/D primer pairs ( Table 2), respectively. These PCR products were mixed, diluted, and used as template in a splicing overlap extension PCR (Horton et al., 1990) with the A/D primers to generate evpB deletion fragment (2,365 bp). The resulting evpB deletion fragment was cloned into pMEG-375 suicide plasmid. The resulting plasmid, pEi evpB, was transformed into E. ictaluri by conjugation, and transformants were selected on BHI agar containing Amp and Col at 30 • C for 2 days. To allow the second homologous recombination, a single Amp-resistant merodiploid colony was plated on BHI agar with sucrose and mannitol and incubated at 30 • C for 3 days. Amp-sensitive colonies were AATCTAGAATCGGCGACCAAACGTAAAG XbaI a Bold letters at the 5 end of the primer sequence represent restriction enzyme (RE) site added. AA nucleotides were added to the end of primers containing a RE site to increase the efficiency of enzyme cut. Underlined bases in primer C indicate reverse complemented primer B sequence. b RE stands for restriction enzyme.
screened by colony PCR using the A/D primers, and further confirmation was done by sequencing of A/D fragment.

Complementation of the evpB Gene
The 1,464 bp open reading frame of the evpB gene was amplified using primers listed in Table 2. The amplicon was cloned into a pBBR1-MCS4 plasmid (Kovach et al., 1995) at the SmaI and XbaI restriction sites. The resulting plasmid, pEievpB, was transferred to Ei evpB by conjugation. Successful transformation was verified by observing plasmid profile of Ei evpB. The resulting strain was designated as Ei evpB+pEievpB.

Determination of Safety and Efficacy of Ei evpB in Catfish Fingerlings
All fish experiments were approved by the Institutional Animal Care and Use Committee at Mississippi State University (protocol numbers: 12-042, 15-043, and 17-288). Virulence and vaccine efficacy of the Ei evpB strain were assessed, as described (Abdelhamed et al., 2013). Briefly, 240 specificpathogen-free (SPF) channel catfish fingerlings (13.88 ± 0.27 cm, 27.77 ± 1.04 g) were stocked in 40 l flow-through tanks (20 fish/tank) with constant aeration and allowed to acclimate for 1 week. Water temperature was maintained at 26 ± 2 • C during the experiment. The tanks were randomly assigned into three groups, and each group contained four replicates. The three groups were Ei evpB, EiWT (positive control), and BHI (sham control). The fish were vaccinated by immersion for 1 h in water containing approximately 3.32 × 10 7 CFU/ml, and then flow-through conditions were resumed. Mortalities were recorded daily, and the presence of E. ictaluri was confirmed by streaking anterior kidney onto BHI plates. At 21-days postimmunization, the vaccinated fish were challenged with EiWT (3.83 × 10 7 CFU/ml in water) by immersion for 1 h as described above. Mortalities were recorded daily.

Determination of Safety and Efficacy of Ei evpB in Catfish Fry
Nine hundred 14-day-old SPF channel catfish fry were stocked in 18 tanks (50 fry/tank). Tanks were randomly assigned to six treatment groups with three replicates per group. Treatment groups consisted of high (3.32 × 10 7 CFU/ml in water) and low (3.32 × 10 6 CFU/ml in water) doses of Ei evpB, EiWT, and BHI. Immersion vaccination was conducted same as fingerling challenge described above. At 21 days post-vaccination, fry were challenged with EiWT by immersion exposure at approximately 3.10 × 10 7 CFU/ml in water. Mortalities were recorded daily.

Evaluation of Various Challenge Doses of Ei evpB in Catfish Fry
Vaccine efficacies of three separate doses of Ei evpB were evaluated in 7-day-old fry to determine the optimal dose. Briefly, 750 fry were stocked into 15 tanks (50 fry/tank). The tanks were divided into five groups with three replicates per group. Vaccinated groups consisted of three doses of Ei evpB (3.72 × l0 6 , 3.72 × l0 7 , and 3.72 × l0 8 CFU/ml in water), EiWT (positive control), and BHI (sham control). Fish were monitored daily, and mortalities were recorded from each tank. After 30 days post-vaccination, fry were challenged with the EiWT by immersion in water (3.80 × 10 7 CFU/ml) for 1 h. Mortalities were recorded daily.

Comparison of Ei evpB and Aquavac-ESC in Catfish Fingerlings
Vaccine efficacy of Ei evpB strain was compared with Aquavac-ESC in catfish fingerlings. Briefly, 320 channel catfish fingerlings (7.75 ± 0.08 cm, 4.50 ± 0.014 g) were stocked into 16 tanks (20 fish/tank). Each group included four replicate tanks. Vaccination groups consisted of Aquavac-ESC, Ei evpB, EiWT (positive control), and BHI (sham control). Fingerlings were vaccinated by immersion in water containing approximately 4.5 × 10 7 CFU/ml for 1 h. Fish were monitored, and dead fish were removed daily. After 21 days, immunized fish were challenged with EiWT by immersion in water with 3.80 × 10 7 CFU/ml for 1 h. Mortalities were recorded daily.

Comparison of Ei evpB and Aquavac-ESC in Catfish Fry
Vaccine efficacy of Ei evpB was compared with Aquavac-ESC in 14-days post-hatch fry. Briefly, 800 channel catfish fry were stocked into 16 tanks (50 fish/tank). Each group included four replicate tanks. Vaccination groups consisted of Ei evpB strain, Aquavac-ESC, EiWT (positive control), and BHI (sham control). Fry were vaccinated by immersion (3.72 × 10 7 CFU/ml in water) for 1 h. Fish were monitored, and dead fish were removed daily. After 21 days, immunized fish were challenged with EiWT by immersion (3.80 × 10 7 CFU/ml in water) for 1 h. Mortalities were recorded daily.

Statistical Analysis
The percent mortality and survival values were arcsine transformed, and pairwise comparison of the means was performed with Tukey procedure. Analysis of variance (ANOVA) was done using PROC GLM in SAS for Windows v9.4 (SAS Institute, Inc., Cary, NC, United States) to assess significance. An alpha level of 0.05 was used in all analyses.

T6SS in E. ictaluri Genome
Analysis of the E. ictaluri genome revealed the presence of a 20,724 bp operon containing 16 genes (evpP, plus genes from NT01EI_RS11890 to NT01EI_RS11960) that encode for the T6SS apparatus, chaperones, effectors, and regulators (Figure 1).

Construction of the Ei evpB Mutant
We successfully introduced an in-frame deletion to the evpB gene in the E. ictaluri chromosome. The resulting Ei evpB strain contained a deletion of 1,167 bp out of 1,488 bp open reading frame (78.42%), resulting in loss of 389 amino acids from the E. ictaluri evpB gene. This in-frame deletion was verified by PCR and sequencing of the amplified fragment from the Ei evpB strain.

Virulence and Efficacy of Ei evpB in Fry
No mortality was observed in the fry vaccinated with Ei evpB at 10 6 and 10 7 CFU/ml in water. In contrast, 98.67 and 100% mortalities were observed in the fry exposed to EiWT at 10 6 and 10 7 CFU/ml in water doses, respectively ( Figure 3A). The fry vaccinated with Ei evpB showed 34.24% survival at 10 6 CFU/ml in water, and 80.34% survival at 10 7 CFU/ml in water at 21 days post-vaccination. On the contrary, the sham-vaccinated group had 1.79% survival ( Figure 3B).

Optimal Vaccine Dose of Ei evpB in Catfish Fry
Mortality in 7-day old fry vaccinated with Ei evpB (10 6 , 10 7 , and 10 8 CFU/ml in water) and sham group ranged between 2.53 and 5.13%, which was not statistically different. Dead fish collected from these treatments did not show any pathology, and E. ictaluri was not present in the fish. On the other hand, very high mortality (95.71%) was observed in EiWT exposed (10 7 CFU/ml in water) fry (Figure 4A). At 30-day post-vaccination with three different doses (10 6 , 10 7 , and10 8 CFU/ml in water), the percent survival in fry challenged with EiWT were significantly higher (57.74, 62.74, and 71.06%, respectively) compared to the sham-vaccinated fry (12.16% survival) (P < 0.05) (Figure 4B).

Comparison of Ei evpB to Aquavac-ESC in Catfish Fry
Ei evpB and Aquavac-ESC showed no mortalities, while negligible mortality was observed in the sham group (0.50%).

DISCUSSION
The primary objective of this study was to develop a live attenuated E. ictaluri vaccine strain based on mutation of the evpB gene, which is the second gene in the T6SS operon. The E. ictaluri EvpB protein was previously annotated as Eip55 (Moore et al., 2002), but it was not known that Eip55 was part of the T6SS. Eip55 is expressed during E. ictaluri infection and is antigenic to channel catfish. The percent sequence identity at the amino acid level between E. ictaluri and E. tarda EvpB (Rao et al., 2004) is high (96.5%). The first step of our work was construction of the Ei evpB strain by in-frame deletion of the evpB gene leaving 189 bp at the 5 end and 132 bp at the 3 end. Mutant construction did not introduce any selective antibiotics to the Ei evpB strain. Introduction of extraneous antibiotic resistance in vaccine strains is not desirable to avoid spread of antibiotic resistance genes. The Ei evpB strain was completely avirulent in catfish fingerlings. Moreover, vaccination of fingerlings with Ei evpB provided full protection against subsequent challenge with EiWT at 21 days post-vaccination. However, U.S. catfish production practices limit immersion vaccination to 7-14 days post-hatch. At this age, 1000s of fry are housed in hatchery tanks and can be vaccinated cost-effectively. To conform to industry practices, the Ei evpB was assessed in 14-day-old fry by immersion at two different doses (10 6 and 10 7 CFU/ml in water). The result demonstrated that the Ei evpB strain was completely attenuated in channel catfish fry at both doses. Following vaccination, we found 80.34% survival at the 10 7 CFU/ml in water dose and 34.24% survival at the 10 6 CFU/ml in water dose. These results demonstrate that 10 7 CFU/ml in water dose of Ei evpB could provide excellent protection levels in catfish fry against ESC.
Immunization of 7-day-old catfish fry with increasing doses of Ei evpB (10 6 , 10 7 , and 10 8 CFU/ml in water) demonstrated that the protection levels of Ei evpB, although was not statistically FIGURE 1 | Type VI secretion system organization in the E. ictaluri genome. Arrows indicate the direction of transcription, and numbers at the beginning and the end indicate genomic coordinates.     significant, tend to be higher with increasing vaccine dose. Challenge doses as high as 10 8 CFU/ml in water were safe under our experimental conditions, and the higher dose of Ei evpB elicited better protection against EiWT. However, subsequent immunizations were conducted using a dose of 10 7 CFU/ml in water, which is a more achievable dose for commercial vaccine manufacturing.
We also compared vaccine efficacy of Ei evpB to the commercial vaccine Aquavac-ESC in channel catfish fingerlings and fry by immersion. In both trials, Ei evpB provided better protection than Aquavac-ESC. Besides the superior performance of Ei evpB compared to Aquavac-ESC, Ei evpB does not have an added antibiotic resistance gene while Aquavac-ESC is rifampicin resistant. Also, Ei evpB has a known genotype, while the genetic basis for Aquavac-ESC attenuation is not described.
The results from fry and fingerling experiments showed that evpB is vital in E. ictaluri virulence, which is consistent with findings in E. tarda PPD130/91, where deletion of evpB led to reduced virulence in blue gourami and impaired replication in gourami phagocytes (Rao et al., 2004). Several T6SS proteins are important for bacterial pathogenesis. However, the function of most T6SS proteins remains unknown (Filloux et al., 2008;Silverman et al., 2012). This is the first report to our knowledge that evpB is required for E. ictaluri virulence. This is also the first study that linked T6SS and virulence in E. ictaluri.
The results shown here suggest that deletion of evpB gene provides an excellent live attenuated vaccine candidate. Live attenuated bacterial vaccines activate immune responses by mimicking the route of natural infection, possess intrinsic adjuvant properties, and can be administrated as mucosal vaccines. In the present study, only the immersion route of exposure was tested, which is the preferred vaccination method in commercial settings, because large numbers of small fish can be vaccinated quickly and cheaply (Stevenson, 1997;Chettri et al., 2013).
Live attenuated vaccines must achieve a precise balance between lack of pathogenicity and sufficient immunogenicity to provide protective immunity. An important consideration for the success of vaccination in catfish fry is that the immune system of fry may not be fully developed (Ellis, 1988). Therefore, fry may not be able to respond to vaccination effectively. The protective immunity in channel catfish is determined by cellular immune mechanisms, which generally precedes the development of humoral immunity Thune et al., 1997;Petrie-Hanson and Ainsworth, 2001). In previous studies, it has been reported that catfish fry failed to produce a significant antibody response before 3 weeks of age due to the poor organization of secondary lymphoid tissue (Patrie-Hanson and Jerald Ainsworth, 1999;Petrie-Hanson and Ainsworth, 2001). This could elucidate why attempts of early vaccination of fry are less likely to succeed. However, our results showed that vaccination of 2-week-old catfish fry with Ei evpB could provide very high protection. It is clear that use of Ei evpB at three or 4-week-old fry may provide full protection against EiWT, but current catfish practices do not permit housing fry in hatcheries beyond 10-14 days.

CONCLUSION
In conclusion, Ei evpB described in this work is entirely safe and provides full protection for catfish fingerlings. Also, Ei evpB is safe and highly protective in catfish fry. Further, Ei evpB is well characterized genetically and does not carry any additional antibiotic resistance. Field trials in earthen ponds will allow us to assess the commercial potential of Ei evpB under production conditions. Understanding the responses of the catfish immune system to Ei evpB vaccination will help us develop a better vaccination strategy.

AUTHOR CONTRIBUTIONS
AK and ML supervised the study. HA, ML, and AK designed the experiments and analyzed and interpreted the data. HA and AK performed the experiments. All authors wrote and approved the manuscript.