Universal Stress Proteins Contribute Edwardsiella ictaluri Virulence in Catfish

Edwardsiella ictaluri is an intracellular Gram-negative facultative pathogen causing enteric septicemia of catfish (ESC), a common disease resulting in substantial economic losses in the U.S. catfish industry. Previously, we demonstrated that several universal stress proteins (USPs) are highly expressed under in vitro and in vivo stress conditions, indicating their importance for E. ictaluri survival. However, the roles of these USPs in E. ictaluri virulence is not known yet. In this work, 10 usp genes of E. ictaluri were in-frame deleted and characterized in vitro and in vivo. Results show that all USP mutants were sensitive to acidic condition (pH 5.5), and EiΔusp05 and EiΔusp08 were very sensitive to oxidative stress (0.1% H2O2). Virulence studies indicated that EiΔusp05, EiΔusp07, EiΔusp08, EiΔusp09, EiΔusp10, and EiΔusp13 were attenuated significantly compared to E. ictaluri wild-type (EiWT; 20, 45, 20, 20, 55, and 10% vs. 74.1% mortality, respectively). Efficacy experiments showed that vaccination of catfish fingerlings with EiΔusp05, EiΔusp07, EiΔusp08, EiΔusp09, EiΔusp10, and EiΔusp13 provided complete protection against EiWT compared to sham-vaccinated fish (0% vs. 58.33% mortality). Our results support that USPs contribute E. ictaluri virulence in catfish.


INTRODUCTION
Enteric septicemia of channel catfish (ESC) is one of the most prevalent diseases of cultured catfish, causing significant losses (USDA, 2014). The most common practice in ESC treatment is use of feed medicated with oxytetracycline, sulfadimethoxine, or florfenicol. However, one of the earliest clinical signs of ESC is reduced appetite. Thus, these antimicrobials are only useful in limiting the spread of an outbreak and rather than treating the disease. Also, medicated feed may lead to the emergence of resistant Edwardsiella ictaluri strains (Tu et al., 2008).
The universal stress proteins (USP) have a conserved domain of 140-160 amino acids, and are present in archaea, bacteria, and plants (Nachin et al., 2005), but not in animals and human (Siegele, 2005). In Escherichia coli usp are involved in various functions from oxidative stress to adhesion and motility (Nachin et al., 2005). Under stress, USPs are overproduced and through a variety of mechanisms aid the survival of organism in stressful conditions (Heermann et al., 2009b). The uspA mutation caused decreased survival in E. coli (Tkaczuk et al., 2013). It is known that USPs are needed by pathogens (Hensel, 2009). USPs affect persistence and survival of Mycobacterium tuberculosis (Hingley-Wilson et al., 2010), and cause growth arrest and reduce the virulence in Salmonella typhimurium C5 (Liu et al., 2007) and Burkholderia pseudomallei (Al-Maleki et al., 2014). USPs are also necessary for the intracellular growth adaption of Listeria monocytogenes (Chatterjee et al., 2006). Similarly, Staphylococcus aureus virulence factors were downregulated in vivo while expression of uspA increased (Chaffin et al., 2012). Acinetobacter baumannii uspA is essential in pneumonia and pathogenesis (Elhosseiny et al., 2015).
Although increased expression of several usp genes in E. ictaluri under various stressors has been reported (Akgul et al., 2018), the role of USPs in E. ictaluri virulence is not known yet. Therefore, in this study, 10 E. ictaluri usp genes were studied by introducing in-frame deletions and determining their survival under acidic and oxidative stress conditions. Also, the virulence and protective properties of mutants against ESC infection were tested in catfish fingerlings.

Animals
All fish experiments were performed based on a protocol approved by the Mississippi State University Institutional Animal Care and Use Committee (protocol number 15-043). Channel catfish fingerlings were obtained from the fish hatchery at the College of Veterinary Medicine, Mississippi State University, and maintained at 25-28 • C during experiments. Tricaine methanesulfonate (MS-222, Western, Chemical, Inc.) was used to sedate (100 mg/ml) or euthanize (400 mg/ml) the catfish.

Construction of In-Frame Deletion Mutants
The nucleotide sequences of 10 E. ictaluri usp genes were obtained from the E. ictaluri 93-146 genome (GenBank accession: CP001600), and four primers were designed for each gene (Tables 2, 3). Restriction sites were included in forward and reverse primers. Overlap extension PCR was used to delete the functional usp genes from the E. ictaluri genome (Horton et al., 1990). Genomic DNA was isolated from E. ictaluri using a DNeasy Blood & Tissue Kit (Qiagen, Valencia, CA, United States) and used as template in PCR. The upstream and downstream

3810F tcagctgtgtgggtagactg
Primers A, B, C, and D were used for mutant construction. Bold letters show restriction enzymes added to A and D primers. Underlined letters in primer C indicate reverse complemented primer B sequence. The last primer in each group used for sequence confirmation. regions of each gene were amplified, and products were gelextracted using a QIAquick Gel Extraction Kit (Qiagen). The amplified upstream and downstream fragments were mixed equally and used as a template in the subsequent overlap extension PCR to generate the in-frame deletion fragment for each gene. The in-frame deletion fragments were digested with appropriate restriction enzymes (NEB) ( Table 2) and cleaned up. The suicide plasmid pMEG-375 was purified from an overnight FIGURE 2 | Confirmation of E. ictaluri USP mutants by using forward (A) and reverse (D) primers. Fragments were amplified from mutant and wild-type strains and separated on two different 1% agarose gels, which were then combined (white lines above indicate joints  The resulting plasmids named as pEi usp02-10 and pEi usp13 were transferred into E. coli SM10λpir or BW19851 by chemical transformation and mobilized into E. ictaluri WT by conjugation. First integration was selected by ampicillin, and ampicillin resistant colonies were propagated on BHI agar to allow for the second crossover allelic exchange. After this step, colonies were streaked on counter selective BHI plates with 5% sucrose, 0.35% mannitol, and colistin to allow loss of pMEG-375. Potential mutant colonies were tested for ampicillin sensitivity to ensure the loss of the plasmid, confirmed by PCR, and sequencing.

Construction of Bioluminescent USP Mutants
The constructed USP mutants were made bioluminescence using pAKgfplux1 plasmid as described previously (Karsi and Lawrence, 2007). Briefly, the overnight culture of both recipient (USP mutants) and donor cells (E. coli SM10λpir carrying pAKgfplux1) were mixed at 1:2 ratio (donor : recipient) and centrifuged briefly. Pellet was transferred onto sterile 0.45 µM filter papers placed on a BHI agar and incubated at 30 • C for 24 h. Bacteria on the filter paper were collected in BHI broth with ampicillin and colistin and then spread on BHI plates containing ampicillin and colistin. After incubation at 30 • C for 24-48 h, ampicillin resistant bioluminescent E. ictaluri colonies carrying pAKgfplux1 appeared on plates.

Growth Kinetics of the E. ictaluri USP Mutants in BHI
Growth kinetics of the ten E. ictaluri USP mutants was compared to E. ictaluri WT in BHI medium as previously described (Abdelhamed et al., 2016 FIGURE 3 | Growth of E. ictaluri USP mutants and WT in BHI broth. The data represent means of four replicates. Ei usp03 and Ei usp04 have a significantly (p < 0.001) higher growth rate than EiWT and other USP mutants, which indicated by a " * ." No significant differences were observed in the growth kinetics of EiWT and Ei usp02, Ei usp05, Ei usp06, Ei usp07, Ei usp08, Ei usp09, Ei usp10, and Ei usp13 strains.

Survival of E. ictaluri USP Mutants in Oxidative Stress
The survival of the ten USP mutants in BHI supplemented with 0.1% of H 2 O 2 were determined as previously described (Seifart Gomes et al., 2011). The experiment was performed in 96 well plates with four replicates under oxidative stress and normal conditions. The plates were incubated in Cytation 5 Cell Imaging Multi-Mode Reader, and the photon emissions were collected for 3 h at 30 • C.

Virulence and Efficacy of E. ictaluri USP Mutants in Catfish Fingerlings
Virulence and vaccine efficacy trials were conducted as reported by our group (Karsi et al., 2009). Approximately 720 channel catfish fingerlings (average: 13.728 cm, 10.544 g) were stocked into 36 tanks at a rate of 20 fish/tank. Tanks were divided into twelve groups with three replicate tanks each group. The experiment included 10 E. ictaluri USP mutants, positive control (EiWT), and negative control (BHI exposed). After 1 week of FIGURE 5 | The survival assay of E. ictaluri WT and USP mutants exposed to 0.1% H 2 O 2 . (A) Each strain had four replicates (column A-D). Strains start with E. ictaluri WT, Ei usp02-13 and BHI control. (B) The bars show the difference between bioluminescence of USP mutants and WT. * indicates a significant difference between stress and non-stress at P < 0.01.
acclimation, fish were challenged/vaccinated by immersion with 1.3 × 10 7 CFU/ml water for 1 h. Mortalities were recorded daily for 21 days, and the mean percent mortalities were calculated for each treatment group. Protective properties of USP mutants against EiWT infection was determined by challenging vaccinated catfish with EiWT (2.8 × 10 7 CFU/ml water). Fish mortalities were recorded daily, and the percent mortality was calculated for each group.

Statistical Analysis
For the growth kinetic experiment, significant differences between EiWT and USP mutants were determined by Student's t-test. For acid and hydrogen peroxide assays, photon counts were log 10 transformed t-tests were conducted. Percent reduction in bioluminescence was calculated by dividing mean photon emissions of USPs to mean photon emission of EiWT. For fish experiments, percent mortalities were arcsine transformed, and analysis of variance (ANOVA) was carried out using PROC GLM of SAS v9.4 (SAS Institute, Inc., Cary, NC, United States). In virulence/vaccination trial, the percent mortalities of USP mutants were compared to that of EiWT, while in efficacy trail, the comparisons were made to the sham-vaccinated group at the alpha level of 0.05.

Construction of the E. ictaluri USP Mutants
Thirteen universal stress proteins were identified in the E. ictaluri genome (Williams et al., 2012) by sequence similarity (Figure 1). They were scattered through the chromosome, and no operon structure was observed. We were able to delete 10 E. ictaluri usp genes in-frame, and mutants were verified by PCR (Figure 2) as well as sequencing. Properties of wild-type and mutated usp genes are shown in Table 3. In-frame deletion resulted in removal of a large portion (86-99%) of the wild-type usp genes (Table 3).

Growth Kinetics of the E. ictaluri USP Mutants in BHI
The growth of EiWT and USP mutants in BHI broth indicated that Ei usp03 and Ei usp04 have a significantly (p < 0.001) higher growth rate than EiWT. After 20 h incubation, the growth of EiWT was 23.6 and 17.42% lower than Ei usp03 and Ei usp04, respectively (Figure 3). Whereas, no significant differences were observed in the growth kinetics of EiWT and Ei usp02, Ei usp05, Ei usp06, Ei usp07, Ei usp08, Ei usp09, Ei usp10, and Ei usp13 strains at all tested time points.

Survival of E. ictaluri USP Mutants in Low pH Stress
To evaluate the role of usp genes in survival and growth of E. ictaluri at low pH, mutants and EiWT were exposed to acidic pH (5.5) and neutral pH, and bacterial growth (quantified by bioluminescent signal) were calculated. The growth rate (photon numbers) of the all USP mutants in low pH was significantly lower than that of in neutral pH. In contrast, the growth of EiWT at low pH was lower but not significant (Figures 4A,B). The strongest effect of low pH was observed in Ei usp03 growth (62% reduction) compared to EiWT. The order of susceptibility of USP mutants in low pH as follows: The reduced growth of the USP mutants indicates that usp genes contribute E. ictaluri survival under acidic conditions.

Survival of E. ictaluri USP Mutants in Oxidative Stress
Exposure to hydrogen peroxide (0.1% H 2 O 2 ) significantly reduced growth of Ei usp05 and Ei usp08 compared to no stress group (91 and 35% reduction, respectively), while growth of Ei usp02 and Ei usp03 increased under oxidative FIGURE 7 | Overall summary of results. Survival percent under acidic (pH) and oxidative stress (H 2 O 2 ) conditions was calculated based on changes in bioluminescence signal. The downward direction arrow indicates reduction in survival percent between mutant strain compared with wild type. The upward direction indicates increase in survival percent. Virulence percent is based on catfish mortality after immersion challenge with USP mutant strains. Efficacy perecent is based of mortality after re-challenge the immunized fish with E. ictaluri WT at 21 day post-immunization. stress (Figures 5A,B). No differences for Ei usp04, Ei usp06, Ei usp07, Ei usp09, Ei usp10, and Ei usp13 strains were observed.

DISCUSSION
Several previous studies reported that universal stress proteins (USPs) play a role in different bacteria to respond to different stress conditions, such as heat, substrate starvation, exposure to antimicrobial agents, acidic stress, and oxidative stress (Seifart Gomes et al., 2011). The objective of this study was to determine the role of E. ictaluri usp genes in acidic and oxidative stresses as well as in virulence. Also, mutants' vaccine potentials were determined. The uspA gene among usp genes has been studied in different bacterial strains. Deletion of the uspA genes resulted in decreased virulence in Salmonella typhimurium C5, Listeria monocytogenes, and Acinetobacter baumannii (Liu et al., 2007;Seifart Gomes et al., 2011;Elhosseiny et al., 2015). Also, uspA affected the host invasion and survival in Salmonella enterica and Mycobacterium tuberculosis (Hensel, 2009;Hingley-Wilson et al., 2010). In the present study, there were four usp genes (usp05, usp06, usp08, and usp09) with high similarity to uspA. The growth rate of Ei usp05, Ei usp06, Ei usp08, and Ei usp09 were similar to E. ictaluri WT. However, Ei usp05 and Ei usp08 showed reduced growth in oxidative and acidic stresses compared to EiWT. Virulence data showed that Ei usp05, Ei usp08, and Ei usp09 were significantly attenuated compared to E. ictaluri WT. However, Ei usp06 was not attenuated. These results are consistent with a previous study in L. monocytogenes where not all uspA are involved in reduced virulence (Seifart Gomes et al., 2011). Previously, our group reported that transposon insertion mutants in usp05 reduced E. ictaluri virulence in catfish and provided better protection against ESC (Kalindamar, 2013). Additionally, expressions of usp05 were very high in response to host stress or high level of H 2 O 2 in E. ictaluri (Akgul et al., 2018). The usp05 gene (uspA) is an important regulator of survival and virulence in many pathogens (Tkaczuk et al., 2013). In E. coli, uspA mutant caused a survival defect under a variety of growtharrested conditions, whereas overexpression induced growth in the growth-arrested state. Our data suggest that usp05, usp08, and usp09 are important virulence genes in E. ictaluri.
We demonstrated that Ei usp03 and Ei usp04 have a faster growth rate than EiWT and other USP mutants. However, lack of usp genes did not cause growth differences in Listeria monocytogenes (Seifart Gomes et al., 2011), E. coli (Nystrom and Neidhardt, 1993) or other bacteria when cultured in conventional media (Liu et al., 2007;Hingley-Wilson et al., 2010). Indeed, Ei usp03 and Ei usp04 did not show any virulence attenuation in E. ictaluri, which was similar to USP mutant Rv2623 in Mycobacterium tuberculosis (Hingley-Wilson et al., 2010). This study suggested that usp genes might play a role in latency and persistence of chronic TB infection. We think that usp03 and usp04 are not involved in virulence but may play other roles in stress responses in E. ictaluri.
Edwardsiella ictaluri can survive and continue growth in up to 3 mM of H 2 O 2 and low acidic pH 5.5. When the USP mutants and EiWT exposed to low pH, growth rates did not change significantly. As shown previously, L. monocytogenes ATP Binding USPs exhibited role in the response to acid stress during exponential growth phase (Tremonte et al., 2016).
Our results indicated that E. ictaluri usp07 contributes to virulence of E. ictaluri. Mortality was significantly decreased in the Ei usp07 mutant compared to EiWT strain. The usp07 is a KdpD protein, and it contains a uspA domain (Heermann et al., 2009a). We included whole KdpD as usp07 because USP domain is located between the N-terminal sensor domain and C-terminal catalytic domain of this Osmo-sensitive K + channel histidine kinase. Mutant KdpD in Salmonella typhimurium is attenuated in animal infection model and macrophage survival experiments. It also promotes resistance to osmotic, oxidative and antimicrobial stresses (Alegado et al., 2011). KdpD is also involved in oxidative-osmotic stress, response to host, and virulence (Freeman et al., 2013). In our gene expression study after host stress, usp07 showed a very high expression level (Akgul et al., 2018). It is important to note that usp07 involved in E. ictaluri virulence and acid stress response.
The usp13 was described as a universal stress protein and extra cytoplasmic adaptor protein (CpxP) like protein (Williams et al., 2012). The usp13 (CpxP) is placed in the inner membrane with histidine kinase CpxA and CpxR, a response regulator (Vogt and Raivio, 2012;Debnath et al., 2013). CpxP is the most highly inducible member of the Cpx regulon, and it has elevated expression in response to both envelope stress and entry into stationary phase growth (Motohashi et al., 1999;DiGiuseppe and Silhavy, 2003). The CPX system is important and required for virulence in both Gram-negative and -positive bacteria (Raju et al., 2012). Previously, we determined that E. ictaluri, usp13 is highly expressed when exposed low acidic pH (5.5) and the catfish invasion (Akgul et al., 2018). The usp13 (cpxP) is an essential regulator of cell membrane stress in bacteria during host infection. Therefore, it is involved in the virulence of E. ictaluri with a very high reduction in virulence (Figure 6).
The expression of E. coli usp genes is controlled by some effector proteins and signaling molecules, such as SOS repose proteins (Gustavsson et al., 2002;Kvint et al., 2003;Persson et al., 2007). However, mechanisms of USPs in other bacterial species are not known entirely. Overall our results are in line with studies from various species that USPs were crucial for protecting the cells from the damaging effects of reactive oxygen species (ROS) (Nachin et al., 2005;Liu et al., 2007;Seifart Gomes et al., 2011;Elhosseiny et al., 2015; Figure 7).

CONCLUSION
Our lab aims to develop live attenuated vaccines to protect catfish against E. ictaluri infections. Live attenuated bacterial should be both safe and confer full protection against wild-type infections. This study identified that Ei usp05, Ei usp08, Ei usp09, and Ei usp13 strains have vaccine potential and further efforts, such as constructing double mutants to improve their safety, could be pursued. The data presented in this study display that USPs are essential for both stress physiology and pathogenesis in E. ictaluri.

ACKNOWLEDGMENTS
We thank the Laboratory Animal Resources and Care at the College of Veterinary Medicine for providing the SPF channel catfish. AA was supported by a fellowship from the Republic of Turkey.