Stress-related genes promote Edwardsiella ictaluri pathogenesis

Edwardsiella ictaluri is a Gram-negative facultative anaerobic rod and the causative agent of enteric septicemia of channel catfish (ESC), which is one of the most prevalent diseases of catfish, causing significant economic losses in the catfish industry. E. ictaluri is resistant to complement system and macrophage killing, which results in rapid systemic septicemia. However, mechanisms of E. ictaluri stress responses under conditions of host environment are not studied well. Therefore, in this work, we report E. ictaluri stress responses during hydrogen peroxide, low pH, and catfish serum stresses as well as during catfish invasion. E. ictaluri stress responses were characterized by identifying expression of 13 universal stress protein (USP) genes (usp01-usp13) and seven USP-interacting protein genes (groEL, groES, dnaK, grpE, and clpB, grpE, relA). Data indicated that three usp genes (usp05, usp07, and usp13) were highly expressed in all stress conditions. Similarly, E. ictaluri heat shock proteins groEL, groES, dnaK, grpE, and clpB were highly expressed in oxidative stress. Also, E. ictaluri grpE and relA were highly expressed in catfish spleen and head kidney. These findings contribute to our understanding of stress response mechanisms in E. ictaluri stress response, and stress-related proteins that are essential for E. ictaluri could be potential targets for live attenuated vaccine development against ESC.


Identification of stress-related proteins
Sequences of the E. ictaluri USP genes and proteins were obtained from the E. ictaluri genome [34], and six heat shock proteins (GroL, GroS, DnaJ, DnaK, ClpB, and GrpE) interacting with USPs were identified using STRING v10 [35]. The STRING database (http://string-db.org) provides pre-calculated protein-protein interactions for organisms, including E. ictaluri. By querying protein sequences of the E. ictaluri USPs, the interacting proteins were identified, and the interaction network was visualized. Because of its regulatory role for some USPs, ppGpp synthetase (relA) was also included in our analysis.

Acid stress (low pH exposure)
An E. ictaluri strain 93-146 colony was inoculated in 5 ml of BHI broth, followed by [16][17][18] h incubation at 30˚C with shaking. The next day, 40 ml of BHI broth was inoculated and grown to an optical density (OD 600 ) of 0.4. Then each culture was divided into four aliquots, 10 ml each, and bacteria were harvested by centrifugation at 6,000 x g for 15 m. Supernatant was removed, and bacteria were resuspended in acidic BHI broth (pH 4.0, acidified with 6 N HCl solution) and standard BHI broth (pH 7.0). Cultures were incubated with shaking (180 rpm) at 30˚C for 30 m., and pellets were collected by centrifugation (7000 x g for 15 m). RNAlater was added, and bacteria were stored for a week at -20˚C until RNA was isolated.

Oxidative stress
E. ictaluri 93-146 was streaked on BHI agar containing 12.5 μg/ml of colistin and incubated at 30˚C for 48 h. Small cultures were prepared by inoculating 5 ml of BHI broth with 12.5 μg/ml colistin and incubating for 16-18h at 30˚C with shaking at 200 rpm. 40 ml of BHI broth were inoculated and grown to an optical density (OD600) of 0.4. Each culture was divided into four aliquots, 10 ml each, and bacteria were harvested by centrifugation at 6,000 x g for 15 m. Supernatant was removed, and bacteria were resuspended in two types of BHI broth: 20 ml standard BHI broth and 20 ml BHI broth supplemented with 1.5 mM (0.05%) H 2 O 2 . Cultures were incubated with shaking (180 rpm) at 30˚C for 30 m. Bacteria were harvested by centrifugation at 7000 x g for 15 m., and pellets were suspended in RNAlater and stored a week at -20˚C until RNA isolation.

Serum stress
E. ictaluri were exposed to naïve catfish serum and heat-inactivated catfish serum. Each treatment included four biological replicates. E. ictaluri cultures were washed three times using 1.25 ml of cell wash buffer (10 mM TrisCl and 5 mM magnesium acetate). Either normal serum or heat-inactivated serum (1.25 ml) was added to each E. ictaluri pellet. Tubes were inverted to mix bacteria and serum thoroughly followed by incubation for 30 m at 30˚C. Serum-bacteria mixture was used for total RNA isolation.

In vivo gene expression
Approximately 18-month-old SPF channel catfish (19.7 cm, 18.5 g) were stocked into two tanks at a rate of 7 fish/tank. After one week of acclimation, fish were anesthetized in water containing 100 mg/L tricaine methanesulfonate (MS-222), and bioluminescent E. ictaluri were injected into intraperitoneal space in 100 μl PBS (approximately 1×10 4 CFU). Negative control fish were injected with 100 μl PBS.
Bacteria were visualized in live catfish using an IVIS Lumina XRMS In Vivo Imaging System (PerkinElmer, Waltham, MA) by following the procedures described previously [36]. Bioluminescence imaging (BLI) was conducted at 3, 6, 12, 24 and 30 h post-infection. At each time point, catfish were anesthetized in water containing 100 mg/L MS222 and transferred immediately to the photon collection chamber. Total photon emissions from the whole fish body were collected at an exposure time of 30 s. Following BLI imaging, fish were returned to well-aerated water for recovery. Bioluminescence was quantified using Living Image Software v 4.2 (Caliper Life Sciences., Hopkinton, Massachusetts), and mean photon intensity for each treatment was used in statistical analysis.
At 30 h post-exposure, fish were euthanized with a high dose of MS-222 (300 mg/L), and anterior kidney and spleen tissues were collected immediately in RNAlater Stabilization Solution. Then, RNA was isolated from four fish with an RNeasy Mini Kit (Qiagen).

Total RNA isolation
In each experiment, total RNA was isolated from four biological replicates using RNeasy Protect Bacteria Mini Kit (Qiagen). Contaminating bacterial DNA was eliminated by DNase I treatment with RNase-Free DNase Set (Qiagen). The concentration and quality of the isolated total RNA were measured by NanoDrop 1000 (Thermo Scientific). First-strand cDNA was produced from 1 μg total RNA using Maxima First Strand cDNA Synthesis Kit for RT-qPCR (Thermo Scientific) by following manufacturer's instructions.

Real-time qPCR
Primers sequences and expected PCR fragment sizes are shown in Table 2. Optimal annealing temperatures of primers were determined by gradient PCR. Primer specificities were determined by melting curve analysis and visual inspection of PCR products on agarose gel. After assessment of primers' quality, quantitative real-time PCR (qRT-PCR) was performed using a Mx3005P qPCR System (Agilent Technologies, CA, USA) and DyNAmo SYBR Green qPCR Kit (Finnzymes Oy, Espoo, Finland). Each 20 μl PCR reaction contained 10 μl SYBR Green 2X mix, 0.2 μM each of forward and reverse primers, and one μl of 100 x diluted cDNA. The PCR was set to initial denaturation at 95˚C for 3 m, 45 cycles of denaturation at 95˚C for 15 s, annealing at 60˚C for 30 s, and extension at 72˚C for 30 s, and a final extension at 72˚C for 3 m. At the end of the PCR, a melting curve program from 60˚C to 95˚C with 0.5˚C increase every 15 s was run. 16S RNA gene was used as internal control. Bacteria grown in BHI broth were used as experimental control while control group in serum experiment was bacteria exposed to heat-treated serum.

Statistical analysis
Relative expression values were calculated by the threshold cycle (Ct) changes in sample and control using the ΔΔCt method [37]. All expression values were normalized against 16S rRNA. Relative expression was determined by the comparative Ct method of relative quantification (RQ), which was calculated with the formula 2 -ΔΔCt . ΔΔCt was calculated by ΔΔCt = ΔCt (stress condition) -ΔCt (non-stress condition), where ΔCt is the normalized signal level in a sample (ΔCt = Ct of target gene-Ct of reference gene).
One-way Analysis of Variance (ANOVA) was used to compare gene expression among different stress conditions with SAS 9.1.4 (SAS Institute, Cary, NC). P value cutoff for statistical significance was < 0.05. To increase statistical reliability, four replicates were run for each sample, and the mean Ct considered for each gene.

Gene Name
Gene ID Primer Size USPs were also identified by using STRING10 [35] (Fig 1B). ppGpp synthetase (RelA) was also included because of its regulatory role for some USPs.

Bioluminescence imaging (BLI) of live catfish
Imaging of bioluminescent E. ictaluri in live catfish allowed us to determine the best time point for tissue collection (Fig 2A and 2B). After a brief decline at 6 h post-injection, average photon intensity, and hence bacteria number, increased continuously, which was approximately 5-fold higher at 30 h compared to initial imaging at 3 h post-injection. BLI showed clearly that E. ictaluri colonized most internal organs of catfish at 30 h post-injection, and it was an appropriate time for collection of anterior kidney and spleen tissues.

Expression analysis of stress-related genes in oxidative stress
In 0.05% H 2 O 2 , groES and dnaK expression increases were 18-and 11-fold, respectively, which was significantly greater than other genes. usp05, groEL, and grpE showed more than 8-fold; clpB showed 4.5 fold; and usp08, usp09, and usp11 exhibited more than 2-fold increased expression (Fig 4).

Expression analysis of stress-related genes in catfish
Expression of the 20 stress response genes was determined in catfish spleen and head kidney at 30 h post-exposure relative to E. ictaluri grown in BHI broth. The usp05 and usp13 genes showed the highest expression levels in the spleen with 191-and 137-fold increase, respectively. Expression increases of other genes in the spleen were: usp07 18-fold; usp04, usp06, usp10, and grpE more than 7.8 fold; and usp08, usp09, usp11, and relA more than 2-fold (Fig 6). In head kidney, usp07 showed the highest expression level with 139-fold increase relative to control (Fig 7). Increased expressions of other genes were usp05 and usp13 more than 22-fold; usp10 more 10-fold; usp04, usp06, usp09, grpE, and relA more than 4-fold; and usp08, usp11, usp12, dnaK, groEL, and groES more than 2 fold. A summary of all experiments was provided in Fig 8.

Discussion
In this research, we aimed to determine the expression of 20 stress-related genes under different stress conditions. There are different types of universal stress proteins (USP) in archaea, bacteria, and plants [22]. E. coli has six USP groups (USPA, USPB, USPC, USPD, USPE, and USPF), which are involved in several functions including oxidative stress, adhesion, and motility [21]. Following bacterial phagocytosis by macrophages, fusion of the lysosome with the phagosome causes release of acid-activated enzymes and toxic metabolites and acidification of the phagolysosome to pH 4-5. As a result, bacteria are typically inactivated and broken down by the enzymes and toxins. E. ictaluri can resist macrophage killing and replicate in phagocytes, which is considered one of the key mechanisms for rapid spread of the pathogen in the host. The low pH and oxidative stress we applied mimic the conditions present in the phagolysosome [38]. Expression increase in usp05, usp13, dnaK, groEL, groES, and grpE may indicate their involvement in E. ictaluri survival in harsh phagolysosomal conditions. Edwardsiella ictaluri stress-related genes Expression of usp genes in serum was upregulated, while expression of HSPs was mostly down-regulated. In particular, usp13 expression was 62.2-fold compared to control. The complement system in serum lyses and opsonizes bacteria for phagocytosis. The complement system of fish can also inactivate bacterial exotoxins [39]. In fish, three pathways of complement  activation have been identified: the classical complement pathway (CCP), alternative complement pathway (ACP), and the lectin complement pathway (LCP) [40]. E. ictaluri is resistant to serum [41]. In E. ictaluri, sialic acid-induced suppression of ACP activation was thought to be the mechanism for E. ictaluri resistance to normal catfish serum [41]. Besides, a larger amount of surface proteins and polysaccharide material were present compared to virulent strains [6]. These surface components help prevent complement and lysozyme in normal serum, from contacting vulnerable sites on the bacterial surface [39]. Because of E. ictaluri's ability to survive in serum, we were not expecting a high expression of stress response genes in normal serum. Only usp13 showed significant up-regulation, suggesting that USP13 might be associated with serum tolerance in E. ictaluri. Edwardsiella ictaluri stress-related genes Imaging of live catfish injected with bioluminescent E. ictaluri indicated a similar infection pattern shown in our previous studies [17,36]. Bioluminescence imaging shows a brief decline of photon intensity at early hours of infection, which is probably due to strong immune responses of catfish against E. ictaluri. Later, photon intensity increases faster due to rapid colonization of catfish by the pathogen. This rapid increase may indicate overwhelming of catfish immune system by the pathogen, which results in death of catfish in a couple of days. Therefore, 30 h post-infection seems a suitable time point for tissue collection because E. ictaluri is in a rapid increase stage while catfish immune responses are not overwhelmed completely.
Whereas usp05, usp08, usp09, and usp11 were highly expressed under in vitro stress conditions, usp04-08, usp10, usp13, grpE, and relA were upregulated in the host. Overall, usp05, usp07, usp13, and grpE may be critical for E. ictaluri to cope with several different stressors. Previously, our group identified a transposon insertion mutant of usp05 that showed attenuated virulence in catfish and provided protection against ESC. The usp05 gene is known as uspA, and it is an important regulator of survival and virulence in many pathogens [23]. An E. coli uspA mutant caused a survival defect under a variety of growth-arrested conditions, whereas overexpression induces growth in the growth-arrested state. Our data suggest that usp05 could be an important virulence gene in E. ictaluri.
Kdp is a K + transporter system in E. coli, and KdpD/E control the expression of the kdpABC operon, which is one of two-component sensor-regulator [42]. USP07 is a KdpD protein, and it contains a USPA domain [43]. We included the whole KdpD as a USP07 because USP domain is located between the N terminal sensor domain and C-terminal catalytic domain of this Osmosensitive K+ channel histidine kinase. Biochemical studies revealed that USPC interacts specifically with the USP domain in the stimulus perceiving N-terminal domain of KdpD, and UspC stabilized the KdpD/KdpE, and DNA complex to act as a scaffolding protein under salt stress in Staphylococcus aureus [44]. KdpD functions as a membrane-associated protein kinase that phosphorylates KdpE in response to environmental signals [43]. Salmonella enterica serovar Typhimurium kdpD mutant is attenuated in host and macrophage survival experiment. Also, it is promoting resistance to osmotic, oxidative, and antimicrobial stresses [45]. KdpD is involved oxidative-osmotic stress response to host and bacterial virulence [46]. In our study, usp07 showed very high expression, especially when E. ictaluri is in the catfish head kidney and spleen, indicating that it may involve in E. ictaluri survival in the host.
USP13 was described as a universal stress protein and CpxP like protein in NCBI. USP13, extracytoplasmic adaptor protein (CpxP), is placed in the inner membrane with histidine kinase CpxA and CpxR, a response regulator [47,48]. CpxP is a highly inducible member of the Cpx regulon, and its expression was elevated in responses to envelope stress and entry into stationary phase growth [49,50]. Also, CpxP functions as an adaptor protein, carrying misfolded periplasmic proteins to the DegP protease for degradation [51,52]. CpxP is degraded along with its misfolded substrate, which suggests that bacteria can posttranslationally modulate CpxP levels [47]. Cpx system is important and required for virulence in both Gram-negative and -positive bacteria [53]. CpxP also modulates the activity of the Cpx system by dynamic interaction with CpxA in response to specific stresses [54]. On the other hand, CpxP overexpression results in a reduced Cpx-response [55], hence interfering with the induction of envelope stress response. Thereby, CpxP inhibits autophosphorylation of reconstituted CpxA [56]. According to the current model, the inhibitory and supporting functions of CpxP for envelope stress response are linked. In unstressed cells, CpxP associates with CpxA to shut off the Cpx-TCS. Envelope-stress conditions induce the displacement of CpxP from CpxA resulting in Cpx-TCS activation [54]. In E. ictaluri, CpxP showed very high expression in all stress conditions except oxidative stress. CpxP is an important regulator of cell membrane stress in bacteria during host infection, and it might promote the virulence and survival mechanisms related to distinct salt and acidic conditions.
Other seven stress proteins were chosen by a possible regulation of USPs or interaction with USPs in E. ictaluri. Protein-protein interactions play critical roles in protein function, and in a large scale protein-protein interaction study in E. coli showed that a USPA has 83 target proteins, which mostly overlap with DnaK, DnaJ, and GroEL, also interactions with GroEL and DnaK (HSP70) [57]. GroEL and DnaK may require both systems for folding, degradation, or translocation [57]. The function of GroEL/ES is in protein folding and possibly in intercellular signaling, and prevention of misfolding under stress conditions was determined [58]. DnaK/J plays an essential role in DNA replication and participates actively in response to hyperosmotic shock. DnaK is also involved in refolding heat-damaged proteins [19]. GroEL was previously determined with various other proteins as a related protein to the bacterial virulence in E. tarda [59]. In another proteomics study comparing two strains of E. tarda with different degrees of virulence, authors showed that GroEL was highly expressed in pathogenic strain compared to nonpathogenic strain [60]. Our result indicated that expression of groEL and groES were high in the oxidative stress, and dnaK was highly upregulated. Therefore, highly expressed heat shock proteins in oxidative stress should further investigated to explain the folding and unfolding process in oxidative stress of E. ictaluri. ClpB plays an important role in protein control to protein disaggregation in collaboration with the DnaK, DnaJ, and GrpE [50], and it is sensitive to heat shock and other extreme stresses [61,62]. In E. tarda, protein expression of ClpB was higher in virulent strain than non-virulent [60]. In our study, the expression of clpB was high in oxidative stress like dnaK, groEL, and groES. ClpB is part of a stress-induced system involved in the recovery of the cell from damage in cooperation with DnaJ, DnaK, and GrpE. The grpE was also shown similar high expression in oxidative stress and the internal organs. Further, it has very high expression in low pH. Thus, grpE may be an important heat shock protein to cope with stress in E. ictaluri.
The stringent response (SR) coordinates adaptations to nutritional starvation and various stress conditions [63,64]. SR relies on the 'alarmones' ppGpp, and pppGpp synthetases [65]. The ppGpp acts by modifying the activity of many cellular targets including DNA replication, transcription, translation, ribosome assembly and metabolism [63,65,66]. Also, it is recently found that it has a role in DNA repair mechanism [67]. The SR is in the center bacterial survival and virulence [66,68,69]. In our study, we determined that the expression of ppGpp synthetase (relA) was increased when E. ictaluri was in kidney and spleen. Similarly, serum affects the expression of relA. Therefore, we think that relA could be essential for E. ictaluri adaptation in the host. Further studies should be conducted in relA to better understand its role in E. ictaluri virulence and the regulation of USPs.
In this study, we identified that three usp genes (usp 05-07-13) were highly expressed in many stress conditions, and five heat shock proteins were highly expressed in oxidative stress (groEL/ES, dnaK, grpE and clpB), and grpE was highly expressed in spleen, kidney, and low pH. Finally, relA was highly expressed in kidney and spleen. As a result, we expect these findings will help us understand the role of stress proteins in E. ictaluri virulence. Furthermore, essential USPs and stress regulators in E. ictaluri can be potential targets for live vaccine development against ESC.