Transposon mutagenesis and identification of mutated genes in growth-delayed Edwardsiella ictaluri

Background Edwardsiella ictaluri is a Gram-negative facultative intracellular anaerobe and the etiologic agent of enteric septicemia of channel catfish (ESC). To the catfish industry, ESC is a devastating disease due to production losses and treatment costs. Identification of virulence mechanisms of E. ictaluri is critical to developing novel therapeutic approaches for the disease. Here, we report construction of a transposon insertion library and identification of mutated genes in growth-delayed E. ictaluri colonies. We also provide safety and efficacy of transposon insertion mutants in catfish. Results An E. ictaluri transposon insertion library with 45,000 transposants and saturating 30.92% of the TA locations present in the E. ictaluri genome was constructed. Transposon end mapping of 250 growth-delayed E. ictaluri colonies and bioinformatic analysis of sequences revealed 56 unique E. ictaluri genes interrupted by the MAR2xT7 transposon, which are involved in metabolic and cellular processes and mostly localized in the cytoplasm or cytoplasmic membrane. Of the 56 genes, 30 were associated with bacterial virulence. Safety and vaccine efficacy testing of 19 mutants showed that mutants containing transposon insertions in hypothetical protein (Eis::004), and Fe-S cluster assembly protein (IscX, Eis::039), sulfurtransferase (TusA, Eis::158), and universal stress protein A (UspA, Eis::194) were safe and provided significant protection (p < 0.05) against wild-type E. ictaluri. Conclusions The results indicate that random transposon mutagenesis causing growth-delayed phenotype results in identification bacterial virulence genes, and attenuated strains with transposon interrupted virulence genes could be used as vaccine to activate fish immune system.


Background
Enteric septicemia of catfish (ESC) is a devastating disease that causes significant production loss and treatment cost for the catfish aquaculture industry [1]. A few antimicrobials and a commercial live attenuated vaccine are available for treatment of ESC. However, treatment of sick catfish by medicated feed is not effective due to early onset of anorexia. The extensive use of antimicrobials can induce the appearance of resistant strains [2,3]. The commercial ESC vaccine Aquavac-ESC has been available for the catfish industry for more than 15 years [4], but ESC is still one of the major diseases in the US catfish industry.
Random transposon insertion is a high-throughput genetic manipulation tool that allows random mutation of genes at the genome level. Mariner family transposon Himar1 inserts itself randomly into "TA" nucleotide sequences [18,19]. Mariner family transposons have been widely used to generate random mutagenesis in fish pathogen Mycobacterium marinum, and also human pathogens such as Pseudomonas aeruginosa, Campylobacter jejuni, Leptospira interrogans, and Rickettsia prowazekii [20][21][22][23][24].
In this research, MAR2xT7 transposon, a Himar1 derivative [20], was used to identify genes required for E. ictaluri growth on a solid complex medium. We expect that colonies exhibiting attenuated growth on solid media will have transposon insertions in important bacterial genes, and these mutants may also show attenuated virulence in the catfish host and potentiate catfish immune responses [14]. Therefore, attenuation and vaccine efficacy of 19 transposon mutants were evaluated in channel catfish.

Transposon insertion library
By using MAR2xT7 transposon, an E. ictaluri transposon insertion library containing 45,000 transposants was constructed. Colonies with transposon insertion and delayed growth were observed on the BHI agar media after 48 h (Fig. 1). The initial overnight growth of these small colonies in BHI broth was also very slow compared to wild type, but this difference disappeared in later broth cultures (data not shown).
The complete genome size of E. ictaluri strain 93-146 is 3,812,301 bp, which contains 3597 total genes.

Gene identification
Transposon end amplification by single primer PCR yielded 151 samples with PCR products, of which 94 were sequenced successfully. After analysis, 56 unique genes containing transposon insertions were identified (Table 1). These unique genes contained a total number of 2235 MAR2xT7 transposon insertion sites, and the exact number of MAR2xT7 transposon insertion site in each gene was indicated in Table 1.

Functional annotation
Protein sequences of all 56 genes were annotated functionally and assigned to biological process (localization, cellular process, metabolic process, response to stimulus, biological regulation, signaling, multi-organism process, single-organism process, and biogenesis), cellular component (cell, macromolecular complex, and extracellular region), and molecular function (binding, transporter activity, catalytic activity, and nucleic acid binding transcription factor) (Fig. 2).

Subcellular localization
The locations of 15 proteins were unknown. Of the 41 proteins with known subcellular location, most were localized to the cytoplasm (20 proteins) and cytoplasmic membrane (16 proteins). Extracellular space, outer membrane, and periplasm contained very few proteins (3, 1, 1 proteins, respectively).

Proteins involved in bacterial virulence
Out of 56 unique proteins, 30 matched significantly to known virulence-associated proteins from other Gramnegative and Gram-positive pathogenic bacteria in MVirDB (Table 3).

Discussion
The bioinformatics analyses of 56 unique genes with transposon insertions showed that more than half (54%) were potential virulence factors in other pathogenic bacteria. Among the virulence factors, Type III secretion system (T3SS), twin-arginine translocation pathway (Tat), and ATP-binding cassette transporter (ABC) seem to be important for E. ictaluri virulence and invasion of the channel catfish [25][26][27].
The functional gene ontology analysis with Blast2GO indicated that most of the proteins participate in cellular and metabolic networks in the biological process while their molecular functions frequently matched to binding and catalytic activity. The proteins located in extracellular regions are part of the signaling process or response to any stimulus sensing bacteria in the biological process. Although several proteins showed transporter activity, these proteins account for localization and are in the cytoplasm.
The subcellular locations predicted by PSORTb revealed that most of the identified proteins are found in the cytoplasm and cytoplasmic membrane. Although many well-known virulence proteins are located in the outer membrane or periplasm in Gram-negative bacteria, only three proteins of T3SS are located in extracellular space and outer membrane, and one of the ABC transporter proteins was found in the periplasmic space.
The host-pathogen interaction examined by HPIDB proved that many proteins have a high similarity to other virulence-associated proteins in different pathogenic bacteria including Y. pestis, S. flexneri, and B. anthracis. The pathogenic Gram-negative bacteria Y. pestis and S. flexneri share the same evolutionary lineage with Edwardsiella sp. in Enterobacteriaceae [28]. Thus, most of the virulence-associated proteins may have a similar role in E. ictaluri. Two T3SS effector proteins EseJ and EseM have a predicted interaction with the channel catfish ubiquitin-conjugating enzyme E2 (XP_017323313). These two T3SS-related effector proteins known as E3 ubiquitin ligase play an important role in manipulation of host ubiquitination pathways.
The interrupted genes eseJ (Eis::037), eseM (Eis::038), esaC (Eis::041), eseL (Eis::110), and esaT (Eis::176) are part of T3SS, which are involved in export of proteins inside the host immune cells [29]. EsaC and EsaT are the structural membrane associated proteins of T3SS. Eis::41, YscC ring-shaped structure protein in the outer membrane, is required for a stable oligomeric complex to shape a T3SS in the outer membrane [30]. YscT inner membrane-embedded component is located in the cytoplasm, which has extended T3SS that provides a strategy to exploit host cell ubiquitin pathway [32]. EseM, T3SS leucine rich repeat protein (SlrP), is also required to form a complex ubiquitin ligase enzyme [33]. T3SS effector protein mutants eseJ, eseM, and eseL, and T3SS structural mutants esaC and esaT showed significantly decreased virulence. However, in comparison of protection level of those two main groups, T3SS structural proteins EsaC and EsaT have been caused less protection in catfish. Mutation in T3SS effector proteins provides better protection against pathogenic bacteria [34][35][36]. EseL has provided significant protection among other T3SS related effector proteins. T3SS effector proteins could contribute the bacterial survival inside host immune cells [37,38]. Transport processes in bacterial cells through outer membrane and periplasmic space are linked to E. ictaluri metabolism to survive in the host environment as well as switching between various biochemical processes during different stages of ESC. Eis::157, tatB, is located in the periplasmic space and is involved in the translocation of proteins including the components of respiratory complexes using a proton gradient as an energy source [39]. tatB mutant exhibited slow growth under low-iron conditions and observed a 10-fold decrease in Legionella pneumophila growth [40]. Eis::086, amiA, is a Tat pathway dependent substrate encoding a cell wall amidase. Tat pathway mutant causes mislocalization of AmiA protein, preventing translocation in the periplasm [41,42]. Eis::011, ABC transporter periplasmic amino acid binding protein, is an important antigenic factor involved in adhesion and aspartate/glutamate transport in the microaerobic environment in Campylobacter jejuni [43]. Eis::207, potB, encodes a protein associated with spermidine/putrescine transport system. Polyamines are mostly involved in stabilization of DNA for stress resistance, intracellular signaling processes, and swarming motility [44,45]. Polyamines are also associated with the virulence in the intracellular pathogen Salmonella enterica [46]. Eis::002, PTS system IIB component, is a cytoplasmic component of the major carbohydrate transport system highly conserved through bacteria [47]. PTS system participates in a variety of virulence mechanisms including biofilm formation, modulating the virulence gene expression, and regulating carbohydrate metabolism in pathogenic bacteria [48][49][50]. Eis::171, magnesium-translocating P-type ATPase, is an inducible magnesium transport system when bacteria grow at the low concentration of magnesium. Although Mg 2+ is not essential for virulence, it participates in many cellular activities as a cofactor [51]. Magnesium is the part of the regulatory network that regulates the virulence-associated mechanisms in S. enterica [52]. Eis::195, dcuA, is encoded with aspartase in the same operon that is determined as an antiporter mechanism involved in the transport of aspartate under the anaerobic conditions [53]. DcuA function in the metabolic pathway under anaerobic conditions contributes the pathogenicity for the colonization in the lower oxygen level [54].
Pathogenic bacteria adapted different carbohydrate metabolism, which is activated by oxygen presence in the host environment. Eis::018, aspartate ammonia-lyase, is involved in the production of fumarate activated specifically under anaerobic conditions while there is no available electron acceptor. Bacteria encodes aspartate ammonia-lyase to utilize alternative carbon sources in the host environment if there are no available carbon sources [55,56].
Bacterial stress related proteins induce the protective mechanisms under a variety of stress conditions to protect the bacterial cell inside or outside of the host [57]. Universal stress protein A (UspA) in Eis::194, is one of the stress proteins found in intracellular pathogenic bacteria. uspA expression reaches a high level when bacteria are exposed to heat, starvation, antimicrobial, and oxidative agents [58,59]. UspA is a conserved protein that presents in Eubacteria, Archaea, plants, and fungi and the expression of UspA is triggered by exposure to oxidative agents in growth arrested cells [60][61][62]. UspA plays a significant role in the pathogenicity of bacteria, and uspA mutants are less virulent and sensitive to changes in the host environment. Mutation of S. typhimurium C5 uspA resulted in less virulence and more susceptibility to nutrient starvation oxidative agents [59]. In Listeria monocytogenes, uspA mutants were shown to have impaired activity in oxidative agent's exposure to low pH conditions [58]. Deletion of uspA gene in Acinetobacter baumannii revealed that it has a significant role in protecting the bacteria from H 2 O 2 and low pH [63]. IscX in Eis::S039, acts as a regulator for the Fe-S (iron-sulfur) cluster, which encodes proteins essential for cell activities [64]. FeS assembly protein IscX (YfhJ) is a part of the iron-sulfur cluster (ISC) mediated FeS cluster, which is a small acidic protein that binds IscC and Fe, and acts as a Fe donor in FeS cluster [65,66]. ISC mediated FeS biogenesis is involved in survival of bacteria that face with iron starvation and oxidative stress. In S. flexneri, ISC mutants were less invasive and cannot form plaques on Henle cells monolayers [67]. ISC transcriptional regulator iscR mutant in Pseudomonas aeruginosa caused more susceptibility to oxidative agents and a significant decrease in virulence [68]. The importance of ISC system in bacterial virulence has been emphasized in different studies. However, limited information is known about the role of IscX in bacterial virulence.
Hypothetical protein in Eis::004 is located in the cytoplasm. There is no available information about the function of this hypothetical protein in any virulence related mechanisms. However, decreased virulence and significant protection against ESC revealed that Eis::004 mutant could be considered as a vaccine candidate for live attenuated vaccine development.

Conclusions
In summary, these results showed that random transposon mutagenesis in the E. ictaluri genome resulted in colonies with delayed growth on complex solid media, and many of the disrupted genes have important functions

Construction of transposon insertion library
Transposon insertion library was constructed by conjugation using the donor E. coli SM10λpir carrying pMAR2xT7 and the recipient E. ictaluri 93-146 wild type (WT) containing pAKgfplux1 [69]. Transposon insertion mutants were selected on selective BHI agar plates containing 100 μg/ml of   [20] ampicillin, 12.5 μg/ml of gentamicin, and 25 μg/ml colistin. Various sizes of gentamicin resistant transposon insertion colonies were observed on the selective BHI plates, and 250 smallest colonies compared to normal colony size were cultured in the BHI broth with colistin and gentamicin at 30°C for 2 days. Finally, bacterial stocks were prepared in 20% glycerol and stored at − 80°C freezer.

Transposon end mapping
Genomic DNA was isolated from the frozen E. ictaluri transposon insertion mutants using the heat denaturation method. Briefly, 100 μl frozen culture were added in 1 ml ddH 2 O and mixed well. Bacteria were collected by centrifugation and water was removed completely. After dissolving the bacterial pellet in 100 μl ddH 2 O, each sample was transferred to 200 μl PCR tubes and tubes were incubated at 100°C for 10 min by using an Applied Biosystems 2720 Thermal Cycler (Life Technologies, Grand Island, NY). Samples were mixed well by vortexing, and bacterial cell debris was pelleted by centrifuging at 14,000 rpm for 5 min. The supernatant containing the genomic DNA was used as template in subsequent PCR reactions. Single primer PCR was performed by using a transposon-specific R1 primer (5`-CCGTATGCCCAACT TTGTATAGA-3`) to amplify the transposon end and flanking bacterial DNA [70]. Before sequencing, the PCR products were cleaned by using ExoSAP-IT for PCR Product Cleanup (Affymetrix, Santa Clara, CA). Sequencing was conducted at Eurofins MWG Operon LLC (Huntsville, AL) using a transposon-specific nested R3 primer (5`-TCTC GGCTTGAACGAATTGTT-3`).

Bioinformatics analyses
Transposon sequence removal and sequence trimming based on sequence quality scores were done by using the Sequencher DNA sequence analysis software v4.10.1 (Gene Codes Corp., Ann Arbor, MI). Trimmed sequences were searched against the available E. ictaluri 93-146 genome [29] by using basic local alignment search tool (Blast) at the National Center for Biotechnology Information (NCBI) for gene identification. Using the GI numbers, a FASTA file containing all protein sequences were downloaded from the Batch Entrez database of NCBI and used for downstream analysis. Gene Ontology (GO) annotation, visualization, and metabolic and cellular processes were determined by using Blast2GO [71] at the cut-off level 2. Subcellular localization of proteins was predicted by using PSORTb version 3.0.2 [72]. E. ictaluri proteins involving in host-pathogen interactions were identified by using the Host-Pathogen Interaction Database (HPIDB) at the cut-off level 0.0001. Bacterial proteins interacting with channel catfish proteins were determined at the cut-off level 0.00001, at identity filter 50% in bacterial proteins, and 70% in channel catfish proteins. [73]. The potential E.
ictaluri virulence proteins were identified using the Microbial Virulence Database (MVirDB) at the cut-off level 0.5 [74]. TA sequence frequencies in the entire E.
ictaluri genome, open reading frames, and genes with transposon insertion were calculated using CLC genomics workbench 11.0.1 (Qiagen, Redwood City, CA). ictaluri wild-type), and negative (BHI) controls were assigned to three or four tanks randomly. Catfish were challenged/vaccinated by immersion exposure using transposon mutants or wild type (3.09 × 10 7 CFU/ml of water) using published procedures [15]. Catfish mortalities were recorded for 21 days. After 21 days of the first vaccination, both vaccinated, and sham-vaccinated catfish were infected with E. ictaluri wild type by immersion exposure (3.27 × 10 7 CFU/ml of water). Catfish mortalities were recorded for two weeks.

Statistical analysis
We used SPSS V25 (IBM Corp., Armonk, NY) to conduct statistical analysis. For each strain, mean percent mortalities were calculated and arcsine-transformed. The one-way analysis of variance at significance level 0.05 was conducted using the "Univariate" function, in which strains were independent and arcsine-transformed mortalities were dependent variables. Because our data included different sample sizes, and variances were not equal, Games-Howell post hoc test was selected to identify significant differences between mutants and wild type or mutants and sham vaccinated group in virulence and efficacy experiments, respectively.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Authors' contributions
Conception and design of the study: MLL, AK. Performed experiments: SK, JL, HA, HCT. Analyzed data: SK. Wrote the manuscript: SK, MLL, AK. All authors read and approved the final manuscript.
Ethics approval and consent to participate All fish experiments were conducted under a protocol approved by the Institutional Animal Care and Use Committee at Mississippi State University (protocol number 12-042).

Consent for publication
Not applicable.

Competing interests
The authors declare that they have no competing interests.