Identification and Characterization of gtf, norB, and trx Genes in Flavobacterium columnare

Three genes from a shotgun genomic library of Flavobacterium columnare virulent strain ALG-00-530 were identified, characterized, and selected for differential expression analysis based on sequence similarity to putative virulence genes from related species. These genes were: glycosyltransferase (gtf), nitric oxide reductase (norB) and thioredoxin (trx). A collection of 30 F. columnare strains, including strains from genomovars I and II, were tested for the presence of these genes. Distribution patterns of gtf, norB, and trx across the species were not uniform. Nucleotide sequence variation was observed between genomovars for each gene; however, strains within the same genomovar shared identical gene sequences. Nine strains of F. columnare, from both genomovars, were chosen for gene expression analysis. The expression profile of these genes varied when selected strains were grown in vitro under identical conditions. ALG-00-530, a high virulent strain, was chosen for gene expression comparison under standard growing conditions, iron-limited conditions and in the presence of skin explants from channel catfish. NorB, and trx gene expression levels varied when ALG00-530 was incubated under different conditions. Identification and Characterization of gtf, norB, and trx Genes in Flavobacterium columnare Yinfeng Zhang, and Covadonga R. Arias* Department of Fisheries and Allied Aquacultures, Auburn University, Auburn, AL, USA Journal of Microbial & Biochemical Technology Open Access JMBT/Vol.1 Issue 1 *Corresponding author: Cova R. Arias, 203 Swingle Hall, Auburn University, Auburn, AL-36849, USA, Tel: (334) 844 4382; E-mail: ariascr@acesag.auburn.edu Received December 10, 2009; Accepted December 26, 2009; Published December 26, 2009 Citation: Zhang Y, Arias CR (2009) Identification and Characterization of gtf, norB, and trx Genes in Flavobacterium columnare. J Microb Biochem Technol 1: 064-071. doi:10.4172/1948-5948.1000013 Copyright: © 2009 Zhang Y, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.


Introduction
Flavobacterium columnare, the etiologic agent of columnaris disease, is a Gram-negative bacterium that can infect most freshwater fish species. Differences in virulence are known to exist among F. columnare isolates, resulting in variable mortality of fish (Pacha and Ordal, 1970;Suomalainen et al., 2006). Understanding the pathogenesis of F. columnare is critical for fish health, but our knowledge of F. columnare virulence factors is far from being complete.
Only a few genes from F. columnare have been described to date, and most known sequences from this pathogen correspond to ribosomal genes. Among the non-ribosomal sequences identified, the chondroitin AC lyase gene in F. columnare has been previously sequenced and characterized (Xie et al., 2005). The enzyme codified by this gene is able to degrade acidic polysaccharides, such as hyaluronic acid and chondroitin sulfates (Griffin, 1991; Teska, 1993;Stringer-Roth et al., 2002). Differences in chondroitin AC lyase activity have been observed between F. columnare isolates (Stringer-Roth et al., 2002). Relation of chondroitin AC lyase activity to virulence in F. columnare has been recently reported (Suomalainen et al., 2006). Genes encoding outer membrane proteins, such as zinc metalloprotease and prolyl oligopeptidase, have also been characterized in a virulent F. columnare isolate (Xie et al., 2004). In addition, several proteases have been identified in F. columnare, but their specific roles in columnaris pathogenicity are still unknown (Bertolini and Rohovec, 1992;Newton et al., 1997).
The ability of F. columnare to attach to fish tissues is thought to be a factor correlated to virulence (Decostere et al., 1999). Bader et al. (2005) selected for an adhesive-defective F. columnare strain that exhibited reduced virulence to channel catfish. Unfortunately, no adhesin gene has been identified in F. columnare.
The aim of this work was to identify and characterize putative virulence genes in this species and to further investigate the genetic diversity of F. columnare using non-ribosomal markers. To achieve this objective, a partially sequenced shotgun genomic library constructed from a virulent F. columnare strain was used. Three putative virulence genes were selected for further analysis. Gene presence and expression patterns were analyzed across the species. Two in vitro experiments aimed to mimic some of the environmental conditions F. columnare encounters during infection (limited iron and fish skin presence) were used for gene expression analysis.  (Table 1). Cells were cultured in modified Shieh broth  at 28ºC.

Putative virulence gene identification
A shotgun genomic library of the F. columnare virulent strain ALG-00-530 was constructed. Briefly, F. columnare DNA was extracted and purified following standard protocols (Sambrook and Russell, 2001). Total DNA was partially digested with Sau3A I. Digested DNA ranging from 1 to 1.25 kb was purified by double size selection and ligated to the digested pBluescript II (KS-) vector (Stratagene, Cedar Creek, TX) at a molar ratio of 1:2. The cloning site was BamH I. Library titer was estimated at 35 recombinant clones per microliter. Quality control tests showed 1% or less empty clones and an expected genomic coverage of 20X. Three thousand clones of this library have been sequenced to date at the USDA-ARS MSA Genomics Laboratory (Stoneville, MS) following standard procedures.
Sequences from the shotgun genomic library were compared with sequences in the GenBank database available at the NCBI (National Center for Biotechnology Information http:// www.ncbi.nlm.nih.gov/) using the BLASTX algorithm (Altschul et al., 1990). Sequences with more than 35% identity were recorded. Among the identified genes, three were chosen for further analysis due to their high identity (≥65%) with known genes described as virulence factors in other bacteria. Open reading frames (ORF) of the three sequences were identified using Vector NTI ® Suite 8 software package (Invitrogen, Carlsbad, CA).
ORF were translated into amino acid sequences by using Vector NTI ® Suite 8 and the GenBank database was searched for protein sequences using the BLASTP tool (Altschul et al., 1990). Protein-protein identity percentage was recorded for each chosen ORF. Specific primers for each gene were designed using Vector NTI ® Suite 8. Primer sequences for each gene were shown in Figure 1. These primers were tested on 30 strains of F. columnare by PCR. Unless otherwise stated, all PCR reagents were purchased from Promega (Madison, WI). Each 50 µL PCR reaction included 2.5 µM MgCl 2 , 1X buffer, 0.2 µM of both primers, 0.2 µM of dNTPs, 1.7 unit of Taq polymerase, and 60 ng of DNA template. The PCR amplification profile was 5 min hot start at 95ºC; 35 cycles of 30 s at 94ºC, 45 s at 58ºC, and 1.5 min at 72ºC; and 10 min at 72ºC.

Nucleic acids extraction and RT-PCR
Total bacterial DNA was extracted using the Qiagen DNeasy Tissue kit (Qiagen, Valencia, CA) following the manufacturer's instructions. RNA was extracted from 1.5 mL of bacterial culture using RNeasy plus Mini Kit (Qiagen). Turbo DNA-free kit (Ambion, Austin, TX) was used to eliminate DNA contamination in RNA samples. cDNA synthesis and RT-PCR were performed using the Reverse Transcription System (Promega). One microgram of total RNA was used to synthesize the cDNA in 20 µL of reaction. cDNA synthesis reaction was diluted to 100 µL, and 10 µL were used as template for PCR using gene specific primers, as described above.   Cloning and sequencing analysis PCR products were resolved through standard agarose gel electrophoresis. Amplified products were purified using the High Pure PCR Product Purification Kit (Roche Diagnostic Corporation, Indianapolis, IN) and cloned into pGEMTEasy (Promega). To ensure sequence accuracy, two clones from each strain were sequenced at the Auburn University Sequencing Core (Auburn, AL). Nucleotide sequence alignments and deduced amino acid sequence alignments were performed by CLUSTAL W algorithm (Chenna et al., 2003).

Gene expression
Expression of gtf, norB, trx was tested under standard conditions, iron-limited conditions, and in the presence of catfish skin explants. Following RNA extraction and cDNA synthesis, genes were amplified from cDNA by the designed primers as mentioned above. 16S rRNA cDNA was amplified as an internal control with universal primers UFUL (5'-GCCTAACACAT GCAAGTCGA-3') and URUL (5'-CGTATTACCGCGGCT GCTGG-3') (Chen et al., 2004). Positive (genomic DNA) and negative (total RNA) controls were included for each gene amplification. Amplified products were electrophoresed at 100 V for 30 min and visualized under UV light with ethidium bromide added to a 1% gel to a final concentration of 0.5 µg mL -1 .

Gene expression under standard conditions
Nine strains of F. columnare, representing genomovars I and II, were cultured in modified Shieh broth at 28ºC for 20 h. Bacterial cells were harvested from one and a half milliliters broth culture followed by RNA extraction and RT-PCR.

ALG-00-530 gene expression under iron-limited conditions
Two iron chelators, 2, 2-dipyridyl and transferrin, were used to remove free iron from the growth medium. 2, 2-dipyridyl was added to 50 mL of Shieh broth to a final concentration of 50 µM. Preparation of the transferrin solution followed the protocol described by Biosca et al., (1993). Briefly, human apo-transferrin (Sigma, St. Louis, MO) was dissolved at a concentration of 1 µM in a solution containing 100 µM Tris, 150 µM NaCl, and 50 µM NaHCO 3 (pH 8.0). This solution was sterilized by filtration. Prior to inoculation with F. columnare, transferrin was added to 50 mL Shieh medium to achieve a final concentration of 10 µM and incubated at 37ºC for 60 min. Cultured ALG-00-530 bacterial cells were collected by centrifugation at 2,000 g for 20 min. Pellets were washed once in PBS and centrifuged again. Cells were resuspended in approximately 1 mL Shieh broth and then 1/3 mL (of this broth) was transferred into 50 mL Shieh broth containing 2, 2-dipyridyl, or apo-transferrin. Fifty milliliters of Shieh broth was inoculated as control. One and a half milliliters of bacterial cells were collected at time 0 h, 2 h, 4 h, 6 h, 8 h, and 24 h after inoculation. Cells were centrifuged and pellets were resuspended in 100 µL of RNA later (Qiagen) and frozen at -80ºC until used. RT-PCR was performed as explained above.

ALG-00-530 gene expression in the presence of catfish skin explants
Skin explants from channel catfish that tested negative for F. columnare antibody presence were obtained according to the procedure by Xu and Klesius, (2002). One milliliter of Medium 199 (Sigma) supplemented with fetal bovine serum at 2.5% was added into each well of a 24-well culture plate (Costar, Cambridge, MA). One cm 2 of fish skin explant was placed onto the bottom of 12 wells. One and a half milliliters of an overnight culture of ALG-00-530 bacterial cells was collected for the baseline control (C 0 ); one and half milliliters of the same bacterial culture was added into each of the 24 wells. Wells without fish skin explants were regarded as controls, while wells containing the fish explants were referred to as treatments. Bacterial cells were collected from each individual well at 30 min, 1 h, 3 h, 6 h, 10 h, and 24 h from both control and treatment wells. After centrifugation, bacterial pellets were suspended in RNA later (Ambion, Austin, TX) and frozen at -80ºC until used. RNA extraction and RT-PCR were performed as described above.

Putative virulence gene identification
Out of 3,000 clones sequenced from the ALG-00-530 shotgun library, 73 sequences displayed a high nucleotide similarity with known genes present in GenBank (based on BLASTX search, data not shown). Within this subset, three sequences were identified from published literature as putative virulence genes and therefore selected for this study: glycosyltransferase gene (gtf) (GenBank accession number DQ911521), nitric oxide reductase gene (norB) (GenBank accession number DQ911522), and thioredoxin gene (trx) (GenBank accession number DQ911523). ORF were translated into amino acid sequences which were then compared to GenBank sequences using BLASTP. Identity percentages of ORF based on protein-protein alignment were as follows: Gtf was 81% identical to F. johnsoniae sequence accession number ZP_01245614, NorB was 77% identical to an unidentified Flavobacterium species sequence accession number ZP_01106280, while Trx was 83% identical to F. johnsoniae sequence accession number ZP_01243853. Based on the PCR results, all strains contained norB and trx genes (Table 1). However, no PCR amplification for gtf was observed in four strains (GA-02-14, ARS-1, BM, and HS).

Gene expression under standard conditions
The 16S rRNA gene, used as internal control, was expressed at any given time in all experiments. Gene expression patterns of nine strains of F. columnare varied at 20 h post inoculation under standard growing conditions and are summarized in Table  2. Gene expression differed not only between genomovars but also within each genomovar. Genomovar I displayed three expression patterns. Strains GA-02-14 and HS shared an identical expression profile, showing no expression of both norB and trx gene. Strains BM and ARS-1 also shared a similar expression profile with no expression of norB but positive expression of trx; however, the expression of trx gene in BM was stronger than in ARS-1. FC-RR (an avirulent mutant) displayed a distinct pattern with expression of norB but no expression of the other two genes. Genomovar II strains showed differences in expres-sion patterns, as well. Stronger expression of norB and trx was observed in ALG-00-530 compared to other strains. LSU strain showed weak expression of trx gene, while BZ-1-02 and ALG-00-527 failed to express any of the tested genes.

ALG-00-530 gene expression under iron-limited conditions and in the presence of skin explants
Iron-limited conditions did not interfere with F. columnare housekeeping gene expression ( Table 3). The 16S rRNA gene was uniformly expressed throughout the experiment. Overall, transferrin seemed to have little effect on norB and trx gene expression when compared to standard conditions, although trx transcript levels seemed lowered when transferrin was incorporated into the medium. 2, 2-dipyridyl had an immediate repression effect on norB and trx expression, while the 16S rRNA gene was not affected by this chemical (Figure 2). However, transcript levels for norB and trx went back to high levels 2 h after transfer to iron-limited conditions, remaining stable until 8 h and 24 h for norB and trx, respectively.
When F. columnare cells were incubated in cell culture medium with or without skin explants, a distinct expression pattern was observed (   . However, beyond ribosomal variability, no study has been conducted to investigate the intra-species variation of F. columnare at the single gene level, mainly due to the lack of genetic information available for this species. In the current study, glycosyltransferase gene (gtf), nitric oxide reductase gene (norB), and thioredoxin gene (trx) have, for the first time, been identified and characterized in F. columnare. These genes presented a high similarity to homologous genes within the Flavobacterium genus, showing the highest similarity with F. johnsoniae gene sequences. Although strains within the same genomovar showed identical gene sequences, there was a 4 to 7% nucleotide sequence variation observed between genomovars for each gene. The variability found at the nucleotide level was also translated to the amino acid level. Lee et al., (1998) suggested that single amino acid substitutions could change the biological activity of proteins. We found some nonconserved amino acid substitutions between the genomovar I ALG-00-530 and the genomovar II ATCC 49512 in Gtf, and NorB protein sequences (data not shown). Therefore, the activity of these proteins could differ between F. columnare genomovars.
Our data agreed with the genomovar segregation previously reported for F. columnare. Some examples have been published of such correlation between ribosomal gene-based variability and non-ribosomal gene-based variability i.e. F. psychrophilum (Soule et al., 2005). Non-ribosomal variation divided F. psychrophilum into two lineages that are associated with different host species. This is not the case for F. columnare, since channel catfish is susceptible to both F. columnare genomovars. However, preliminary studies showed genomovar II strains were more virulent to channel catfish than genomovar I strains (Craig A. Shoemaker, USDA-ARS, Auburn, AL, personal communication).
Gtf, norB, and trx genes characterized in the present study have been described as virulence factors in other bacterial species ( . As a result, it is possible that F. columnare continues to express both norB and trx in order to reduce oxidative stress due to iron deficiency or limitation. Gene expression was weak when cells were transferred to cell culture medium, regardless of the presence of catfish skin explants. This may be the result of the relative high salinity in this medium. During the catfish skin explants experiment, bacterial cells were cultivated in cell culture medium 199 diluted 1:1 with bacterial broth (final salinity was about 0.45%). This salinity was required to keep the skin cells alive. It has been reported that increased salinity significantly reduced growth and adhesion ability of F. columnare (Altinok and Grizzle, 2001). The general trend observed was a reduction in gene expression that might be due to the high salinity levels; however, expression levels of the 16S rRNA gene seemed to be unaffected.
In conclusion, our data confirmed the genomic diversity of F. columnare at the single gene level. Nucleotide sequences of gtf, norB, and trx of F. columnare differed between genomovars I and II. Multiple gene expression patterns existed both between and within genomovars. Although the relation between gene expression pattern and virulence is unclear, this study addressed for the first time in vitro gene expression in F. columnare. Further studies are ongoing in order to confirm the difference in virulence between genomovars I and II. Nevertheless, and due to the clear genetic division between genomovars, we strongly recommend the inclusion of more than one genomovar in future F. columnare studies.