GO term enrichment analysis.

Gene Ontology (GO) term enrichment analysis of the differentially expressed genes in the two strains revealed a consistent pattern of functional enrichment. 15 enriched terms were identified for CMCP6 and 16 for YJ016, with the primary differences being the ranking of terms. In the cellular component term hierarchy, which describes the localization of the gene, the only significantly enriched term for both CMCP6 and YJ016 was GO:0016020 (membrane). In the molecular function hierarchy, which describes molecular activities, the only significantly enriched term, again appearing in both species, was GO:0060089 (molecular transducer activity). All remaining significantly enriched terms identified were within the biological process hierarchy, and included such biological roles as cellular response to stimulus, signal transduction, cellular communication, formamide and other amide metabolism, and (in the case of YJ016), taxis (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Enriched Gene Ontology Categories of Differentially Expressed Genes in human serum and artificial seawater.

https://doi.org/10.1371/journal.pone.0114376.t002

Substrate binding and transport.

In the human host, V. vulnificus must overcome a number of physiological challenges to survive and multiply. In vivo growth requires the acquisition of scarce but essential nutrients, such as iron and sulfur, which are critical to cell function. Indeed we identified genes involved in iron acquisition to be upregulated in human serum. Iron is an essential micronutrient for V. vulnificus, and is critical for growth and virulence [17], thus this bacterium employs siderophore-based systems to chelate iron from high-affinity binding proteins such as ferritin, transferrin, and lactoferrin. V. vulnificus also makes use of an OM receptor, HupA, to acquire iron from heme-containing compounds such as hemoglobin and cytochromes [18]. In our transcriptome analysis, we found the receptor and transport genes for the hydroxamate siderophore, the biosynthetic, transport, and receptor genes for the catechol siderophore (vulnibactin), and hupA to be induced upon exposure to human serum. This finding corroborates a previous study in which V. vulnificus expression of vuuA and fhuA (encoding the vulnibactin and hydroxamate siderophore receptors, respectively), was significantly enhanced in human serum compared to natural seawater [19].

V. vulnificus utilizes three TonB systems (TonB1, TonB2, and TonB3) to transport both the hydroxamate-type siderophore and vulnibactin, as well as heme and hemoglobin [20]. The TonB systems consist of three integral inner membrane proteins (TonB, ExbB, and ExbD), which generate the energy needed to activate the OM TonB receptor proteins. The operons coding for TonB1 and TonB2 were found to be significantly upregulated in HS, however TonB3 was not differentially expressed. This finding strongly supports a previous study from the Crosa lab which found V. vulnificus TonB1 and TonB2 systems to be induced under iron-limiting conditions whereas TonB3 transcription did not change in response to iron concentrations [21].

Our study found the operon (potABCD) encoding spermidine transport to be up-regulated in human serum. Polyamines such as spermidine, are a class of small organic compounds with a hydrocarbon backbone and multiple amino groups, are known to play a crucial role in maintaining optimal conformation of nucleic acids, and are therefore essential for normal cellular growth and multiplication [22]. However, research has unveiled some novel functions of polyamines in microorganisms linking these compounds with biofilm formation, escape from phagolysosomes, bacteriocin production, toxin activity, and protection from oxidative and acid stress [22]. Upregulation of spermidine transport genes in HS indicates that polyamines could serve an important function for V. vulnificus within the human host and this proposition deserves further attention.

Clinical isolates of V. vulnificus harbor a 33-kb genomic island (region XII) located on chromosome II, which contains an arylsulfatase gene cluster, a sulfate reduction system, two chondroitinases, and an oligopeptide ABC transport system [10]. It has been proposed that acquisition of this genomic island may serve to enhance fitness in the aquatic environment and/or human host, however studies to validate this proposition for V. vulnificus have yet to be reported. We found a subset of genes within the genomic region XII to be upregulated in human serum, particularly those within the arysulfatase gene cluster. Arysulfatases hydrolyze arylsulfate ester bonds to release free sulfate, which is essential for bacterial growth and survival [10], [23]. These are typically expressed under conditions of sulfur starvation, but have also been shown to be regulated by monoamine compounds such as norepinephrine, which has been implicated in quorum sensing signaling within the human host [24].

Intracellular signaling.

Cyclic-di-GMP is an intracellular second messenger that integrates environmental signals and has been demonstrated to regulate several distinct cellular processes in vibrios such as motility, biofilm formation, virulence, and rugose colony morphology [25]–[27]. This signaling molecule has been proposed to be a critical component for the transition of V. cholerae from the aquatic environment to the host [28]. The intracellular concentrations of this second messenger are regulated by diguanylate cyclases (DGCs) and phosphodiesterases (PDEs) which synthesize and degrade c-di-GMP, respectively. Studies have shown that phosphodiesterase activity of the response regulator, VieA, reduces intracellular c-di-GMP levels which allows for optimal gene expression of virulence factors, specifically toxT, the transcriptional activator of toxin-coregulated pili (tcp) (the major colonization factor of V. cholerae) as well as cholera toxin (ctx) [28]. Furthermore, through the action of c-di-GMP, VieA negatively regulates expression of vps genes which contribute to exopolysaccharide production and biofilm formation in V. cholerae [29]. A similar regulatory response system (RocR/SadR) has been identified in Pseudomonas aeruginosa which was also shown to control biofilm and virulence phenotypes in a reciprocal manner in response to environmental signals [30]. Interestingly, our study found the sensor histidine kinase vieS and the cognate response regulator vieA to be significantly upregulated in V. vulnificus when exposed to human serum. Considering that V. vulnificus appears to lack toxT, tcp and ctx homologs, the downstream effect of this two-component system deserves further investigation as it may play a critical role in the transition from environment to host, as is the case for V. cholerae.

In V. cholerae, the ToxR regulon encodes over 20 genes that aid in intestinal colonization, toxin production and survival within the host [31]. Homologs of the toxRS operon have been identified in V. vulnificus and the ToxRS signal transduction system has been shown to stimulate hemolysin production and alter the outer membrane protein expression profile [32]. In addition to upregulation of toxRS expression in HS, we found increased expression of three hemolysins, along with differential expression of several outer membrane proteins.

Intracellular signaling also occurs through quorum sensing (QS) in which signaling molecules (autoinducers) accumulate extracellularly, bind a membrane-bound sensor to effect a collective change in gene expression patterns within the population. This results in the modulation of processes such as bioluminescence, biofilm formation, virulence, and fitness factor production (see [33] for a review on QS). Three systems of bacterial communication exist in Gram-negative bacteria and each system utilizes chemically different autoinducers. Only one functional quorum sensing pathway (autoinducer-2, involving interspecies communication) has been described in V. vulnificus, but it has been proposed that this bacterium possesses other, not yet identified, compensatory channels for quorum sensing [34]. The AI-3 system involves inter-kingdom cross-signaling in which bacteria can sense and respond to host-derived signals [35]. The sensor kinase, QseC, of this two-component signaling system acts as an “adrenergic receptor” that senses and responds to epinephrine, norepinephrine, and AI-3 in enterohemorrhagic E. coli, subsequently activating virulence genes [35], [36]. Interestingly, we identified the qseBC homolog in V. vulnificus to be upregulated in cells exposed to HS. Whether the QseBC system functions as a genuine inter-kingdom signaling network for V. vulnificus is an area of research that deserves prompt investigation as this system may be used to detect and respond to eukaryotic adrenergic signals present in the GI tract and/or human serum, thereby assisting the bacterium's navigation throughout the human host.

It is unclear whether V. vulnificus employs AI-1 (intraspecies) signaling as a LuxI/LuxR pair has yet to be genetically identified in this species [37], [38]. However, we identified a gene encoding a LuxR-family transcriptional regulator to be upregulated in human serum. In AI-1 signaling, LuxR serves as the response regulator which binds AHL signals and subsequently interacts with DNA to effect changes in gene expression in response to the stimulus. Interestingly, many bacteria (both with and without AI-1 signaling capabilities) have been found to possess LuxR-family proteins for which there is no obvious cognate LuxI synthase [39]. Recent research suggests these “LuxR-family orphans” may help in fine tuning of existing QS regulatory networks, control independent regulons, or allow bacteria to detect and respond to neighboring interspecies or inter-kingdom signals [39], [40]. Indeed, these propositions suggest the need to investigate the relevance of these LuxR-family orphans in pathogens such as V. vulnificus.

Response to heat/protein stabilization.

Exposure to temperature stress can have deleterious effects on cellular components, particularly due to protein denaturation. In response, bacterial cells will activate the heat shock response (HSR) in order to protect the cell from physiological stress. This signaling pathway is governed by the transcription factor, σ³², which rapidly induces transcription of several heat-shock proteins (HSPs) including molecular chaperones and ATP-dependent proteases that aid in repair or disposal of damaged proteins. Exposing V. vulnificus to HS at 37°C (compared to ASW at room temperature) resulted in a remarkable upshift in σ³², hsp chaperones, and proteases. Interestingly, this effect was still prominent at two hours of incubation in serum. HSR typically occurs within minutes of temperature upshift and reaches a new steady-state level within ca. 30 m [41]. Considering the cross-protective qualities of HSPs [42], it is plausible that these genes also promote cellular stability in the presence of antimicrobial constituents within human serum.

Toxins and exoenzymes.

V. vulnificus produces cytotoxic effects to human cells by secreting hemolysins and extracellular proteases [3]. Hemolysins, such as the vvhA cytolysin, cause cellular damage by forming pores in the cell's membrane which can lead to apoptosis, vascular permeability and hypotension [43]. Extracellular proteases, such as the metalloprotease vvpE and elastase, cause tissue necrosis and cutaneous lesions as well as vascular permeability and edema. While the pathogenic effects of these putative virulence factors have been thoroughly investigated, some studies indicate that they may not play a significant role in virulence [44]–[46]. This conclusion is supported in our current study, which found no change in expression of vvhA between ASW and HS, and down-regulation of vvpE and elastase in HS. However, we found three other putative hemolysins and several proteases (including two zinc-dependent proteases and two collagenases) to be upregulated in human serum. Further evaluation of these putative toxins is necessary to assess if and how they contribute to virulence.

Transcription factors.

In V. vulnificus, the alternative sigma factor, RpoS, aids in adaptation to environmental stress (such as osmotic shock, nutrient stress, and oxidative stress) and has been shown to be required for full motility [47]. We found rpoS expression to be enhanced in ASW, highlighting the necessity of the general stress response in the environment. Additionally, we found the Crl transcriptional regulator to be upregulated in ASW. In E. coli, Crl is an auxiliary factor that binds to σ^S and enhances its affinity for certain promoters, including the genes necessary for curli fimbriae formation [48]. These extracellular matrix components facilitate surface colonization and biofilm formation, and expression is positively regulated by curli subunit gene D (CsgD) [48], [49]. Previous studies have identified CsgD homologs in V. vulnificus, V. parahaemolyticus, and V. cholerae [50]. In V. cholerae, the CsgD homolog, VpsT, has been shown to enhance biofilm formation and activate genes involved in VPS exopolysaccharide production [51]. In our study, expression of a VpsT homolog was upregulated in V. vulnificus cells (CMCP6 only) in ASW.

Expression of the gene encoding the alternative sigma factor RpoN (σ⁵⁴) was also upregulated in ASW. This transcription factor regulates over 30 operons in E. coli, nearly half of which are involved in nitrogen assimilation and metabolism [52]. The remaining σ⁵⁴-dependent genes have been proposed to neutralize adverse conditions and/or prevent depletion of metabolites and energy resources in limiting environments [52]. In some Vibrio spp., σ⁵⁴ has been documented to upregulate genes involved flagellar synthesis and motility [53]–[55]. Interestingly, deletion of rpoN in V. cholerae reduces the competitive fitness of cells (relative to the wild-type) when colonizing the infant mouse model, however this deletion enhances fitness in intestinal colonization by V. parahaemolyticus in the adult mouse model [53], [55]. Furthermore, this sigma factor has been implicated in regulating EPS in V. vulnificus and V. fischeri [56], [57]. Thus, in ASW, RpoN may be regulating a number of processes including nitrogen metabolism, motility, colonization, and biofilm formation and further investigation into this regulatory network in V. vulnificus is prompted.

IscR is a transcription factor that regulates the isc operon in E. coli and controls the expression of over 40 genes including those involved in Fe-S cluster biosynthesis, anaerobic respiration, and biofilm formation [58]. Recently, an IscR homolog was identified in V. vulnificus that regulates a number of genes possibly involved in pathogenesis, including genes involved in motility and adhesion, hemolytic activity, and oxidative stress response. Interestingly, our study revealed upregulation of five out of eight genes within the isc operon, including iscR, when cells were exposed to ASW. Additionally, IscR regulated genes involved in chemotaxis, methyl-accepting chemotaxis proteins, and a glutaredoxin were also upregulated in our study, reaffirming the relationship identified by Lim et al. (2014). While IscR appears to play a critical role in protecting the cell from reactive oxygen species generated by host cells during infection [59], it may be that IscR also plays an important role in environmental survival.

Adherence and Biofilm formation.

The tad (tight adherence) genes encode the machinery needed to assemble Flp pili (fibrils) that mediate adherence to surfaces. The tad genes have been found to be essential for adherence, biofilm formation, colonization, and pathogenesis in a number of genera, and are considered to be instrumental in the colonization of diverse environmental niches [60]–[63]. These genes are present on a genomic island referred to as the “widespread colonization island” which is present in a wide variety of bacteria, including several human pathogens [61]. Some species harbor more than one distinct tad locus, including V. vulnificus, which contains two loci on chromosome I and a third locus on chromosome II. One of these three tad loci was differentially expressed in our study with expression being enhanced in ASW relative to HS. The function(s) of the tad loci in V. vulnificus have yet to be investigated but it is reasonable to conjecture that this widespread colonization island provides important functions for this organism specifically regarding adherence and colonization. Indeed, a previous RNAseq study found that (when compared to growth in a nutrient-rich medium), a V. vulnificus biotype 3 strain isolated from a wound infection expressed the Flp-coding region [64].

Additionally, we found several genes involved in biogenesis of the type IV mannose-sensitive haemagglutinin (MSHA) pilus to be upregulated in ASW. In V. cholerae, MSHA facilitates attachment to chitinous substrates, such as zooplankton exoskeletons, allowing for subsequent biofilm formation [65], [66]. In contrast to the toxin-coregulated pilus (a predominant virulence factor in V. cholerae), MSHA biogenesis has been shown to be repressed in vivo, acting as an “anticolonization factor” in the human host [67]. Our results suggest that MSHA may play a similar role in V. vulnificus in which this pilus mediates attachment to substrates in the aquatic environment.

V. vulnificus produces extracellular surface polysaccharides (EPS) and capsular polysaccharides (CPS) both of which are involved in biofilm formation. CPS is recognized as an essential virulence factor for V. vulnificus and negatively impacts biofilm formation [68], [69], whereas EPS production is essential for bacterial attachment and biofilm formation [57]. V. vulnificus has at least three kinds of EPS and we found one of these polysaccharide biosynthetic loci to be significantly upregulated in ASW. This set of EPS genes (homologous to the “syp genes” originally identified in V. fischeri) appears to be transcriptionally regulated by σ⁵⁴ [56], and produces wrinkled colonies, pellicle formation, and matrix production when induced [70]. Upregulation of syp-like genes in ASW suggests this EPS locus may play an important role in environmental persistence.

Intracellular signaling and chemotaxis.

As previously mentioned, the signaling pathway regulator c-di-GMP controls phenotypes associated with biofilm formation and planktonic growth in a reciprocal manner. Intracellular levels of c-di-GMP increase through the action of diguanylate cyclases (DGCs) which synthesize this second messenger. Subsequently, c-di-GMP levels can be reduced through the action of phosphodiesterases (PDEs) which degrade this molecule. We identified 17 DGCs and six PDEs to be upregulated in ASW, as well as two PDEs to be downregulated in ASW. This result suggests that cells in this environment have higher intracellular levels of c-di-GMP. In V. vulnificus, c-di-GMP positively regulates biofilm formation by regulating production of EPS, thus cells in ASW would presumably exhibit enhanced biofilm formation and possibly decreased virulence potential [25], [71]. This speculation is further supported through the observed upregulation of rpoS expression in ASW, as RpoS has been shown to modulate c-di-GMP levels by enhancing expression of DGCs.

Chemotaxis is initiated by membrane-bound chemoreceptors called methyl-accepting chemotaxis proteins (MCPs) which bind a ligand resulting in a signal transduction cascade to direct the flagellum's activity. The role of chemotaxis in the process of infection has been examined in several bacteria/enteric pathogens and for many of these, motility is an important virulence factor. Interestingly, chemotaxis in V. cholerae inhibits its ability to colonize the small intestine of infant mice and is inversely regulated with expression of virulence traits [72].

V. cholerae has three chemotaxis gene clusters, denoted cheA-1, cheA-2, and cheA-3, although only cheA-2 has been found to be essential for chemotaxis in growth media [73]. In the current study, we identified two of these homologs (cheA-2 and cheA-3) in V. vulnificus and found the cheA-3 homolog (located on chromosome II) to be upregulated in ASW. Additionally, many of the MCPs located throughout the genome were significantly expressed in this condition. In V. cholerae cheA-3 is not required for chemotactic control of flagellar motility, thus it has been suggested that this operon could regulate flagellum-independent motility (e.g. twitching motility) [73], a proposition that could be extended to V. vulnificus. The intimate relationship between chemotaxis and virulence expression that has been documented in V. cholerae prompts further investigation into these mechanisms in V. vulnificus. For both of these pathogens, chemotaxis is assuredly a vital fitness factor in both environmental and pathogenic settings.

V. vulnificus C-genotypes possess an operon homologous to the RsbRST stress module (or stressosome) which is classically found in Gram-positive bacteria such as Bacillus subtilis and Staphylococcus aureus [11] and governs the “general stress response”, sometimes playing a role in virulence. This supramolecular signaling complex perceives signals related to environmental stress or nutrient limitation, triggering a signal transduction phospho-relay that ultimately activates the sigma factor (σ^B), which will then bind to the core RNA polymerase to direct transcription of over 150 genes involved in stress adaptation [74]. For organisms such as V. vulnificus that possess stressosome orthologs, but lack σ^B, the stressosome has been proposed to be involved in regulating aerotaxis, two-component signaling systems, and the biosynthesis of secondary messenger signaling molecules [74]. We assessed 20 C-genotypes and 29 E-genotypes for the presence of this operon and found this module to be specific to C-genotypes, with 75% of C-genotypes and none of E-genotypes possessing the entire operon (S1 Table). Furthermore, we found these genes to be upregulated in ASW relative to HS, suggesting that the stressosome serves an important function for the organism.

We examined gene expression of the rsbRST operon using relative qRT-PCR and found these genes to be repressed more than 100-fold in HS relative to ASW (Fig. 3), validating our RNAseq transcriptome findings. To further characterize expression of the stressosome we examined how this operon is regulated temporally using relative qRT-PCR. Interestingly we found stressosome genes to be upregulated over time in ASW (Fig. 4A), as well as in HS (Fig. 4B), albeit to a lesser extent. The enhanced expression of stressosome-related genes in ASW relative to HS suggests an ecological function for this system, thus we used membrane diffusion chambers to examine temporal gene expression of rsbR and rsbT, in situ. When placed in their native estuarine environment, cells upregulated rsbR and rsbT over 10-fold within 6 hours and continued to express these genes for up to 48 hrs (Fig. 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Gene expression of the stressosome in human serum relative to artificial seawater.

Relative qRT-PCR was performed on CMCP6 to examine expression of select stressosome genes (rsbRST and a downstream two-component sensory protein, here denoted “TCSP”) confirming RNAseq results. GAPDH was used to normalize expression data. Asterisks represent significant differences in expression p<0.0001 (One-way ANOVA).

https://doi.org/10.1371/journal.pone.0114376.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Temporal gene expression of the stressosome during exposure to A) artificial seawater or B) human serum.

Relative qRT-PCR was performed on CMCP6 to examine expression of select stressosome genes (rsbRST and a downstream two-component sensory protein, here denoted “TCSP”) after incubation in each condition for 2 hrs relative to 30 min. GAPDH was used to normalize expression data. Asterisks represent significant differences in expression p<0.0001 (One-way ANOVA).

https://doi.org/10.1371/journal.pone.0114376.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. Temporal gene expression of the stressosome in natural estuarine waters.

Cells of CMCP6 were incubated in natural estuarine waters for 2 days using membrane diffusion chambers. Relative qRT-PCR was performed on samples collected at several time points and expression of select stressosome genes (rsbR, circles; rsbT, squares) was analyzed relative to 0 hrs of incubation (prior to in situ exposure). GAPDH was used to normalize expression data. Asterisks represent significant differences in expression (One-way ANOVA).

https://doi.org/10.1371/journal.pone.0114376.g005

Further studies are warranted to assess if and how the stressosome module contributes to pathogenicity and/or niche colonization in strains of V. vulnificus harboring these genes. It may be that this system has been optimized to allow for integration of various signals into a single output, thereby allowing for fine-tuning of stress responses in a particular environment. Expression of this operon in situ also suggests a role for this system in the environment and it may have been acquired to aid in niche expansion.

Experimental design.

Two clinically isolated strains of V. vulnificus (CMCP6 and YJ016) were selected due to availability of their genome sequences. Strains were grown in heart infusion (HI) broth (BD, New Jersey) at 30°C overnight at 200 rpm and then diluted 1∶100 v:v into fresh HI and grown at 30°C for ca. 90 m at 200 rpm until cells reached logarithmic phase (OD₆₁₀ 0.15–0.25). Cells were pelleted at 5000×g at room temperature for 10 m, resuspended in 10 parts per thousand (ppt) artificial seawater (ASW), and inoculated into either 10 ppt ASW or normal human serum (MP Biomedicals) to a final concentration of ca. 1e8 CFU/ml. Cells were incubated in 10 ppt ASW at room temperature (22°C), or human serum at 37°C, on a rotisserie for 2 hours. ASW at 22°C was chosen to represent a controlled version of the natural estuarine waters and temperature this bacterium encounters in its native environment. Human serum at 37°C was chosen to reflect the host environment and human body temperature. Pooled human serum (HS) used in this study was collected from male donors and contained nutrients in the form of glucose (96 mg/dl) with total protein levels of 5.5–7.5 g/dl [76].

RNA extraction and purification.

Two biological replicates were performed for each condition and after 2 hours of incubation cells were treated with 2 volumes of RNAprotect (Qiagen) following manufacturer's instructions. Preserved pellets were resuspended in TE buffer, pH 8.0 (Ambion) with 1 mg/ml lysozyme (Sigma Aldrich) and vortexed at medium speed for 30 m. RNA was extracted using RNeasy Midi Kit (Qiagen) following protocol with on-column DNase I treatment. RNA was eluted from the column using nuclease-free water and a second post-extraction DNase treatment using TURBO DNA-free (Life Technologies) following the rigorous DNase treatment protocol. RNA quality and quantity was assessed using a NanoDrop spectrophotometer (Thermo Scientific) and RNA samples having a 260/280 ratio <1.7 were not used in downstream applications.

To confirm the complete removal of DNA, endpoint PCR was performed on RNA samples using Promega's Go-Taq DNA polymerase, 5X Green GoTaq Reaction Buffer, 10 mM dNTP mix, and primers targeting the species-specific vvhA gene [7]. Cycling parameters were according to the manufacturer's recommendations, with an annealing temperature of 53.1°C and 40 cycles of amplification. Any amplification of the vvhA gene was indicative of DNA contamination in which case the RNA was not used for downstream processes.

RNA integrity was assessed using an Agilent 2100 Bioanalyzer and 6000 Nano kit (Agilent Technologies). The 23S/16S rRNA ratio and RNA integrity number (RIN) were confirmed to be >1.8 and close to 10, respectively. Due to the enhanced sensitivity of this technique, the electropherogram for each sample was also screened for residual DNA contamination. Samples yielding bands >4000 nt were not used in downstream applications. Total RNA samples were quantified using a NanoDrop Spectrophotometer (Thermo Scientific).

mRNA Enrichment and cDNA Library Preparation.

To enhance the sequencing coverage of mRNA transcripts, we depleted highly expressed rRNA transcripts using Epicentre's Ribo-Zero kit (Epicentre Biotechnologies). Remaining mRNA was assessed for quality and levels of rRNA contamination using the Agilent Bioanalyzer and RNA 6000 Pico Chip (Agilent Technologies). Samples having >1% rRNA contamination were not used for downstream applications. mRNA was quantified using Qubit's 2.0 Fluorometer and RNA High Sensitivity Assay Kit (Life Technologies).

DNA libraries complementary to mRNA sequences were prepared for each sample using ScriptSeq v2 RNA-Seq Library Preparation Kit (Epicentre Biotechnologies). This kit, compatible with Illumina Technologies, enabled directional, paired-end sequencing. Libraries were prepared following manufacturer's instructions. The size and quality of each cDNA library was assessed using Agilent Bioanalyzer's High Sensitivity DNA kit (Agilent Technologies). Insert sizes were approximately 400 bp+/−35 bp. cDNA library quantification was performed using Qubit's 2.0 Fluorometer and Quant-iT dsDNA High Sensitivity Assay Kit (Life Technologies). Samples were stored at −20°C and transported to the David H. Murdock Research Institute (Kannapolis, N.C.) for sequencing using Illumina's Genome Analyzer.

Data preprocessing.

Approximately 61 to 79 million paired-end raw reads were generated for each set of biological replicates (S2 Table). RNA sequencing data is publicly available in NCBI's BioProject database (BioProject ID PRJNA252365). Raw Illumina sequence reads were filtered to remove reads which fell below the sequencer quality threshold. A second round of filtering using fastq-mcf [77] removed low-quality regions of sequence using a quality trimming window of size 4 and a minimum quality score of 15. The resulting reads were of high average quality and read lengths were tightly clustered around 100 nucleotides. The remaining filtered reads were aligned to reference genomes using Bowtie2 [78]. The alignment rate using Bowtie2 with default settings for paired-end reads was greater than 95% for all samples in the study indicating high quality and reproducibility of the data set (S2 Table). The view and sort functions of SAMTools [79] were used to convert the aligned read data into the correct format for subsequent analysis.

Expression analysis.

Read counts were summarized against the CMCP6 and YJ016 reference genomes using FeatureCounts [80]. Differential gene expression analysis was performed using the EdgeR package [81] following the approach outlined by Jenkins [82]. EdgeR identifies significant differentially expressed genes out of genome-scale count data using exact tests based on the negative binomial distribution. Replicate data for each condition was used for normalization using the TMM method [83]. The artificial sea water (ASW) condition was treated as the baseline condition for analysis. Genes identified as differentially expressed in human serum (HS) relative to ASW, with a p-value less than 0.0001 were selected for further analysis.

Gene Ontology (GO) enrichment analysis.

Gene category enrichment analysis was performed using the Ontologizer software [84]. The study sets of differentially expressed genes generated by EdgeR were compared to the population set of GO-annotated genes in the published CMCP6 and YJ016, respectively. The Ontologizer makes several different enrichment analysis methods available. The Parent-Child Union method, which avoids detection artifacts due to the hierarchical structure of the GO, was used to identify enriched GO term categories, and the Benjamini-Hochberg [85] method for multiple-testing correction was applied in order to minimize the false discovery rate. Reference Gene Ontology OBO file [86] and genome-specific GO annotations for V. vulnificus CMCP6 and YJ016 were downloaded from UniProt-GOA [87] on April 1, 2014. 3066 genes in CMCP6 and 3107 genes in YJ016 had available GO term mappings, resulting in 331 of 469 DE genes in CMCP6 and 406 of 661 DE genes in YJ016 being included in the enrichment analysis. Differential expression is summarized by GO terms identified as enriched in the study set relative to the population set, with a p-value <0.05.

PCR analysis of RsbRST stressosome module.

Using PCR, 20 V. vulnificus C-genotypes and 29 E-genotypes were assessed for the presence of stressosome-associated genes (rsbR, rsbS, rsbT, rsbU, and a downstream two-component sensor protein). Primers (Sigma Aldrich) were designed for each gene using all three sequenced C-genotype strains of V. vulnificus (CMCP6, YJ016, and M06-24) reported in NCBI's database. Primer sequences are listed in S3 Table. Optimal primer quality and fidelity were assessed using IDT OligoAnalyzer 3.1 software, and primer specificity was initially analyzed using in silico PCR [88]. DNA extracted from each strain was subjected to 30 cycles of PCR with Promega's 5X Green GoTaq Reaction Buffer, 10 mM PCR Nucleotide Mix, 1.25 U GoTaq DNA Polymerase, and 0.5 µM of each primer. Thermal cycling parameters were followed according to manufacturer's recommendations (Promega) and PCR products were visualized by gel electrophoresis on 1% agarose gels stained with ethidium bromide.

Gene expression of RsbRST stressosome module.

Gene expression of selected genes associated with the stressosome module was examined in human serum relative to 10 ppt ASW using relative quantitative reverse transcription-PCR (qRT-PCR) as previously described [89]. Primers were designed to be suitable for qRT-PCR and PCR amplification efficiency was assessed using parameters previously described [89]. Cells of CMCP6 were treated and RNA was extracted as described in the “experimental design and sample preparation” section above. Total RNA (1 µg) was reverse transcribed using qScript cDNASuperMix (Quanta Biosciences) and 50 ng of cDNA template was carried over for quantitative PCR (qPCR). qRT-PCR was performed on three technical replicates for each sample using PerfeCTa SYBR green FastMix, LowROX (Quanta Biosciences). Negative controls and “no-RT” controls were employed to rule out the influence of DNA contaminants or residual genomic DNA, respectively. Expression levels of each gene were normalized by using an endogenous control gene (glyceraldehyde-3-phosphate dehydrogenase [GAPDH]) to correct for sampling errors. Fold changes in expression levels were measured using the Pfaffl equation [90], taking into account the differences in PCR efficiencies between primer sets. Significant differences in gene expression were assessed using a one-way analysis of variance (ANOVA) followed by Bonferroni's post hoc test for multiple comparisons. Significance was determined by using a 95% confidence interval. Gene expression data were analyzed by using GraphPad Prism (version 5.0; GraphPad Software Inc.).

To further investigate the expression pattern of the stressosome module, gene expression was analyzed over time in human serum and 10 ppt ASW (2 hrs relative to 30 min) using the protocol described above. Additionally, in situ gene expression by cells incubated in natural estuarine waters was also performed as previously described [91]. Briefly, laboratory grown cells were diluted into 15 ppt ASW, injected into sterile (0.22 µm) membrane diffusion chambers, and deployed in estuarine waters along Core Creek in Beaufort, North Carolina. Chambers were deployed in September of 2012 and samples were collected periodically for up to 2 days. For the duration of the experiment, estuarine water temperatures ranged from 13.8°C to 17.2°C and salinity ranged from 17 ppt to 22 ppt. Samples were treated with RNAprotect (Qiagen) within 10 minutes of collection. RNA extraction followed by qRT-PCR was performed as previously mentioned.