Role of Anaerobiosis in Capsule Production and Biofilm Formation in Vibrio vulnificus

Abstract

Vibrio vulnificus, a pervasive human pathogen, can cause potentially fatal septicemia after consumption of undercooked seafood. Biotype 1 strains of V. vulnificus are most commonly associated with human infection and are separated into two genotypes, clinical (C) and environmental (E), based on the virulence-correlated gene. For ingestion-based vibriosis to occur, this bacterium must be able to withstand multiple conditions as it traverses the gastrointestinal tract and ultimately gains entry into the bloodstream. One such condition, anoxia, has yet to be extensively researched in V. vulnificus. We investigated the effect of oxygen availability on capsular polysaccharide (CPS) production and biofilm formation in this bacterium, both of which are thought to be important for disease progression. We found that lack of oxygen elicits a reduction in both CPS and biofilm formation in both genotypes. This is further supported by the finding that pilA, pilD, and mshA genes, all of which encode type IV pilin proteins that aid in attachment to surfaces, were downregulated during anaerobiosis. Surprisingly, E-genotypes exhibited distinct differences in gene expression levels of capsule and attachment genes compared to C-genotypes, both aerobically and anaerobically. The importance of understanding these disparities may give insight into the observed differences in environmental occurrence and virulence potential between these two genotypes of V. vulnificus.

INTRODUCTION

Vibrio vulnificus, an opportunistic human pathogen, is the leading cause of seafood-related deaths in the United States (1). Ubiquitous in estuarine and coastal waters, V. vulnificus has been isolated from a multitude of sources, including sediment, oysters, water, shrimp, clams, and fish (2, 3). Consumption of raw or undercooked seafood, particularly oysters, causes rapid septicemia, carrying a 50% mortality rate (1).

V. vulnificus is an extremely diverse species comprised of three distinct biotypes based on specific host ranges and phenotypic differences (4, 5). Although all three biotypes have been associated with human pathogenicity, biotype 1 is the major cause of primary septicemia in humans. Biotype 1 isolates can be further subdivided into two distinct genotypes, clinical (C) and environmental (E) (6). These genotypes can be easily and rapidly determined by utilizing endpoint PCR to amplify the species-specific virulence-correlated gene (vcg), specifically targeting polymorphisms within this allele (4). Of V. vulnificus strains isolated from clinical cases, 90% have the vcgC allele (C-genotype), while 87% of environmental isolates have the vcgE (E-genotype) allele (4, 7). Although genotype correlates with virulence, it is not necessarily predictive, as shown by Thiaville et al. They found that 3 of the 9 most virulent strains, when inoculated subcutaneously in iron dextran-treated mice, possessed the vcgE allele, which confirms that not only those strains with the vcgC allele are capable of causing disease (8).

Multilocus sequence typing of six highly conserved housekeeping genes also supports the division of biotype 1 strains of V. vulnificus into two distinct lineages, which correlates with the C- and E-genotyping system (9, 10). Additionally, 16S rRNA typing has shown that C-genotypes correlate highly (66.7%) with B-type 16S rRNA, whereas E-genotypes correlate (80%) with A type 16S rRNA (11, 12). Despite having molecular markers for differentiating between strains associated with disease and those typically considered less virulent, the mechanisms for how V. vulnificus survives and colonizes the host still need more investigation.

The human gut is described as being anaerobic; however, it has been shown that there is in fact a gradient of oxygen present in mammalian hosts (13, 14). He et al. described an oxygen gradient present in the intestines of mice which showed a decrease in oxygen tension from the midstomach, midduodenum, midcolon/small intestine, and finally the sigmoid colon-rectal junction (13). Orally ingested V. vulnificus must travel through these oxygen gradients, attach to intestinal epithelial cells, and be able to penetrate into the bloodstream for primary sepsis to occur. Although anaerobiosis and subsequent genotypic changes have not been characterized in V. vulnificus, there have been multiple studies showing the effect of oxygen on V. cholerae (15–18). Various proteins have been differentially expressed in proteomic evaluations of V. cholerae incubated with and without oxygen (19). Six proteins highly expressed during anaerobiosis were located on the Vibrio pathogenicity island (20), and five of these were involved in intestinal colonization (19). Other notable proteins involved in iron acquisition, metabolism, exopolysaccharide production, and biofilm maturation were enhanced by anaerobiosis (19). Understanding the differences between C- and E-genotypes in response to anaerobiosis may help elucidate their specific virulence potentials.

Capsule production in V. vulnificus is an important virulence determinant and has been shown to significantly enhance pathogenicity (21–23). Encapsulated (opaque colonies on solid media) strains can undergo a reversible phase transition to the nonencapsulated (translucent colony) phenotype (24). Translucency negatively correlates with virulence, as shown by decreased serum resistance, lowered lethality in mice, and susceptibility to phagocytosis (23, 25–27). V. vulnificus has given rise to multiple CPS types (28–30), two of which show significant homology with the group I and group IV CPS operon in Escherichia coli (31–33). Encoded in these operons are three highly conserved genes, wza wzb and wzc, involved in transport of polymers across the outer membrane of the cell (31, 33, 34). Environmental conditions have been previously shown to initiate differential regulation of group 1 CPS genes and play a role in phase variation in this bacterium as well as other human pathogens (29, 32, 35, 36). It is important to understand the regulation of CPS genes in V. vulnificus during anaerobiosis, a condition it encounters in its natural environment and in the human host (14, 15, 37).

Polysaccharide production has been previously implicated in the ability for multiple Vibrio species to form biofilms, i.e., surface-associated bacterial communities enclosed by a polysaccharide matrix (38–43). A correlation between polysaccharide production and biofilm formation has been demonstrated in V. vulnificus, V. cholerae, V. parahaemolyticus, and V. fischeri under aerobic conditions; however, this relationship has yet to be investigated anaerobically (40, 43–46). The genome of V. vulnificus contains multiple highly conserved polysaccharide loci, including those that produce exopolysaccharides (EPS), lipopolysaccharides (LPS), and capsular polysaccharides (CPS) (31, 33, 34, 36, 47–49). EPS has been shown to correlate positively with biofilm formation in V. vulnificus (50), but previously a negative relationship between CPS and biofilm formation aerobically in this bacterium, as well as other Vibrio species, has been described (43, 44, 51). However, the effect of oxygen availability on biofilm formation has not been examined in V. vulnificus.

Additionally, type IV pili have been previously shown to contribute to attachment to both abiotic and biotic surfaces, as well as having a role in the initial stages of biofilm formation in other Vibrio species (40, 41, 43, 44, 52). Specifically, pilA, pilD, and mshA, encoding a pilin protein subunit, prepilin peptidase, and mannose-sensitive hemagglutinin, respectively, have been shown to be important for biofilm formation as well as attachment to chitinous substrates in V. vulnificus (39, 53, 54). Again, the role of oxygen on expression of type IV pili and the subsequent effects on biofilm formation in V. vulnificus are poorly understood.

Along with type IV pili, flagella have exhibited a role in the formation of biofilms in multiple human pathogens (40, 55). The role of flagella in V. cholerae varies between strains, but lack of functional flagella has been associated with an overall decrease in attachment (40, 42). V. parahaemolyticus also exhibits decreased biofilm formation when genes for multiple proteins involved in flagellum production are mutated (43). As with other Vibrio species, V. vulnificus utilizes a single polar flagellum that aids in motility and attachment. Aerobically, nonflagellated mutants of V. vulnificus exhibit an inability to attach to polystyrene and glass wool (56), and we further examined its role in attachment during anaerobiosis.

As a facultative anaerobe, V. vulnificus has the capability to metabolize under both aerobic and anaerobic conditions. Such phenotypic plasticity allows V. vulnificus to rapidly adapt to changing environments and is an important factor in the colonization and growth in these various niches. The primary goals of this study were to understand the relationship between CPS and biofilm formation in V. vulnificus during anaerobiosis and to uncover the molecular underpinnings that are likely responsible during these processes.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

V. vulnificus strains utilized in this study are listed in Table 1 and were stored at −80°C in Bacto Luria-Bertani broth (LB; BD Difco, NJ) containing 20% glycerol. These strains were chosen because they have been well characterized by genome sequencing and multilocus sequence typing (MLST) analysis, as well as virulence potential analysis based on serum survival (8-11, 48, 49, 57-60). All strains were grown in Bacto heart infusion (HI) broth (BD Difco, NJ) overnight (ON) at 30°C with aeration in a rotary incubator. For type IV pilin mutants, media were supplemented with both 25 μg/ml and 50 μg/ml of streptomycin (S9137; Sigma) and spectinomycin (S4014; Sigma), respectively. ON cultures of all strains were inoculated at a 1:100 ratio into HI broth containing 0.5 mg/liter resazurin (Applied Chemical), an indicator of redox potential. This reversible indicator, which is pink at a redox potential above −51 mV and is colorless at −110 mV, was used to indicate when culture media were aerobic or anaerobic, respectively (61). Cultures were incubated for 24 h at 37°C with and without oxygen. All samples were cultured in glass borosilicate tubes with loose-fitting caps to allow oxygen to freely enter the tubes as well as efficient removal of oxygen in the anaerobe chamber. Additionally, anaerobic samples were grown in anaerobe chambers utilizing the GasPak EZ anaerobe container system with BBL GasPak anaerobic indicators.

TABLE 1.

V. vulnificus strains used in this study

Strain	Isolation source	Genotypea	16S rRNAf
CMCP6	Clinical, blood	C	B
MO6-24b	Clinical, blood	C	B
JY1701	Environment, oyster	E	A
JY1305	Environment, oyster	E	A
CVD737c	Translucent mutant	C	B
C7184d	Clinical, blood	C	B
C7184ΩAe	pilA mutant	C	B
C7184ΩDe	pilD mutant	C	B

^aGenotypes were determined by identification of the virulence-correlated gene (4).

^bParental wild-type strain of translucent CPS transposon mutant (21).

^cTranslucent transposon mutant that cannot synthesize CPS (21).

^dParental wild-type strain of type IV pilus mutants.

^eType IV pilin mutant (39, 52).

^fTyping according to polymorphisms in the 16S rRNA gene (9, 12).

RNA harvesting.

Anaerobic cultures were removed using a 3-ml sterile syringe that was prefilled with a 2:1 ratio of RNAprotect (Qiagen) to cell culture, following the manufacturer's protocol. The use of the syringe allowed for rapid uptake of culture into the RNAprotect immediately following the opening of the anaerobe chamber, thereby minimizing exposure to oxygen. Aerobic cultures were removed and RNA protected by using a similar method, omitting the use of the syringe, as the cultures were already freely exposed to oxygen. Cells were stored at −80°C until lysis and extraction of RNA. RNA extraction was performed as described by Williams et al. (54). Briefly, RNA-protected cell pellets were lysed by gently vortexing with 10 mg/ml lysozyme in Tris-EDTA (TE) buffer at pH 8.0 for 30 min. RNA extraction was performed on lysed cells using the RNeasy minikit (Qiagen) following the manufacturer's instructions, with the addition of the optional on-column DNase I treatment. RNA was eluted twice with nuclease-free water, and a final postextraction DNase treatment was performed using the “rigorous” treatment with Turbo DNA-free (Ambion), according to the manufacturer's protocol. The total amount and quality of final RNA were determined using a NanoDrop spectrophotometer (Thermo), and samples having an A₂₆₀/A₂₈₀ ratio of ≥1.7 were stored at −80°C.

PCR to detect DNA contamination.

RNA was analyzed for DNA contamination by using endpoint PCR to amplify vvhA, a species-specific gene target (4). Primers for vvhA were used with Promega Go-Taq DNA polymerase, 5× Green GoTaq reaction buffer, and 10 mM deoxynucleoside triphosphate (dNTP) mix. An annealing temperature of 53.1°C and 40 cycles of amplification were used, following the manufacturer's instructions. Contamination was indicated by any amplification of the vvhA gene, and such samples were not utilized for further experiments.

Primer design.

Primers for quantitative reverse transcription-PCR (qRT-PCR) were designed using the NCBI Primer-BLAST software against three sequenced C-genotype strains of V. vulnificus (CMCP6, YJ016, and MO6-24) and three sequenced E-genotype strains (JY1701, JY1305, and E64MW). Primers optimized for this study are listed in Table 2. The IDT OligoAnalyzer 3.1 software was used to evaluate primer quality, and specificity was determined using in silico PCR. Primer pair efficiency was further estimated by using an in silico PCR estimation tool, and those yielding efficiencies of ≥1.5 were purchased from Sigma-Aldrich. These underwent further validation by utilizing endpoint PCR for the genes of interest in each strain.

TABLE 2.

Primers designed for qRT-PCR

Gene	Primer target	5′–3′ sequencea	Product size (bp)b	Strain(s) amplified
Glyceraldehyde phosphate dehydrogenase	GAPDH	F: TGAAGGCGGTAACCTAATCG	97	All
		R: TACGTCAACACCGATTGCAT
Outer membrane lipoprotein	wza	F: TTGCCGGTTGTACTGTACCC	79	All
		R: TCTTGGCTGCTATCGTTCAC
Cognate phosphatase	wzb	F: GTAAGCCAGCCGATGCAATG	98	All
		R: GCTACACAGCGCAGAGGTAA
Tyrosine autokinase	wzc	F: GCTGCTTGGCATTCTTTTGGA	93	All
		R: GTTGAAAGTAGCGCAACCGC
Type IV pilin protein subunitc	pilA	F: TCATTGGTGTGTTAGCCGCA	73	JY1305 only
		R: GCTGAGGCAGCTTCTGACTT
Type IV pilin protein subunitc	pilA	F: TTGCTAAAAGTGAAGCCGCA	169	JY1701 only
		R: CCAGCTGCACTAGTAGGGTT
Type IV pilin protein subunitc	pilA	F: GCACAGCTCCAACCAGTAGT	57	CMCP6, MO6-24
		R: TTGGCGGCACTTCAACAATG
Type IV pilin protein prepilinc peptidase	pilD	F: GGCTTACTGGTAGGCAGCTT	126	JY1701, JY1305
		R: GGTTTCTGTCGGCGGTGATA
Type IV pilin protein prepilinc peptidase	pilD	F: TTGGCTTACTGGTAGGCAGC	128	CMCP6, MO6-24
		R: GGTTTCTGTCGGTGGTGTGA
Mannitol-sensitive hemagglutinin	mshA	F: CAAGGCGGTTTCACCCTGAT	90	All
		R: CAGATTTAGAAAACGCGGAGCC
Flagellar monomeric hook subunit	flgE	F: TCTAACTGAGCTGCGGACAA	145	All
		R: TACTGCTCAATCTGGCTGGC

^aF, forward primer target; R, reverse primer target.

^bProducts amplified between sequenced strains in silico.

^cPrimers designed for this gene were different between strains/genotypes.

Relative qRT-PCR.

An optimized qRT-PCR protocol, described by Williams et al. (54), was used in this study with slight modifications. Briefly, 1 μg of total RNA was reverse transcribed with qScript cDNA supermix (Quanta Biosciences) to measure relative expression of target genes. cDNA was then diluted to 50 ng/μl for relative qRT-PCR. PerfeCTa SYBR green FastMix, low ROX (Quanta Biosciences) was utilized for expression with at least two biological and three technical replicates for each strain. This resulted in at least six threshold cycle (C_T) values per target gene. To correct for sampling error, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), an endogenous control, was used for normalization of target genes. Fold change in expression levels was measured in two ways: (i) in anaerobically grown relative to aerobically grown cells within the same genotype, and (ii) in E-genotype strains relative to C-genotype strains. Fold change in gene expression was calculated using the Pfaffl equation (62), and previously described PCR efficiency analysis was utilized to account for difference in the primer sets (63).

Biofilm quantification.

Biofilms were quantified using previously described assays with slight modifications (64). Briefly, all strains were grown as described above and inoculated at a 1:100 ratio into 2 ml autoinducer bioassay (AB) medium supplemented with 1% fumarate as a sole carbon source (65, 66). Borosilicate tubes were used for all biofilm experiments, with cells incubated for 24 h at 37°C under aerobic and anaerobic conditions. After 24 h, supernatants from these tubes were removed, followed by quantification of the CFU/ml. CFU/ml were quantified by serial dilutions, followed by plating of 100 μl of supernatant onto HI plates for both aerobic and anaerobic cultures. The remaining attached biofilm was washed 3 times by the addition of 10 ml of phosphate-buffered saline (PBS) with a serological pipette, subsequent removal, and repeated with fresh PBS. Washed tubes were then stained with 1% crystal violet for 15 min and then washed 3 times as described above with PBS to remove excess stain, and the dye was eluted with 95% ethanol. Eluted crystal violet was quantified by spectrophotometry at 550 nm. Absorbance values were normalized by dividing the optical density at 550 nm (OD₅₅₀) for each sample by its corresponding log CFU/ml.

Capsule production.

For phenotypic evaluation of V. vulnificus CPS variants, all strains were grown in HI broth overnight at 30°C with aeration in a rotary incubator. These cultures were then aseptically streaked on HI plates supplemented with 1.5% agar. Plates were then incubated for 24 h at 37°C under either aerobic or anaerobic conditions and then examined for the presence or absence of capsule by qualitative visualization of colony opacity (22).

Statistical analysis.

All data were analyzed by using Prism (version 5.0; GraphPad Software Inc.). Two-way analysis of variance (ANOVA) was utilized for biofilm formation analysis, followed by Sidak's post hoc test for multiple comparisons. Gene expression results were analyzed in the following two ways: (i) for anaerobically grown cultures relative to aerobic cultures, and (ii) for E-genotypes (JY1701 and JY1305) relative to C-genotypes (CMCP6 and MO6-24). Significant differences between target transcripts were evaluated using Mann-Whitney (nonparametric) rank-sum tests with adjusted P values calculated using the Bonferroni method.

RESULTS AND DISCUSSION

Role of anoxia on CPS production.

Aerobically, V. vulnificus strains grown on solid media had opaque colony morphologies. However, all strains grown for 24 h on solid media without oxygen resulted only in colonies exhibiting the translucent phenotype, indicating a reduction in CPS biosynthesis. Additionally, all strains remained translucent on solid media with a prolonged anaerobic incubation of up to 72 h. However, when anaerobic plates were removed from the anaerobe jar and exposed to oxygen, the colonies regained opacity after approximately 24 h, suggesting that these colonies regained their ability to produce capsule when oxygen was available. When grown anaerobically for 24 h in liquid culture, the medium was only slightly turbid compared to that of the aerobic cultures. This corresponded to a significant difference (P < 0.001) between the OD₆₁₀ of aerobic cultures compared to anaerobic cultures (Table 3), which normally is indicative of less growth. However, we determined that after 24 h under either condition, anaerobic CFU/ml were not significantly different (C-genotypes, P = 0.20; E-genotypes, P = 0.23) than in aerobic cultures when plated on solid media for both genotypes (Table 3). This disparity suggests that although translucency by definition is described by growth on solid media, anaerobically grown cells might in fact be translucent in liquid culture as well. Another possibility is that these cells are smaller, which may also produce similar results when comparing OD₆₁₀ to CFU/ml. This was important, in that it required all normalizations to be based solely on plate counts rather than optical density readings, which did not reflect anaerobic CFU/ml accurately. Further investigation into the biological function of translucency in V. vulnificus is needed, as its implication is not clear in response to oxygen availability in the environment or the human host.

TABLE 3.

Quantification of V. vulnificus cells in liquid culture after 24 h of incubation under different conditions

Genotypea	Condition	OD610b,c	Log CFU/mlb
C	Aerobic	0.40 ± 0.02*	8.48 ± 0.08
C	Anaerobic	0.14 ± 0.02*	8.05 ± 0.17
E	Aerobic	0.491 ± 0.02*	8.91 ± 0.16
E	Anaerobic	0.193 ± 0.01*	8.50 ± 0.12

^aStrains with C-genotypes were CMCP6 and MO6-42; strains with E-genotypes were JY1701 and JY1305.

^bData are means ± standard errors.

^cAsterisks indicate a significant (P < 0.05) difference (based on Mann-Whitney nonparametric rank-sum tests) between the OD₆₁₀ of aerobically versus anaerobically grown cultures within the same genotype.

To investigate the relationship between anoxia and loss of CPS, we examined expression of several of the genes responsible for capsule production, wza, wzb, and wzc, in V. vulnificus under aerobic and anaerobic environments. In E. coli as well as V. vulnificus, wza is involved in transporting the polymers to the outer membrane of the cell by forming a multimeric structure composed of outer membrane lipoprotein (34, 55, 67). In E. coli, the coordinated function of both a cognate phosphatase (wzb) and a tyrosine kinase (wzc) is also required for the formation of the CPS outside the cell (21, 31, 32, 51), but the specific functions of these genes have not been characterized for V. vulnificus. Previous studies have shown that a mutation in any one of these genes results in a translucent phenotype. This results in an accumulation of these polymers inside the periplasmic space, leading to cells lacking external CPS (24).

We found that, relative to aerobically grown cultures, all three genes were significantly downregulated (P < 0.001) under anaerobic conditions for V. vulnificus, regardless of genotype (Fig. 1). Although both C- and E-genotypes significantly downregulated genes responsible for CPS production anaerobically, E-genotypes exhibited higher expression levels than C-genotypes under both conditions (Fig. 2). Upon entry into the anaerobic environment of the small intestines within the human host, the bacterium must evade a number of physiological threats. Considering the essential role of capsule in pathogenicity, the finding that CPS production was greatly reduced during anaerobiosis is an interesting result. It may be that the presence of capsule is not advantageous at the early stages of colonization in the human host but may be essential for later stages of disease progression, i.e., evasion of phagocytosis in the bloodstream.

FIG 1.

Effect of anaerobiosis on CPS gene expression in both C-genotypes (CMCP6 and MO6-24) (A) and E-genotypes (JY1701 and JY1305) (B) of V. vulnificus. Each graph represents relative expression levels (under anaerobic [AN] relative to aerobic [AE] conditions) of genes involved in CPS production (wza, wzb, and wzc). Error bars represent the standard errors of five biological replicates, comprising three technical replicates averaged for each strain (two of each genotype). Asterisks represent statistically significant differences between aerobic and anaerobic expression of each gene transcript, based on the nonparametric Mann-Whitney test with corrected P value utilizing the Bonferroni calculation. Both the C-genotype (A) and E-genotype (B) significantly downregulate wza, wzb, and wzb in response to anaerobiosis: ***, P < 0.001; ****, P < 0.0001.

FIG 2.

Expression of genes involved in capsular polysaccharide production by two E-genotypes (JY1701 and JY1305) relative to two C-genotypes (CMCP6 and MO6-24) of V. vulnificus under aerobic (A) and anaerobic (B) conditions. Error bars represent the standard errors of five biological replicates, comprising three technical replicates averaged for each strain (two of each genotype). Asterisks represent statistically significant differences (**, P < 0.01; ****, P < 0.0001) between E-genotypes relative to C-genotypes of each gene transcript, based on the nonparametric Mann-Whitney test with corrected P value utilizing the Bonferroni calculation. Anaerobically, E-genotypes express significantly higher levels of wza, wzb, and wzc than C-genotypes; under aerobic conditions, significant differences were observed only for wzb and wzc.

Role of CPS in biofilm formation.

Studies have shown a negative relationship between CPS and biofilm formation in V. vulnificus and that translucent strains form more robust biofilms (51). Thus, we hypothesized that anaerobiosis would lead to an increase in biofilm formation due to the reduction in CPS exhibited under this condition. To determine whether translucency plays a role in biofilm formation under anaerobiosis, we allowed all strains listed in Table 1 to form biofilms for 24 h under aerobic and anaerobic conditions. While both C- and E-genotypes were able to form biofilms under these conditions, there was a significant (P < 0.05) decrease in biofilm formation under anaerobic conditions (Fig. 3). Our findings suggest that the reduction in CPS during anaerobiosis may play a role in the decreased ability of V. vulnificus to form biofilms under this condition, differing from what has been shown previously under aerobic conditions (51). Lack of available oxygen may play a larger role in the formation and/or dispersal of these biofilms, which may override any effect the opacity of the cells may have on biofilm formation.

FIG 3.

Biofilm formation of clinical (CMCP6 and MO6-24) and environmental (JY1701 and JY1305) genotypes of V. vulnificus after 24 h of incubation under both aerobic and anaerobic conditions. Different letters indicate significance (P < 0.001) between genotypes and culture conditions (n = 26). Two-way analysis of variance showed that oxygen has a role in limiting biofilm formation for both C- and E-genotypes and that E-genotypes form significantly more biofilm aerobically than C-genotypes.

To examine the specific role of CPS in biofilm formation, we utilized a transposon mutant, CVD737, which is unable to synthesize CPS because of a disruption in the wza gene (21). V. vulnificus C-genotype strain MO6-24 is the parental strain from which CVD737 was derived (21). Aerobically, our results supported previous findings (51), in which the complete absence of CPS was found to significantly (P < 0.0001) enhance biofilm formation (Fig. 4). However, we found that during anaerobiosis, both CVD737 and the parent strain formed significantly less biofilm than did cells grown aerobically. Interestingly, there was not a significant difference (P > 0.05) between the two strains in their ability to form biofilm anaerobically. Thus, the downregulation of CPS genes anaerobically in the wild-type parent led to a very similar phenotype as in the mutant strain, resulting in similar levels of biofilm formation. This further suggests that decreased capsule production, during anaerobiosis, may not be the only determinant in the ability for V. vulnificus to form robust biofilms.

FIG 4.

Biofilm formation of a transposon mutant of V. vulnificus (CVD737) lacking functional CPS biosynthesis, compared to that of the wild-type parental strain (MO6-24). Quantification of biofilm after 24 h of incubation under both aerobic and anaerobic conditions was conducted for all replicates. Different letters indicate significance between the parent and mutant strain when incubated under aerobic and anaerobic conditions (n = 18). Two-way analysis of variance showed that capsule has a negative role in aerobic biofilm formation but no significant role under anaerobic conditions.

Role of anoxia on surface attachment proteins.

Since V. vulnificus formed significantly less biofilm anaerobically than under aerobic conditions (Fig. 3), we investigated expression of genes responsible for the production of cell surface proteins implicated in biofilm formation. Anaerobically, expression of both pilA and pilD was significantly downregulated (P < 0.0001) in both C- and E-genotypes relative to expression levels in aerobically grown cultures (Fig. 5). Additionally, mutants lacking type IV pili in V. vulnificus have previously shown an inability to form biofilm efficiently (39, 52). We found that, although under both conditions mutants formed significantly less biofilm than the parental strain (Fig. 6), there was still an overall decrease in formation during anaerobiosis. Anaerobically, this decrease was consistent with lower expression of pilA and pilD in the wild-type strain, which led to poor biofilm formation (Fig. 3).

FIG 5.

Effect of anoxia on expression of multiple surface attachment genes involved in biofilm formation in C-genotype (CMCP6 and MO6-24) (A) and E-genotype (JY1701 and JY1305) (B) V. vulnificus strains. Error bars represent standard errors of five (two strains in each group) biological replicates with three technical replicates. Asterisks represent a significant decrease in anaerobic (AN) expression of pilA, pilD, and mshA (type IV pili) compared to aerobic (AE) expression in both genotypes and a significant increase in flgE (flagellar hook protein) expression; the asterisks represent statistically significant differences between aerobic and anaerobic expression of each gene transcript, based on the nonparametric Mann-Whitney test with corrected P value utilizing the Bonferroni calculation. ***, P < 0.001; ****, P < 0.0001. Both C-genotype (A) and E-genotype (B) strains show a significant effect of anaerobiosis on type IV pilus and flagellum expression.

FIG 6.

Biofilm formation of a type IV pilin A (pilA) mutant (C7184ΩA) and prepilin peptidase (pilD) mutant (C7184ΩD) of V. vulnificus compared to that of the parental wild-type strain (C7184). Quantification of biofilm after 24 h of incubation under either aerobic or anaerobic conditions was conducted for all replicates (n = 12). Different letters represent statistical significance (P < 0.001) between the mutant and wild-type under the two conditions. Two-way analysis of variance showed that both the wild-type and the mutant strains produced significantly less biofilm under anaerobic conditions.

We also found that lack of oxygen significantly decreased (P < 0.0001) expression of mshA in both genotypes (Fig. 5). This suggests an inability for both genotypes to make functional type IV pili anaerobically, possibly indicating that they are not required in this environment. Interestingly, during in vivo growth of V. cholerae, transcriptome analysis revealed that mshA was highly expressed compared to levels during in vitro growth (16). However, others have shown that MshA is not required for V. cholerae to colonize the gastrointestinal tract but that it aids in effective attachment to zooplankton and other biotic surfaces in the environment (54, 68, 69).

We examined flgE, a hook protein subunit, as a determinant for flagellum production, since flgE mutants do not have a functional flagellum (56, 70). Additionally, flgE mutants of V. vulnificus have previously shown an inability to form robust biofilms compared to their wild-type parental strains (35). Since the flagella in V. vulnificus are important in initial attachment to a multitude of surfaces (35) and we found that the biofilm is impaired anaerobically (Fig. 3), we hypothesized that genes involved in its production would also be decreased anaerobically. Interestingly, during anaerobiosis, V. vulnificus significantly increased (P < 0.001) expression of flgE (Fig. 5) and was significantly more motile in sloppy agar (data not shown). Rather than aiding in initial biofilm formation, this anaerobic upregulation of flgE implies another role for the flagellum, such as chemotaxis, in this environment.

Our lab has previously described that E-genotypes intrinsically express greater amounts of pilA, pilD, and mshA when incubated aerobically in artificial seawater (ASW) than do C-genotypes (54). In that study, the increase in E-genotype expression relative to C-genotype was implicated in the differential ability of V. vulnificus to attach to chitin and was suggested to be an important determinant for environmental persistence (54). We found the expression levels of all four genes were significantly higher (P < 0.01) in E-genotypes than in C-genotypes, aerobically as well as anaerobically (Fig. 7). Higher expression levels aerobically (Fig. 7A) coincided with an increase in biofilm formation (Fig. 3), and this may indicate an important role for type IV pili in the formation of biofilms under this condition. However, we did not find a similar relationship for cells grown anaerobically. This may be due to the diversity of these two genotypes, and the function of these surface proteins may be more important in the environment, as described for other Vibrio species (68, 71). As with our previous expression results, flgE expression was also significantly increased (P < 0.01) in the E-genotypes (relative to C-genotypes) under both conditions (Fig. 7).

FIG 7.

Expression of genes involved in biofilm production by two E-genotype strains (JY1701 and JY1305) relative to that of two C-genotype strains (CMCP6 and MO6-24) of V. vulnificus under aerobic (A) or anaerobic (B) conditions. E-genotypes expressed significantly more type IV pili (pilA, pilD, and mshA) as well as flagellum hook protein (flgE) under both aerobic and anaerobic conditions. Error bars represent the standard errors of five biological replicates, comprising three technical replicates averaged for each strain (two of each genotype). Asterisks represent statistically significant differences: **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 between E-genotypes relative to C-genotypes of each gene transcript, based on the nonparametric Mann-Whitney test with corrected P values utilizing the Bonferroni calculation.

Pseudomonas aeruginosa biofilm dispersal has previously been shown to be induced by lack of available oxygen, which triggers a change in expression of genes involved in sessile versus planktonic lifestyles (72). We propose that when oysters harboring V. vulnificus are ingested, they traverse to the gut, which subsequently results in exposure to anaerobiosis. Presumably in an attached state, V. vulnificus would respond to the anaerobic conditions and possibly switch to a planktonic lifestyle. Our results support this switch by the downregulation type IV pili and the upregulation of flagella, which is indicative of nonsessile living and why biofilm formation is impaired anaerobically. Although the gut is anaerobic, oxygen is present in various amounts in the tissues surrounding the lumen (14). The upregulation of flgE we described under this condition may aid in such an ability of V. vulnificus to move toward a more favorable environment where the cells may then attach and subsequently penetrate through the gut lining and into the bloodstream. As we observed a reversion to an encapsulated phenotype after removal from anaerobic conditions, we suggest that once in the oxygenated bloodstream, genes for CPS are upregulated. This encapsulated phenotype can then evade host immune responses, replicate, and cause potentially fatal septicemia, as the ability to transition from opaque to translucent phenotypes and vice versa is an important virulence determinant (21, 23, 24, 73).

Additionally, this bacterium is part of the normal flora in oysters and is frequently exposed to various levels of oxygen during low tide and seasonal drought. This exposure to anaerobiosis in the environment, and the prevalence of E-genotypes in oysters as well as the water column, may result in the differences between C- and E-genotype numbers of V. vulnificus (74). Our lab has previously shown an innate difference in expression levels of type IV pili aerobically (54), and these results further validate those findings as well as expound upon this under conditions lacking oxygen.

There are still many unanswered questions about how V. vulnificus is able to successfully cause disease in the human host and why some genotypes correlate more highly with pathogenesis. Responses to environmental parameters encountered in the host are important indicators of virulence potential, and this study reports for the first time the role of anaerobiosis for V. vulnificus.

More investigation is needed before we can elucidate if cellular physiology during anaerobiosis has a role in predicting pathogenic potential in both C- and E-genotypes. We are currently studying the molecular mechanisms involved in metabolic regulation, exoenzyme production, and stress response during anaerobiosis. Such studies will help determine if this condition enhances pathogenicity in one genotype over the other. If we can understand the role of oxygen on V. vulnificus and uncover its impact on global gene regulation, this could potentially improve current methods for effective prevention, harvesting, and storage of potentially harmful oysters.