Heterogeneous Growth Enhancement of Vibrio cholerae in the Presence of Different Phytoplankton Species

ABSTRACT

Vibrio cholerae is a ubiquitously distributed human pathogen that naturally inhabits marine and estuarine ecosystems. Two serogroups are responsible for causing cholera epidemics, O1 and O139, but several non-O1 and non-O139 V. cholerae (NOVC) strains can induce cholera-like infections. Outbreaks of V. cholerae have previously been correlated with phytoplankton blooms; however, links to specific phytoplankton species have not been resolved. Here, the growth of a NOVC strain (S24) was measured in the presence of different phytoplankton species, alongside phytoplankton abundance and concentrations of dissolved organic carbon (DOC). During 14-day experiments, V. cholerae S24 was cocultured with strains of the axenic phytoplankton species Actinocyclus curvatulus, Cylindrotheca closterium, a Pseudoscourfieldia sp., and a Picochlorum sp. V. cholerae abundances significantly increased in the presence of A. curvatulus, C. closterium, and the Pseudoscourfieldia sp., whereas abundances significantly decreased in the Picochlorum sp. coculture. V. cholerae growth was significantly enhanced throughout the cogrowth experiment with A. curvatulus, whereas when grown with C. closterium and the Pseudoscourfieldia sp., growth only occurred during the late stationary phase of the phytoplankton growth cycle, potentially coinciding with a release of DOC from senescent phytoplankton cells. In each of these cases, significant correlations between phytoplankton-derived DOC and V. cholerae cell abundances occurred. Notably, the presence of V. cholerae also promoted the growth of A. curvatulus and Picochlorum spp., highlighting potential ecological interactions. Variations in abundances of NOVC identified here highlight the potential diversity in V. cholerae-phytoplankton ecological interactions, which may inform efforts to predict outbreaks of NOVC in coastal environments.

IMPORTANCE Many environmental strains of V. cholerae do not cause cholera epidemics but remain a public health concern due to their roles in milder gastrointestinal illnesses. With emerging evidence that these infections are increasing due to climate change, determining the ecological drivers that enable outbreaks of V. cholerae in coastal environments is becoming critical. Links have been established between V. cholerae abundance and chlorophyll a levels, but the ecological relationships between V. cholerae and specific phytoplankton species are unclear. Our research demonstrated that an environmental strain of V. cholerae (serogroup 24) displays highly heterogenous interactions in the presence of different phytoplankton species with a relationship to the dissolved organic carbon released by the phytoplankton species. This research points toward the complexity of the interactions of environmental strains of V. cholerae with phytoplankton communities, which we argue should be considered in predicting outbreaks of this pathogen.

INTRODUCTION

Vibrio cholerae is an aquatic Gram-negative gammaproteobacterium that can induce the highly infectious gastrointestinal disease cholera (1). Only two V. cholerae serogroups are responsible for causing outbreaks of cholera (O1 and O139) (2), but although members of the other 206 serogroups that have been described to date do not cause cholera epidemics, they can induce other milder forms of gastroenteritis (3). It has been estimated that these non-O1, non-O139 V. cholerae (NOVC) strains cause 1 to 3.4% of acute diarrheal episodes in developed and developing countries (4), with health infections related to NOVC in the United States alone contributing to an annual burden of $20 million (5). Recent evidence indicates that these strains are increasingly causing human illness related to either direct exposure or via seafood consumption, with putative links to shifting environmental conditions that may be influenced by climate change (2, 3). Through the mechanism of horizontal gene transfer, these strains also have the capacity to acquire pathogenicity factors, virulence regulators and antibiotic-resistant transposons from the environment or from other environmental bacteria (3, 6).

V. cholerae is endemic to aquatic ecosystems, including surface coastal waters, estuaries, and freshwater habitats, where the abundance of this bacterium is governed by a range of environmental conditions, including temperature, salinity, seasonality, and chlorophyll a levels (1, 7). Emerging links have been established between phytoplankton blooms and abundances of toxigenic V. cholerae (8–10), as well as links between V. cholerae abundance and concentrations of dissolved organic carbon (DOC) derived from cyanobacteria (11, 12). Outbreaks of V. cholerae have been spatially correlated with chlorophyll a using remote sensing data sets (13, 14), with strong evidence that this bacterium is associated with phytoplankton biomass (1, 7) and often involved in attachment to phytoplankton cells (15, 16). However, the precise nature of the ecological associations (e.g., extent of growth enhancement) between V. cholerae and specific phytoplankton species has not been closely examined (7, 10).

Growing evidence suggests that the phytoplankton-associated microbiome has played a central role in the fitness and evolution of diverse phytoplankton species (17). For example, it is becoming clear that several species of diatoms heavily rely on relationships with bacteria (18). In turn, these phytoplankton-bacterium interactions can shape biogeochemical cycling, aquatic food web structure, and productivity (19, 20). Metabolic exchanges between bacteria and phytoplankton can be mutualistic, antagonistic, parasitic, and even competitive, whereby these exchanges can be dynamic in space and time (20, 21). The DOC exuded by different phytoplankton species can vary significantly in both biochemical composition and quantity (22, 23); thus, different phytoplankton species are expected to establish varied ecological relationships with other microorganisms (24).

Relationships between microorganisms can be dependent upon the chemical composition of the metabolites released by each partner (24). Different groups of heterotrophic bacteria have been shown to develop close ecological interactions with specific phytoplankton associates (21), which are likely governed by different metabolic requirements and responses. For instance, members of the Rhodobacterales are primarily thought to utilize low-molecular-weight phytoplankton-derived metabolites, while members of the Flavobacteriales more effectively transform high-molecular-weight carbohydrate polymers (25). Gammaproteobacteria, including Vibrio spp., are often generalists that can competitively utilize both high- and low-molecular-weight molecules, ranging from DOC to chitin derived from phytoplankton (26, 29).

The ecological importance of the associations occurring between heterotrophic marine bacteria and phytoplankton is receiving growing recognition (20, 21). These associations can be species specific and involve complex and often reciprocal exchanges of chemicals affecting the growth of one or both partners (18, 27, 28). It is, however, currently not known to what extent V. cholerae growth is influenced by different phytoplankton species, if the nature of these interactions depends on the phytoplankton species, or if reciprocal chemical exchanges occur between the partners. Answers to these questions may provide insight into how this bacterium can rapidly increase in abundance in the marine environment, which may favor its ability to infect humans (8). Here, we conducted a series of coculture experiments involving ecologically significant and diverse phytoplankton and an NOVC strain to determine the differential effects on growth between these microorganisms, whereby the growth of the V. cholerae strain was hypothesized to vary according to different sources of phytoplankton-derived DOC.

RESULTS

Cocultures.

(i) Actinocyclus curvatulus coculture. Over the 14-day cogrowth experiment, Vibrio cholerae abundance was significantly higher when grown in the presence of A. curvatulus relative to the f/2-medium control (P < 0.05, Fig. 1A). Indeed, in the presence of A. curvatulus, the abundance of V. cholerae was 3 times higher than that of the control by day 2, with this difference increasing to 10-fold by day 14 (P < 0.05, Fig. 1A). Within the axenic culture of A. curvatulus, dissolved organic concentrations were approximately 2 times higher than those in the control on day 14 (analysis of variance [ANOVA], P < 0.05, Fig. 1C), and there was a significant correlation between DOC concentrations and V. cholerae abundance (r = 0.87, P < 0.005). The normalized DOC concentrations were high (relative to other treatments) throughout the 14-day experiment (Fig. 1D).

FIG 1.

A. curvatulus and V. cholerae coculture. (A) Cell abundance (cells mL⁻¹) of V. cholerae with A. curvatulus (pink) and V. cholerae in f/2-medium (gray) over a 14-day period. (B) A. curvatulus abundance (cells mL⁻¹) between the axenic control (purple line) and coculture (pink) over a 14-day period. Data for panels A and B are the mean ± standard error (n = 4). Significant differences between control and coculture abundances are denoted with an asterisk (*) (repeated measure ANOVA, P < 0.05). (C) DOC (ng C mL⁻¹) of A. curvatulus axenic cultures was sampled on days 0, 8, and 14 (purple) alongside a control of f/2-medium (pink). (D) DOC per cell (pg C cells⁻¹) of A. curvatulus axenic cultures. Data for panels C and D are the mean ± standard error (n = 3). *, significant differences (repeated measure ANOVA, P < 0.05).

Cell abundances of the diatom A. curvatulus also reached significantly higher levels in the presence of V. cholerae relative to the axenic culture (P < 0.05, Fig. 1B). The maximum phytoplankton growth enhancement occurred on day 6, at the end of the exponential phase of A. curvatulus phytoplankton growth, where the abundance of the diatom was 35.8% higher than that within the axenic control (Fig. 1B).

(ii) Cylindrotheca closterium coculture. V. cholerae abundances were significantly higher than those of the control when grown in coculture with the diatom C. closterium, although this growth enhancement was largely restricted to the last time-point, from a notable 6-fold increase relative to the control on day 14 (P < 0.05, Fig. 2A). This peak in V. cholerae abundance corresponded temporally to a significant decrease in the abundance of C. closterium (Fig. 2B). In contrast to the patterns observed in the A. curvatulus experiment, the overall DOC levels in the C. closterium culture did not differ significantly from those of the control (P > 0.05, Fig. 2C). However, on day 14, when both C. closterium phytoplankton cell abundances declined and V. cholerae cell abundance peaked, DOC levels increased by 22.7% above those of the control (Fig. 2B and C). In line with this pattern, V. cholerae abundance significantly correlated with DOC concentrations (r = 0.98, P < 0.005), with maximum cell abundances corresponding with high normalized DOC concentrations (Fig. 2D).

FIG 2.

C. closterium and V. cholerae coculture. (A) Cell abundance (cells mL⁻¹) of V. cholerae with C. closterium (yellow) and V. cholerae f/2-medium PHSW (gray) over a 14-day period. (B) C. closterium abundance (cells mL⁻¹) between the axenic control (orange line) and coculture (yellow) over a 14-day period. Data for panels A and B are the mean ± standard error (n = 4). *, significant differences (repeated measure ANOVA, P < 0.05). (C) DOC (ng C mL⁻¹) of C. closterium axenic cultures (orange) was sampled on days 0, 8, and 14 alongside a control of f/2-medium (pink). (D) DOC per cell of C. closterium (pg C cells⁻¹) axenic cultures. Data for panels C and D are the mean ± standard error (n = 3).

Overall, the abundance of V. cholerae cells reached in the coculture with C. closterium was significantly lower (P < 0.05, Fig. 3A) than the V. cholerae abundances reached with the A. curvatulus coculture. Similarly, throughout the experiment, normalized DOC levels were significantly lower in the C. closterium coculture compared to A. curvatulus (P < 0.05, Fig. 3B). In contrast to the A. curvatulus coculture, the presence of V. cholerae did not significantly increase abundances of C. closterium (repeated measure ANOVA, P > 0.05, Fig. 1B, Fig. 2B).

FIG 3.

Abundance of V. cholerae and DOC across the different phytoplankton cultures. (A) Cell abundance (cells mL⁻¹) of V. cholerae in coculture with A. curvatulus (pink), C. closterium (yellow), Picochlorum sp. (green), Pseudoscourfieldia sp. (blue), and V. cholerae in f/2-medium PHSW (gray) (mean ± standard error; n = 4). *, significant differences (repeated measure ANOVA, P < 0.05). (B) DOC per cell (pg C cells⁻¹) for A. curvatulus (purple), C. closterium (orange), Picochlorum sp. (green), and Pseudoscourfieldia sp. (blue) (mean ± standard error; n = 3). *, significant differences (repeated measure ANOVA, P < 0.05).

(iii) Pseudoscourfieldia sp. coculture. When grown in the presence of the Pseudoscourfieldia sp., the abundance of V. cholerae was not significantly different from that of the control (P > 0.05, Fig. 4A), apart from the last time point (day 14), when abundances significantly increased by 40.8% relative to the bacterial control (P < 0.05, Fig. 4A). Similarly to the patterns observed in the C. closterium experiment, this late increase in V. cholerae abundance corresponded with a significant decline in the abundance of the Pseudoscourfieldia sp. (Fig. 2B, Fig. 4B). The DOC concentrations in the Pseudoscourfieldia sp. axenic culture did not differ significantly from those observed in the f/2-medium control throughout the experiment (P > 0.05, Fig. 4C) but were significantly correlated with V. cholerae abundances (r = 0.71, P < 0.05). Normalized to cell abundances, the DOC concentrations were variable throughout the 14-day experiment (Fig. 4D).

FIG 4.

Pseudoscourfieldia sp. and V. cholerae coculture. (A) Cell abundance (cells mL⁻¹) of V. cholerae with Pseudoscourfieldia sp. (blue) and V. cholerae f/2-medium PHSW (gray) over a 14-day period. (B) Pseudoscourfieldia sp. abundance (cells mL⁻¹) between the axenic control (light blue line) and coculture (blue) over a 14-day period. Data for panels A and B are the mean ± standard error (n = 4). *, significant differences (repeated measure ANOVA, P < 0.05). (C) DOC (ng C mL⁻¹) of Pseudoscourfieldia sp. axenic cultures (light blue) was sampled on days 0, 8, and 14 alongside a control of f/2-medium. (D) DOC per cell (pg C cells⁻¹) of Pseudoscourfieldia sp. axenic cultures. Data for panels C and D are the mean ± standard error (n = 3).

Throughout the experiment, mean V. cholerae abundance was approximately 3 times lower in the Pseudoscourfieldia sp. coculture compared to the A. curvatulus coculture (Fig. 3A). The presence of V. cholerae had no significant impact on the growth of the Pseudoscourfieldia sp. (P > 0.05, Fig. 4B).

(iv) Picochlorum sp. coculture. The abundance of V. cholerae cells grown in coculture with the Picochlorum sp. was not significantly different from that of the control during the first 6 days of the experiment and then decreased significantly for the final stages of the experiment (P < 0.05, Fig. 5A). When grown with the Picochlorum sp., the V. cholerae abundance was 2 times lower at the start of the experiment (day 2) and 11 times lower at the end of the experiment (day 14) relative to the A. curvatulus coculture experiment (Fig. 3A).

FIG 5.

Picochlorum sp. and V. cholerae coculture. (A) Cell abundance (cells mL⁻¹) of V. cholerae with Picochlorum sp. (light green) and V. cholerae f/2-medium PHSW (gray) over a 14-day period. (B) Picochlorum sp. abundance (cells mL⁻¹) between the axenic control (dark green) and coculture (light green line) over a 14-day period. Data for panels A and B are the mean ± standard error (n = 4). *, significant differences (repeated measure ANOVA, P < 0.05). (C) DOC (ng C mL⁻¹) of Picochlorum sp. axenic cultures (dark green) was sampled on days 0, 8, and 14 alongside a control of f/2-medium (pink). (D) DOC per cell (pg C cells⁻¹) of Picochlorum sp. axenic cultures. Data for panels C and D are the mean ± standard error (n = 3).

While the Picochlorum sp. had a negligible impact on V. cholerae growth, the presence of V. cholerae intriguingly enhanced the growth of the Picochlorum sp., with a 44% increase in Picochlorum sp. cells relative to the control. The greatest increase in Picochlorum sp. cell abundances occurred during days 4 to 8 and day 14 of the experiment (P < 0.05, Fig. 5B). DOC levels in the Picochlorum sp. culture did not differ significantly from the f/2-medium control, and when normalized per cell, they were significantly less than those measured in the A. curvatulus culture and displayed a variable concentration throughout the 14-day experiment (ANOVA, P > 0.05, Fig. 5C and D, Fig. 3B). The DOC of the Picochlorum sp. and bacterial abundance of V. cholerae in the Picochlorum sp. coculture had a significant, high correlation (r = 0.99, P < 0.005).

(v) Bioinformatics of the Vibrio cholerae S24 genome. There were no hits between Vibrio cholerae S24 and the virulent genes such as ctxA, the cholix toxin, the type VI secretion system (T6SS), and the heat-stable enterotoxin NAG-ST. The environmental strain did, however, have successful hits to genes associated with cholera pathogenesis, as well as genes that assist in gastrointestinal illness and infections (see Table S1 in the supplemental material).

DISCUSSION

Non-O1 and non-O139 V. cholerae do not cause cholera on an epidemic scale, but they can still pose a significant human health risk through gastroenteritis, sepsis, or infections caused either by seafood consumption or through recreational bathing (2). While Vibrio cholerae abundances in coastal environments have can be correlated with chlorophyll a concentrations (7), the potential ecological associations between V. cholerae and individual phytoplankton species are poorly characterized. However, other marine bacteria display species-specific interactions with different phytoplankton species (24). V. cholerae abundance has been previously established through the remineralization of chitin from zooplankton (29) alongside a range of zooplankton taxa (30) in the presence of phytoplankton blooms (8) and even through attachment to phytoplankton and zooplankton exuviae (15). Therefore, a more precise understanding of how V. cholerae interacts with specific phytoplankton populations may provide a heightened capacity to predict outbreaks of this pathogen in coastal ecosystems to reveal potential links between ecological drivers and human infections (14).

The environmental strain V. cholerae S24 contains genes that can induce pathogenesis (Table S1); for example, the repeat in toxin gene is involved in cytotoxic activity (rtxC) (31), the transmembrane gene is involved in pathogenesis (toxR) (32), the hemolysin gene is involved in cytotoxic and enterotoxic cell vacuolation (hlyA) (33), the outer membrane gene (ompU) is positively regulated by the toxR gene (34), and the mannose-sensitive hemagglutinin gene (mshA), is responsible for colonization in all strains (16). The recA gene enables DNA transformation through remediated homologous recombination (35), with mobile genetic elements in the environment also increasing the dangers of environmental strains to cause epidemic-scale infectious disease (3, 6, 36).

Our results demonstrate that while there was a link between phytoplankton-derived DOC and V. cholerae growth, as has been observed in other Vibrio spp. (29), the strength and nature of this enhancement varied markedly depending upon the phytoplankton species. In our experiments, the diatom Actinocyclus curvatulus enhanced the growth of V. cholerae S24 the most, whereby cell counts reached up to 10 times those observed in the control and were between 3 and 6 times higher than the levels observed in the other phytoplankton species tested. These results are consistent with increased abundance of V. cholerae observed in association with increasing DOC concentrations in previous research (12). The sustained growth promotion of V. cholerae during the coculture with A. curvatulus was concomitant with the relatively high DOC concentrations released by the diatom during the experiment. Furthermore, the enhancement of the growth of A. curvatulus when V. cholerae was present points toward a potentially mutualistic exchange of metabolites which is perhaps akin to the reciprocal partnerships observed between other phytoplankton species and heterotrophic bacterial species (27). Further research is needed to determine the specific chemical currencies involved in the V. cholerae-A. curvatulus interaction and how it could mirror the types of exchanges observed in other diatom-bacterium mutualisms (37).

Significant increases in the abundance of V. cholerae were also evident in the Cylindrotheca closterium and Pseudoscourfieldia sp. coculture experiments. However, in contrast to the patterns observed with A. curvatulus, increases in V. cholerae abundance only occurred during the final stages of these experiments, at a time that coincided with both significant declines in the abundance of the phytoplankton cells and an increase in the DOC levels. This pattern is consistent with the V. cholerae potentially benefiting from the decline in the phytoplankton population when cell lysis or senescence leads to increased DOC release (38). This pattern would be consistent with the increase in bacterial biomass regularly observed at the termination of phytoplankton blooms in the environment (39), where Gammaproteobacteria are often one of the main groups of bacteria that increase in abundance in response to algal-derived carbon (25, 29).

Finally, the picoeukaryote Picochlorum sp. did not promote a significant increase in V. cholerae growth, whereby it in fact led to a decrease in bacterial abundance relative to the control. This result may be explained by several factors, including the release of specific antagonistic compounds by the algae (21, 38, 40). Indeed, our observations are consistent with a previous study that showed that the haptophyte Isochrysis galbana inhibits the growth of other pathogenic Vibrio species through the synthesis of antibacterial fatty acids (41). Intriguingly, despite the insignificant impact of the Picochlorum sp. on V. cholerae growth, our experiments indicate that the presence of the bacterium enhanced the growth of the Picochlorum sp. Notably, this pattern is also directly analogous with the previous study examining the relationships between I. galbana and the pathogenic Vibrio sp., where despite suppression of the growth of the Vibrio sp. by I. galbana, the bacteria enhanced the growth of the microalgae (41). Our findings highlight the diversity of the nature of interactions between V. cholerae and phytoplankton and open the door for further comparative studies, which should include comparisons across marine, estuarine, and freshwater strains of phytoplankton and different strains of V. cholerae.

Overall, the highly heterogeneous set of ecological interactions observed between V. cholerae and the four phytoplankton species tested here points toward a likely diverse array of relationships between this bacterium and different phytoplankton species in the environment. We acknowledge that these results are based on a single V. cholerae strain, and interstrain differences in the nature of relationships between bacteria and phytoplankton are likely to exist (15). Our results indicate that in the presence of phytoplankton-derived DOC, NOVC growth responses are highly heterogeneous and governed by the identity of specific phytoplankton species, but further research is needed to characterize the chemical currencies involved in these interactions. This implies that attempts to link the environmental dynamics of this potential human pathogen to coarse indices such as chlorophyll a will not necessarily be illuminating, and a more precise characterization of the ecological networks that underpin blooms of this species will benefit efforts to predict acute V. cholerae outbreaks and potential subsequent human infection events.

MATERIALS AND METHODS

Bacterial strain and maintenance.

A strain of non-O1, non-O139 Vibrio cholerae environmental serogroup (V. cholerae S24) was previously isolated from Georges River, Sydney, Australia, in an urban region (33°57.9′S, 150°58.9′E) and preserved in 20% glycerol at −80°C (35, 42). The V. cholerae isolate S24’s genomic characteristics have previously been described in detail (35), where it was shown to lack the major virulence factors: the cholera toxin, the seventh pandemic islands, and the toxin coregulated pilus, which are associated with cholera epidemics. A phylogenetic analysis of recA sequences from the Vibrionaceae indicated that S24’s recA gene is characteristic of recA genes found within the V. cholerae clade, with it clustering near other non-O1/0139 strains of V. cholerae (35). Phytoplankton strains were isolated from an oceanographic time-series site (Port Hacking; 34°07.06′S, 151°13.09′E) located approximately 27 km from the point of isolation of V. cholerae S24 and were selected for cogrowth experiments.

Phytoplankton cultures.

Taxonomically and morphologically diverse phytoplankton strains that were isolated from the same geographic region as the V. cholerae strain examined here were selected for experiments and consisted of two large diatoms, A. curvatulus (30 μm in diameter) and C. closterium (50 μm in length, 3 μm in diameter), and two picoeukaryotes, a Pseudoscourfieldia sp. (3 μm in diameter) and a Picochlorum sp. (2.5 μm in diameter). The diatom A. curvatulus is a centric diatom that is a globally ubiquitous species in marine environments, including the Australian coastal waters where the V. cholerae strain targeted here was isolated (43). C. closterium cells are straight and sigmoidal shaped (44), are also widely distributed in high and low latitudes in marine and brackish waters, and have been found to be one of the most prevalent species in bloom dynamics in southern Australian estuaries and marine water (45). The Pseudoscourfieldia sp. is a unicellular flagellate with a cell shape that can range from oblong to truncate ovate. The cells contain 2 unequal flagella and are globally distributed in marine environments (46). Finally, Picochlorum sp. cells are coccoid in shape with a wide environmental range of freshwater, brackish environments, and marine waters (47).

Phytoplankton maintenance.

Phytoplankton cultures were made axenic according to the methods of Shishlyannikov et al. (48), with modifications. First, cells were concentrated using gravity filtration of 50-mL culture volumes (5 μm for A. curvatulus and C. closterium, 2 μm for the Pseudoscourfieldia sp., and 0.22 μm for the Picochlorum sp.; Durapore, Millipore, USA). Cells were then gently resuspended from the filter in 25 mL sterile seawater with a final concentration of 20 μg/mL Triton X-100 and were gently inverted for 60 s. The cultures were then immediately washed with sterile seawater to remove all traces of Triton X-100 from the culture. The algae were resuspended from the filter again and diluted 50:50 into f/2-medium and inoculated with a cocktail of antibiotics at the following final concentrations: ciprofloxacin (20 μg/mL), ampicillin (100 μg/mL), gentamycin (66.6 μg/mL), and streptomycin (50 μg/mL) for 30 h. After 30 h they were transferred into fresh f/2-medium without antibiotics and tested by various methods, including drop plating on marine agar (18.7 g Difco marine broth 2216, 9 g NaCl, supplemented with 15 g Difco agar in 1 L of distilled water), screening algae using liquid marine broth (8.7 g Difco marine broth 2216, 9 g NaCl, supplemented in 1 L of distilled water), and enumeration on flow cytometry to confirm that phytoplankton species were axenic. All cultures were subsequently maintained in f/2-medium (49), which was made using filter-sterilized (0.22-μm filter; Durapore, USA) autoclaved seawater from Port Hacking seawater (PHSW) (35 ppt) and incubated under fluorescent light at an average intensity of 95 μmol m⁻² s⁻¹, measured using a 4-pi sensor, under a 12:12 light-dark cycle at 23°C.

Coculture experiments.

Phytoplankton-V. cholerae S24 coculture experiments were established according to Bramucci et al. (50). First, V. cholerae S24 was cultured in 10% marine broth (diluted in filtered autoclaved PHSW; Difco marine broth 2216; BD, USA), and the bacterial concentration was diluted to 100,000 cells mL⁻¹ (here, called bacterial stock). Then, the cell density in each axenic phytoplankton culture was diluted to a concentration of 20,000 cells mL⁻¹ (here, called phytoplankton stock). Phytoplankton-bacterial cocultures were established by mixing each phytoplankton stock with the bacterial stock in a 1:1 ratio, resulting in concentrations of 50,000 bacterial cells mL⁻¹ to 10,000 cells mL⁻¹ of phytoplankton. For each phytoplankton species, four 50-mL replicates were incubated in tissue culture flasks and maintained under the same temperature and light conditions as the phytoplankton. In addition, three types of controls were set up alongside the cocultures: (i) a bacterial control treatment, where the bacterial stock was diluted 1:1 with f/2-medium and then aliquoted into four replicate 50-mL tissue culture flasks, (ii) phytoplankton control treatments, where each axenic phytoplankton was diluted 1:1 with f/2-medium and then aliquoted into seven 50-mL replicates in tissue culture flasks where four 50- mL replicates were reserved for coculture treatments and three 50-mL flasks were reserved for dissolved organic carbon (DOC) measurements, and (iii) a medium control treatment, where f/2-medium used in the experiment was aliquoted into three 50-mL replicates reserved for DOC measurements. Each treatment was monitored over a period of 14 days.

Algal and bacterial enumeration.

Aliquots (200 μL) were taken from each treatment every 2 days over the 14-day experiment. From these, two 100-μL aliquots were sampled to enumerate phytoplankton and bacterial microorganisms separately. Coculture samples were fixed with 2 μL of glutaraldehyde for 20 min (final concentration, 0.25%), whereas axenic phytoplankton and the bacterial control were sampled in 200-μL aliquots and fixed with 4 μL glutaraldehyde for 20 min (final concentration, 0.25%). Samples were then enumerated on a CytoFLEX LX flow cytometer (Beckman Coulter, USA). Phytoplankton cells were analyzed unstained and enumerated based on chlorophyll a fluorescence (692 nm excitation) and forward scatter (FSC). To quantify V. cholerae S24 cells, samples were stained with the nucleic acid stain SYBR green I for 20 min in the dark (final concentration, 1:25,000) and enumerated using SYBR green fluorescence (488 nm excitation) and side scatter (SSC).

DOC measurements.

Samples were collected for the quantification of dissolved organic carbon (DOC) for axenic phytoplankton cultures and the medium control at three time points during the experiment (days 0, 8, and 14). DOC measurements were made from axenic cultures of the 4 phytoplankton strains, to avoid the confounding effects of bacterial consumption or release of DOC expected in the cocultures. At each time point, 50-mL samples were divided into three 12.5-mL technical replicates, diluted 1:3 with MilliQ water, and filtered through a 0.22-μm filter (Millipore, USA). A blank, comprising of MilliQ water filtered through a 0.22 μm filter was employed. After filtration, all samples were frozen at –80°C. Prior to analysis, samples were thawed, and DOC concentrations were measured using a TOC-L series analyzer (Shimadzu, Japan).

Statistical analysis.

A two-way repeated measure analysis of variance (ANOVA), with P-values adjusted using the Bonferroni multiple testing correction method, was used to compare bacterial abundance between treatments over time with the software R Studio (51). This method was also applied to the phytoplankton abundance and DOC concentrations during the coculture experiments. One outlier was removed on day 0 for the f/2-medium control within the concentrations of DOC. A Pearson correlation coefficient was also determined to examine relationships between bacterial abundance with DOC concentrations within each treatment over time, using the statistical software PAST (52).

Bioinformatics of the Vibrio cholerae S24 genome.

Using the Pathosystems Resource Integration Center (PATRIC) all-bacterial Bioinformatics Resource Center (BRC) (http://www.patricbrc.org) (53) to locate genes associated with pathogeneses in V. cholerae and using BLAST on NCBI (https://www.ncbi.nlm.nih.gov), a nucleotide query was used to formally identify the V. cholerae S24 genome.