Rugosity in Grimontia hollisae

Abstract

Grimontia hollisae, formerly Vibrio hollisae, produces both smooth and rugose colonial variants. The rugose colony phenotype is characterized by wrinkled colonies producing copious amounts of exopolysaccharide. Cells from a rugose colony grown at 30°C form rugose colonies, while the same cells grown at 37°C form smooth colonies, which are characterized by a nonwrinkled, uncrannied appearance. Stress response studies revealed that after exposure to bleach for 30 min, rugose survivors outnumbered smooth survivors. Light scatter information obtained by flow cytometry indicated that rugose cells clumped into clusters of three or more cells (average, five cells) and formed two major clusters, while smooth cells formed only one cluster of single cells or doublets. Fluorescent lectin-binding flow cytometry studies revealed that the percentages of rugose cells that bound either wheat germ agglutinin (WGA) or Galanthus nivalis lectin (GNL) were greater than the percentages of smooth cells that bound the same lectins (WGA, 35% versus 3.5%; GNL, 67% versus 0.21%). These results indicate that the rugose exopolysaccharide consists partially of N-acetylglucosamine and mannose. Rugose colonies produced significantly more biofilm material than did smooth colonies, and rugose colonies grown at 30°C produced more biofilm material than rugose colonies grown at 37°C. Ultrastructurally, rugose colonies show regional cellular differentiation, with apical and lateral colonial regions containing cells embedded in a matrix stained by Alcian Blue. The cells touching the agar surface are packed tightly together in a palisade-like manner. The central region of the colony contains irregularly arranged, fluid-filled spaces and loosely packed chains or arrays of coccoid and vibrioid cells. Smooth colonies, in contrast, are flattened, composed of vibrioid cells, and lack distinct regional cellular differences. Results from suckling mouse studies showed that both orally fed rugose and smooth variants elicited significant, but similar, amounts of fluid accumulated in the stomach and intestines. These observations comprise the first report of expression and characterization of rugosity by G. hollisae and raise the possibility that expression of rugose exopolysaccharide in this organism is regulated at least in part by growth temperature.

Grimontia hollisae is a human pathogen, originally described by Hickman et al. (11) in 1982, which was isolated from clinical samples from patients displaying acute diarrhea and named at that time as a new species, Vibrio hollisae. Recently Thompson, et al. (30) placed the organism into a new genus, Grimontia, when tests revealed that the 16S rRNA gene sequence of the organism showed similarities of only 90.8%, 91.2%, 93%, and 94.5% to the 16S rRNA gene sequences of the organism's nearest neighbors, Vibrio cholerae, Salinivibrio sp., Photobacterium spp., and Enterovibrio norvegicus, respectively. In addition, G. hollisae differs from other members within the Vibrionaceae by failing to grow on thiosulfate-citrate-bile-sucrose agar and by producing greater percentages of the cell wall fatty acids C_18:1ω9c and C_16:1ω9c (30).

While environmental samples occasionally contain G. hollisae (6, 23), the organism is found with much greater frequency in samples taken from clinical cases of gastroenteritis, acute diarrhea, bacteremia, and septicemia (1, 4, 9, 16, 18, 21). G. hollisae produces a hemolysin (Vh-rTDH) which differs somewhat from the thermostable hemolysin of Vibrio parahaemolyticus (12, 13, 23, 24, 35, 38, 39), and the organism can invade cells in tissue culture (19). The organism also produces an enterotoxin that causes the elongation of Chinese hamster ovary cells in tissue culture and induces fluid accumulation in suckling mice (14, 15). The enterotoxin differs structurally and functionally from both cholera toxin and the heat-labile toxin of Escherichia coli (14). In addition, in response to iron limitation, G. hollisae produces aerobactin, a dihydroxamate siderophore, as well as its receptor (25). A PCR method developed by Vuddhakul et al. (31) specifically targets two genes of G. hollisae, gyrB and toxR, facilitating both the isolation and identification of the organism from environmental or clinical samples.

In addition to the foregoing characteristics, G. hollisae possesses another feature that is likely to contribute to its ability to persist and survive in the environment: rugosity. Rugosity, or wrinkling, of colonies is accompanied by the production of voluminous amounts of carbohydrate-enriched extracellular material which is thought to protect the colonies from environmental hazards or stressful growth conditions. Rugosity occurs, and has been documented, in both biotypes of V. cholerae 01 (26, 32-34, 36, 37), V. cholerae 0139 (20), Vibrio vulnificus (10), and Salmonella enterica (3). In the present work and for the first time, information on the expression of rugosity in G. hollisae is presented.

MATERIALS AND METHODS

Stock cultures of G. hollisae strains.

See Table 1 for a list of strains (and properties) which were used in these studies. G. hollisae ATCC 33564 was obtained from the American Type Culture Collection and came as a mixed culture containing both rugose and smooth colonies. PCR analysis of the strains (both rugose and smooth variants) showed that they possessed the G. hollisae species-specific gyrB and toxR loci (31; data not shown). All of the strains, including the colony variants of ATCC 33564, were frozen in tryptic soy broth (BBL, Becton Dickinson Microbiology Systems, Cockeysville, MD) containing 1% salt and 25% glycerol, and maintained at −84°C. Though data on all of the strains in our collection are presented here, much of the work is from results obtained from studies with G. hollisae ATCC 33564. Separate rugose and smooth colonies of strain ATCC 33564 were serially passaged four times onto tryptic soy agar containing 1% salt (TSA-S) and were maintained as described above.

TABLE 1.

Phenotypic properties of G. hollisae strains used in this study

Strain no.	Phenotypea	CV staining (OD570 units)b		EPS by EMc	Source
		30°C	37°C
ATCC 33564 (SR501)	R	2.46 ± 0.47	1.51 ± 0.45	+	ATCCd
ATCC 33564 (SR501)	S	0.83 ± 0.26	0.31 ± 0.07	−	This study
SR507	M	2.72 ± 0.21	1.59 ± 0.17	+	WFUe
SR511	S	0.08 ± 0.02	0.06 ± 0.01	NTf	WFU
SR516	R	2.84 ± 0.05	2.56 ± 0.14	+	WFU
SR516	S	1.95 ± 0.18	1.42 ± 0.14	+	WFU

^a“Phenotype” refers to whether the colonies of a culture growing on TSA-S at 30°C produced rugose (R), smooth (S), or mixed (M) rugose and smooth colonies. See the text (Materials and Methods section) for an explanation of each of the tests.

^bRugose colonies produced more crystal violet (CV)-stained biofilm material than smooth colonies at both growth temperatures (P < 0.001). Rugose colonies grown at 30°C produced more biofilm material than rugose colonies grown at 37°C (P < 0.001). Statistical analysis was performed by using the Student t test.

^cEM, electron microscopy.

^dStrain 33564 was received from ATCC as a mixed rugose/smooth culture. Rugose and smooth colony variants were isolated during this study.

^eG. hollisae strains from reference 15; WFU, S. Richardson, Wake Forest University, Winston-Salem, NC.

^fNT, not tested.

Preparation of whole colonies of G. hollisae for fixation and examination by light and electron microscopy.

Whole rugose and smooth colonies were overlaid with molten and tempered agarose (Histogel; Richard-Allan Scientific, Kalamazoo, MI). After solidification of the agarose, the colonies were cut out of the agar and the samples were processed according to the procedure described by Fassel et al. (8). The colonies were processed as small cubes throughout fixation, postfixation, dehydration, infiltration with epoxy plastic, and final embedding. The specimens were trimmed to remove excess agar and dehydrated gradually through a graded series of cold 30%, 50%, 70%, and 95% ethanol solutions (4°C), three changes of 100% ethanol, and three changes of propylene oxide (room temperature) and slowly infiltrated with a plastic mixture containing Eponate (Ted Pella; Redding, CA). These specimens were embedded in flat molds so that they could be oriented approximately as they were in the living state. For light microscopy, the polymerized specimens were removed from the molds, trimmed, sectioned at 0.5 to 1.0 μM using glass knives, mounted on glass slides, and stained with a mixture of 1% toluidine blue and 1% sodium tetraborate. After rinsing in water, the slides were dried, dipped briefly in 100% ethanol, dried again, and cleared using xylene. Finally, coverslips were added using the mountant TBS Shurmount (Electron Microscopy Sciences; Fort Washington, PA). Sections were photographed using an Olympus model BX51 microscope with a Spot 2.3.1 digital camera interfaced to a Dell Optiplex GX260 computer (OPELCO, Dulles, VA). Alcian Blue-stained colonies processed according to the procedure described by Fassel et al. (8), as described above, were trimmed and sectioned at an ultrathin level using diamond knives. Sections were mounted on 300-mesh copper grids, doubly stained using 2% uranyl acetate and Reynolds' lead citrate as described by Tall et al. (29), and examined using a Zeiss EM-10 electron microscope.

Preparation of G. hollisae suspensions for examination by transmission electron microscopy.

Small samples of smooth and rugose cultures grown on agar for 36 to 48 h were removed from the agar using wooden sticks and immediately suspended in 1.5-ml aliquots of a cold fixative containing 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.2). To reveal acidic carbohydrates on the surfaces of smooth and rugose cells, cells from the same plates were prefixed and fixed in mixtures containing Alcian Blue using the above-mentioned procedure described by Fassel et al. (8). All samples were postfixed for 1 to 2 h in an ice-cold solution of 1% osmium tetroxide in 0.1 M sodium cacodylate buffer (pH 7.2), rinsed briefly in ice-cold 0.1 M cacodylate buffer (pH 7.2), and embedded in 3% Noble agar. The specimens were trimmed to remove excess agar, dehydrated, infiltrated, and embedded as described above but were placed in BEEM capsules for polymerization. The capsules were placed in an oven set at 60°C for 48 h. The polymerized samples were removed from the capsules, trimmed using razor blades and glass knives, sectioned using a Leica Ultracut S ultramicrotome, and processed as described above for electron microscopy.

Preparation of living cells of G. hollisae for examination by Nomarski differential interference microscopy.

After 24 and 36 h of growth on TSA-S plates, cells were removed from rugose colonies using fine pipette tips and delivered to very small droplets (25 μl) of 0.9% saline on glass slides. Coverslips were added, and the preparations were examined using an Olympus BX51 microscope.

Procedure for obtaining quantitative estimates of effect of temperature on biofilm formation by G. hollisae.

This series of experiments was undertaken to compare the amount of biofilm material produced by rugose organisms over a 24-h period with the amount produced by smooth organisms under the same conditions and to determine the extent to which growing the organisms at two different temperatures, 30°C and 37°C, affected biofilm formation. Four replicate experiments were conducted to compare the biofilm growth at 37°C, and eight replicate experiments were performed to compare the biofilm growth at 30°C. Each of these 12 experiments involved taking smooth and rugose cells from frozen stocks and culturing them onto TSA-S for 24 h; inoculating two 5-ml volumes of M9 medium supplemented with 1% salt, one with 24-h smooth cells and the other with 24-h rugose cells; incubating the two cultures at 30°C with shaking at 100 rpm for approximately 10 h; adjusting the optical densities at 650 nm (OD₆₅₀) of both cell suspensions to 0.4; and finally inoculating 500 μl of marine broth (1:100 dilution of M9 culture) into each of 12 13- by 100-mm glass tubes. The cultures were grown statically for 24 h at either 30°C or 37°C. The cell suspensions were discarded; each tube was rinsed six times with running deionized water and allowed to drain; and 600-μl aliquots of freshly prepared and filtered 0.1% crystal violet were added and left for 30 min. The stain was discarded, and each tube was rinsed 12 times with running deionized water, drained, and allowed to dry in a hood. The material stained with crystal violet in each tube was solubilized over a period of 6 h using 1-ml aliquots of dimethyl sulfoxide; and 900-μl aliquots of the samples were transferred to plastic 1-ml cuvettes and measured at a wavelength of 570 nm using an Ultraspec 2000 spectrophotometer (Pharmacia Biotech). The data were evaluated using Student t tests and by applying a statistical analysis system procedure known as the general linear model.

Bleach stress test.

Four replicate experiments were performed to compare the responsiveness of smooth and rugose cells to stress imposed by the treatment with bleach. Each experiment involved growing rugose and smooth cells from frozen stocks on TSA-S for 24 h, inoculating smooth and rugose cells into 5-ml aliquots of M9-1% NaCl medium, and incubating the tubes for approximately 10 h at 30°C with 100-rpm shaking. The optical densities of the cultures were adjusted to approximately 0.400 (OD₆₅₀), and 1-ml aliquots of the smooth and rugose cultures were added to 9-ml volumes of M9-1% salt containing 0.0002% sodium hypochlorite (Sigma-Aldrich, St. Louis, MO). Immediately the cells were subjected to bleach for 5 and 30 min, the suspensions were diluted serially to 10⁻⁷ by sequentially adding 0.1-ml aliquots to small plastic tubes containing 0.9-ml volumes of 0.9% NaCl. The 10⁻³, 10⁻⁵, and 10⁻⁷ samples were plated in triplicate on TSA-S. Control rugose and smooth samples, which were not exposed to bleach, were diluted and plated similarly. The total elapsed time from plating the 5-min and 30-min samples were 18 min and 53 min, respectively. All of the plates were incubated at 30°C for 48 h, and the colonies were counted with the assistance of a Biologics AccuCount 1000 plate counter (28). These counts were used to estimate the recovery rate for each group. The recovery rate was calculated for a given group by dividing the average plate count obtained from the bleach-treated cells by the average plate count obtained from cells that were not exposed to bleach. This value was multiplied by 100 to yield the mean percentage recovery for the group. The data were evaluated using analysis of variance.

Flow cytometric analysis and direct fluorescent lectin staining.

A single colony from TSA-S overnight cultures of rugose and smooth variants of ATCC strain 33564 was inoculated into marine broth and allowed to grow. Flow cytometric analysis of grown rugose and smooth marine broth (MB) cultures was done by diluting each culture 1:50 in MB and then vortexing each broth culture dilution to make an evenly dispersed cell suspension. The cell suspensions were sorted according to size using an EPICS Elite flow cytometer (Beckman/Coulter, Miami, FL). Sorted cells were then analyzed by phase-contrast microscopy to determine cell clump size. Lectin binding studies were accomplished by diluting the MB cultures 1:50 in HEPES buffer containing 10 μg/ml of wheat germ agglutinin (WGA) or Galanthus nivalis lectin (GNL). Cells were sorted by fluorescent signal.

Suckling mouse assay.

Inocula were cultured on TSA-S medium and tested for their ability to cause fluid accumulation in sealed suckling mice as described previously by Kothary and Richardson (15). Pregnant ICR mice were obtained from Harlan Sprague Dawley (Indianapolis, Ind.), and studies were carried out in accordance with Institutional Animal Care and Use Committee-approved protocol number 413. Suckling mice weighing 2.5 to 4.0 g were starved for 6 h and were orally fed 50 μl of a diluted culture, and mice were sacrificed 6 h postchallenge. The fluid accumulation (FA) ratio was expressed as the ratio of the weight of the stomach plus intestine to the remaining body weight. The FA ratios were analyzed as previously described by using the Student t test (28, 29).

RESULTS

Phenotypic expression of smooth and rugose colony variants of G. hollisae.

Smooth and rugose variants of G. hollisae produce phenotypically different colonies. Smooth colony variants grown at 30°C on TSA-S form colonies that are flattened, translucent, and homogeneous in appearance (Fig. 1A). Rugose colony variants, in contrast, produce colonies that are wrinkled, opaque, and possessed colony margins which are regionally heterogeneous. Moreover, rugose colonies growing on the same agar plate show variations in size and appearance (Fig. 1B to D). Most rugose colonies display shallow, radially distributed marginal folds and a few larger folds centrally (Fig. 1B and C); but some rugose colonies show fine, intricate folding that is best revealed using image enhancement software (Fig. 1D). The latter colony type does not appear to progress to a more highly folded form. Occasionally, rugose colonies will have small amounts of smooth colony material (“oozing”) radiating from their margins (Fig. 1C). Oozing is a relatively infrequent but quite persistent occurrence in rugose-expressing G. hollisae cultures. For some strains, e.g., ATCC 33564 and SR516, rugose variants could be cultured separately from smooth-variant counterparts. It is noteworthy as well that not all of the G. hollisae strains produced rugose colony variants (Table 1), and some strains produced mixed rugose/smooth colony variants, while other strains presented only either rugose or smooth colonies. Strain SR507 presented cultures with mixed phenotypes, despite many attempts to isolate and separate the rugose and smooth variants. Strain 511 grew in culture as a smooth phenotype only.

FIG. 1.

Colonies of G. hollisae ATCC 33564 grown on TSA-S at 30°C. Panel A represents a smooth-colony variant, and panels B to D represent the rugose colony variant. The arrow in panel C shows small amounts of smooth colony material radiating from the margin. Bar markers represent 1 mm.

Phenotypic expression of rugosity is affected by temperature.

When cultured on TSA-S agar at 30°C, rugose- or smooth-variant strains of 33564 and SR516 produced predominantly rugose or smooth colonies, respectively, whereas strain SR507 when grown at 30°C produced both rugose and smooth colonies. When cultured on TSA-S at 37°C, all of these strains formed smooth colonies that were grossly indistinguishable from each other. In contrast, strain 511 produced only smooth colonies at both growth temperatures.

Morphological changes occurring in living smooth and rugose cells during colony formation at 30°C and 37°C.

Normarski differential interference contrast (DIC) microscopy was employed to follow cell-to-cell interactions during colony formation. During colony formation at 30°C, rugose cells underwent considerably more morphological change than smooth cells. In living samples examined after 24 to 48 h of growth at 30°C, cells associated with smooth colonies are rod-like or vibrioid in shape (Fig. 2A), and some are motile. Cells associated with rugose colonies at 24 h are predominantly nonmotile, rod-like or vibrioid cells; but unlike smooth cells, they often align themselves in chains which aggregate laterally to form larger multicellular cell masses (Fig. 2B).

FIG. 2.

Nomarski differential interference microscopy of smooth cells of G. hollisae (A) and rugose cells (B) at 36 h of growth at 30°C. Magnifications of photographs are approximately 100×.

Cells isolated from both smooth and rugose colonies grown at 37°C for 24 to 36 h were predominantly motile and rod-like or vibrioid shaped (data not shown). Cells associated with rugose colonies had characteristics similar to those seen for smooth cells cultured at 30°C. However, occasionally larger cells, which are absent in cell suspensions from smooth colonies, were seen (data not shown).

Features of fixed, embedded, and sectioned whole smooth and rugose colonies revealed by light microscopy.

The cell types found in fixed, intact colonies closely resemble cells in living colonies that were visible by Nomarski DIC microscopy. In transverse 0.5-μm plastic sections stained with toluidine blue, smooth colonies grown at 30°C for 36 h display unwrinkled free surfaces, acutely angular margins, and a closely packed population of cells (Fig. 3A and B). The cells within a colony may show focal clustering, but otherwise they display little or no consistent regional variation in orientation, size, or staining properties. In contrast, transverse sections of rugose colonies grown at 30°C for 36 h reveal distinct regional differentiation in organization (Fig. 3C). The free apical surface of the colony displays a variable number of folds. Cells found immediately below the colony surface are closely packed and form a prominently stained layer that continues laterally to the basal margins of the colony. The cells in this layer vary widely in size and appearance from rods to elongated cells, and the orientation of the cells appears to be random (Fig. 3D). The margins of rugose colonies are slightly blunted, and the cells comprising them are continuous with cells in the basal region. Many of the basal cells in direct contact with the agar surface are elongated and oriented vertically (Fig. 3E), but cells of various sizes are also present. In the interior of the rugose colony (Fig. 3C to E), many of the cells appear to be rods or cocci, arranged in short chains, less heavily stained, and less closely packed within the matrix than cells found elsewhere in the colony.

FIG. 3.

Transverse 0.5-μM plastic sections stained with toluidine blue of fixed smooth (A and B) or rugose (C to E) colonies grown at 30°C for 36 h. Panel A is a low-magnification photomicrograph of margins of a smooth colony. Panel B is a photomicrograph of a higher-magnification image of the colony shown in panel A, showing both the apical and basal regions of the colony. Panels C to E represent photomicrographs of different regions of a rugose colony, showing its apical and basal margins (C), a higher-magnification image of a region of demarcation between light and heavy cell concentrations within a colony fold (D), and a higher-magnification image of the basal margin of the colony. Magnifications of photomicrographs are approximately 20× to 400×.

Ultrastructural features of smooth and rugose colonies.

Ultrathin sections of intact smooth colonies grown at 30°C show mainly single, rod-like or vibrioid cells, some of which bear single polar flagella (Fig. 4A and B). The cells also appear to be randomly oriented within the colony and lack tight associations with the basal agar surface (Fig. 4A). The surfaces of individual cells isolated from smooth colonies display a small amount of acidic mucosubstances, revealed by Alcian Blue staining (Fig. 4C). On the other hand, rugose colonies grown at 30°C consist of cells in the marginal apical layer that vary in shape from coccoid and rod-like to elongated cells (Fig. 4D). Staining with Alcian Blue reveals that the cells in this region are embedded in a prominent extracellular matrix displaying an electron-dense, fibrous or webbed appearance (Fig. 4D and E). Cells associated with the basal surface of the colony are often elongated; and, as indicated earlier, these cells tend to align vertically in relation to the agar surface (Fig. 4F). Cells that are located in the interior of the colony, just below the apical layer, are mainly a mixture of rod-like and coccoid cells embedded in copious amounts of a fine fibrous extracellular matrix (Fig. 4G), similar to what was seen by light microscopy. These cell types are also the ones found most commonly in the region immediately interior to the elongated cells at the base of the colony. Coccoid cells display small, irregularly shaped masses of electron-dense material associated with their cell membranes, and scattered tiny vesicles are found frequently in the matrix immediately surrounding them (Fig. 4G, inset). Similar vesicles are also found, but less frequently, throughout the colony in the vicinity of the larger cell types. Also found in all regions of the colony are variable numbers of degenerating cells, often immediately adjacent to intact cells (Fig. 4H). In some locations deep within the colony, cells can be very sparse and widely separated from each other by large expanses of matrix. Cells obtained from rugose colonies grown at 30°C for 36 h and processed as a cell pellet looked similar to those cell types found associated with a rugose colony (data not shown). Rugose colonies grown at 37°C show features closely resembling those of smooth colonies.

FIG. 4.

Electron photomicrographs of ultrathin sections of fixed smooth (A and B) or rugose (D to H) colonies grown at 30°C for 36 h. Panel A is a low-magnification electron photomicrograph of the basal margin of a smooth colony. Panel B is an electron photomicrograph of a higher-magnification image of the colony shown in panel A, showing cells that are flagellated. Panel C represents smooth cells stained with Alcian Blue, showing expression of exopolysaccharide. Panels D to H represent photomicrographs of different regions of a rugose colony, showing its apical (D) and basal (F) margins, a higher-magnification image of a region of heavy cell concentration within a colony fold (E), showing cells embedded in copious amounts of exopolysaccharide, and a higher-magnification image of a region of low (G) or high (H) cell concentration within a colony fold, showing cells embedded in copious amounts of exopolysaccharide. The arrows in panel F and H are pointing to elongated and degenerate cells, respectively. Bar markers represent 1 μM except for the bar marker in the panel G inset, which represents 0.2 μm.

Biofilm formation by G. hollisae over 24-h period.

Rugose cells (evidenced by a greater accumulation of crystal violet-stained material) cultured at 30°C or 37°C produced significantly more biofilm than smooth cells cultured at the same temperature (Table 1). Significantly more biofilm material was established by rugose cells grown at 30°C than by rugose cells grown at 37°C. Smooth cells grown at 30°C also produced significantly more biofilm material than smooth cells grown at 37°C, but the amount was significantly less than that obtained for rugose cells. The biofilm material associated with rugose variants of G. hollisae grew as a pellicle which was clearly visible on the upper surface of the broth culture and clung to the inner wall of culture tube (data not shown). In fact, growth was so concentrated to the pellicle that the liquid portion of the culture lacked any significant turbidity. In contrast, the smooth variant grew less as a pellicle and produced a clearly visible and highly turbid cell suspension in the liquid portion of the broth culture. The amount of crystal violet accumulated by the rugose strains also correlated nicely with the observation of extracellular polysaccharide or exopolysaccharide (EPS) by electron microscopy examination. Though G. hollisae strain SR516 is represented in Table 1 as producing both rugose and smooth variants, smooth-variant cultures always produced a higher percentage of rugose colonies than did the smooth variant of strain 33564. As shown in Table 1, strain SR516 also produced more EPS than strain 33564; therefore, it may have a greater propensity to form more rugose colonies than strain 33564. This was also evident in electron microscopy examination of smooth colony variants of cells from SR516, which also possessed a greater amount of EPS.

Bleach kill kinetics with G. hollisae strain 33564.

Results of bleach kill kinetic experiments are as follows. By 5 min (elapsed time, 20 min) of 0.0002% sodium hypochlorite exposure, 60.58% (range, 3.25 to 2,865.33%) of the median number of rugose cells within the treatment population were still alive, compared to only 0.20% (range, 0.04 to 9.41%) of the smooth cells. The difference in percent recovery between the two groups at this time point was statistically significant (P < 0.01, as determined by using Dunn's method of one-way analysis of variance). The percentage of the rugose cells sampled at 30 min (elapsed time, 53 min) still alive by the end of the experiment had dropped to 22.21% (range, 0.73 to 990.68%), whereas the percentage of the smooth cells recovered was statistically significantly different (P = 0.022) and was approximately 0.2% (0.19%; range, 0.00003 to 13.33%). There was no statistical difference in percent recoveries between rugose cells after 5 (20 min of total exposure) min of exposure to bleach, and the percentage of rugose cells recovered after 30 min of exposure (53 min of total exposure) to bleach. These results show that the population of smooth cells was significantly reduced within 5 min of exposure to bleach, whereas the population of rugose cells survived exposure to bleach throughout the experiment. There was also no evidence that the stress of bleach exposure during the experiment caused smooth cells to phase switch to the rugose phenotype.

Flow cytometry analysis of rugose and smooth-variant cells.

Analysis of the cells, sorted by size by flow cytometry, displayed in Fig. 5A, showed that the population of rugose cells (Fig. 5A2) grown at 30°C was comprised of two major clusters (labeled peak 1 and peak 2), compared to a single cluster of cells observed for smooth cells (Fig. 5A1). Fluorescent lectin-binding studies, shown in Fig. 5B and C, revealed that the percentages of clustered rugose cells binding either WGA (5B2) or GNL (5C2) were greater than the respective percentages of smooth cells (WGA, 35% versus 3.5%; GNL, 67% versus 0.21%) binding the lectins (5B1 and C1). Together, these results suggest that the exopolysaccharide expressed on the rugose cells bound the fluorescently labeled lectins and is composed of both N-acetylglucosamine and mannose.

FIG. 5.

Analysis of smooth and rugose colony variants using flow cytometry. Data are shown as three-dimensional plots with forward scatter (FS) versus side scatter (SS) versus number of cells (# of cells). (A) Cell sorting by size of cells or clumps. A1, smooth cells; A2, rugose cells. Note that peaks one and two denote the two major clusters of cells found in the rugose population. (B) Cell sorting by bound WGA lectin-FITC signal. B1, smooth cells; B2, rugose cells. (C) Cell sorting by bound GNL lectin-FITC signal. C1, smooth cells; C2, rugose cells.

Suckling mouse assay.

The results of suckling mouse infectivity studies are as follows. Both rugose and smooth variants of G. hollisae ATCC 33564 produced a significant fluid accumulation ratio response (0.11 ± 0.006 for the rugose variant and 0.11 ± 0.009 for the smooth variant), which statistically was not different from the FA ratio response obtained for V. cholerae O139 strain 1837 (0.10 ± 0.01), the positive control sample, yet was significantly different from that for the negative control sample consisting of charcoal-yeast extract (CYE) alone (P < 0.001). (Fifteen OD₆₅₀ units of the rugose and smooth colony variants of G. hollisae in CYE broth represented approximately 7.75 × 10⁶ and 8.0 × 10⁸ CFU/mouse, respectively; 50 OD₆₅₀ units of the V. cholerae strain 1837 represented approximately 8.75 × 10⁹ CFU/mouse.)

DISCUSSION

The current understanding of the expression of the extracellular polysaccharide leading to the rugose phenotype by marine vibrios is that not only does this trait impart properties of enhanced resistance to stress agents and greater cell adherence attributes displayed in pellicle or biofilm formation (20, 26, 32-34, 36, 37), but it also confers drastic and complex structural modifications in colony morphology. Inferentially, these properties are due to differences in the cells comprising such colonies as well. It is thought that the adeptness of Vibrio cells at changing back and forth from either phenotypic variant would favor the variant derivative that is more fit for survival under the prevailing environment conditions, such as those encountered in a human host or on an environmental surface. Regulation of rugosity seems to be multifactorial and highly complex (5, 17, 26, 36, 37). Although there are some similarities, great differences are observed among organisms capable of expressing rugosity (3, 36, 37). Even though Hickman et al. (11) originally described mixed opaque and nonopaque colonies for V. hollisae, now named Grimontia hollisae, we feel that the present study is the first complete description of the rugose phenotype expressed by this organism.

Descriptions of the rugose phenotype for rugose-expressing organisms, such as V. cholerae (20, 22, 32-34), V. vulnificus (10), and Salmonella enterica serovar Typhimurim (3), and now G. hollisae, include recapitulations of two different colony variants: smooth colonies, which are comprised of cells which do not express an exopolysaccharide, and rugose colonies, which consist of cells which are expressing copious amounts of the rugose exopolysaccharide. The rugose colony variant displayed by ATCC strain 33564 was typical of that observed by the other rugose-expressive G. hollisae strains, displaying colonies with shallow, radially distributed marginal folds mixed with larger, centrally located folds. However, high-frequency rugose switching experiments as reported by Ali et al. (2) were not performed. Additionally, the observation of reversion or “oozing” of smooth colonies arising from rugose colonies in G. hollisae is identical to similar descriptions of “oozing” reported by White (33) for rugose-expressive V. cholerae strains. In fact, White's original description of rugose-expressive V. cholerae cells includes his observations that “while the general character of selected rugose races is readily maintained by selection, there is perpetual reversion to the normal growth habitat” and that “sooner or later during incubation of a plate culture reverting outgrowths develop at the margins of the rugose colonies, the degree of reversion varying from culture to culture.”

During colony formation at 30°C, rugose cells undergo considerably more morphological change than smooth cells, and the cell types found in fixed colonies closely resemble cells that are visible by differential interference microscopy. The observations by microscopy of living rugose cells in this study showed that over time the cellular behavior of cells was to align themselves in chains, which then eventually formed laterally arranged aggregates or large multicellular cell masses. This description is markedly similar to what Yildiz and Schoolnik reported for V. cholerae O1 El Tor strains (36). However, one noted deviation in the results from these two studies is in the time frame that the aggregate formations occurred between these two Vibrio species. Aggregate formation by V. cholerae O1 occurred more quickly than that formed by G. hollisae: 6 h for V. cholerae versus 24 h for G. hollisae. Similarly, the Nomarski DIC microscopy behavior of the living smooth cells showed little or no distinctive regional variation in orientation, size, or staining properties and was identical to that reported by Yildiz and Schoolnick (36) for smooth V. cholerae O1 cells. In contrast, cells obtained from rugose colonies grown at 30°C for 36 h revealed distinct regional differences in organization. These observations are similar to what both White (33) and Rashid (26) had reported for V. cholerae. In fact, the observation of elongated cells by DIC microscopy within cell suspensions obtained from rugose colonies at this time, as well as elongated cells associated with the basal surface of rugose colonies, is also reminiscent of White's (33) original description of rugose-expressive V. cholerae cells. The presence of elongated cells within a rugose colony is also suggestive that these cells have down-regulated the need to divide and form cell wall septa. Rashid et al. (26) reported that elongated cells of V. cholerae O1, which were deficient in recycling of cell wall material due to the loss of N-acetylmuramoyl-l-alanine amidase activity, also lacked the ability to form rugose colonies. Could these cells than be the source of the cells which make up the “ooze” material that we occasionally see with G. hollisae rugose colonies?

Phenotypic expression of rugosity in G. hollisae was also affected by growth temperature. When grown at 30°C, G. hollisae colonies were rugose in appearance, while when the organism was grown at 37°C, the colonies were smooth in appearance. This observation is different from what has been described for V. cholerae (33) but is similar to descriptions of Salmonella serovar Typhimurium (3) and V. vulnificus (10). Ultrathin sections of smooth colonies grown at 30°C show almost exclusively small, flagellated, vibrioid cells, whereas ultrathin sections of rugose colonies grown at 30°C show populations of cells varying in size within the superficial layer.

Results from bleach stress experiments showed that smooth cells are more sensitive than rugose cells to bleach treatment. The kinetics observed in the experiments reported here are similar to those reported for V. cholerae O1 (22, 27, 36). Rugose cells cultured at 30°C or 37°C attach and produce significantly more biofilm than smooth cells cultured at the same temperatures. Both rugose and smooth cells grown at 30°C produce significantly more biofilm material than their counterpart rugose and smooth cells grown at 37°C. The observation that rugose cells produce more biofilm than do smooth cells is similar to what has been shown for rugose-expressing organisms (26, 36).

Fluorescence-activated cell sorter analysis shows that the population of rugose cells was presented as cells associated in large clumps, compared to the single cells and cells in doublets observed for smooth cells. This observation is similar to those described by Morris et al. for V. cholerae O1 (22). Fluorescence-activated cell sorter analysis also showed that WGA and GNL, two lectins specific for terminal N-acetylglucosamine and (α-1,3)mannose polysaccharide residues bound better to rugose cells than to smooth cells. Together, these results suggest that the exopolysaccharide has both terminally arranged N-acetylglucosamine and mannose residues that are associated with the outermost meshwork of the exopolysaccharide expressed on the cells. This observation is different from that reported by Yildiz and Schoolnik (36), who found that the rugose polysaccharide expressed by an El Tor, Inaba strain of V. cholerae O1 was composed primarily of equal amounts of glucose and galactose, with smaller amounts of N-acetylglucosamine and mannose. However, these results are similar to those reported by Wai et al. (32) for an El Tor biotype, Ogawa serotype strain of V. cholerae O1. Thus, across serotypes of V. cholerae O1 and now G. hollisae, there seems to be divergency in the exopolysaccharide composition. Studies designed to isolate the exopolysaccharide of G. hollisae to quantitatively assess its composition are in progress.

Suckling mouse infection studies showed that both rugose and smooth colony variants of G. hollisae produced a significant fluid accumulation ratio response, which statistically was not different from the FA ratio response obtained for V. cholerae O139 strain 1837, the positive control sample, yet was significantly different from the negative control sample consisting of CYE alone. These results are consistent with those reported by Rashid et al. (26) for V. cholerae O1. The biological basis for the lack of any increased virulence observed for the rugose morphology-expressing strains is not clear. The inability to observe a statistically significant increase in virulence could be due the fact that in the mouse, rugosity may not have been expressed by G. hollisae because of the fact that the mouse's body temperature (35 to 37°C) may not have been conducive for rugose expression. More importantly, these results also support the contention put forth by Faruque et al. (7) for V. cholerae, and possibly for all rugosity-expressing organisms, that rugosity may be more involved in persistence, transmissibility, and survival than in causing enteropathogenicity. For example, it has been noted by these researchers that in V. cholerae intestinal infections, cells are shed as planktonic material as well as cells in clumps of biofilm. This could also explain why for V. cholerae there is no temperature regulation of rugosity (33). Further studies on additional strains of Grimontia and with other infection models will be necessary to determine if there is any relationship of rugosity to virulence.

In summary, Hickman et al. (11) in 1982 described two distinct colony types growing within a culture of a marine vibrio then named V. hollisae, which is now named G. hollisae. Little information was given except that one colony type was described as an opaque type of colony and the other as nonopaque. Here we report on a new phenotypic form of G. hollisae, the rugose variant, and its counterpart, the smooth variant, which are expressed predominantly during culture at 30°C, and we feel that the description reported here is a more refined description of the two colonies originally described by Hickman et al. (11). The unique expression of the rugose phenotype, which is controlled in part by the growth temperature, suggests that the expression of the exopolysaccharide may be important and more beneficial for cell survival. Indeed, in this study no difference in ability to cause fluid accumulation in suckling mice was noted between the two G. hollisae phenotypes. The observation that rugosity in G. hollisae promoted survival of the organisms under stress suggests that expression of the exopolysaccharide is important in persistence and adaptation of the organism in a variety of econiches. Understanding the mechanisms that regulate and promote the persistence of G. hollisae is a key factor in understanding this pathogen and might lead to better systems for prediction and detection of the organism within seafoods and its ultimate elimination from such foods and prevention of disease.