Download Vibrio cholerae 01 Can Assume a Chlorine

1364
Vibrio cholerae 01 Can Assume a Chlorine-Resistant Rugose Survival Form
that Is Virulent for Humans
J. Glenn Morris, Jr., Marcelo B. Sztein, Eugene W. Rice,
James P. Nataro, Genevieve A. Losonsky,
Pinaki Panigrahi, Carol O. Tacket, and Judith A. Johnson
Center for Vaccine Development, Divisions of Infectious Diseases
(Department of Medicine) and of Neonatology (Department of
Pediatrics), and Department of Pathology, University of Maryland
School of Medicine, and Veterans Affairs Medical Center, Baltimore,
Maryland; Drinking Water Research Division, Risk Reduction
Engineering Laboratory, US Environmental Protection Agency,
Cincinnati. Ohio
Vibrio cholerae can shift to a "rugose" colonial morphology associated with expression of an
amorphous exopolysaccharide that promotes cell aggregation. Flow cytometric studies indicated
that up to 3% of particles in rugose cultures represented aggregates of >5 bacterial cells. Rugose
variants of our test strains displayed resistance to killing by chlorine, with viable cells persisting for
>30 min in 2 mg/L free chlorine; strains also showed resistance to killing by complement-mediated
serum bactericidal activity. Six volunteers fed 106 cfu of a rugose variant of V. cholerae 01 EI Tor
Inaba N16961 developed symptoms typical of cholera, with a mean diarrheal stool volume of 2.2
L (range, 1.4-4.3). Isolates recovered from the stool of infected volunteers retained the rugose
phenotype. The data suggest that rugose strains cause human disease. The role of these strains in
the epidemiology of cholera remains to be determined.
Researchers working with Vibrio cholerae have traditionally
selected smooth colonies from culture plates for study; little
attention has been given to the pathogenesis and survival of
other colonial morphologies, most of which have been dismissed as avirulent rough variants. In 1938, Bruce White at
the National Institute for Medical Research (London) provided
a detailed description of "rugose," or wrinkled, colonies that
appeared on serial passage of smooth V. cholerae strains [1]
(figure 1). In a series of studies conducted in the 1930s and
1940s, he demonstrated that cells from rugose colonies were
embedded in an amorphous intercellular matrix (which he
termed zoogloea) and were distinct morphologically and immunologically from "rough" variants, which have a modified
lipopolysaccharide (LPS) [1-3].
Interest in rugose forms of V. cholerae was revived during
the recent South American epidemics with the demonstration
by E. W. Rice and colleagues at the US Environmental Protection Agency that bacteria within rugose cultures remained via-
Received 5 March 1996; revised 16 July 1996.
Presented in part: 29th Joint Conference on Cholera and Related Diarrheal
Diseases, Asilomar, Calfomia, 1-3 December 1993.
All studies were approved by the Institutional Review Board, University of
Maryland at Baltimore. Informed consent was obtained from all study participants.
Financial support: Thrasher Research Fund (to J.G.M.) and Department of
Veterans Affairs (to J.A.J.) for laboratory studies; NIH (AI-15096) for volunteer studies.
Reprints or correspondence: Dr. J. Glenn Morris, Jr., Infectious Diseases
Section, Veterans Affairs Medical Center, 10 N. Greene St., Baltimore, MD
21201.
The Journal of Infectious Diseases 1996; 174:1364-8
© 1996 by The University of Chicago. All rights reserved.
0022-1899/96/7406-0035$01.00
ble in the presence of chlorine. Both smooth and rugose forms
were found to be adherent to Caco-2 cells (a model for intestinal
adherence), and both generated a striking fluid accumulation
response in ligated rabbit ileal loop models, suggesting that
smooth and rugose variants had comparable virulence [4].
However, the rugose phenotype remained poorly characterized,
and there were uncertainties about the ability of rugose strains
to actually colonize and cause illness in humans.
Materials and Methods
Characterization of rugose variants. Initial studies were done
with 48 V cholerae strains, including 25 in a group 1 (V cholerae
01),9 in 0 group 139 (V cholerae 0139 Bengal), and 14 in other
a groups (non-Ol V cholerae). Three strains were studied in
more detail: V cholerae N16961, an EI Tor Inaba strain well
characterized in animal models and human volunteers [5]; V cholerae C6706, an EI Tor Inaba strain from a cholera patient in Peru
[4]; and V cholerae NRT-36S, an encapsulated 031 strain that
we have shown to be pathogenic in animals and volunteers [6]. To
isolate spontaneous rugose variants, bacteria from smooth colonies
were passed in alkaline peptone water for 2-4 days and then plated
on Luria agar at 37°C.
Hiss staining was used to examine cells by light microscopy.
For electron microscopy, cells were grown on CF A agar at 25°C
(conditions that have previously been shown to optimize expression of pili in V. cholerae [7]) and stained with phosphotungstic
acid. Cell aggregation was quantitated by flow cytometry (EPICS
ELITE flow cytometer/cell sorter; Coulter Cytometry, Hialeah,
FL). In these latter studies, data for 200,000-500,000 bacterial
particles/sample were analyzed in a 128 X 128 matrix by using
the Multi-2D software package (Phoenix Flow Systems, San
Diego) and shown as posterior views of rotated forward-scatter
versus side-scatter isometric displays.
JID 1996; 174 (December)
Concise Communications
1365
Figure 1. Photomicrograph of rugose colony of V. cholerae.
Chlorine inactivation experiments (in triplicate) were done at
20 ± 2aC in pH 7.0 chlorine demand-free phosphate buffer at
various concentrations of free chlorine [4, 8]. To evaluate resistance of rugose variants to complement-mediated killing by normal
human serum, bacteria were incubated for 30 min with 65% pooled
normal (nonimmune) human serum at 37°C in the presence of
guinea pig complement [9].
Volunteer studies. Studies were done under quarantine on the
inpatient Research Isolation Ward of the Center for Vaccine Development , University Hospital, following previously described protocols [6, 10]. Mean age of volunteers was 28 years (range, 19-39)
and all were in excellent health as determined by comprehensive
physical and laboratory examinations.
The challenge strain was administered to the volunteers with 2
g of sodium bicarbonate to neutralize stomach acidity . All stools
passed by the volunteers were characterized, weighed, and cultured
for V. cholerae by direct plating after enrichment [6, 10]. Therapy
with oral rehydration solution (1.5 mU1.0 g of liquid stool) was
initiated immediately after passage ofthe first diarrheal stool. Therapy with tetracycline (to which both smooth and rugose variants
were susceptible) was started 24 h after the onset of diarrhea.
In preparing the bacterial inoculum, we picked 3 rugose colonies
that were agglutinable with Inaba antiserum from brain-heart infusion (BHI) plates that had been inoculated with a pure frozen stock
of a spontaneous rugose variant of V. cholerae strain N 16961.
After overnight incubation at 37°C, the bacterial growth was harvested from all three BHI plates, washed twice in PBS, and resuspended in PBS, pH 7.4, with vigorous vortexing for I min. The
concentration of the resultant suspension was adjusted spectrophotometrically, and counts were verified by direct plating.
In samples collected on days 7 and 10, antibody-secreting cells
were identified from among peripheral blood mononuclear cells
by ELISPOT as previously described [10]; cells were screened for
expression of IgG, IgM, and IgA directed against cholera toxin
(CT), V. cholerae Inaba and Ogawa LPSs , and the rugose exopolysaccharide (isolated and partially purified by a modification of the
method originally described by White [I D. In samples collected
on days 7, 10, 21, and 28, specific serum IgG responses to CT
and IgG, IgM, and IgA responses to Inaba and Ogawa LPSs, were
measured by ELISA. Sera were also tested for vibriocidal activity
against standard test strains (V. cholerae 01 El Tor Ogawa 3008
and V. cholerae 0 I classical Inaba VO I) and the challenge strain,
V. cholerae N 16961/Ru . Methods and definitions for positive responses have been reported [6, 10, II].
Results
Characterization of rugose variants. It was possible to
identify rugose variants of all 48 V. cholerae strains studied.
On light microscopy with a Hiss stain, rugose variants appeared
as small coccoid shapes aggregated in a background of faintly
stained material. In contrast, the smooth variants had normal
vibrio morphology and little background material. When examined by electron microscopy, type C pili [7] were present on
smooth and rugose variants of C6706; no pili were seen on
Concise Communications
1366
JID 1996; 174 (December)
N16961/Ru
1
1
SS Log
FS Log
N16961
Figure 2. Light-scatter properties of V
cholerae strains N 16961 ( smooth) and
N16961/Ru (rugose), as determined by
flow cytometry. Data are shown as posterior views of rotated forward-scatter vs.
side-scatter isometric displays . In experiment shown, % of aggregates was 0.4%
for NI6961 and 2.8% for NI6961 /Ru .
1
SS Log
FS Log
either variant ofNRT-36S. Electron microscopic studies further
demonstrated aggregation of bacterial cells in the rugose cultures, as well as a suggestion of an intercellular matrix material
linking the cells.
Flow cytometric techniques were used to quantitate cell aggregation. While aggregates were virtually nonexistent in smooth V.
cholerae suspensions, up to 3% of the events collected from
rugose bacterial suspensions represented bacterial aggregates
(figure 2). To confirm that the major peak near the origin in the
isometric display represented single bacteria, while the higher
intensity forward-scatter and side-scatter particles represented bacterial aggregates, bacteria in the different regions were sorted
(i.e., physically separated) and examined by Gram's stain. Small
particles were within an electronic gate centered on the singles/
doubles peak shown in figure 2; aggregates were within a gate
centered on the aggregate region. While the main peak near the
origin in the isometric displays consisted primarily of single and
double cells, most bacteria in the aggregates region consisted of
6 to >50 bacteria in an aggregate form; 0.9% of the aggregate
particles counted had > 50 cells.
Concise Communications
JID 1996; 174 (December)
1367
bacterial count (log10)
7r----------------------------------.
6
5
4
3
2
1
O---.
o
-----.------l.--
5
..l--
10
--L
15
...i-
20
...--J'----
25
-.L.
30
..........--J
35
minutes
- - - N16961 smooth
Figure 3.
-+- N16961 rugose
Survival of V cholerae El Tor N1696l (smooth) and N16961/Ru (rugose) in presence of 0.5 mg/L free chlorine.
As shown in figure 3, cultures of smooth variants were consistently inactivated in <20 s when exposed to 0.5 mg/L free
chlorine. In contrast, disinfection of rugose variants displayed
a deviation from first-order kinetics, with an initial 2-3 order
of magnitude decrease in the number of viable bacteria, followed by persistence of a subpopulation of cells. In an effort
to totally eradicate rugose strains, bacteria were exposed to 2
mg/L free chlorine at pH 7.0 and 20°C; starting with an inoculum of ,....., 106 cfu, V. cholerae were still recoverable after 30
min. Smooth variants were completely killed when incubated
for 30 min with normal (non immune) human serum in the
presence of complement. Rugose strains survived in serum,
although counts decreased from an initial inoculum of 107 . 1 to
104 . 1 for C6706 and from 106 .9 to 104 .3 for N16961.
Volunteer studies. The spontaneous rugose variant of V.
cholerae strain Nl6961 was administered to 6 volunteers at an
inoculum of ~ 106 cfu/mL; when plated out, the inoculum was
99.83% rugose (1/600 colonies was smooth). All volunteers
receiving the rugose strain developed profuse, watery diarrhea,
with a mean incubation period of 19 h. Mean stool volume
was 2.2 L, with volunteers having an average of 15.3 diarrheal
stools. Peak excretion rates of V cholerae ranged from 2.4 X
107 to 1.1 X 109 cfu/g of stool. Of the V. cholerae colonies
recovered from stool, >99.8% had a rugose morphology.
Stool volumes for our volunteers did not differ significantly
from those of past volunteers challenged with 106 cfu of a
smooth N16961 variant; however, stool volumes of persons
receiving rugose strains were significantly different from those
reported after challenge with ~ 104 cfu of smooth forms of the
strain (P < .02, Wilcoxon rank sum test). Peak V. cholerae
counts in stool after rugose challenge were comparable to those
reported for volunteers after challenge with smooth N 16961
variants.
With the exception of the response to the matrix material
itself, immunologic responses were similar to those seen in
volunteers receiving smooth V. cholerae variants. In the ELISPOT assays, cells secreting IgG and IgA antibodies directed
against CT were identified in all 6 volunteers. Antibody-secreting cells directed against Inaba and Ogawa LPSs and the
matrix material were demonstrable in 1-5 volunteers, depending on the assay. All volunteers showed a strong serum
IgG anti-CT response (mean peak optical density = 2.6). They
also demonstrated vibriocidal responses to the smooth Inaba
and Ogawa test strains (reciprocal geometric mean peak antibody titers of 6303 and 1810, respectively). However, when
the rugose challenge strain was used in the vibriocidal assay,
typical responses were not seen, and when contents of test
wells were plated out, viable organisms could be demonstrated
at all serum dilutions.
Discussion
Rugose strains represent an interesting and still poorly understood morphologic variant of V. cholerae. Our data suggest
1368
Concise Communications
that all V. cholerae can assume this form, with conditions
present in alkaline peptone water (a standard enrichment broth
medium for V cholerae) selecting for or promoting the shift
to rugose morphology. Rugose strains appear to produce an
exopolysaccharide that promotes cell aggregation. Reminiscent
of the protective effect of biofilms [12], this aggregation may
shield individual cells from killing by disinfectants, such as
chlorine, or lysis by complement. Vibrios are an important
component of marine biofilms, and it is possible that the exopolysaccharide produced by V cholerae plays a role in marine
biofilm formation; this, in tum, may contribute to attachment of
bacteria to marine organisms, such as plankton. In this context,
rugosity may represent a normal biologic adaptation of a species (V cholerae) that originates from marine and estuarine
environments. Analogies to rugosity can be found in a number
of other bacterial species, including the expression of alginate
by mucoid strains of Pseudomonas aeruginosa and expression
of an adhesive exopolysaccharide by the marine genus Hyphomonas.
In contrast to rugose, rough V. cholerae strains have a modified or 0 antigen-deficient LPS [1, 13, 14]. It has generally
been accepted (although without overwhelming data) that
rough strains have decreased virulence [14, 15], and there has
been a tendency to assume that virulence was confined to
smooth V. cholerae variants. In prior studies at the Center for
Vaccine Development, great care has always been taken to
select smooth colonies for use in preparation of inocula for
cholera challenges. As a result, we felt comfortable in assuming
that prior challenges, conducted with N 16961 under identical
experimental conditions, were representative of response to the
smooth form of the strain. While numbers in volunteer studies
of this type are of necessity small, our data suggest that the
clinical syndrome and host immunologic response elicited by
smooth and rugose variants are directly comparable. Since vibriocidal activity is dependent in part on complement-mediated
killing (to which rugose forms show resistance), it is not unexpected that vibriocidal results differed in assays that incorporated the rugose challenge strain.
In his 1940 paper on the rugose morphology, White [2] states
that "it is difficult to escape the conclusion that the rugose
substance is a protective secretion with a role in assisting the
survival of the race [V cholerae] in nature." This comment
has particular relevance in light of the apparent ability of rugose
forms to survive in the presence of chlorine, which forms one
of the first-line defenses against cholera. Chlorination is an
effective intervention in controlling cholera [16]. However, if
rugose strains are present in a water system, chlorination may
simply decrease the number of viable V. cholerae rather than
totally eradicate the organism. The persistence of viable cholera
organisms in this setting may, in turn, contribute to further
JID 1996; 174 (December)
spread of the infection. There are limited observations suggesting that rugose forms are present in samples from clinical
and environmental sources [2]. However, we still know little,
if anything, about the impact of rugose strains on survival of
V. cholerae in the environment and the epidemiology of the
disease in human populations. Further studies in these areas
are warranted.
Acknowledgment
We thank Michael Tanner for his excellent technical assistance
in the flow cytometric analyses.
References
1. White PB. The rugose variant of vibrios. J Pathol Bacteriol 1938;46: 1-6.
2. White PB. The characteristic hapten and antigen of rugose races of cholera
and El Tor vibrios, J Pathol Bacteriol 1940; 50: 160-4.
3. White PB. A note on the globular forms of Vibrio cholerae. J Gen Microbiol 1950;4:36-7.
4. Rice EW, Johnson CH, Clark RM, et al. Vibrio cholerae 01 can assume
a "rugose" survival form that resists killing by chlorine, yet retains
virulence. Tnt J Environ Health Res 1993;3:89-98.
5. Levine MM, Black RE, Clements ML, Nalin DR, Cisneros L, Finkelstein
RA. Volunteer studies in development of vaccines against cholera and
enterotoxigenic Escherichia coli; a review. Tn: Home T, Holmgren J,
Merson MH, Mollby R, eds. Acute enteric infections in children. New
prospects for treatment and prevention. Amsterdam: Elsevier/NorthHolland Biomedical Press, 1981:443-59.
6. Morris JG Jr, Takeda T, Tall BD, et al. Experimental non-O group I Vibrio
cholerae gastroenteritis in humans. J Clin Invest 1990; 85:697 - 705.
7. Hall RH, Vial PA, Kaper JB, Mekalanos JJ, Levine MM. Morphological
studies on fimbriae expressed by Vibrio cholerae 01. Microb Pathog
1988;4:257 -65.
8. American Public Health Association. Standard methods for the examination of water and wastewater, 17th ed. Washington, DC: APHA, 1989.
9. Johnson JA, Panigrahi P, Morris JG Jr. Non-Ol Vibrio cholerae NRT36S
produces a polysaccharide capsule that determines colony morphology,
serum resistance, and virulence in mice. Infect Tmmun 1992;60:684-9.
10. Morris JG Jr, Losonsky GE, Johnson JA, et al. Clinical and immunologic
characteristics of Vibrio cholerae 0139 Bengal infection in North American volunteers. J Infect Dis 1995; 171:903-8.
11. Levine MM, Young CR, Black RE, Takeda Y, Finkelstein RA. Enzymelinked immunosorbent assay to measure antibodies to purified heatlabile enterotoxins from human and porcine strains of Escherichia coli
and to cholera toxin; application in serodiagnosis and seroepidemiology.
J Clin Microbiol 1985;21: 174-9.
12. Marshall KC. Biofilms: an overview of bacterial adhesion, activity, and
control at surfaces. ASM News 1992;58:202-7.
13. Kaper JB, Morris JG Jr, Levine MM. Cholera. Clin Microbiol Rev 1995;
8:48-86.
14. Barna D. Laboratory diagnosis of cholera. In: Barna D, Burrows W, eds.
Cholera. Philadelphia: WB Saunders, 1974:85-126.
15. Sack RB, Miller CEo Progressive changes of vibrio serotypes in germ-free
mice infected with V. cholerae. J Bacterial 1969; 99:688-95.
16. Deb BC, Sircar BK, Sengupta PG, et al. Studies on interventions to prevent
El Tor cholera transmission in urban slums. Bull WHO 1986;64: 127-131.