Microbiol. Immunol., 37(9), 679-688, 1993

Immunogenicity of Vibrio cholerae 01 Fimbriae in Animal and Human Cholera

Masahiko Ehara *1, Yoshio Ichinose 1, Mamoru Iwami 1, Akiyoshi Utsunomiya 1, Shoichi Shimodori 2, Stanley K. Kangethe 3, Bianca C. Neves 4, Krongkaew Supawat 5, and Satoshi Nakamura 6

1Department of Bacteriology, Institute of Tropical Medicine, Nagasaki University, Nagasaki, Nagasaki 852,

Japan, 2Department of Microbiology, School of Health Sciences, Kyushu University, Fukuoka, Fukuoka 812, Japan, 3Kenya Medical Research Institute (KEMRI), Centre for Microbiology Research, P.O. Box 54840, Nairobi, Kenya, 4Department of Bacteriology, Adolfo Lutz Instituto, Avenida Dr. Arnaldo 351, Brazil, 5National Institute of Health, Nonthaburi 11000, Thailand, and 6Department of Bacteriology, School of Medicine, University of the Ryukyus, Okinawa 903-01, Japan

Received February 22, 1993; in revised form, May 27, 1993. Accepted June 16, 1993

Abstract: Parenteral immunization with either formalin-fixed whole cells of the fimbriate Bgdl7 strain or purified fimbriae protected against Vibrio cholerae 01 infection in rabbits, independent of biotype and serotype. Parenteral immunization of adult rabbits with purified fimbriae prior to V. cholerae 01 challenge resulted in a reduction of 2 to 3 orders of magnitude in the number of bacteria recovered from the small intestines of immunized rabbits in comparison to non-immunized controls. IgG and IgA antibodies against fimbrillin of V. cholerae 01 were detected in the convalescent sera of patients with cholera; however, little fimbrial antigen was detected in the commercially available cholera vaccines when examined by polyclonal and monoclonal antibodies against fimbriae. These data suggest that fimbrial hemagglutinin is a major adhesin of V. cholerae 01 and that parenteral immunization with fimbriae generates a specific immune response in the gut that may serve as one means of mitigating subsequent V. cholerae 01 gut infection.

Key words: Cholera, Fimbriae, Vaccine

Cholera is an acute diarrheal disease that is caused by the Gram-negative bacillus Vibrio chol-erae 01. V. cholerae 01 infects via uptake of contaminated water and foods and preferentially colonizes the upper small intestine.

Fimbriae of V. cholerae 01 consist of cell-as-

sociated hemagglutinin that is proposed to be one of the major adhesins mediating the interaction of V. cholerae 01 and human epithelial cells in the small intestine (1). Passive protection tests with anti-mannose-binding hemagglutinin pili anti-bodies (polyclonal and monoclonal) have been

proven effective in protection against experimental cholera caused by El Tor vibrios in infant mouse and in rabbit intestinal loop models (7): but to date, there has been no correlation between serum antibody titers to known antigens purified from V. cholerae 01 and clinical protection. Naturally

occurring V. cholerae 01 infection confers long-lasting protection against reinfection, but such effective V. cholerae 01 antigens remain to be elucidated. Experimentally infecting rabbits with V. chol-

erae 01 using the ligated ileal loop test has shown that the bacteria are associated with small intestinal epithelial cells (4, 5, 8). Therefore, small intestinal infection of rabbits may be used to analyze paramet-ers of immunity that interfere with the persistence of V. cholerae 01 at the intestinal epithelium. As

previously reported (2, 3), fimbriate V. cholerae 01 strain of Bgd17 (classical biotype, Inaba serotype) was selectively induced and hydrophobic fimbriae were purified from the fimbriate strain.

The purpose of this study was to determine whether or not rabbits immunized parenterally

* Address correspondence to Dr . Masahiko Ehara,

Department of Bacteriology, Institute of Tropical Medicine,

Nagasaki University, 1-chome 12-4, Sakamoto, Nagasaki,

Nagasaki 852, Japan.

Abbreviations: AT-Broth, alkaline tryptone broth;

CFU, colony-forming units; ICDDR, B, International

Centre for Diarrheal Disease Research, Bangladesh; ORS,

oral rehydration salts; PBS, phosphate-buffered saline;

SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel

electrophoresis; VCC, viable cell counts.

679

680 M. EHARA ET AL

with 1% formalin-fixed whole cells of the fimbriate Bgd17 strain or with purified fimbriae are protected from V. cholerae 01 infection. Commercially available cholera vaccines and convalescent sera of

patients with cholera were also analyzed to deter-mine whether cholera vaccines contain any fimbrial antigens (polymer or oligomer) or whether anti-fimbrillin antibodies are developed in the convales-cent sera of patients with cholera.

Materials and Methods

Animals. New Zealand white breed rabbits of both sexes weighing 2.0-2.5 kg were obtained from a single breeder. They were acclimatized to the housing facility for at least 1 week prior to the study. Bacteria. Vibrio strains used in ligated ileal loop

tests were isolated at The International Centre for Diarrhoeal Disease Research, Bangladesh

(ICDDR, B) from patients with diarrhea. The strains are listed in Table 1. Eighty-six B series strains were from M. Iwanaga, University of the Ryukyus, and Bgd17 was from Y. Takeda, Kyoto University. Medium. Vibrio cells were grown in alkaline tryptone broth (AT-broth, 0.1% sodium bicarbo-nate, 1% Bactotryptone, 0.2% yeast extract, 0.5% NaC1) at 37 C under static conditions.

Antigens. a) Formalin-fixed whole cells. Fim-briate cells of strain Bgd17 were grown at 37 C overnight in 5 horizontal Roux bottles containing 60 ml of AT-broth. Whole cells were harvested by centrifugation and suspended in 20 ml of phos-

phate-buffered saline (10 mm PBS, pH 7.4) to a density of about 2 X 1010 colony-forming units

(CFU) per ml. The cell-suspension was washed once in PBS by centrifugation and the resultant

pellet was mixed with 50 ml of 1% formalin and kept overnight at 4 C. Formalin-fixed whole cells were washed 3 times with PBS by centrifugation and resuspended in 20 ml of PBS. After thoroughly mixing, the cell suspension was divided into 20 vials, then lyophilized and kept at — 30 C until use.

b) Purified fimbriae. Fimbriae of V. cholerae 01 were purified according to the method reported

previously (3) with a minor modification. Fim-briate vibrios of the strain of Bgd17 were cultured in 100 horizontal Roux bottles containing 60 ml of AT-broth for 8 hr. The inoculum (seed culture) was also prepared from an 8 hr-culture of the stock kept at —80 C.

Concerning the final yield of the hydrophobic

fimbriae, we examined the culture conditions. The

same seed culture, used in the above static culture,

was inoculated into 6 Erlenmeyer flasks (for

5 liters) containing 1 liter of AT-broth and cultured

with shaking at 37 C overnight. The cells were

harvested and processed for purification. The final

yields of the hydrophobic fimbriae were examined

on sodium dodecyl sulfate-polyacrylamide gel

electrophoresis (SDS-PAGE) by comparing the

fractions obtained from sucrose-linear density

gradient centrifugation.

Immunization. a) Immunization of rabbits with

formalin-fixed whole cells of fimbriate Bgd17

strain. Four rabbits were immunized with formalin-

fixed whole cells (2 X 101° CFU/ml) mixed with an

equal volume of Freund's complete adjuvant (1 ml

of the cell-suspension in PBS and 1 ml of the

adjuvant). The rabbits were injected subcutaneous-

ly at two sites in the back and intramuscularly at

two sites in the thigh every 2 weeks substituting

incomplete adjuvant for complete adjuvant.

The development of antibody against fimbriae

was monitored by an enzyme-linked immunosor-

bent assay (ELISA) and Western blotting.

b) Immunization of rabbits with purified fim-

briae. Three rabbits were immunized with purified

fimbriae (200 ƒÊg/ml) mixed with an equal volume

of Freund's complete adjuvant (0.5 ml of fimbrial

suspension in PBS and 0.5 ml of adjuvant). The

rabbits were injected three times every 2 weeks as

described in a). Development of a specific antibody

against fimbriae was examined by Western blotting.

Challenge test. Challenge experiments were done

1 week after the last injection in the adult rabbit

ligated ileal loop test.

Adult rabbit ligated ileal loop test. Adult rabbits

were anesthetized intravenously with Nembutal

(Abbot, Chicago, Ill., U.S.A.) and intramuscularly

with Ketalar 50 (Sankyo Co., Tokyo). The abdo-

men was opened, and the small intestine was tied

into 6-cm segments with 2-cm spacer loops between

the experimental loops. Eleven loops were made per

rabbit. The large loops were injected with 2 ml of a

fivefold-diluted bacterial suspension cultured over-

night in AT-broth. Control loops (6 cm) were

injected with 2 ml of PBS. Three classical strains

(86B1, 86B28, wild Bgd17) and two El Tor strains

(86B63, 86B64) were injected in duplicate into

random sites in the loops.

At the end of the experiment (8 hr after injection

of bacterial suspension), the rabbits were sacrificed

and the intestines were removed immediately. For

the rabbits immunized with formalin-fixed whole

AN EFFECTIVE VACCINE ANTIGEN: FIMBRIAE 681

cells, the differences in ileal colonization by vibrio

cells were compared with those of non-immunized

rabbits by electron microscopy. For the rabbits

immunized with purified fimbriae, intestinal seg-

ments (ca. 1 cm2) were removed from the loops.

Each segment was washed by vigorous shaking in

two small bottles of PBS to remove debris and any

vibrios not firmly bound. Each piece of the intestine

was homogenized in ceramic bowls containing 2 ml

of PBS. Viable cell counts (VCC) were determined

on TCBS (Eiken Chemical Co., Ltd.,Osaka, Japan)

agar plates after a serial tenfold dilution of the

homogenate in PBS. The VCCs of vibrio cells

colonized onto the small intestine were compared

between immunized and non-immunized control

rabbits.

Analysis of serum immunoglobulin. Rabbits

immunized with formalin-fixed whole cells or

purified fimbriae were bled from the auricular vein

every 2 weeks just before the boosting injection. The

development of specific IgG antibody against

fimbriae in the sera was examined by ELISA.

Microtiter plates, coated with 100 ƒÊ1 of purified

fimbriae (5 ƒÊg/m1) overnight, were incubated with

200-fold-diluted rabbit sera for 1 hr at 37 C. After

washing, plates were incubated for 1 hr with

peroxidase-conjugated goat anti-rabbit IgG (Cap-

pel, Organon Teknika Corp., Westchester, Pa., U.S.

A.). After washing, the plates were incubated in the

dark for 1 hr after the addition of peroxidase

substrate solution (0.5 mg/ml of o-phenylene

diamine and 0.02% H202 in 0.05 M citrate-phos-

phate buffer, pH 5.0). The reaction was stopped by

adding 4 N H2SO4, then the plates were read with

Toyo ELISA analyzer, model ETY-96.

The specific antibodies elicited by immunization

with the fimbriate cells or purified fimbriae were

determined by immunoblotting. Paired sera,

obtained from 8 cholera patients, bacteriologically

diagnosed and admitted at Bamrasnaradura Infec-

tious Disease Hospital, Nonthaburi 11000, Thai-

land, were also examined to determine whether or

not IgG and IgA antibodies against fimbriae were

developed in the peripheral blood. Immuno-

blotting was performed as described by Towbin et

al (9). Crude fimbrial fractions, obtained by su-

crose-linear density gradient centrifugation, were

resolved by SDS-PAGE in duplicate, one of which

was stained with Coomassie blue and the other

electroblotted onto nitrocellulose membrane (Bio-

Rad, Richmond, Calif., U.S.A.) using a Bio-Rad

electroblotting apparatus (70 V for 1 hr). SDS-

PAGE was performed in a 1 mm thick mini-slab gel

according to the system of Laemmli (6). Samples

were electrophoresed through a 5% stacking gel

followed by a 15% running gel for 1.5 hr at 15 mA

constant current per slab. Rabbit and human

convalescent sera were diluted 200-fold and 20-fold,

respectively. Sheets of nitrocellulose membrane

were incubated with either polyclonal antibody

against fimbriae as a positive control for fimbrillin,

or with convalescent sera or sera from rabbits

immunized with formalin-fixed whole cells or

purified fimbriae. They were then developed with

peroxidase-conjugated goat anti-rabbit IgG, anti-

human IgG or anti-human IgA.

Analysis of cholera vaccine. Commercially

available cholera vaccines were purchased from

Takeda Pharmaceuticals, Osaka, Japan and Denka

Seiken, Tokyo. Vaccines were also donated by

Biken, Kagawa, Japan and The Kitasato Institute,

Tokyo. Two vaccine vials containing 10 ml of 109

CFU/ml of both serotypes were combined and

centrifuged at 10,000 •~ g for 30 min at 4 C. Each

pellet was solubilized with 50 ƒÊ1 of Laemmli's

sample buffer for SDS-PAGE and boiled for 5 min.

After a 10-min centrifugation at 12,000 •~ g at room

temperature, 15 ƒÊl of the supernatant was loaded

onto mini-slab SDS-PAGE. The presence of the

fimbrial antigen was examined by Western blotting

with polyclonal and monoclonal antibodies against

fimbriae (3). Standard strains, such as 35A3 and

NIH41 used as cholera vaccine strains, were cul-

tured overnight in AT-broth under static condi-

tions. After adjusting the cell density to 109 CFU/

ml, whole cells were analyzed for the presence of the

fimbrial antigen by Western blotting together with

the strains listed in Table 1.

Electron microscopy. For negative staining, one

drop of the sample was placed on a sheet of

PARAFILM "M" (American Can Co., Greenwich,

Conn., U.S.A.) and a Formvar-coated copper grid

was floated on the drop for 2 min. The excess

solution was removed with filter paper. The speci-

men was washed three times with distilled water,

then stained with 1% uranyl acetate for 30 sec. The

excess stain was removed by blotting with a filter

paper. The specimen was examined with a JEM

100CX electron microscope operated at 80 kV. For

scanning electron microscopy, the pieces of the

intestine (3 to 5 mm) were extensively washed by

shaking consecutively in two bottles of PBS to

remove debris and any bacteria not firmly bound.

The samples were fixed in small vials with 2%

glutaraldehyde at 4 C overnight and postfixed for

1 hr with 1% osmium tetroxide in PBS. The samples

682 M. EHARA ET AL

were dehydrated by a stepwise replacement of the

water with ethanol. The samples were then dried in

a critical-point drying apparatus (Polaron E3000),

trimmed, mounted on copper stubs, coated with

gold-palladium, then examined under a JSM-840A

scanning electron microscope operated at 10 kV.

Statistical analysis. Bacterial counts were con-

verted by log transformation to stabilize variances.

The analysis of variances was performed using

Genstat V. Data are expressed as the mean values •}

the standard deviations. Data obtained from im-

munized rabbits and control rabbits were examined

by means of unpaired t-test.

Results

Purification of Hydrophobic Fimbriae Hydrophobic fimbriae of V. cholerae 01 were

purified from the fimbriate Bgd17 strain cultured in Roux bottles for 8 hr at 37 C under static conditions (Fig. 1). The properties of the fimbriae (HA activity sensitive to D-mannose and D-glucose, molecular weight of the subunit, fimbrillin) were the same as those previously reported (3).

The final yield of the hydrophobic fimbriae from the 8-hr culture was 3 times higher than that of the

overnight culture (data not shown). When the fimbriate cells were grown in shaking flasks, the final yield of the hydrophobic fimbriae was reduced markedly, as shown in Fig. 2.

Challenge Test in Rabbits Immunized Parenterally with Fixed-Whole Cells of the Fimbriate Bgd17 Strain Four rabbits immunized with the fixed-cells had

a significantly high titer of IgG antibody against fimbriae. No. 1 rabbit was a poor responder but the IgG antibody titer was significantly elevated, as shown in Fig. 3. When examined by immunoblot-ting, the serum (200-fold diluted) of the No. 1 rabbit was negative (Fig. 4). The other three rabbits were positive and the 6-week sera were comparable with the positive control (hyper-immune serum against fimbriae) (Fig. 4). These rabbits together with the four controls were challenged with the vibrio strains listed in Table 1. Each ileal loop specimen from both of immunized and unimmun-ized rabbits was examined under a scanning elec-tron microscope for colonization by vibrio cells. All the specimens prepared from the immunized rabbits showed complete absence of vibrios, how-ever localized colonization was seen in every speci-

Fig. 1. An electron micrograph of the hydrophobic fimbriae of V. cholerae 01. Vibrio cells of fimbriate strain Bgdl7 were

cultured for 8 hr in AT-broth under static conditions. Fimbriae were purified as mentioned in "Materials and Methods." The

fimbrial specimen was stained negatively with 1% uranyl acetate.

AN EFFECTIVE VACCINE ANTIGEN: FIMBRIAE 683

A B

men taken from the control rabbits. Representative

pictures taken from a control rabbit and from immunized rabbit No. 1, challenged with 86B28

strain are shown in Fig. 5. Despite the weak

response to the immunization, No. 1 rabbit was

completely protected from colonization by the

challenge vibrio strains. The mucous layer on the

villus surface was completely digested in the control

specimen, however the mucous layer of the immun-

ized rabbit remained intact.

Challenge Test in Rabbits Immunized with Purified Fimbriae (100 jig/dose)

Three immunized and three control rabbits were challenged with the strains described above. Devel-opment of the specific IgG antibody against fimbriae in three rabbits is shown in Fig. 6. Vibrios colonizing the upper small intestine were quantita-tively cultivated and compared between immunized and control rabbits (Table 2). Parenteral immuni-zation with purified fimbriae prior to V. cholerae 01 intestinal infection resulted in a reduction of about 2 to 3 orders of magnitude in the number of recovered bacteria independent of biotype and serotype of challenged V. cholerae 01 strains.

Fig. 2. Effects of culture conditions on the production of hydrophobic fimbriae of V. cholerae 01. SUS-PAUL, profiles of the fractions obtained by sucrose density gradient centrifugation were compared. A: Vibrio cells of the fimbriate Bgd17 strain were cultured by shaking in 6 Erlenmeyer flasks (for 5 liter) containing I liter of AT-broth at 37 C overnight. B: Vibrio cells of the same strain described above were cultured in 100 Roux bottles containing 60 ml of AT-broth at 37 C for 8 hr under static conditions. The arrowhead indicates the fimbrillin band at 18 kDa.

Fig. 3. Mean net O.D. of samples from four rabbits tested

for IgG anti-fimbriae antibody by ELISA by weeks post-

injection with formalin-killed whole cells of the fimbriate

Bgd17 strain. The broken line indicates "Negative con-

trol •} 3 standard deviations." Rabbit numbers are indicated

on the right.

Table 1. Bacterial strains

684 M. EHARA ET AL

When evaluated by the unpaired t-test, the dif-ferences in the average numbers of the colonized vibrios between immunized and non-immunized rabbits were statistically significant (P G 0.001).

The fluid accumulation ratio (FA ratio) was also obtained from these challenge tests. The FA ratios of the rabbits immunized with fimbriate whole cells or purified fimbriae were not markedly different from those of control rabbits (data not shown). In these animal models, vibriocidal activity, anti-chol-era toxin antibody, and anti-proteinase antibody may affect the FA ratio more than anti-fimbriae antibody.

Analysis of Convalescent Sera of Cholera Patients When 200-fold diluted sera was used to demon-strate the presence of a specific antibody against fimbriae, fimbrillin bands were undetectable by Western blotting. Fimbrillin bands were demon-strated faintly by 20-fold diluted sera. IgA antibody against fimbriae was detectable in the blood speci-mens 2 to 5 weeks after the onset of diarrhea (Fig. 7). IgG antibody against fimbriae was undetectable in the blood until the 3rd week after the onset of cholera, but lasted longer than the IgA antibody in the peripheral blood (Fig. 7). There was no detect-able antibody against fimbriae in the sera obtained on the third day after the onset of cholera (data not shown).

Analysis of Commercially Available Cholera Vac-cines Four commercially available cholera vaccines in

Japan were analyzed by Western blotting to deter-mine whether or not they contained any cell-as-sociated hemagglutinin (fimbriae). When reacted with the polyclonal antibody against fimbriae, only one of the four vaccines had a very small amount of fimbrial antigen (Fig. 8B, lane c). However, when reacted with monoclonal antibody No. 42, none of the four vaccines appeared to have the fimbrial antigen (Fig. 8C). As the fimbrial antigen was not detected clearly in vaccine samples, the standard strains used for cholera vaccine in Japan (35A3 and NIH41) were examined for the presence of the fimbrial antigen. All the strains tested had the fimbrial antigen when reacted with monoclonal antibody No. 42 (Fig. 9).

Discussion

We determined by challenge tests that hydro-phobic fimbriae of V. cholerae 01 are colonization factors and that they constitute an effective antigen. Protective immunity against the challenge strains independent of biotype and serotype was elicited in the gut of the rabbits immunized parenterally with purified fimbriae.

Culture conditions were examined to increase the final yield of the hydrophobic fimbriae. Prolonged

Fig. 4. Western blot showing the development of anti-fimbrillin antibody in four rabbits by weeks post-injection with formalin-killed whole cells of the fimbriate Bgdl7 strain. Lanes: pre, pre-immune serum; p.c., positive control (reacted with anti-fimbriae polyclonal antiserum). The arrowhead indicates the fimbrillin band.

AN EFFECTIVE VACCINE ANTIGEN: FIMBRIAE 685

Fig. 5. Representative electron micrographs showing the colonized vibrios and intact mucus layer. A: The.villus surface of

an unimmunized rabbit challenged with strain 86B28. The mucus layer is absent on the surface of the epithelial cells. B: The

villus surface of rabbit No. I immunized with formalin-fixed whole cells of fimbriate Bgd17. The mucus layer is kept intact

covering the microvilli. •~ 5,000.

686 M. EHARA FT AL

culture of vibrio cells resulted in a decreased recovery of the hydrophobic fimbriae (data not shown). The final yield decreased dramatically when vibrios were grown in AT-broth with shak-ing. The most suitable conditions seemed to be an 8-to 10-hr culture after the inoculation with a seed which had been cultured for 8 to 10 hr under static conditions. The hydrophobic fimbriae purified as described

above were sensitive to D-mannose, D-glucose, and D-fructose as reported previously (3). The fimbriae lose the HA activity completely in the presence of 0.25% of these monosaccharides. These biological

properties of the fimbriae support the efficacy of oral rehydration salts (ORS) for the early interven-tion of cholera, because it contains 2% glucose. Therefore, ORS plays two important roles in the Na+-glucose cotransport system and in inhibiting the colonization of V. cholerae 01 onto the epi-thelial cells in the small intestine.

Rabbits immunized parenterally with formalin-

Fig. 6. Western blot showing the development of anti-

fimbrillin antibody in three rabbits by weeks post-injection

with purified fimbriae (100 ,ƒÊg/dose). Lane: p.c., positive

control reacted with anti-fimbriae polyclonal antibody. The

arrowhead indicates the fimbrillin band.

Table 2. Bacterial recovery after intestinal challenge with V. cholerae 01

Fig. 7. Western blot of the convalescent sera showing the development of the IgA and IgG anti-fimbrillin antibodies by days

post-onset of cholera. Lane: p.c., positive control reacted with anti-fimbriae polyclonal antibody. The arrowhead indicates the fimbrillin band.

AN EFFECTIVE VACCINE ANTIGEN: FIMBRIAE 687

A B C

fixed whole cells of the fimbriate strain Bgd17

developed the antibody against fimbriae. This has

not yet been documented.

Rabbits immunized parenterally with purified

fimbriae showed a reduction in the number of

bacteria recovered from the intestine of 2-3 orders

of magnitude in comparison to non-immunized

controls, independent of the biotype and serotype

of the challenge strains. These findings indicate that

the hydrophobic fimbriae of V. cholerae 01 consti-

tute a colonization factor and that they are a

protective antigen indispensable to vaccine devel-opment.

Western blotting showed that the convalescent

sera contained specific IgA and IgG antibodies

against fimbriae. This evidence shows that fimbriae

of V. cholerae 01 are expressed in vivo during

colonization.

As IgA and IgG antibody titers against fimbrillin

were not high in the convalescent sera, fimbrillin

was not detected by 200-fold diluted sera. Further-

more, No. I rabbit immunized with formalin-fixed

whole cells of the fimbriate Bgd17 strain, which

showed a poor response to immunization, was

protected from colonization by challenge vibrios. These data suggest that even a low immune response

to fimbrillin leads to the protection from vibrio

colonization.

Western blotting revealed that commercially

available cholera vaccines in Japan did not contain

the fimbrial antigen. However, these vaccines are

mixtures of strains 35A3 and NIH41 which have the

fimbrial antigen. This discrepancy may derive from

the difference in the form of the fimbrial antigen on

the outer membrane. Most of the fimbrial antigens

Fig. 8. SDS-PAGE and Western blot analyses of commercially available cholera vaccines. A: SDS-PAGE, B: Western blot with anti-fimbriae polyclonal antibody, C: Western blot with anti-fimbriae monoclonal antibody No. 42. Lanes: M, Low molecular weight marker proteins (Pharmacia): a, purified fimbriae; b-e, cholera vaccines produced by The Kitasato

Institute, Biken, Seiken, and Takeda, respectively. One of four vaccines contains a very faint amount of the fimbrial antigen when reacted with the polyclonal antibody (see B, lane c) and none of the four vaccines contained fimbrial antigen when reacted with monoclonal antibody No. 42.

Fig. 9. Detection of the 18 kDa protein (fimbrillin) in whole-cell lysates from biotype classical strains of V. cholerae 01. Lanes: 1-7, wild Bgd17, Vc12, 35A3, Vcl6l,

NIH41, Inaba761, and H218, respectively. All samples were boiled in the presence of 2% SDS and 5% 2-mercaptoethanol for 5 min, then centrifuged at 12,000 X g for 10 min. Fifteen

microliters of each supernatant was loaded onto SDS-PAGE, transferred onto nitrocellulose membrane, then reacted with monoclonal antibody No. 42. All strains tested

had the fimbrial antigen, although fimbriae were not observ-ed morphologically in these strains under EM and they were not agglutinated by anti-fimbriae polyclonal antiserum.

688 M. EHARA ET AL

of non-fimbriate strains of V. cholerae 01 seem to localize in the hydrophobic domain of the outer membrane, forming oligomers. Under denaturing conditions, these oligomers may be easily detached from the membrane.

We propose that the fimbriae purified from a fimbriate strain as a colonization factor, are suit-able for use as an effective cholera vaccine, as opposed to using wild (non-fimbriate) strains iso-lated from stool specimens.

We are grateful to M. Ehara's associates Drs. Ohmiya, T., Kajiwara, Y., Kawane, K., Hokao, A., Shinzato, T. for their financial support of our work. Thanks are also due to Dr . Tansupousawadicul, S., Bamrasnaradura Infectious Dis-ease Hospital, Nonthaburi, Thailand for his cooperation in collecting blood specimens from cholera patients.

References

1) Ehara, M., Ishibashi, M., Ichinose, Y., Iwanaga, M., Shimodori, S., and Naito, T. 1987. Purification and

partial characterization of fimbriae of Vibrio cholerae 01. Vaccine 5: 283-288.

2) Ehara, M., Iwami, M., Ichinose, Y., Shimodori, S ., Kangethe, S.K., and Honma, Y. 1991. Selective induc-

tion of fimbriate Vibrio cholerae 01. Trop. Med. 33: 93-107.

3) Ehara, M., Iwami, M., Ichinose, Y., Shimodori, S ., Kangethe, S.K., and Nakamura, S. 1991. Purification

and characterization of fimbriae from fimbriate Vibrio cholerae 01 strain Bgd17. Trop. Med. 33: 109-125.

4) Freter, R. 1969. Studies on the mechanism of action of intestinal antibody in experimental cholera. Tex. Rep. Biol. Med. 27 (Suppl.): 299-316.

5) LaBrec, E.H., Sprinz, H., Scneider , H., and Formal, S. B. 1965. Localization of vibrios in experimental chol-era: a fluorescent antibody study in guinea pigs , p. 262-276. In Proceedings of the cholera research sym-

posium, Honolulu, Hawaii, U.S., Public Health Ser-vice Publication no. 1328, U.S. Government Printing Office, Washington, D.C.

6) Laemmli, U.K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680-685.

7) Osek, J., Svennerholm, A.-M., and Holmgren , J. 1992. Protection against Vibrio cholerae El Tor infection by specific antibodies against mannose-binding hemag-

glutinin pili. Infect. Immun. 60: 4961-4964. 8) Schrank, G.D., and Verwey, W.F . 1976. Distribution of

cholera organisms in experimental Vibrio cholerae infections: proposed mechanisms of pathogenesis and antibacterial immunity. Infect . Immun. 13: 195-203.

9) Towbin, H., Staehelin , T., and Gordon, J. 1979. Electrophoretic transfer of proteins from polya-crylamide gels to nitrocellulose sheets . Proc. Natl. Acad. Sci. U.S.A. 76: 4350-4354.