INFECION AND IMMUNITY, Feb. 1971, p. 243-248Copyright i 1971 American Society for Microbiology

Vol. 3, No. 2Printed in U.S.A.

Antigens of Bordetella pertussisV. Separation of Agglutinogen I and Mouse-Protective Antigen

R. F. ROSS' AND J. MUNOZRocky Mountain Laboratory, National Institute of Allergy and Infectious Diseases, Hamilton, Montana 59840

Received for publication 21 September 1970

Agglutinogen I of Bordetella pertussis strain 353/Z (serotype 1) was separatedfrom protective antigen and histamine-sensitizing factor by starch-block electro-phoresis. Most of the agglutinogen migrated towards the cathode in starch-blockelectrophoresis, although some remained near the origin. Fractions containing mostof the agglutinogen I were free of detectable mouse-protecting or histamine-sensitiz-ing activities. Agglutinogen 1 from a serotype 1,3 B. pertussis strain (J20) migratedsimilarly to the agglutinogen I from strain 353/Z. All agglutinogen 3 activity was

found at the point of application in the starch block. No clear relationship was foundbetween agglutinogen I and mouse-protecting antigen or histamine-sensitizingfactor.

Eight different agglutinogens have been de-scribed for Bordetella pertussis (1, 5, 6). Sixof these (1, 2, 3, 4, 5, 6) are species specific,and two (7, 13) are shared with other speciesof the genus Bordetella. Agglutinogen 1 iscommon to all strains of B. pertussis, whereasagglutinogens 2, 3, 4, 5, and 6 are found invarious combinations as strain-specific antigens.Agglutinogens 2, 3, and 5 resist 100 C for 2.5 hr,but 1, 6, and 7 are destroyed by this treatment(4).The relationship of these agglutinogens to

immunity induced by pertussis vaccines is notclear. Work with purified fractions of B. per-tussis has indicated that at least some of theagglutinogens are separated from the antigenthat is responsible for immunity in the mouseprotection test of Kendrick et al. (10). More-over, no difference in mouse-protective activityhas been demonstrated among vaccines pre-pared from cultures possessing different com-binations of agglutinogens (2, 7, 8, 24). Theseobservations indicate that the specific agglutino-gens 2, 3, 4, 5, and 6 do not play a significantrole in active immunization of mice. This can-not be said with certainty of agglutinogen 1,because this antigen is present in all smoothstrains of B. pertussis (5), and definite separationof this agglutinogen and mouse-protective an-tigen (PA) has not been reported.

Recent epidemiological findings (3, 17) andresults of passive mouse protection tests (18,

' Present address: Associate Professor, Veterinary MedicalResearch Institute, Iowa State University, Ames, Iowa 50010.

19) have suggested to the authors that type-specific immunity to B. pertussis may exist.These studies have emphasized the need for moreprecise definition of the role of the agglutinogensin immunity to B. pertussis.We previously showed that the histamine-

sensitizing factor (HSF) and PA can be foundin the same fractions obtained from starch-block electrophoresis (15), and that the physicaland chemical properties of PA and HSF aresimilar if not identical. However, PA or HSFhas not been obtained in a pure form. Because ofthe uncertainty regarding the role of agglutino-gen I in protection and the unavailability of apure preparation of PA, it was important to de-termine whether these antigens could be sepa-rated.

MATERIALS AND METHODSSerotype I B. pertussis strain 353/Z and serotype

1, 3 strain J20 were obtained from N. Preston, Uni-versity of Manchester, Manchester, England (17).Strain 04965, another serotype I strain, was ob-tained from Parke Davis, Detroit, Mich., throughthe courtesy of Henry Devlin. B. pertussis strain2753, serotypes 1, 2, 3, 4, and 5, which was used tochallenge mice, was obtained from Merck & Co.,West Point, Pa. It was originally obtained from P.Kendrick as strain 18-323. B. bronchiseptica strain22067 (serotype 7, 8, 9, 12, 13) was obtained fromG. Eldering, Michigan Department of Public Health,Grand Rapids, Mich.

B. pertussis cells were grown in and harvested fromliquid medium (23) as previously described (15).Dried whole cells for agglutinin absorption were

243

on January 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

ROSS AND MUNOZ

prepared by dialyzing freshly grown cells againstcold deionized water and lyophilizing them.

Antigens used for immunization of rabbits and foragglutination tests were saline suspensions of cellsgrown on Bordet-Gengou agar with 15% horse bloodand preserved in 1:10,000 Merthiolate. In some cases,0.3% Formalin was used instead of Merthiolatefor cell suspensions used in agglutination tests.The plate agglutination test was performed by

mixing one drop of antigen (40 billion cells per ml)with one drop of each serum dilution; the plate wasrocked several times, incubated for 5 min in a par-tially closed box to prevent evaporation, rockedagain several times, and read for agglutination over afluorescent light-illuminated viewer.

Antiserum specific for agglutinogen I was pro-duced by immunizing rabbits with a series of intra-venous (iv) and subcutaneous (sc) injections ofMerthiolate-killed B. pertussis 353/Z cells. The anti-serum was made monospecific by absorption with 40mg of dried B. bronchiseptica 22067 cells and 5 mgof autoclaved B. pertussis J20 cells per ml of serum.A sample of agglutinogen I serum was also suppliedby G. Eldering.

Alkaline saline extract (SE) from B. pertussisstrains 353/Z, 04965, and J20 and from B. bronchi-septica strain 22067 were prepared by a procedurepreviously described (15). The SE from strain 353/Zhad a histamine-sensitizing dose (SD50) of <5 &gper mouse and a protecting dose (PD50) of 5 &g permouse.

Acetone-dried cells, SE, and MgSO4-precipitatedHSF were prepared as previously described (16a).The sediments after extraction of SE and after pre-cipitation of HSF by MgSO6 were dialyzed againstdistilled water and lyophilized.

Endotoxin was prepared by the trichloroaceticacid extraction method of Boivin as described byKabat (9).Mouse protection tests were performed as pre-

viously described (16), by using intracranial (ic)challenge with B. pertussis 2753. Histamine sensi-tization tests were performed by giving the sensitiz-ing materials iv and challenging the mice intraperi-toneally (ip) 24 hr later with 0.5 mg of histaminebase given as histamine diphosphate (14).

Agglutinin absorption tests to determine amountsof agglutinogen present in various fractions wereperformed as follows. Saline dilutions of antigenwere mixed with an equal volume of antiserum di-luted to a concentration twice its plate agglutinationtiter, and the mixture was incubated for 15 min atroom temperature, mixed again, and allowed to in-cubate for an additional 15 min. The mixture wascentrifuged, and the absorbed serum was tested foragglutinins as described above. Nonspecific absorp-tion of antibodies was controlled by absorbing thesera with SE prepared from B. bronchiseptica strain22067, and specific absorption was controlled byabsorbing with SE from strain 353/Z. Consistently,0.25 to 0.5 mg of SE from strain 353/Z absorbedantibodies from 1 ml of diluted serum.

Starch-block electrophoresis was performed asdescribed previously (15). The starch block (50 by

9.5 by 1.2 cm) was made in phosphate buffer (pH6.2; A = 0.02). SE (100 mg) was applied at the centerof the block. Electrophoresis, performed at 6 to 10ma and 200 v, was carried out at 4 C for 40 hr. Theblock was cut into 25, 2-cm strips, each strip wasthen eluted with the phosphate buffer, and the eluateswere dialyzed against water and lyophilized.

RESULTS

SE from B. pertussis 353/Z was fractionatedby starch-block electrophoresis, and the frac-tions were subsequently tested by agar diffusion.This SE exhibited an antigenic complexity simi-lar to that of SE prepared from another strainof B. pertussis (15). Sixteen different bands de-veloped with the anti-B. pertussis serum (Fig. 1).Only four bands developed when factor I anti-serum was employed (Fig. 2). One band waslocated at fractions 7 and 8, a second in fraction10, and a third in fractions 9 to 15 which corre-sponded closely to the distribution of agglutino-gen 1. A fourth was a weak band close to thesample application in fractions 14 and 15 (thismay be due to protective antigen).The relationships of optical density, agglutino-

gen I activity, mouse-protecting activity, andhistamine-sensitizing activity of the variousstarch-block electrophoresis fractions are shownin Fig. 3.The HSF and PA activities remained near the

point of application in fractions 13 and 14.Some PA was also found in fraction 15. Most ofthe factor I agglutinin-absorbing activity mi-grated toward the cathode and was found infractions 8 to 11. Also, some of this activity wasusually found close to the origin (fractions 13to 16; Fig. 3). This distribution of agglutinogen1, HSF, and PA occurred in many different elec-trophoretic runs; only once was all of the agglu-tinogen activity found in the region of fractions8 to 11. Fractions 9 to 15 produced a line of pre-cipitate which coincided with agglutinogen Iactivity. Fraction 10 had, in addition, anotherunidentified band (Fig. 2).

Starch-block electrophoresis was also donewith SE from B. pertussis strains 04965 (sero-type 1) and J20 (serotypes 1, 3). SE from strain04965 gave similar results to those obtained withstrain 353/Z SE. With J20, it was found thatagglutinogen 3 did not move from the point ofapplication, whereas agglutinogen I had a dis-tribution similar to that found with SE fromstrains 353/Z and 04965.

Agglutinogen I activity of various prepara-tions of B. pertussis 353/Z was compared withthe PA and HSF of each preparation. Levels of

244 INFEC. IMMUN.

on January 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

ANTIGENS OF B. PERTUSSIS

0 ® AB~~~~~_~~~~~~~~~~ B

FiG. 1. Immunodiffusion test with eluates from starch-block electrophoresis fractions of SE from B. pertussis353/Z. The numbers refer to the various fractions tested. The fractions were labeled starting at the most cathodalportion of the block. The wells labeled AR contained rabbit anti-B. pertussis serum. The well labeled Endo con-tained endotoxin from B. pertussis 3S3/Z. At least 16 bands can be distinguished.

PA and HSF did not correlate with content ofagglutinogen I (Table 1). For example, 0.125mg of SE absorbed all of the agglutinins fromantiagglutinogen I serum and had a PD50 of14.7 pg/mouse, whereas 0.5 mg of MgSO4-precipitated HSF was required to absorb thesame amount of agglutinin, yet it had a PD50virtually the same as the parent SE. The lackof correlation of agglutinogen and HSF activityis even more striking. Endotoxin from 353/Zdid not contain agglutinogen, HSF, or PA.

DISCUSSIONSince the development of the mouse protection

test by Kendrick and co-workers (10) for theevaluation of pertussis vaccines, efforts have beenmade to determine which component of the B.pertussis cell is responsible for the induction ofimmunity in children. Heat-labile toxin (22),

agglutinogens (17), and hemagglutinin (11)were considered important in induction of spe-cific immunity to whooping cough. Studies pub-lished by the British Medical Research Council(12) showed a direct correlation between PAactivity of pertussis vaccines and their effective-ness in production of immunity to whoopingcough. Since this report, it has been assumedthat the antigen measured by the mouse pro-tection test is identical to that responsible forimmunization of children. With this mouseprotection test, heat-labile toxin, heat-stabletoxin, some "agglutinogen" preparations, andhemagglutinin have failed to produce activeimmunity (13, 20).The substance responsible for mouse protec-

tion has been suspected to be a lipoprotein,the activity of which is destroyed by heating at80 C for 30 min; it is polydispersed and can be

VOL. 3, 1971 245

on January 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

ROSS AND MUNOZ

Q2Fu--_1 t3

Fl FL Fl FL

13 4 !5~~~~ 16

I i (8@ 19C : ao

F (FI Fl ( Fl

20 21 22 23 24 ).v

FIG. 2. Immutnodiffusion test of eluates from starch-block electrophoresis fractions of SE from B. pertussis353/Z developed with specific agglutinogen I anitiserum (Fl). The fractions were labeled startinzg at the mostcathodal portioni of the block. The well labeled Endo contained endotoxin from B. pertussis 353/Z.

isolated in a soluble form by various means (13,14). This substance is different from some agglu-tinogens, heat-labile toxin, hemagglutinin, andendotoxin (13). The results presented here showthat mouse-protective activity of purified frac-tions is not due to agglutinogen 1. Most agglu-tinogen I can be separated by starch-block elec-trophoresis from HSF and protective antigen.This isolated agglutinogen was not serologicallyrelated to PA, HSF, or endotoxin from B. per-tussis and did not protect mice or sensitize themto histamine. Preston has indicated that agglu-tinogens 1, 2, and 3 are important in protectingchildren against whooping cough (17-19). Hisconclusions were based on results obtained withpassive mouse protection tests. Antisera specificfor these factors passively protected mice betteragainst B. pertussis strains containing homolo-gous agglutinogens than against other strains.Indirect evidence also suggested to Preston that

vaccines employed in England suppressed in-fections with strains homologous to the vaccinebut not those due to other serological types (17).However, Eldering et al. (7, 8) were not able todetect a significant difference in active or passiveimmunization studies with strains of differentserotypes, and we, in the present report, couldnot show mouse protection with agglutinogen 1.However, these observations do not explain therole of these agglutinogen factors in the prophy-laxis of whooping cough. Before this question isanswered, it will be necessary to determine therole of the various purified antigens in protectionagainst whooping cough. Intracerebral challengewith B. pertussis is not a close approximation ofinfection in man. It is known that, in the mouse,depending on the route of challenge (ic or intra-nasal), two distinct antigens, and perhaps alsomechanisms, are involved in protection against

246

I!

INEC. IMMUN.-

on January 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

VOL. 3,1971 ANTIGENS OF B. PERTUSSIS 247

.9

.8

.7 | ._ D. AT 280

s | ;°0... AGGLUTINOGEN UNITS

00

H:S~f:.A- - -

OD~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-04~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~

_

2n34 64 Wo1 21 4t 61 8 02 22 422-~~~~~~~~~~~~~~~~~~~~~~~~2i-1 -O ....O... ...O...... ......O..O...-O c

HSF:--+4+

- -Cr--1PA: + + - ----- E

0

2 3 4 5 6 7 8 9 10 12 13 14 15 16 17 18 19 20 21 22 23 24 25

FRACTION NUMBERFiG. 3. Relationship of optical density, agglutinogen 1, HSF, and PA of eluates from starch-block electro-

phoresis fractions from SE ofB. pertussis 353/Z. The sample ofSE was applied at fraction 13. (Cathode on left;anode on right.) The slight difference in distribution notedfor HSF and PA may be due only to the different dosesemployed for the protection test and the histamine sensitization test.

TABLE 1. Agglutinogen content, compared to protective and histamine-sensitizing potency of variouspreparations from 353/Z Bordetella pertussis

Amt of material used to absorb 1 ml ofdiluted antiserum"

Prepn oPDsb SD,oc

0.125 0.25 0.5 1 2 4

Whole cells.+ + _ _ _ 6.26 10.2Acetone-dried cells.+ + - - _ 10.5 2.76Alkaline saline extract.. + _ _ - 14.7 5Sediment from SE prep.+ + + - _ _ 32.9 16.5MgSO4 ppt from SE.+ + + - _ _ 15.69 1.0Supernatant after removal ofHSF by MgSO4.+ + + + + + >125 >25

Trichloroacetic acid extract (en-dotoxin).+ + + + + + >125 >25

a The specific antiagglutinogen 1 serum had a titer of 1:80. Values are expressed as milligrams permilliliter. Symbols: +, agglutination (insufficient absorption); -, no agglutination (complete ab-sorption) .

b PD5, = 50% protective dose. Values expressed as micrograms per mouse.c SD,o = 50% histamine sensitizing dose. Values expressed as micrograms per mouse.

infection by B. pertussis (21). Thus, it is pre- gens, endotoxin, heat-labile toxin, and hemag-mature to state that agglutinogens do or do not glutinins are not involved in the active immuniza-play a role in immunity to whooping cough. It tion of mice challenged ic with virulent B. per-seems well established, however, that agglutino- tussis.

on January 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

ROSS AND MUNOZ

ACKNOWLEDGEMENTS

We thank B. M. Hestekin and J. F. Ayers for their able tech-nical help. R. F. R. was a postdoctoral fellow supported by PublicHealth Service grant IF2-AI-29031-01 from the National Instituteof Allergy and Infectious Diseases.

LITERATURE CITED

1. Andersen, E. K. 1953. Serological studies on Haemophiluspertussis, Haemophilus parapertussis and Haemophilusbronchiseptica. Acta Pathol. Microbiol. Scand. 33:202-224.

2. Andersen, E. K., and M. W. Bentzon. 1958. The failure toshow correlation between type-specificity and protectionin experimental pertussis in mice. Acta Pathol. Microbiol.Scand. 43:106-112.

3. Chalvardjian, N. 1965. The content of antigens 1, 2 and 3 instrains of Bordetella pertussis and in vaccines. Can. Med.Ass. J. 92:1114-1116.

4. Eldering, G. 1962. Attempts to separate the protective antigenand agglutinogen of Bordetella pertussis. Round TableConference on Pertussis Immunization, Prague 1:81-96.

5. Eldering, G., C. Hornbeck, and J. Baker. 1957. Serologicalstudy of Bordetella pertussis and related species. J. Bac-teriol. 74:133-136.

6. Eldering, G., W. C. Eveland, and P. C. Kendrick. 1962.Fluorescent antibody staining and agglutination reactionsin Bordetella pertussis cultures. J. Bacteriol. 83:745-749.

7. Eldering, G., J. Holwerda, and J. Baker. 1966. Bordetellapertussis culture having only species factor 1. J. Bacteriol.91:1759-1762.

8. Eldering, G., J. Holwerda, and J. Baker. 1967. Mouse-protec-tive properties of Bordetella pertussis serotypes in passivetests. J. Bacteriol. 93:1758-1761.

9. Kabat, E. A. 1961. Kabat and Mayer's experimental im-munochemistry. Charles C Thomas, Springfield, Illinois.

10. Kendrick, P. L., G. Eldering, M. K. Dixon, and J. Misner.1947. Mouse protection tests in the study of pertussisvaccine: a comparative series using the intracerebral routefor challenge. Amer. J. Public Health 37:803-810.

11. Keogh, E. V., and E. A. North. 1948. The haemagglutinin ofHemophilus pertussis. L. Haemagglutinin as a protectiveantigen in experimental murine pertussis. Aust. J. Exp.Biol. Med. Sci. 26:315-322.

12. Medical Research Council. 1959. Vaccination against whoop-ing cough. Brit. Med. J. 1:994-1000.

13. Munoz, J. 1963. Symposium on relationship of structure ofmicroorganisms to their immunological properties. L.

Immunological and other biological activities of Bordetellapertussis antigens. Bacteriol. Rev. 27:325-340.

14. Munoz, J., and R. K. Bergman. 1966. Some histamine sen-

sitizing properties of soluble preparations of the histaminesensitizing factor (HSF) from Bordetella pertussis. J.Immunol. 97:120-125.

15. Munoz, J., and B. M. Hestekin. 1963. Antigens of Bordetellapertussis. III. The protective antigen. Proc. Soc. Exp.Biol. Med. 112:799-805.

16. Munoz, J., E. Ribi, and C. L. Larson. 1959. Antigens ofBordetella pertussis. I. Activities of cell walls and proto-plasm. J. Immunol. 83:496-501.

16a. Munoz, J., R. F. Smith, and R. L. Cole. 1970. Some physicalproperties of histamine sensitizing factor. Symp. Ser.Immunobiol. Stand. 13:265-270.

17. Preston, N. W. 1963. Type-specific immunity against whoop-ing cough. Brit. Med. J. 2:724-726.

18. Preston, N. W. 1966. Potency tests for pertussis vaccines:doubtful value of intracerebral challenge test in mice. J.Pathol. Bacteriol. 91:173-179.

19. Preston, N. W., and P. Evans. 1963. Type-specific immunityagainst intracerebral pertussis infection in mice. Nature(London) 197:508-509.

20. Schuchardt, L. F., J. Munoz, W. F. Verwey, and J. F. Sagin.1963. The relationship of agglutinogen to other antigens ofBordetella pertussis. J. Immunol. 91:107-111

21. Standfast, A. F. B. 1958. The comparison between field trialsand mouse protection tests against intranasal and intra-cerebral challenges with Bordetella pertussis. Immunology1:135-143.

22. Strean, L. P., and G. A. Grant. 1940. The preparation andproperties of Hemophilus pertussis endotoxin. Can. Med.Ass. J. 43:528-531.

23. Verwey, W. F., E. H. Thiele, D. N. Sage, and L. F. Schuchardt.1949. A simplified liquid culture medium for the growth ofHemophilus pertussis. J. Bacteriol. 58:127-134.

24. Vigil Lagarde, C. 1961. Nuevas cepas de Bordetella pertussisaisladas en Mexico. Rev. Mex. Lab. Clin. 13:79-83.

248 INFEC. IMMUN.

on January 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from