Vol. 16, No. 2 Printed in U.S.A.
INFECTION AND IMMUNITY, May 1977, p. 471-475 Copyright © 1977 American Society for Microbiology
Lipopolysaccharide Serotyping of Neisseria meningitidis by Hemagglutination Inhibition R. E. MANDRELL* AND W. D. ZOLLINGER Department ofBacterial Diseases, Walter Reed Army Institute of Research, Washington, D.C. 20012
Received for publication 18 January 1977
The method of hemagglutination inhibition was used to investigate the antigenic diversity of lipopolysaccharide (LPS) from Neisseria meningitidis and to develop a serotyping system based on this antigen. The system uses outer membrane complex prepared by a simple extraction procedure to inhibit homologous hemagglutination reactions involving sheep erythrocytes sensitized with purified LPS and rabbit antiserum raised to whole meningococci. Antisera with specificity for eight different LPS determinants were used as typing sera to serotype a cross section of 67 meningococcal strains. Only two strains (both group A) were not typable with the eight sera, and most strains had more than one type. Comparison of LPS type and bactericidal serotype suggests that the LPS and protein serotypes are independent serological markers.
Strains of Neisseria meningitidis are presently classified into serogroups on the basis of their capsular polysaccharide and into serotypes on the basis of bactericidal reactions (2, 5) or precipitin reactions done in capillary tubes or agar gel (3). The serotype antigen in these systems is generally believed to be the major outer membrane protein, although this has only been directly shown for serotype 2(11) (4; F. A. Wyle and D. L. Kasper, Bacteriol. Proc., M210, p. 99, 1971). The lipopolysaccharide (LPS) is an important component of the outer membrane of the meningococcus (9) and has been found to be immunologically active in rabbits and humans (10), but the antigenic diversity of this component and its involvement in meningococcal serotyping has not been systematically studied. Zollinger et al. demonstrated three antigenically different types of meningococcal LPS by indirect hemagglutination (HA) (10), and Davis and Arnold were able to detect differences in the chemical composition and biological activity of LPS from different meningococcal strains (1). The present studies were undertaken to examine by serological methods the antigenic diversity of meningococcal LPS and to test the feasibility of serotyping meningococci on the basis of this antigen. Rabbit sera specific for eight different meningococcal LPS determinants were identified and used in an HA inhibition system to serotype a cross section of menin-
gococcal strains. MATERIALS AND METHODS Bacterial strains. The following strains of N. meningitidis were kindly provided by Carl Frasch: 471
M1080, M981, B16B6, M992, M990, M986, M978, M982, M1011, M136, and S3032. All other strains were from the culture collection of the Department of Bacterial Diseases, Walter Reed Army Institute of Research. Cultures were preserved in the lyophilized state. Media and growth conditions were as previously described (F. A. Wyle and D. L. Kasper, Bacteriol. Proc., M210, p. 99, 1971). Rabbit antisera. Antisera against whole viable meningococci were made in 1.8- to 2.3-kg (4- to 5pound) New Zealand white rabbits as previously described (10). Purification of antigens. Outer membrane complex, outer membrane protein, LPS, and capsular polysaccharide were prepared as previously described (10). The prototype LPS antigens were also prepared by the hot phenol-water method as described by Westphal et al. (8). Indirect HA assays. HA assays were performed as described by Zollinger et al. (10) using sheep erythrocytes sensitized with alkaline-treated LPS at the optimal concentration (usually about 50 ug/ml). Optimal sensitization of the sheep erythrocytes was important in obtaining reproducible titers with the typing sera and consistent results in the serotyping. HA inhibition. The rabbit antisera were diluted to contain 4 units of anti-LPS antibody per ml (1 unit of antibody per ml is the lowest concentration of antibody that produces maximal agglutination). The diluted antiserum was mixed with an equal volume (0.025 ml) of inhibitor (outer membrane complex or LPS) at two standard concentrations: 250 Ag and 31 Ag of LPS per ml. The concentration of purified LPS was based on dry weight, whereas the concentration of LPS in the outer membrane complex was estimated from the protein concentration. The protein concentration was assumed to be twice the LPS concentration in the outer membrane complex preparations (Wyle and Kasper, Bacteriol. Proc., p. 99, 1971). After incubation at 37°C for 1 h,
472
INFECT. IMMUN.
MANDRELL AND ZOLLINGER
0.05 ml of 0.5% sheep erythrocytes sensitized with the homologous LPS was added. The plate was agitated to mix the reagents and left at room temperature for 2 h. The agglutination pattern of the wells was scored on a 0 to +4 scale. The wells containing inhibitor were compared to a negative control well containing homologous antigen as inhibitor and showing 0 or +1 agglutination at both antigen concentrations. An LPS antigen was assigned a given serotype if it inhibited (0 to +2 agglutination) the corresponding HA reaction at both the higher and lower concentrations. LPS antigens that inhibited at the higher concentration, but not at the lower concentration, are presented with that serotype in parentheses (Tables 3-5). Consistency in reading the HA patterns was important for reproducible serotyping. Radioactive bactericidal assay. The radioactive bactericidal assay was performed as described by Kasper and Wyle (7). Absorption of sera with LPS-sheep erythrocytes. A 1:10 dilution of antiserum was absorbed at 37°C for 1 h with 0.1 volume of packed sheep erythrocytes sensitized with LPS. RESULTS LPS typing sera. Antisera to the bactericidal
serotyping strains of Gold and Wyle (5) and Frasch and Chapman (2), as well as several additional laboratory strains, were screened by indirect HA for titer and specificity of anti-LPS antibodies. Antisera having an HA titer 51:64 and having at least a three-tube (eightfold) higher titer against the homologous LPS than
against LPS from most heterologous strains further tested as possible typing sera. The spectrum of specificity observed is illustrated in Table 1. Four rabbit antisera were tested against the homologous LPS as well as LPS from a series of heterologous strains. Antisera against strains M992 and M981 were very specific, exhibiting little HA activity with most heterologous LPS antigens, whereas anti-M136 was very cross-reactive. Although anti-M978 had marginal specificity, it was found to be useful as a typing serum. During the course of these studies, antisera from rabbits immunized with the same strain were compared for titer and specificity. Since differences in both the titer and degree of specificity were observed, several antisera were usually made and screened for the best titer and specificity. Based on this kind of HA data and further testing in grid HA inhibition experiments, eight rabbit antisera with distinctly different specificities toward meningococcal LPS were chosen as prototype antisera. The results of a grid inhibition experiment with the eight typing sera are given in Table 2. The reaction of each antiserum, at the appropriate dilution, with the homologous LPS was tested with the homologous antigen and the seven heterologous antigens. The eight homologous reactions define the LPS serotypes, which are numbered in the order they were discovered. Four of the were
TABLE 1. Indirect HA titers of four rabbit antisera Antiserum vs. strain
Erythrocytes sensitized with LPS from strain: M992
M981
M978
M136
99M
6155
89I
4 4 4 M992 lla 4 3 4 4 4 M981 9 3 2 5 3 M978 6 6 10 6 8 7 5 8 6 M136 7 8 8 8 8 a Expressed as log2 of the reciprocal of the highest dilution producing agglutination.
35E
126E
2 6 6 8
2 2 9 5
6155(B)
M978(B)
TABLE 2. Grid inhibition of the eight LPS serotype reactions Inhibiting antigena
Antiserum vs. strain
LPS serotype
126E(C)b
1
126E(C)5
35E(C)
6275(B)
+ +c
_ _
_ _
89R(C)
M981(B)
M992(B)
2 -++-+ 35E(C) _ _ _ _ __ ++ ++ 3 ---- - 6275(B) _+ 4 ----89I(C) + _d _ ++ ----5 --M981(B) --6 -++ M992(B) _ _ _ _ _ ++ ---7 ---6155(B) + + ------8 M978(B) a Outer membrane complex or purified LPS. b Letter in parentheses indicates serogroup. Inhibition of homologous HA reaction by antigen at 250 ,ug and 31 ,ug of LPS per ml. d Inhibition of homologous HA reaction by antigen at 250 ,g but not 31 ,ug of LPS per ml.
++ +
__ +++
VOL. 16, 1977
LPS SEROTYPING OF MENINGOCOCCI
473
TABLE 3. LPS serotype of bactericidal serotyping antigens only inhibited the homologous serum, strains three inhibited two sera, and one inhibited four of the sera. In addition to these eight specific Bactericidal serotypes, the presence of relatively common serotype LPS type Strain Serogroup determinants could also be demonstrated by System Type inhibition of heterologous reactions involving I C Gold 60E 3, 7, (4) sera with low specificity. One such reaction II C 3 1381 Gold (anti-M136 versus B16B6 LPS) was inhibited by C III Gold 1, 8 126E six of the eight prototype antigens and by about C 118V Gold IV, III 4, (2) 70% of all strains tested. 2 C Gold V 35E To rule out any inhibition by contaminating C Gold 32I VI, IV, I 2, (4) protein or capsular polysaccharide antigens, C WR VII, IV 4 89I each of the eight prototype reactions was tested B WR X 6155 3, 7 with the homologous capsular polysaccharide, B WR XI, III 1, 8 6557 outer membrane protein, and LPS as inhibiB 6940 WR XII, III 1, (8) B Frasch 1 M1080 3, 7, 8, (4) tors. The polysaccharide could not inhibit at B Frasch 2 2, (3) B16B6 1,000 ,ug/ml. The homologous protein showed B 5 Frasch 4 M981 slight inhibition at concentrations between 100 B 6 Frasch 5 M992 and 1,000 ,ug/ml, which is consistent with a 1 to B Frasch 6 3, 7 M990 2% contamination of protein by LPS. The hoB 7 Frasch 2, 3, (7) M986 mologous LPS inhibited at 1 to 3 ,ug/ml. B Frasch 8, (3, 1) 3, 8, (4), (7) M978 For routine serotyping OMC prepared from B 7 Frasch 9 M982 the strains to be tested was used as the inhibiB Frasch 2, 10 3, (7) Mio11 tor. Results obtained with OMC were identical B 4 M136 Frasch 11 to those obtained with purified LPS, thus elimiB S3032 Frasch 12 3, 7 nating the need to prepare LPS from each a The Gold serotype strains were described by strain to be serotyped. Gold and Wyle (5) and the Frasch strains by Frasch Serotyping. The LPS serotype of a total of 67 and Chapman (2), and the WR strains were identimeningococcal strains, mostly group B but rep- fied at Walter Reed Army Institute of Research and resenting a cross section of serogroups and bac- by D. L. Kasper and J. M. Grifflss. tericidal serotypes, was determined. Included among these strains were the bactericidal sero- TABLE 4. LPS typing of representative strains from typing strains of Gold and Wyle (5) and Frasch serogroups other than B and C and Chapman (2) and several serotyping LPS type Strain Serogroup strains identified by D. L. Kasper (6) and J. M. A 8 2E Griffiss at Walter Reed Army Institute of ReA A693840 6, 7 search. The LPS serotypes of these prototype NTa A 121 strains are listed in Table 3. All 21 strains were NT A 106 typable, and 15 of 21 inhibited more than one of Y 135M 3, 4 the LPS serotype reactions. It may be seen from Y 6236 2, (6) Table 3 that in several cases strains having 3 Y 7062 different bactericidal (protein) serotypes have 29E 29E 2, 7 the same LPS serotype. Thus, although some 135 6308 3, 7 Z 7, (3) Z, Slaterus correlations have been observed, the LPS sero6 X X, Slaterus type appears to be distinctly different from the bactericidal serotype as well as the serogroup. a NT, Not typable. The LPS type of representative strains from serogroups other than B and C is given in Table 4. Most of the strains were typable, suggesting having a variety of origins, as indicated, were the presence of these eight LPS serotype deter- serotyped in the LPS system. Four strains isominants on LPS of strains from all major sero- lated at Ft. Campbell, Ky., at the same time all groups. The existence of nontypable group A had the LPS types 2 and 5, whereas a further strains, however, indicates that additional typ- group of nine strains isolated from a variety of ing sera may be required to efficiently type places during the same time period were found group A strains. to have several different LPS serotypes. Of five The usefulness of LPS serotyping in epidemi- carrier strains typed, four were isolated the ological studies has yet to be explored, but the same day from a single company of recruits at data in Table 5 give some indication of its po- Ft. Ord, Calif. These four were all found to have tential utility. A series of 25 group B strains the type 2 LPS determinant, but the fifth strain
INFECT. IMMUN.
MANDRELL AND ZOLLINGER
474
isolated at a different time and place was type 4. Of seven case strains isolated between 1965 and 1968, all but one were found to have LPS types 3 and 7. These data suggest that the LPS serotyping system has the capability of distinguishing among group B strains in a meaningful way. The relative frequency of occurrence of the eight LPS serotypes among the 67 meningococcal strains typed is given in Table 6. The most frequently occurring types were 3 (51%) and 7 (45%), which often occurred together and may be closely related. Types 2 and 4 were present on 32 and 26% of the strains, respectively, whereas types 1, 5, 6, and 8 were only present on 9 to 12% of the strains. Involvement of LPS in bactericidal serotypTABLE 5. LPS serotype of group B strains with different temporal and geographical origins Strain ceived 7608 22-7-74 7609 22-7-74 7610 22-7-74 Date
re-
Place isolated
Ft. Campbell, Ky. Ft. Campbell, Ky. Ft. Campbell, Ky. 7611 22-7-74 Ft. Campbell, Ky. 7510 16-11-74 Ft. Eustis, Va. Ft. Meade, Md. 7523 5-2-74 7549 13-3-74 Ft. Polk, La. 7566 27-4-74 Ft. Eustis, Va. 7569 3-5-74 Ft. Jackson, S.C. 7576 23-5-74 Norfolk, Va. 7583 14-7-74 Ft. McPherson, Va. 7599 24-7-74 Ft. Sill, Okla. 7614 1-8-74 Ft. Benning, Ga. Ft. Ord, Calif. 279 8-1-73 355 8-1-73 Ft. Ord, Calif. Ft. Ord, Calif. 366 8-1-73 372 8-1-73 Ft. Ord, Calif. 7641 3-9-74 Washington, D.C. 53I 29-1-66 Ft. Dix, N.J. 144I 14-5-67 Ft. Dix, N.J. 511 12-6-66 Ft. Dix, N.J. 65V 43II 48I
2981 a
20-6-65 12-5-65 5-1-66 6-10-68
seroSourcea LPS type ? ?
CSF ? ? CSF
Ft. L. Wood, Mo.
Pt. Knox, Ky. Ft. Dix, N. J. Ft. Dix, N. J.
CSF CSF ? Blood CSF Blood CSF Throat Throat Throat Throat Throat CSF CSF CSF or blood CSF CSF Blood CSF
2, 5 2, 5 2, 5 2, (5) 7 2, 6 2, 5 2, 5 2, 3, 7, (1) 1, 8, (6) 3, 7 1, 8, (2) 3, 7 2, 3 2, (7) 2 2, (3) 4 3, 7 2, 3 (7) 3, (7)
3, 7 3, 7 2,3,5,7 1, 8
CSF, Cerebrospinal fluid.
TABLE 6. Frequency of occurrence of 8 meningococcal LPS serotypes among 67 strains No. of strains showtion
ing positive inhibi-
1 2 3 4 5 6 7 8
8 21 33 17 8 6 29 7
Percent
tion
12 32 51 26 12 9 45 11
TABLE 7. Bactericidal activity of rabbit antiserum for two meningococcal strains Organism Antiserum
126E
6557
(C:flI:1,8)a (B:XI,III:1,8) +b + Anti-126E + Anti-126E absorbed with 126E LPSsensitized SRBC a Serogroup:bactericidal serotype:LPS serotype. b +, Antiserum contains bactericidal activity at 51:5 dilution; -, no bactericidal activity at 1:5 dilution. ing. The possible involvement of antibodies to LPS in some bactericidal serotyping reactions was suggested by the experiment summarized in Table 7. Antiserum to strain 126E (group C; bactericidal type III; LPS type 1, 8) was found to have bactericidal activity toward the homologous strain as well as strain 6557 (group B; bactericidal type XI, III; LPS type 1, 8). After absorption with 126E LPS-coated erythrocytes, the serum retained its capacity to kill the homologous strain but no longer killed strain 6557. This suggested that the bactericidal activity toward strain 6557 was due to anti-LPS antibodies rather than antibodies to the serotype protein.
DISCUSSION Using the HA inhibition system, we have thus far identified eight different serological determinants on meningococcal LPS. Although these determinants appear to be shared to some degree by all eight serogroups tested, and at least one was found on 65 of 67 strains typed, additional LPS serotype determinants undoubtedly exist. It is not yet clear whether these LPS determinants are located in the core region of the LPS or on 0 side chains, since the rough or smooth nature of these meningococcal strains has not been determined. In general, the LPS serotype determinants appear to be independent of the bactericidal (protein) serotype determinants, but the exact relationship between the protein and LPS serotypes is currently under investigation. Although bactericidal serotyping principally involves antibodies to the outer membrane protein antigens (4; Wyle and Kasper, Bacteriol. Proc., p. 99, 1971), we have demonstrated the involvement of anti-LPS antibodies in at least one instance. Since no specific measures are normally taken to remove anti-LPS antibodies, most bactericidal typing sera probably contain some anti-LPS antibodies with bactericidal po-
VOL. 16, 1977
tential. The overall specificity of the bactericidal serotyping system for the protein antigens probably derives from the tendency of rabbit antisera to have higher titers of antibody to the protein antigens than to the LPS antigens (unpublished observations). The demonstration of LPS serotype antigens in addition to the bactericidal or protein serotype antigens suggests the need for antigenspecific serotyping to clearly distinguish between the two antigens. It would seem desirable for each serotype to eventually have as its basis a definable chemical structure. ACKNOWLEDGMENTS We are indebted to Doris P. Fisher for invaluable help in the preparation of this manuscript. LITERATURE CITED 1. Davis, C. E., and K. Arnold. 1974. Role of meningococcal endotoxin in meningococcal purpura. J. Exp. Med. 140:159-171. 2. Frasch, C. E., and S. S. Chapman. 1972. Classification of Neisseria meningitidis group B into distinct serotypes. I. Serological typing by a microbactericidal method. Infect. Immun. 5:98-102.
LPS SEROTYPING OF MENINGOCOCCI
475
3. Frasch, C. E., and S. S. Chapman. 1972. Classification of Neisseria meningitidis group B into distinct serotypes. II. Extraction oftype-specific antigens for serotyping by precipitin techniques. Infect. Immun. 6:127-133. 4. Frasch, C. E., and E. C. Gotschlich. 1974. An outer membrane protein ofNeisseria meningitidis responsible for serotype specificity. J. Exp. Med. 140:87-104. 5. Gold, R., and F. A. Wyle. 1970. New classification of Neisseria meningitidis by means of bactericidal reactions. Infect. Immun. 1:479-484. 6. Kasper, D. L., J. L. Winklehake, B. L. Brandt, and M. S. Artenstein. 1973. Antigenic specificity of bactericidal antibodies in antisera to Neisseria meningitidis. J. Infect. Dis. 127:378-387. 7. Kasper, D. L., and F. A. Wyle. 1972. Bactericidal antibody assay using 14C-labeled Neisseria meningitidis. Proc. Soc. Exp. Biol. Med. 139:1175-1180. 8. Westphal, O., 0. Luderitz, and F. Bister. 1952. Uber die Extraktion von Bakterien mit Phenol/Wasser. Z. Naturforsch. Teil B 7:148-155. 9. Zollinger, W. D., D. L. Kasper, B. J. Veltri, and M. S. Artenstein. 1972. Isolation and characterization of a native cell wall complex from Neisseria meningitidis. Infect. Immun. 6:835-851. 10. Zollinger, W. D., C. L. Pennington, and M. S. Artenstein. 1974. Human antibody response to three meningococcal outer membrane antigens: comparison by specific hemagglutination assays. Infect. Immun. 10:975-982.
INFECTION AND IMMUNITY, May 1977, p. 471-475 Copyright © 1977 American Society for Microbiology
Lipopolysaccharide Serotyping of Neisseria meningitidis by Hemagglutination Inhibition R. E. MANDRELL* AND W. D. ZOLLINGER Department ofBacterial Diseases, Walter Reed Army Institute of Research, Washington, D.C. 20012
Received for publication 18 January 1977
The method of hemagglutination inhibition was used to investigate the antigenic diversity of lipopolysaccharide (LPS) from Neisseria meningitidis and to develop a serotyping system based on this antigen. The system uses outer membrane complex prepared by a simple extraction procedure to inhibit homologous hemagglutination reactions involving sheep erythrocytes sensitized with purified LPS and rabbit antiserum raised to whole meningococci. Antisera with specificity for eight different LPS determinants were used as typing sera to serotype a cross section of 67 meningococcal strains. Only two strains (both group A) were not typable with the eight sera, and most strains had more than one type. Comparison of LPS type and bactericidal serotype suggests that the LPS and protein serotypes are independent serological markers.
Strains of Neisseria meningitidis are presently classified into serogroups on the basis of their capsular polysaccharide and into serotypes on the basis of bactericidal reactions (2, 5) or precipitin reactions done in capillary tubes or agar gel (3). The serotype antigen in these systems is generally believed to be the major outer membrane protein, although this has only been directly shown for serotype 2(11) (4; F. A. Wyle and D. L. Kasper, Bacteriol. Proc., M210, p. 99, 1971). The lipopolysaccharide (LPS) is an important component of the outer membrane of the meningococcus (9) and has been found to be immunologically active in rabbits and humans (10), but the antigenic diversity of this component and its involvement in meningococcal serotyping has not been systematically studied. Zollinger et al. demonstrated three antigenically different types of meningococcal LPS by indirect hemagglutination (HA) (10), and Davis and Arnold were able to detect differences in the chemical composition and biological activity of LPS from different meningococcal strains (1). The present studies were undertaken to examine by serological methods the antigenic diversity of meningococcal LPS and to test the feasibility of serotyping meningococci on the basis of this antigen. Rabbit sera specific for eight different meningococcal LPS determinants were identified and used in an HA inhibition system to serotype a cross section of menin-
gococcal strains. MATERIALS AND METHODS Bacterial strains. The following strains of N. meningitidis were kindly provided by Carl Frasch: 471
M1080, M981, B16B6, M992, M990, M986, M978, M982, M1011, M136, and S3032. All other strains were from the culture collection of the Department of Bacterial Diseases, Walter Reed Army Institute of Research. Cultures were preserved in the lyophilized state. Media and growth conditions were as previously described (F. A. Wyle and D. L. Kasper, Bacteriol. Proc., M210, p. 99, 1971). Rabbit antisera. Antisera against whole viable meningococci were made in 1.8- to 2.3-kg (4- to 5pound) New Zealand white rabbits as previously described (10). Purification of antigens. Outer membrane complex, outer membrane protein, LPS, and capsular polysaccharide were prepared as previously described (10). The prototype LPS antigens were also prepared by the hot phenol-water method as described by Westphal et al. (8). Indirect HA assays. HA assays were performed as described by Zollinger et al. (10) using sheep erythrocytes sensitized with alkaline-treated LPS at the optimal concentration (usually about 50 ug/ml). Optimal sensitization of the sheep erythrocytes was important in obtaining reproducible titers with the typing sera and consistent results in the serotyping. HA inhibition. The rabbit antisera were diluted to contain 4 units of anti-LPS antibody per ml (1 unit of antibody per ml is the lowest concentration of antibody that produces maximal agglutination). The diluted antiserum was mixed with an equal volume (0.025 ml) of inhibitor (outer membrane complex or LPS) at two standard concentrations: 250 Ag and 31 Ag of LPS per ml. The concentration of purified LPS was based on dry weight, whereas the concentration of LPS in the outer membrane complex was estimated from the protein concentration. The protein concentration was assumed to be twice the LPS concentration in the outer membrane complex preparations (Wyle and Kasper, Bacteriol. Proc., p. 99, 1971). After incubation at 37°C for 1 h,
472
INFECT. IMMUN.
MANDRELL AND ZOLLINGER
0.05 ml of 0.5% sheep erythrocytes sensitized with the homologous LPS was added. The plate was agitated to mix the reagents and left at room temperature for 2 h. The agglutination pattern of the wells was scored on a 0 to +4 scale. The wells containing inhibitor were compared to a negative control well containing homologous antigen as inhibitor and showing 0 or +1 agglutination at both antigen concentrations. An LPS antigen was assigned a given serotype if it inhibited (0 to +2 agglutination) the corresponding HA reaction at both the higher and lower concentrations. LPS antigens that inhibited at the higher concentration, but not at the lower concentration, are presented with that serotype in parentheses (Tables 3-5). Consistency in reading the HA patterns was important for reproducible serotyping. Radioactive bactericidal assay. The radioactive bactericidal assay was performed as described by Kasper and Wyle (7). Absorption of sera with LPS-sheep erythrocytes. A 1:10 dilution of antiserum was absorbed at 37°C for 1 h with 0.1 volume of packed sheep erythrocytes sensitized with LPS. RESULTS LPS typing sera. Antisera to the bactericidal
serotyping strains of Gold and Wyle (5) and Frasch and Chapman (2), as well as several additional laboratory strains, were screened by indirect HA for titer and specificity of anti-LPS antibodies. Antisera having an HA titer 51:64 and having at least a three-tube (eightfold) higher titer against the homologous LPS than
against LPS from most heterologous strains further tested as possible typing sera. The spectrum of specificity observed is illustrated in Table 1. Four rabbit antisera were tested against the homologous LPS as well as LPS from a series of heterologous strains. Antisera against strains M992 and M981 were very specific, exhibiting little HA activity with most heterologous LPS antigens, whereas anti-M136 was very cross-reactive. Although anti-M978 had marginal specificity, it was found to be useful as a typing serum. During the course of these studies, antisera from rabbits immunized with the same strain were compared for titer and specificity. Since differences in both the titer and degree of specificity were observed, several antisera were usually made and screened for the best titer and specificity. Based on this kind of HA data and further testing in grid HA inhibition experiments, eight rabbit antisera with distinctly different specificities toward meningococcal LPS were chosen as prototype antisera. The results of a grid inhibition experiment with the eight typing sera are given in Table 2. The reaction of each antiserum, at the appropriate dilution, with the homologous LPS was tested with the homologous antigen and the seven heterologous antigens. The eight homologous reactions define the LPS serotypes, which are numbered in the order they were discovered. Four of the were
TABLE 1. Indirect HA titers of four rabbit antisera Antiserum vs. strain
Erythrocytes sensitized with LPS from strain: M992
M981
M978
M136
99M
6155
89I
4 4 4 M992 lla 4 3 4 4 4 M981 9 3 2 5 3 M978 6 6 10 6 8 7 5 8 6 M136 7 8 8 8 8 a Expressed as log2 of the reciprocal of the highest dilution producing agglutination.
35E
126E
2 6 6 8
2 2 9 5
6155(B)
M978(B)
TABLE 2. Grid inhibition of the eight LPS serotype reactions Inhibiting antigena
Antiserum vs. strain
LPS serotype
126E(C)b
1
126E(C)5
35E(C)
6275(B)
+ +c
_ _
_ _
89R(C)
M981(B)
M992(B)
2 -++-+ 35E(C) _ _ _ _ __ ++ ++ 3 ---- - 6275(B) _+ 4 ----89I(C) + _d _ ++ ----5 --M981(B) --6 -++ M992(B) _ _ _ _ _ ++ ---7 ---6155(B) + + ------8 M978(B) a Outer membrane complex or purified LPS. b Letter in parentheses indicates serogroup. Inhibition of homologous HA reaction by antigen at 250 ,ug and 31 ,ug of LPS per ml. d Inhibition of homologous HA reaction by antigen at 250 ,g but not 31 ,ug of LPS per ml.
++ +
__ +++
VOL. 16, 1977
LPS SEROTYPING OF MENINGOCOCCI
473
TABLE 3. LPS serotype of bactericidal serotyping antigens only inhibited the homologous serum, strains three inhibited two sera, and one inhibited four of the sera. In addition to these eight specific Bactericidal serotypes, the presence of relatively common serotype LPS type Strain Serogroup determinants could also be demonstrated by System Type inhibition of heterologous reactions involving I C Gold 60E 3, 7, (4) sera with low specificity. One such reaction II C 3 1381 Gold (anti-M136 versus B16B6 LPS) was inhibited by C III Gold 1, 8 126E six of the eight prototype antigens and by about C 118V Gold IV, III 4, (2) 70% of all strains tested. 2 C Gold V 35E To rule out any inhibition by contaminating C Gold 32I VI, IV, I 2, (4) protein or capsular polysaccharide antigens, C WR VII, IV 4 89I each of the eight prototype reactions was tested B WR X 6155 3, 7 with the homologous capsular polysaccharide, B WR XI, III 1, 8 6557 outer membrane protein, and LPS as inhibiB 6940 WR XII, III 1, (8) B Frasch 1 M1080 3, 7, 8, (4) tors. The polysaccharide could not inhibit at B Frasch 2 2, (3) B16B6 1,000 ,ug/ml. The homologous protein showed B 5 Frasch 4 M981 slight inhibition at concentrations between 100 B 6 Frasch 5 M992 and 1,000 ,ug/ml, which is consistent with a 1 to B Frasch 6 3, 7 M990 2% contamination of protein by LPS. The hoB 7 Frasch 2, 3, (7) M986 mologous LPS inhibited at 1 to 3 ,ug/ml. B Frasch 8, (3, 1) 3, 8, (4), (7) M978 For routine serotyping OMC prepared from B 7 Frasch 9 M982 the strains to be tested was used as the inhibiB Frasch 2, 10 3, (7) Mio11 tor. Results obtained with OMC were identical B 4 M136 Frasch 11 to those obtained with purified LPS, thus elimiB S3032 Frasch 12 3, 7 nating the need to prepare LPS from each a The Gold serotype strains were described by strain to be serotyped. Gold and Wyle (5) and the Frasch strains by Frasch Serotyping. The LPS serotype of a total of 67 and Chapman (2), and the WR strains were identimeningococcal strains, mostly group B but rep- fied at Walter Reed Army Institute of Research and resenting a cross section of serogroups and bac- by D. L. Kasper and J. M. Grifflss. tericidal serotypes, was determined. Included among these strains were the bactericidal sero- TABLE 4. LPS typing of representative strains from typing strains of Gold and Wyle (5) and Frasch serogroups other than B and C and Chapman (2) and several serotyping LPS type Strain Serogroup strains identified by D. L. Kasper (6) and J. M. A 8 2E Griffiss at Walter Reed Army Institute of ReA A693840 6, 7 search. The LPS serotypes of these prototype NTa A 121 strains are listed in Table 3. All 21 strains were NT A 106 typable, and 15 of 21 inhibited more than one of Y 135M 3, 4 the LPS serotype reactions. It may be seen from Y 6236 2, (6) Table 3 that in several cases strains having 3 Y 7062 different bactericidal (protein) serotypes have 29E 29E 2, 7 the same LPS serotype. Thus, although some 135 6308 3, 7 Z 7, (3) Z, Slaterus correlations have been observed, the LPS sero6 X X, Slaterus type appears to be distinctly different from the bactericidal serotype as well as the serogroup. a NT, Not typable. The LPS type of representative strains from serogroups other than B and C is given in Table 4. Most of the strains were typable, suggesting having a variety of origins, as indicated, were the presence of these eight LPS serotype deter- serotyped in the LPS system. Four strains isominants on LPS of strains from all major sero- lated at Ft. Campbell, Ky., at the same time all groups. The existence of nontypable group A had the LPS types 2 and 5, whereas a further strains, however, indicates that additional typ- group of nine strains isolated from a variety of ing sera may be required to efficiently type places during the same time period were found group A strains. to have several different LPS serotypes. Of five The usefulness of LPS serotyping in epidemi- carrier strains typed, four were isolated the ological studies has yet to be explored, but the same day from a single company of recruits at data in Table 5 give some indication of its po- Ft. Ord, Calif. These four were all found to have tential utility. A series of 25 group B strains the type 2 LPS determinant, but the fifth strain
INFECT. IMMUN.
MANDRELL AND ZOLLINGER
474
isolated at a different time and place was type 4. Of seven case strains isolated between 1965 and 1968, all but one were found to have LPS types 3 and 7. These data suggest that the LPS serotyping system has the capability of distinguishing among group B strains in a meaningful way. The relative frequency of occurrence of the eight LPS serotypes among the 67 meningococcal strains typed is given in Table 6. The most frequently occurring types were 3 (51%) and 7 (45%), which often occurred together and may be closely related. Types 2 and 4 were present on 32 and 26% of the strains, respectively, whereas types 1, 5, 6, and 8 were only present on 9 to 12% of the strains. Involvement of LPS in bactericidal serotypTABLE 5. LPS serotype of group B strains with different temporal and geographical origins Strain ceived 7608 22-7-74 7609 22-7-74 7610 22-7-74 Date
re-
Place isolated
Ft. Campbell, Ky. Ft. Campbell, Ky. Ft. Campbell, Ky. 7611 22-7-74 Ft. Campbell, Ky. 7510 16-11-74 Ft. Eustis, Va. Ft. Meade, Md. 7523 5-2-74 7549 13-3-74 Ft. Polk, La. 7566 27-4-74 Ft. Eustis, Va. 7569 3-5-74 Ft. Jackson, S.C. 7576 23-5-74 Norfolk, Va. 7583 14-7-74 Ft. McPherson, Va. 7599 24-7-74 Ft. Sill, Okla. 7614 1-8-74 Ft. Benning, Ga. Ft. Ord, Calif. 279 8-1-73 355 8-1-73 Ft. Ord, Calif. Ft. Ord, Calif. 366 8-1-73 372 8-1-73 Ft. Ord, Calif. 7641 3-9-74 Washington, D.C. 53I 29-1-66 Ft. Dix, N.J. 144I 14-5-67 Ft. Dix, N.J. 511 12-6-66 Ft. Dix, N.J. 65V 43II 48I
2981 a
20-6-65 12-5-65 5-1-66 6-10-68
seroSourcea LPS type ? ?
CSF ? ? CSF
Ft. L. Wood, Mo.
Pt. Knox, Ky. Ft. Dix, N. J. Ft. Dix, N. J.
CSF CSF ? Blood CSF Blood CSF Throat Throat Throat Throat Throat CSF CSF CSF or blood CSF CSF Blood CSF
2, 5 2, 5 2, 5 2, (5) 7 2, 6 2, 5 2, 5 2, 3, 7, (1) 1, 8, (6) 3, 7 1, 8, (2) 3, 7 2, 3 2, (7) 2 2, (3) 4 3, 7 2, 3 (7) 3, (7)
3, 7 3, 7 2,3,5,7 1, 8
CSF, Cerebrospinal fluid.
TABLE 6. Frequency of occurrence of 8 meningococcal LPS serotypes among 67 strains No. of strains showtion
ing positive inhibi-
1 2 3 4 5 6 7 8
8 21 33 17 8 6 29 7
Percent
tion
12 32 51 26 12 9 45 11
TABLE 7. Bactericidal activity of rabbit antiserum for two meningococcal strains Organism Antiserum
126E
6557
(C:flI:1,8)a (B:XI,III:1,8) +b + Anti-126E + Anti-126E absorbed with 126E LPSsensitized SRBC a Serogroup:bactericidal serotype:LPS serotype. b +, Antiserum contains bactericidal activity at 51:5 dilution; -, no bactericidal activity at 1:5 dilution. ing. The possible involvement of antibodies to LPS in some bactericidal serotyping reactions was suggested by the experiment summarized in Table 7. Antiserum to strain 126E (group C; bactericidal type III; LPS type 1, 8) was found to have bactericidal activity toward the homologous strain as well as strain 6557 (group B; bactericidal type XI, III; LPS type 1, 8). After absorption with 126E LPS-coated erythrocytes, the serum retained its capacity to kill the homologous strain but no longer killed strain 6557. This suggested that the bactericidal activity toward strain 6557 was due to anti-LPS antibodies rather than antibodies to the serotype protein.
DISCUSSION Using the HA inhibition system, we have thus far identified eight different serological determinants on meningococcal LPS. Although these determinants appear to be shared to some degree by all eight serogroups tested, and at least one was found on 65 of 67 strains typed, additional LPS serotype determinants undoubtedly exist. It is not yet clear whether these LPS determinants are located in the core region of the LPS or on 0 side chains, since the rough or smooth nature of these meningococcal strains has not been determined. In general, the LPS serotype determinants appear to be independent of the bactericidal (protein) serotype determinants, but the exact relationship between the protein and LPS serotypes is currently under investigation. Although bactericidal serotyping principally involves antibodies to the outer membrane protein antigens (4; Wyle and Kasper, Bacteriol. Proc., p. 99, 1971), we have demonstrated the involvement of anti-LPS antibodies in at least one instance. Since no specific measures are normally taken to remove anti-LPS antibodies, most bactericidal typing sera probably contain some anti-LPS antibodies with bactericidal po-
VOL. 16, 1977
tential. The overall specificity of the bactericidal serotyping system for the protein antigens probably derives from the tendency of rabbit antisera to have higher titers of antibody to the protein antigens than to the LPS antigens (unpublished observations). The demonstration of LPS serotype antigens in addition to the bactericidal or protein serotype antigens suggests the need for antigenspecific serotyping to clearly distinguish between the two antigens. It would seem desirable for each serotype to eventually have as its basis a definable chemical structure. ACKNOWLEDGMENTS We are indebted to Doris P. Fisher for invaluable help in the preparation of this manuscript. LITERATURE CITED 1. Davis, C. E., and K. Arnold. 1974. Role of meningococcal endotoxin in meningococcal purpura. J. Exp. Med. 140:159-171. 2. Frasch, C. E., and S. S. Chapman. 1972. Classification of Neisseria meningitidis group B into distinct serotypes. I. Serological typing by a microbactericidal method. Infect. Immun. 5:98-102.
LPS SEROTYPING OF MENINGOCOCCI
475
3. Frasch, C. E., and S. S. Chapman. 1972. Classification of Neisseria meningitidis group B into distinct serotypes. II. Extraction oftype-specific antigens for serotyping by precipitin techniques. Infect. Immun. 6:127-133. 4. Frasch, C. E., and E. C. Gotschlich. 1974. An outer membrane protein ofNeisseria meningitidis responsible for serotype specificity. J. Exp. Med. 140:87-104. 5. Gold, R., and F. A. Wyle. 1970. New classification of Neisseria meningitidis by means of bactericidal reactions. Infect. Immun. 1:479-484. 6. Kasper, D. L., J. L. Winklehake, B. L. Brandt, and M. S. Artenstein. 1973. Antigenic specificity of bactericidal antibodies in antisera to Neisseria meningitidis. J. Infect. Dis. 127:378-387. 7. Kasper, D. L., and F. A. Wyle. 1972. Bactericidal antibody assay using 14C-labeled Neisseria meningitidis. Proc. Soc. Exp. Biol. Med. 139:1175-1180. 8. Westphal, O., 0. Luderitz, and F. Bister. 1952. Uber die Extraktion von Bakterien mit Phenol/Wasser. Z. Naturforsch. Teil B 7:148-155. 9. Zollinger, W. D., D. L. Kasper, B. J. Veltri, and M. S. Artenstein. 1972. Isolation and characterization of a native cell wall complex from Neisseria meningitidis. Infect. Immun. 6:835-851. 10. Zollinger, W. D., C. L. Pennington, and M. S. Artenstein. 1974. Human antibody response to three meningococcal outer membrane antigens: comparison by specific hemagglutination assays. Infect. Immun. 10:975-982.
