CHEMICAL, IMMUNOCHEMICAL, AND STRUCTURAL
STUDIES OF CROSS-REACTIVE CELL WALL
ANTIGENS OF Streptococcus mutans
THOMAS ALLEN BROWN
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
To my wife, Mary
The author wishes to acknowledge the faculty, students, and staff
of the Department of Microbiology and Cell Science for their friendship,
advice, and encouragement during the preparation of this dissertation.
He wishes to express his deep appreciation to the chairman of
his supervisory committee, Dr. Arnold S. Bleiweis, for his valuable
guidance, criticisms and friendship throughout his graduate career.
He would also like to thank the other members of his committee,
Drs. Edward M. Hoffmann and L. William Clem for their advice and help
in direction of this work and preparation of this manuscript.
He would also like to thank Dr. R. W. King, Department of
Chemistry, for his generous assistance with mass spectrometry, and
Richard M. Vaught, Michael Taylor, and Steven F. Hurst for their
Finally, the author would like to express his most loving
gratitude to his wife, Mary, whose love, support, and encouragement
made this dissertation possible.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS. . ... iii
LIST OF TABLES .. .. ... ... .. v
LIST OF FIGURES ........ ............. vi
ABSTRACT................. .. .... viii
INTRODUCTION. . . 1
MATERIALS AND METHODS . .. 11
RESULTS . . . 24
DISCUSSION ...... ................... 79
LITERATURE CITED ................... .. 87
BIOGRAPHICAL SKETCH ................... 92
LIST OF TABLES
1. SURVEY OF THE VARIOUS CLASSIFICATION SCHEMES
FOR Streptococcus mutans . 5
2. CHEMICAL COMPOSITION OF PURIFIED CELL WALLS OF
S. mutans AHT (a) and B13 (d). . 25
3. CHEMICAL COMPOSITION OF FORMAMIDE EXTRACTS OF
S. Mutans AHT (a) and B13 (d) PURIFIED CELL WALLS. 35
4. COMPOSITION OF LECTIN-PURIFIED ANTIGENS FROM
S. mutans AHT (a) and B13 (d) . ... 56
5. INHIBITION OF THE PRECIPITIN REACTION BETWEEN AHT
ANTIGEN AND ANTI-a SERUM ... .. 62
6. INHIBITION OF THE PRECIPITIN REACTION BETWEEN B13
ANTIGEN AND ANTI-d SERUM . .... .64
7. INHIBITION OF THE PRECIPITIN REACTION BETWEEN AHT
ANTIGEN AND ANTI-a-d SERUM . .. 66
8. INHIBITION OF THE PRECIPITIN REACTION BETWEEN B13
ANTIGEN AND ANTI-a-d SERUM . .... .67
9. METHYLATION ANALYSIS OF THE ART ANTIGEN. ... 68
10. METHYLATION ANALYSIS OF THE B13 ANTIGEN. ... 69
LIST OF FIGURES
1. Gel immunodiffusion experiments demonstrating
the reaction of identity between Lancefield
and formamide extracts of AHT and B13. ..... 27
2. Gel immunodiffusion experiments demonstrating
the patterns obtained between formamide extracts,
unabsorbed sera, and absorbed sera .. .. 30
3. Immunoelectrophoretic patterns obtained with formamide
and Lancefield extracts of AHT and B13 ...... 34
4. Gel filtration profile of AHT formamide extract on
Biogel P-150 . ..... ..... 37
5. Gel filtration profile of B13 formamide extract
on Biogel P-150. . . ... 39
6. Affinity chromatography column profile of B13
formamide extract. . ... 42
7. Gel immunodiffusion experiments using crude extracts
and purified antigens. . ... 45
8. Immunoelectrophoretic patterns obtained with crude
extracts and purified antigens . .. 47
9. Gel immunodiffusion experiments using purified antigens,
unabsorbed sera, and absorbed sera .. 49
10. Gel filtration profile of AHT purified antigen on
Biogel P-150 . .... 53
11. Gel filtration profile of B13 purified antigen on
Biogel P-150 . .... 55
12. Gel immunodiffusion patterns obtained using purified
antigens and cross-reacting sera . 59
13. Quantitative precipitin curves obtained using
the various absorbed sera and the purified
antigens . .... . 61
14. Immunoelectrophoretic study of the degradation
of Lancefield extracts of B13 cell walls at various
extraction times. . . ... 73
15. Immunoelectrophoretic study of the degradation of
Lancefield extracts of AHT cell walls at various
extraction times. . . ... 77
16. Schematic arrangement of the sugar residues in the
AHT and B13 polymers . . 84
Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy
CHEMICAL, IMMUNOCHEMICAL, AND STRUCTURAL
STUDIES OF CROSS-REACTIVE CELL WALL
ANTIGENS OF Streptococcus mutans
THOMAS ALLEN BROWN
Chairman: Arnold S. Bleiweis
Major Department: Microbiology and Cell Science
This dissertation is concerned with the isolation, purification,
and immunochemical characterization of two key polysaccharide antigens
of Streptococcus mutans, the principal etiologic agent of dental caries.
These cell-wall antigens were isolated from two strongly cross-reacting
serotypes (a and d) which have been confused by investigators. The in-
formation gathered here may aid in future reclassification or in the
development of more accurate identification procedures for these oral
pathogens. Also, the development of molecular vaccines composed of
important somatic antigens depends on knowledge of antigen composition
The principal cross-reacting wall antigens of S mutans strains
AHT serotypee a) and B13 serotypee d) were extracted from purified cell
walls by hot formamide. The antigens were purified by affinity chro-
matography using the galactose-binding lectin from Ricinus communis.
The purified antigens appeared to be immunologically identical to crude
antigen extracts of whole cells or cell walls. The antigens were
composed primarily of galactose and glucose in molar ratios of 3.4:1 for
AHT, and 2.1:1 for B13. Rhamnose, which is the major constituent of the
crude formamide extracts, was present in only trace amounts in the lectin
purified preparations. Hapten inhibition using absorbed antisera specific
for the unique antigenic determinants (a or d) and the cross-reacting
determinants (a-d) showed that the a specificity depended on an a-linked
galactose residue, possibly linked to glucose. In the case of the d
determinant both glucose and galactose allowed significant inhibition,
however, glucose gave slightly greater inhibition.
The cross-reacting a-d determinant appears to be galactose, possibly
linked 1-6 to glucose. Methylation analysis of the antigens using com-
bined gas-liquid chromatography-mass spectrometry showed that both antigens
possess identical branch point residues as well as three other identical
residues including terminal galactose moieties which may be responsible for
the cross reaction. The AHT antigen contains a 1-3 linked galactose residue
and B13 contains a 1-3 linked glucose residue which may determine the a and d
specificities respectively. Both antigens contain galactofuranose residues
which may contribute to the observed acid-lability of these antigens.
The data presented in this dissertation provide some clues concern-
ing the possible molecular structures of the a and d antigens. Due to the
complexity of each polymer, it is not yet possible to draw complete
structures. There does appear, however, to be a structure common to each
antigen which will be proposed in this dissertation. Finally, our
immunochemical studies of the a, d, and a-d haptenic sites (by precipitin
inhibition analysis) have failed to confirm published data or have
revealed contrary information.
Streptococcus mutans, a Gram-positive coccus belonging to the
viridans group, is considered to be the principal etiological agent in
dental caries. Its ability to colonize tooth surfaces by the production
of a sticky dextran-like capsule and the ability to produce acid from
fermentable carbohydrates account for the cariogenicity of this organism.
Dental caries is the most prevalent disease of civilized populations and
poses serious health and financial problems. Gibbons (29) states "It has
been estimated that well over two billion dollars are spent each year in
the United States for the treatment of this disease, and thus tooth decay
ranks as one of the most expensive bacterial infections of mankind."
Accordingly, a great deal of work has been done on the classification of
S. mutans and toward the development of vaccines for these organisms.
One of the foremost ways of classifying S. mutans strains is serologically.
This, and the potential development of vaccines, necessitates a good
understanding of the antigens associated with the various strains.
This dissertation will present chemical, immunochemical, and structural
data on purified antigens of two cross-reacting strains of S. mutans of
the serotypes a and d.
Classification of S. mutans strains. The various cariogenic and
most other oral streptococci are classified as belonging to the viridans
group, which give greening on blood agar and are not classifiable by the
Lancefield grouping scheme.
Early investigations in 1960 by Fitzgerald and Keyes (27) and
Fitzgerald et al. (26) showed that cariogenic isolates from rats and
hamsters could be distinguished by their ability to ferment sorbitol,
but antigenic extracts failed to react with antisera used for the identi-
fication of Lancefield groups A-H and K-S (37). Zinner et al. (64) iso-
lated various cariogenic strains from human sources and also found them
to be non-reactive with Lancefield grouping sera.
S. mutans strains form a relatively homogeneous group when classi-
fied according to biochemical reactions. Most strains ferment sorbitol
and mannitol, as well as melibiose, salicin, and inulin, and produce
large amounts of non-dialyzable glucan from sucrose. They also hydrolyze
esculin, grow in 4% NaC1 broth, and on 10% bile agar. Using cluster
analysis, Colman and Williams (15) showed that among 216 non-hemolytic
streptococci the dectran-producing organisms formed a separate cluster
comprised of mostly S. mutans. Guggenheim (30) and Edwardsson (25) also
showed S. mutans to form a homogeneous group by biochemical tests.
Drucker and Melville (22, 23), and Carlsson (13) showed by numerical
taxonomy that S. mutans was a distinct group and could be distinguished
from other common oral streptococci: S. sanguis, S. mitis, and S.
When examined by serological and genetic methods, S. mutans as
a group appears more heterogeneous. Jablon and Zinner (36, 63) examined
eight strains of cariogenic streptococci using immunofluorescence and
were able to place them in four groups, I-IV. However, there were cross-
reactions with other streptococci making the technique less useful than
subsequently developed techniques for the identification of S. mutans.
Perhaps the most useful method to date of classifying strains of
5. mutans was the serological scheme developed by Bratthall (9, 10, 11)
utilizing diffusion in gel and comparative immunoelectrophoretic patterns
of extracts perpared according to Lancefield (37). Although cross-reactions
existed, as were seen by other workers, the use of comparative immuno-
electrophoresis allowed Bratthall to separate 70 strains of S. mutans into
five serotypes: a through e. The serotype e organisms cross-reacted using
anti-Lancefield group E serum. The serotypes a through d did not react
with any of the Lancefield grouping sera. In a subsequent study, Bratthall
(12) used antisera conjugated with fluorescein isothiocyanate (FITC) and
found that unabsorbed, conjugated antisera prepared against reference
strains of the five serotypes showed cross-reactions with members of the
other serological groups of S. mutans as well as members of other strepto-
coccal species. Most notably, anti-AHT or anti-3720 (a) serum showed a
strong cross-reaction (4+) with B13 (d). Anti-B13 serum showed a moderate
cross-reaction with the same serotype a organisms (2+). Anti-JC2 serum
(c) showed a strong cross-reaction (3+ to 4+) with LM7 (e), while anti-
SS9 serum, a Lancefield group E serum, reacted weakly (1+) with JC2 (c)
and B13 (d), and strongly (3+ to 4+) with LM7. Through various whole cell
absorption procedures Bratthall was able to eliminate cross-reactions with
non-S. mutans species and produce specific staining sera for all serotypes
except c, which still reacted with LM7 (e). The S. mutans serotype e
cells could be identified by absorbing antiserum against Lancefield group
E (strain SS9) with S. mutans LM7 (e). Then a S. mutans strain of sero-
type e will react with unabsorbed anti-SS9 serum, with anti-JC2 (c) serum,
but not with the absorbed anti-SS9 serum. On the other hand, Lancefield
group E strains react with both the absorbed and unabsorbed anti-SS9
serum, but not with anti-JC2 (c) serum. The serotype c organisms do not
react appreciably with either anti-group E serum.
de Stoppelaar (20) was able to divide various cariogenic strepto-
cocci into three groups by precipitin analysis. His groups MI, MII, and
MO correspond to Bratthall's serotypes c, a, and e respectively.
Further work by Perch et al. (48) demonstrated two more serotypes,
f and g, using absorbed fluorescine-labelled antisera.
In addition to the serological grouping schemes, four genetic
groups based on DNA-DNA hybridization and % guanine-cytosine content were
shown by Coykendall (16, 17, 19) to parallel four of Bratthall's five
serotypes. Coykendall (18) proposed the adoption of subspecies names for
his genetic groups I through IV: S. mutans var. mutans, var. rattus, var.
sobrinus, and var. cricetus respectively. Shklair and Keene (54) separated
S. mutans into five biotypes, a-e, which correspond directly with
Bratthall's serotypes. Biotypes were determined by fermentation patterns
using mannitol, mannitol + bacitracin, sorbitol, raffinose, melibiose,
and production of ammonia from arginine.
Table 1 summarizes the relationships among the strains of S. mutans,
determined by various workers, using serological methods, DNA-DNA re-
association (genetic groups), guanine-cytosine content (%GC), and bio-
chemical reactions (biotypes).
Immunochemical nature of the serotype antigens. In 1973 Van de
Rijn and Bleiweis (59) isolated a glycerol teichoic acid antigen from
S. mutans AHT (a) by cold 5% trichloroacetic acid (TCA) extraction.
The material was fractionated by agarose gel chromatography and the antigen
fraction contained glycerol and phosphorus in a 1:1 molar ratio. The
fraction also contained glucose, galactose, alanine, and fatty acids.
SURVEY OF THE VARIOUS CLASSIFICATION SCHEMES FOR
Serotypes groups % GC Subspecies Biotypes
(10 )a (48 ) (20 ) (36) (17) (18) (54)
a MII I IV 42-44 var. circetus a
b II II 41-43 var. rattus b
c MI I 36-38 var. mutans c
d MII III 44-46 var. sobrinus d
e MO e
aSee Literature Cited
Partial alkaline hydrolysis yielded a glyceride containing a disaccharide
of glucose and galactose attached to the 2-hydroxyl group of glycerol.
Precipitin inhibition analysis showed that B-galactosides were the
greatest inhibitors (> 75%) while glucose and other glucosides inhibited
to a lesser extent (< 30%). Immunodifussion and immunoelectrophoresis
showed this antigen to be present in all six serotype a organisms tested
and absent in serotypes b, c, and d.
A short while later, in 1973, Mukasa and Slade (46) isolated a
polysaccharide antigen from lyophilized whole cells of a hamster strain
of S. mutans, HS6 (a), using boiling water. The ethanol precipitate of
the extract was purified by chromatography on diethylaminoethyl (DEAE)-
Sephadex A-25, Sephadex G-200, and Carboxymethyl (CM)-Sephadex C-25.
The antigen was primarily composed of galactose (54%), glucose (10.4%),
glucosamine (8.6%), and galactosamine (5.4%), with small amounts of
lipid, phosphorus, and 5% amino acids. The antigen reacted not only
with anti-HS6 serum but also with anti-B13 (d) serum. The cross-
reactivity could be eliminated by absorbing anti-HS6 serum with B13 whole
cells. Cross-reactive ("a-d") serum could be obtained by releasing the
antibody that bound to B13 cells during absorption. In subsequent
papers, Linzer et al. (39, 41) showed that the purified antigen possessed
two sites on a single molecule: the "a" site, unique to the a serotype,
and an "a-d" site, which was responsible for the cross-reaction with d
serotypes. Precipitin inhibition studies with abosrbed anti-HS6 serum
showed the "a" site to be dependent on a-1-4 and -1-6 glucosides,
however, free galactose or galactosides other than lactose were not tested
as potential inhibitors. The latter disaccharide gave less than 10%
inhibition. Maltose (glu a-1-4 glu) and gentiobiose (glu B-1-6 glu) each
gave approximately 50% inhibition, while a and 8-methylglucopyranosides gave
less than 10%. Inhibition studies with "a-d" serum showed 32% inhibition
with galactose, 36% with galactosamine, and 24% with melibiose. No other
inhibitors were tried. Since this dissertation deals with the relation-
ship of the antigens of serotypes a and d in great detail, including
structural analyses of the purified antigens, the reports outlined above
(41, 46) will be further discussed below (see Discussion).
In 1973, Mukasa and Slade (45) isolated two antigens from whole
cells of a serotpye b organism FA1, using cold 10% TCA. One (antigen 1)
was a polysaccharide and the other (antigen 2) was a mucoprotein. Both
antigens possessed the same immunological specificity. Both antigens
contained about 30% galactose and 3% galactosamine, while rhamnose was the
major component (47%) of the carbohydrate antigen (antigen 1) and present
only in traces in antigen 2. Antigen 2 contained about 40% protein while
antigen 1 contained 5% protein. When studied by immunoelectrophoresis,
the two antigens migrated as individual bands, with antigen 2 migrating
the farthest toward the anode. Galactose and galactosamine were the best
precipitin inhibitors (78%) of antigen 2 reactivity. Antigen 1 was not
studied for immunochemical specificity.
In 1974, Vaught and Bleiweis (60) isolated a glycerol teichoic
acid antigen from purified cell walls of S. mutans BHT (b) using cold 10%
TCA. The extract was fractionated by gel filtration and ion exchange
chromatography. Two fractions were obtained, each yielding a single
separate band on immunoelectrophoresis which migrated in the same manner
as the bands obtained from mild acid extracts of whole cells of serotype
(b) organisms. The faster migrating band, fraction C, was similar in
mobility to Slade's antigen 2 except fraction C contained primarily
galactose with glycerol and phosphorus in a molar ratio of 1.0: 1.4.
Only small amounts of rhamnose and glucose were measured. This fraction
resembled a glycerol teichoic acid. Fraction B was very similar to Slade's
antigen 1, both in mobility and in chemical content.
In 1974, Wetherell and Bleiweis (61) isolated a polysaccharide
antigen by hot formamide extraction from purified cell walls of S. mutans
GS5 (c). The major antigen, purified by gel filtration and ion exchange
chromatography, was composed mainly of rhamnose and glucose in a molar
ratio of 1.7: 1.0. Precipitin inhibition studies revealed the major
hapten to be an a-linked glucose moiety. In a subsequent study, Linzer
et al. (40) isolated a nearly identical polysaccharide from hot 5% TCA
extracts of whole cells of strain Ingbritt, another serotype c organism.
Linzer and Slade (42) isolated a polysaccharide antigen from whole
cells of S. mutans strain B13 (d) using cold 5% TCA. After the antigen
was purified by ion exchange and gel filtration chromatography, it was
found to be composed mainly of galactose and glucose in a 2:1 molar ratio.
The antigen was similar to the a antigen in that it was a single molecule
possessing two reactive sites: one specific for the d serotype (d site)
and one responsible for the cross-reaction with the a antigen (a-d site).
Anti-B13 serum could be abosrbed with HS6 (a) cells to yield a specific
d serum. Inhibition studies with anti-d serum (absorbed) showed the
hapten to be galactose, possibly a-linked. The nature of the a-d site
specificity in B13 was not investigated.
The serotype e antigen was purified after hot formamide extraction
of purified cell walls of V-100 by Wetherell and Bleiweis (62).
Two pclysaccharide antigens were resolved by ion-exchange chromatography:
one antigen (I) was specific for serotype e, and the other (II) cross-
reacted with serotype C antisera. Antigen I was composed mainly of
rhamnose and glucose in a 2.3:1 molar ratio and possessed a B-glucosyl
group as its antigenic determinant. Antigen II possessed identical
chemical components and its homologous reactivity with anti-e serum also
was inhibited primarily by B-glucosides. Although not proven by the
authors, it was speculated antigen II may possess a-glucosyl haptenic
moieties to account for cross-reactions with c serum. Hamada and Slade
(33) isolated a similar antigen from strain MT703 (e) by the autoclave
extraction method of Rantz and Randall (51). The major antigen, el was
composed mainly of rhamnose (56%) and glucose (37%) and possessed a
S-glucoside as a haptenic determinant. The e c cross-reactive antigen
was not found.
Hamada et al. (32) isolated a polysaccharide antigen with serotype
f specificity from whole cells of strains OMZ175 and M1557 using hot 5%
TCA. The antigen from OMZ175 was composed of rhamnose (49%) and glucose
(47%), and the antigen from MI557 also contained rhamnose and glucose,
60% and 39% respectively. The best inhibitor of precipitation was
isomaltose, a diglucoside with an l-~6 linkage. The antigen was cross-
reactive with the glucan ("mutan") produced by the same organism when
grown in sucrose.
A polysaccharide antigen from a pH 7.3 buffer-boiled extract of
S. mutans 6715 (g) was isolated by lacono et al. (35) in 1975. At the
time this strain was thought to belong to serotype d, but is now con-
sidered to be a serotype g organism (34, 48). The antigen, purified by
ethanol precipitation and gel filtration, was composed primarily of
galactose (60.6%) and glucose (10.3%) and amino acids (9.5%). The
composition was very similar to the a antigen of Mukasa and Slade (46)
and indeed serotype g strains have been shown to cross-react with both
serotype a and d strains. It is not known whether the molecular site
responsible for the a-a and d-g cross-reaction is the same as that for
the a-d cross-reaction. Precipitin inhibition tests showed that galactose
was the most effective inhibitor, although glucose also gave significant
inhibition. The immunodominant region is thought to involve the 1-6
linkage of galactose and glucose in either the a or B configuration.
The work presented in this dissertation began in 1973, prior to
the publication of work on the polysaccharide antigens of serotypes a
and d. Because of major discrepancies between the published data and our
early findings, and more importantly because of a desire to gain infor-
mation concerning the basic structural nature of these antigens, work was
continued. The findings in this dissertation agree in part with some of
the published data on the a, d, and a-d antigens, but they also refute,
or at least open to further debate, previous findings heretofore unveri-
fied. In addition, the structural data presented below which have not
previously been published, provide further insight into the chemical
structure of these antigens and may provide important information to
future investigators desiring to synthesize immunogens for anti-caries
vaccines or to more fully understand the immunogenicity and antigenicity
of surface polysaccharides of S. nmutans.
MATERIALS AND METHODS
Bacterial strains and growth conditions. S. mutans strain AHT
(a) was obtained from J. Jablon (University of Miami, Miami, Florida),
and strain B13 (d) was obtained from D. Bratthall (University of Goteborg,
Cultures were maintained as lyophilized stocks. Samples to be
lyophilized were grown for 20 h at 37C in Todd-Hewitt broth (Difco
Laboratories, Detroit, Michigan) supplemented with 0.4% glucose (THG),
and adjusted to pH 6.5 with concentrated HC1 before autoclaving. The
pellet from 250 ml of broth was suspended in 2 ml of sterile 20% powdered
milk solution (Pet, Inc., St. Louis, Missouri), and 0.2 ml samples were
lyophilized in glass ampoules on a Virtis freeze drying apparatus (Virtis
Research Equipment Co., Gardiner, New York). The ampoules were sealed
under vacuum and stored at -200C.
Cells for antigen preparation were grown in ten liter batches in
THG. Growth was initiated by inoculating 10 ml of THG from lyophilized
stocks. The entire culture was transferred at 24 h intervals into 250 ml,
1000 ml, and finally 8750 ml of THG giving a total volume of 10 1.
Cells were harvested 24 h after the last transfer in a Delaval gyro-testor
(Delaval Separator Co., Poughkeepsie, New York). Autolysins were in-
activated by heating the cells at 600C for 20 min.
Extraction of antigens for routine identification. Lancefield-
type acid extracts were used for routine identification. Approximately
5 mg of whole cells or cell walls were suspended in 1 ml of saline and
adjusted to pH 2.5 with HC1. The suspensions were boiled for 10 min,
cooled, centrifuged at 1070 x g for 20 min to remove cell debris and the
supernate neutralized with 0.05 N NaOH.
Cell walls used for extraction were prepared by the method of
Bleiweis et al. ( 7) which involves breakage with glass beads in a Braun
tissue homogenizer (Bronwill Scientific, Rochester, New York), followed
by treatments with ribonuclease, deoxyribonuclease, and trypsin.
Extraction of antigens for purification. Antigens for further
purification were extracted by the hot formamide method of Fuller (28).
Cell walls (800 mg) were suspended in 40 ml formamide and heated at 180 C
for 30 min in a paraffin oil bath. The suspension was then washed by
adding 2 vol (80 ml) of an acidic ethanol solution containing 19 parts
absolute ethanol and 1 part 2N HC1 (vol/vol), and centrifuging at 7700 x g
for 20 min in glass centrifuge tubes. The supernate was decanted and
saved and the pellets were washed twice more with 30 ml of a mixture of
the above acid-ethanol solution and deionized water (2:1, vol/vol).
Acetone (900 ml) was added to the combined supernates and the resulting
solution was immediately poured into round bottom centrifuge bottles (250
ml capacity). One milliliter of 1M NaC1 was added to each bottle and the
bottles were stored at 40C for at least 8 h. The bottles then were
centrifuged at 530 x g for 20 min, the supernates were decanted and the
pellets were allowed to dry. The pellets were dissolved in deionized
water, centrifuged at 27,000 x g for 20 min, and the supernates were
dialyzed and lyophilized.
Isolation of castor bean lectin (CBL). The castor bean lectin
from Ricinus communis was isolated by the method of Nicolson and
Blaustein (47) for use in affinity chromatography. Extracts of castor
beans are extremely toxic and allergenic and, accordingly, care must be
taken in handling them and all glassware was detoxified using a dilute
solution of sodium hypochlorite.
R. communis beans, 100 g (kindly provided by the McNair Seed Co.,
Plainview, Texas), were blended in a model 1120 Waring blender (Waring
Products Div., Dynamics Corporation of America, New Hartford, Connecticut)
with 1 1 of 0.005 M sodium phosphate buffer pH 7.2 containing 0.2 M NaCI
and 10-4 M Mg+2, Ca+2, and Mnf2 as chlorides (PBS +). The beans were
blended into a smooth paste (ca. 30 min), then the extract was filtered
through cheesecloth and centrifuged at 10,000 x g for 30 min at 40C.
The supernate was carefully decanted and saved, without including the
floating lipid layer, and centrifuged again. The supernatant solution was
adjusted to 60% saturation with ammonium sulfate and slowly stirred at
40C overnight. After centrifugation at 10,000 x g for 20 min, the pellet
was dissolved in PBS+ and dialyzed against PBS+ to remove the ammonium
sulfate. The dialysate (ca. 50 ml) was then applied to a 4 x 40 cm column
of Sepharose 4B (Pharmacia Fine Chemicals Inc., Piscataway, New Jersey)
and eluted with PBS+ until the absorbance at 280 nm was less than 0.05.
Since Sepharose 4B contains polysaccharides with galactose-terminal groups,
it served as an affinity column to bind the lectin. The lectins ( a
mixture of two, RCA60 and RCA120) then were eluted with 0.2 M D-galactose
in PBS4+. The A280 peak was pooled, dialyzed against PBS++ and then
deionized water, and finally lyophilized.
Preparation of affinity columns with castor bean lectins. An
affinity chromatography gel was prepared using Affi-Gel 10 (Biorad
Laboratories, Richmond, California), an agarose gel containing a 10-A
aliphatic spacer arm terminating with an active carboxy-N- hydroxy-
succinimide ester group. Approximately 350 mg of the lectin was dissolved
in 50 ml of 0.1 M sodium phosphate buffer pH 7.0 containing 10 M metal
ions as before, and 0.05 M D-galactose, to prevent binding of the lectin to
the gel matrix. To each of two vials of gel (approximately 13 ml wet gel
volume per vial), 25 ml of the lectin solution was added and the vials
were gently agitated on a shaker for 24 h at 40C. The coupled gel was
packed into a 1.3 x 30 cm column and eluted with PBS+ plus 0.2 M D-galactose
until the absorbance at 260 nm reached baseline, indicating the N-hydroxy-
succinimide released during coupling was removed. The column was washed
with about 200 ml of PBS+ minus galactose before loading with a sample.
Purification of antigens by affinity chromatography. Samples of
crude formamide extract (50 mg) were loaded on the affinity column at 40C
and eluted with PBS+. The flow rate was 12 ml/h and 2 ml fractions were
collected. After 72 ml were collected to remove non-binding materials,
the elution system was changed to PBS+ plus 0.2 M D-galactose to remove
bound polysaccharides. The column was eluted with at least 100 ml of the
latter buffer before re-equilibrating with the starting buffer. [All
buffers used with the affinity columns were pre-filtered through a 0.45
um membrane filter to prevent clogging of the bed surface].
Fractions were monitored for carbohydrate by the phenol-sulfuric
acid method as described below and for antigenic reactivity by capillary
tube precipitation using the appropriate antiserum.
Gel filtration chromatography. Samples of crude formamide
extract (50 mg) were loaded on a 2.5 x 80 cm column of Bio-Gel P-150
(Biorad) 100-200 mesh and eluted with 0.85% NaCl at a flow rate of 10 ml/h.
Fractions (2 ml) were collected and analyzed for absorbance at 220 nm,
rhamnose as described below, and antigenic reactivity using capillary
Samples of lectin-purified antigen were filtered as above except
on 2.5 x 40 cm columns. Fractions (2 ml) were collected and assayed for
sugars by the phenol-sulfuric assay and for antigenic reactivity by
precipitation in capillary tubes.
Quantitative chemical assays. Phosphorus in cell walls was
measured by the method of Chen et al. (14), and for the purified antigens
by the method of Lowry et al. (43). Both methods depend on the formation
of a blue phosphomolybdate complex, but differ in the ashing procedure
and in sensitivity.
Determination of glucose and galactose in cell walls was performed
using the Glucostat and Galactostat reagents (Worthington Biochemical Corp.,
Freehold, New Jersey). Samples were hydrolyzed in 3 ml of 2 N HC1 for
4 h at 1000C. Hydrolyzed samples were quantitatively transferred to a
10 ml volumetric flask and neutralized to a phenolpthalein endpoint with
5 N NaOH. The volume at this point was 7.6-7.8 ml. The sample was then
brought to 10.0 ml with 0.27 M sodium phosphate buffer, pH 7.0.
Rhamnose in cell walls and column fractions was measured by the
method of Dische and Shettles (21).
Glycerol was determined utilizing the Glycerol Stat-Pack
(Calbiochem, Atlanta, Georgia) according to the manufacturer's instructions.
Samples of cell walls (10 mg) were hydrolyzed for 3 h at 1050C in 1 ml of
2 N HC1. After hydrolysis, the acid was removed by lyophilization and
three 1.5 ml distilled water washes, each followed by lyophilization.
Free glycerol was released by treating the samples with 2 mg alkaline
phosphatase (Worthington) for 24 h at 370C in 3 ml of 0.08 M ammonium
carbonate buffer pH 9.3.
The phenol-sulfuric acid assay was performed according to Dubois
et al. (24) using glucose as a standard. The tubes were read at 485 nm.
Protein was determined by the method of Lowry et al. (44).
Details of the assay are given below.
Amino acids and amino sugars were determined using a JEOL model
JLC-6AH automated amino acid analyzer (JEOL, Inc., Cranford, New Jersey)
according to the method of Spackman et al. (58). Peaks were quantitated
using an Autolab System AA integrator (Spectra-Physics, Mountain View,
California). Samples (10 mg) were hydrolyzed in sealed ampoules with
either 5 ml of 4 N HC1 for 11 h at 1050C or with 5 ml of 6 N HC1 for 18 h
at the same temperature. The second method was necessary for the analysis
of cell walls of serotypes a and d in order to fully hydrolyze an acid
resistant component presumed to be a lysyl-threonine dipeptide (8).
After hydrolysis, acid was removed using a rotary evaporator (Buchler
Instruments, Fort Lee, New Jersey) and the samples were washed three
times with deionized water followed by evaporation. Hydrolyzed samples
were quantitatively transferred to a 25 ml volumetric flask and diluted
to volume using 0.01 N HC1.
Gas liquid chromatography of sugars. Sugars in the crude
formamide extracts and purified antigens were measured by gas liquid
chromatography. Samples (1-5 mg) were hydrolyzed in sealed ampoules with
1 N H2SO4 for 8 h at 100C. After hydrolysis, the samples were cooled,
opened, and 1.0 ml of a mannitol solution of a known concentration (usually
5.5 umole/ml) was added as an internal standard. The hydrolysates were
neutralized with BaCO3, and the precipitate of insoluble BaSO4, as well
as any excess BaC03, was removed by centrifugation at 1000 x g for 20 min.
Supernates were centrifuged again as above and then lyophilized.
Hydrolysates were converted to their trimethylsilyl derivatives using
TRI-SIL (Pierce Chemical Co., Rockford, Illinois) according to the
instructions in the Pierce handbook of silylation (49). The derivatives
were separated on 6 ft x 4 mm glass columns packed with SE-30 Ultraphase,
3%, on Chromosorb W (HP) 80/100 mesh (Pierce) on a Packard model 803 gas-
liquid chromatograph (Packard Instrument Co., Downers Grove, Illinois).
The operating temperature was 165 C isothermall), the carrier gas (nitrogen)
flow rate was 30 cc/min, and a hydrogen flame ionization detector was used.
Peaks were quantitated using an Autolab Minigrator electronic digital
integrator (Spectra-Physics, Santa Clara, California). Calibration curves
were constructed by injecting samples containing varying amounts of the
standard sugars deriviatized as above and a constant amount of the internal
standard, mannitol. The ratio of area mannitol/area standard was plotted
vs. moles mannitol/imoles standard.
Methylation analysis of polysaccharide antigens. Purified antigens
were methylated using the method of Hakomori (31) according to Bjorndal
et al. (4) and Lindberg (38). The procedure involves treating the poly-
saccharides with methylsulfinyl sodium in dimethyl sulfoxide (DMSO),
followed by methylation with methy iodide. The methylated product is
hydrolyzed, reduced, acetylated and analyzed by gas liquid chromatography
and mass spectrometry. All solvents were re-distilled and stored in glass-
stoppered bottles and care was taken to prevent contamination with plastics.
The methysulfinyl carbanion was prepared according to Sanford and
Conrad (53) and stored at -20 C under mineral oil.
Samples of antigen (1.5 mg) were placed in small serum vials and
capped with rubber septa. The vials were flushed with dry nitrogen through
an 18-gauge hypodermic needle and the samples dissolved in 1.0 ml of DMSO
which had been dried over 3 A molecular sieves (Davison Chemical,
Baltimore, Maryland). Methylsulfinyl sodium, 1.0 ml, was introduced into
the vials with a glass syringe and the mixture was treated for 1 h with
ultrasound at room temperature. The reaction was allowed to proceed for
an additional 6 h with stirring, and then methyl iodide (1.0 ml) was added
dropwise with cooling in an ice bath. The solution was agitated in an
ultrasonic bath for 30 min, diluted with 10 ml of deionized water, dialyzed
overnight against 10 1 of deionized water, and evaporated to dryness on a
rotary evaporator at 40 C. The above procedure was repeated using the
product of these reactions to ensure complete methylation.
The methylated polysaccharides were hydrolyzed in two steps.
The residue was dissolved in 3 ml of 90% formic acid and treated at 1000C
for 2 h. The samples were evaporated to dryness at 400C and then dissolved
in 1 ml of 1 N H 2SO4 and hydrolyzed for 8 h at 1000C. The hydrolysate
was neutralized with barium carbonate, clarified by centrifugation and
evaporated to dryness at 40 C.
The residues were dissolved in 5 ml of deionized water containing
15 mg sodium borohydride and allowed to stand at room temperature for 2 h.
To the mixture was added 1 ml (wet volume) of AG 50W-X8, H+ form, (Biorad).
The mixture was agitated for 5 min, filtered, and evaporated to dryness
at 400C. The residual boric acid was removed by codistillation with
methanol (5 x 5 ml) on the rotary evaporator. Finally, the mixture of
reduced, methylated sugars was acetylated with acetic anhydride-pyridine,
1:1, for 15 min at 1000C. The reagents were removed by distillation with
toluene (5 x 5 ml) on the rotary evaporator at 400C. The residue was
dissolved in 0.1-0.2 ml acetone for analysis.
Analysis of methylated sugar derivatives. The partially-methylated
alditol acetate derivatives were separated by gas liquid chromatography on
6 ft x 2 mm silanized glass columns packed with 3% OV-225 or 3% ECNSS-M
on Gas-Chrom Q,100-120 mesh (Applied Science Laboratory, Inc., State College,
Pennsylvania). Temperature for determination of relative retention times
was 1700C, and 155C for measurement of peak area. The nitrogen carrier flow
rate was 20 cc/min and a hydrogen flame ionization detector was used.
Analysis was carried out on a Packard model 803 gas-liquid chromatograph
(Packard Instrument Co.). Retention times relative to 1,5-di-O-acetyl-2,
3,4,6-tetra-O-methyl-D-glucitol were determined by interpolation between
known standards prepared from a mixture of 2,3,4,6-tetra-0-methyl-D-glucose
(Sigma Chemical Co., St. Louis, Missouri) and methylated, hydrolyzed, lactose
(Sigma) and oyster glycogen (Nutritional Biochemical Corp., Cleveland, Ohio).
This mixture provides standards of 1,5-di-0-acetyl-2,3,4,6-tetra-0-methyl-D-
glucitol, 1,5-di-0-acetyl-2,3,4,6-tetra-O-methyl-D-glactitol, 1,4,5-tri-O-
acetyl-2,3,6-tri-0-methyl-D-glucitol, and 1,4,5,6-tetra-0-acetyl-2,3-di-O-
Confirmation of peak identity was obtained by combined GLC-mass
spectrometry on a Pye model 104 gas chromatograph (Pye Unicam Ltd.,
Cambridge, England) interfaced with an Associated Electrical Industries
model MS 30 double-focusing mass spectrometer with a DS-30 data system
(AEI, Scientific Apparatus Ltd., Manchester, England). Peaks were detected
by following total ion current. Gas chromatographic conditions were as
described above, using OV-225 as the stationary phase and a column temper-
ature of 170 or 2000C. Helium was used as the carrier.
Preparation of antisera. Cells of each strain (AHT, B13) were
grown for 20 h in 250 ml of THG in plastic centrifuge bottles. Cells were
pelleted by centrifugation at 10,000 x g for 15 min and washed twice with
sterile saline. The cells were resuspended in 100 ml of saline containing
0.6% formalin and incubated for 24 h at room temperature. The cells were
washed twice with sterile saline and diluted in saline to a reading of 360
using a #66 filter on a Klett-Summerson model 800-3 colorimeter (Klett
Manufacturing Co., Inc., New York, New York). After noting the volume, the
cells were pelleted as above and resuspended in one-tenth of the original
volume in sterile saline. The suspensions were stored at -200C and used
as immunogens for antiserum production.
Male New Zealand white rabbits (3 kg) were injected intravenously
on Monday, Wednesday, and Friday according to the following regimen:
during the first week 0.1 ml was injected, the second week 0.2 ml, and
the third week 0.3 ml. Rabbits were given a prophylactic intramuscular
injection of 1 ml (50 mg/ml) of Liquamycin (oxytetracycline-HCl, Pfizer,
Inc., New York, New York) after each immunization. On Monday of the
fourth week animals were test-bled from the ear artery and the serum was
tested for ability to precipitate with the homologous, cell-wall
Lancefield extract in capillary tubes, and for cross-reactivity by
Ouchterlony gel-diffusion analysis. Animals with suitable titers were
anesthetized with pentothal sodium and exsanguinated by heart puncture.
Sera were stored in small aliquots at -200C with 0.01% sodium azide.
Gel diffusion and immunoelectrophoresis. For both gel diffusion
and immunoelectrophoresis (IE), gels containing 0.75% agarose (Sigma) in
0.05 M sodium barbital/HCl pH 8.2 were used. Slides for IE (7.6 x 3.8 cm)
were cleaned in chromic acid, dipped in 0.2% agarose in deionized water
and allowed to dry. The buffered, molten 0.75% agarose solution (3 ml)
was pipetted onto each slide and allowed to solidify. Wells and troughs
were cut in the slides using a Shandon gel cutter (Shandon Scientific Co.
Ltd., London, England). Electrophoresis was performed at 40C at 10 ma
per slide for 120 min using a VOKAM SAE 2761 power supply (Shandon) and
employing the above barbital buffer. After electrophoresis of the antigen,
the troughs were cleared of agarose, filled with the appropriate antiserum,
and the plates were incubated at 40C in a moist chamber.
For gel diffusion analysis, plastic petri dishes, 50 x 12 mm,
(Falcon No. 1006, Falcon Plastics, Oxnard, California) were filled with
3 ml of the same buffered agarose solution and the appropriate number of
wells (50 ul capacity) were cut. After addition of antigen and antiserum
preparations to the wells, plates were incubated at 20 C until development
of precipitin bands.
Absorption of antisera. Whole antisera were absorbed with an
equal volume of heat-killed, washed, THG-grown cells of the heterologous
strain for 4 h at 4 C with stirring. The cells were pelleted by centri-
fugation at 8000 x g for 20 min. The supernates contained antibodies
specific for the a (AHT) or the d (313) determinants and were found to be
free of the cross-reactive (a-d) antibodies by gel diffusion analysis.
These absorbed sera were designated "anti-a" and "anti-d", respectively.
The B13 (d) cells used to absorb the anti-AHT (a) serum were
used as a source of cross-reactive (a-d) antibodies. After absorption as
above, the cells were washed 3 times with cold 0.02 M sodium phosphate
buffered saline (PBS) pH 7.2, suspended in 3 ml PBS, adjusted to pH 2.2
with 1 N HC1 and stirred for 1 h at 40C. The cells were pelleted at 8000
x g for 20 min. The supernate was neutralized with 1 N NaOH, dialyzed
against one liter half-strength PBS overnight and lyophilized. The
residue was dissolved in one half of its original volume with deionized
water. This preparation was designated "anti a-d"serum.
Quantitative precipitin analysis. Quantitative precipitin
analyses were performed to determine equivalence for each antigen-antibody
system that was studied. A stock solution of each antigen (ca. 50 ug/ml)
was made in 0.85% saline containing 0.01 M disodium ethylenediaminetetraacetate
(sEDTA), adjusted to pH 7.0 with 4 N NaOH. Antisera (20 pl anti-d diluted
1:2 with sEDTA, 25 ul anti-a, or 20 ul anti a-d) were added to 13 x 100 mm
tubes. Because of the problem of insolubility of inhibitors at the higher
concentrations used in the a antigen-anti-a, and the d antigen-anti-d
systems, a volume of 100 ul of sEDTA (corresponding to the volume of inhib-
itor used) was added to the appropriate tubes, while a volume of 50 pl
of sEDTA was added to the a antigen-anti-a-d system, and to the d antigen-
anti-a-d system. Various dilutions of antigen (50 4l in sEDTA) were added
and the tubes were incubated at 370C for 1 h followed by incubation at
40C for 7 days. Following incubation, 200 ul of cold sEDTA was added to
each tube and the tubes were centrifuged at 1500 x g for 1 h at 40C.
The supernates were carefully decanted and the tubes were allowed to drain
for 10 min onto absorbent paper. The samples were washed twice as above
except with 200 4l of saline without EDTA. After thoroughly draining,
the samples were dissolved in 100 pl 1 N NaOH and held for 30 min at room
temperature. Folin reagent C (alkaline copper tartarate minus NaOH),
1.0 ml, was added and the tubes were incubated at 370C for 10 min.
Commercial Folin phenol reagent (Harleco, Gibbstown, New Jersey) was
diluted 1:2 (1 N after dilution) and 100 ul was added to each tube with
constant mixing. Tubes were incubated at 37 C for 30 min and read at
750 nm. A protein standard curve of rabbit gamma globulin (Calbiochem)
was used to quantitate antibody protein in samples. All tubes were done
Quantitative precipitin inhibition. Inhibition of precipitation
was performed in the same manner as the quantitative precipitin analysis
except various concentrations of potential inhibitors in the above volumes
of sEDTA were incubated with the antiserum for 30 min at 370C prior to
the addition of the antigen solutions. All antigen-antibody systems were
reacted at the estimated equivalence point. Control tubes containing only
antiserum and the appropriate volume of sEDTA were run and protein values
subtracted from the readings of the experimental tubes. All inhibitions
were done in triplicate.
Synthesis of methyl furanosides. Two inhibitors, a and S-methyl-
D-galactofuranoside, were synthesized by the method of Augestad and Berner
(1, 2). Since it was not possible to crystallize the derivatives, their
concentrations were determined by the phenol-sulfuric acid assay, the
solutions were evaporated to the consistency of a syrup and then diluted
to the desired concentration in sEDTA.
Chemicals. Most common salts, organic solvents and reagents, and
acids were obtained from Mallinckrodt (Scientific Products, Chamblee,
Georgia). All carbohydrate inhibitors were obtained from Sigma Chemical
Co. (St. Louis, Missouri).
Composition of purified cell walls of S. mutns AHT and B13.
Purified cell walls of AHT and B13 were examined for chemical composition
and for purity as determined by the absence of non-peptidoglycan amino
acids (Table 2). The major sugars present in both walls are rhamnose,
glucose and galactose in molar ratios of 3.8:1.3:1 in AHT and 7.6:4.2:1
in B13. In both cases these three sugars make up nearly 50% of the wall
by dry weight. One unique feature of these walls is the presence of
threonine, which is not found in the walls of other S. mutans serotypes.
The threonine appears to be linked to lysine since upon hydrolysis with
4 N HC1 at 1050C for 11 h, an unknown dipeptide peak is present in the
amino acid profile; and upon employing harsher conditions (6 N HC1, 105 C,
18 h) this peak disappears, and the relative molar amounts of lysine and
threonine increase with respect to the other amino acids. The presence of
threonine, probably as an alanyl-threonine transpeptide was shown in the
peptidoglycan of these serotypes in previous investigations (5, 6, 8, 57).
Extraction of antigens. Preliminary studies showed that antigeni-
cally reactive extracts could be obtained with mild Lancefield-type acid
extractions of whole cells or cell walls, and by extraction with hot
formamide. Figure 1 shows a comparison by gel diffusion of Lancefield and
formamide extracts of AHT and B13 cell walls using unabsorbed, homologous
sera. In both cases identity is seen between the two extracts.
Gel diffusion analysis of the formamide extracts of each strain
against unabsorbed sera prepared against each organism shows a reaction
Table 2. CHEMICAL COMPOSITION OF PURIFIED CELL WALLS OF
S. mutans AHT (a) and B13 (d)
aAmino sugars are reported as acetylated derivatives
b Trace amount not measured
Gel immunodiffusion experiments demonstrating the
reaction of identity obtained between Lancefield
extracts of AHT and B13 (aLE and dLE) and formamide
extracts (aFE and dFE), and the homologous, unabsorbed
anti-AHT serum (Fig. la) and anti-B13 serum (Fig. lb).
of partial identity in each case (Fig. 2a and 2b). Absorption of each
antiserum with the heterologous strain eliminates the cross-reaction
(Fig. 2c and 2d). The cross-reacting antibodies can be released from
the cells used for absorption by incubation at pH 2.2 for 1 h. Figure 2e
shows the reaction of antibodies released from B13 whole cells used to
absorb anti-AHT serum, showing a line of identity between the two extracts.
On immunoelectrophoresis, both extracts show single bands with
cathodic mobility (Fig. 3). Lancefield extracts of B13 generally show a
double-arc "gull-wing" pattern, particularly when extracted for longer
than 10 min (not shown here), but this and the apparent faster mobility
of formamide extracts may be due to size degradation or charge alteration
of the polymers as will be shown below.
For extraction of the antigens from purified cell walls, the
formamide technique was used because it yielded an antigenically reactive
product which was free of peptidoglycan components. Table 3 gives the
compositions of the formamide extracts of both strains. The extracts are
almost totally composed of carbohydrates with less than 1% phosphorus
suggesting the absence of lipoteichoic acid or teichoic acid polymers as
well as nucleic acid contaminants. Amino acid analysis (not shown) re-
vealed only trace amounts of amino acids and amino sugars indicating the
absence of peptidoglycan components or proteins. Routine yields of the
antigenic formamide extract were about 33% of the dry weight of the cell
walls used for extraction.
Purification of antigens. Initially, attempts were made to purify
the antigens by gel filtration of formamide extracts. Figures 4 and 5 show
the profiles of the formamide extracts on Bio-Gel P-150 which has an
Gel immunodiffusion experiments demonstrating the
precipitin'patterns obtained between the formamide
(aFE or dFE) extracts and various antisera: unabsorbed
anti-AHT serum (Fig. 2a), unabsorbed anti-Bl3 serum
(Fig. 2b), anti-AHT serum absorbed with B13 cells (anti-a,
Fig. 2c), anti-B13 serum absorbed with AHT cells (anti-d,
* Fig. 2d), and immunoglobulins released from B13 cells used
to absorb anti-AHT serum (anti-a-d, Fig. 2e).
FE. d FE
la"mA H T
Figure 3. Immunoelectrophoretic patterns obtained with formamide (FE)
and Lancefield (LE) extracts of AHT (Fig. 3a) and B13 (Fig.
3b). Troughs were filled with unabsorbed anti-ART serum
(Fig. 3a) and unabsorbed anti-B13 serum (Fig. 3b).
Table 3. CHEMICAL COMPOSITION OF FORMAMIDE EXTRACTS OF
S. mutans AHT (a) AND B13 (d) PURIFIED CELL WALLS
Major Component AHT B13
umol/mg Ug/mg Imol/mg ug/mg
Rhamnose 3.11 510.4 3.24 532.2
Galactose 1.39 250.3 0.34 62.0
Glucose 1.08 193.7 2.05 369.5
Phosphorus 0,27 8.4 0.25 7.8
Total Recovery 962.8 971.5
n to Cd
- LfN W
0 CC 3
-4 Qg CM r
wO 0) 0a
s- < *-
*u o -'i-
. ... .J 0
e <^ -
'Yj -I-.,.- 4
c. ...... .....
exclusion limit of 150,000 daltons. The profiles were quite similar.
Fractions with antigenic reactivity emerged shortly after the void volume
and trailed into the large rhamnose-containing peak. It appeared that
the majority of these antigenic materials, because they emerged before
the rhamnose peak, did not contain a large amount of that common methyl
pentose. It was also apparent that it would be difficult to cleanly sep-
arate the antigenic materials from the non-antigenic rhamnose-containing
material by gel filtration. For this reason, an affinity chromatography
approach was devised employing the galactose-binding lectin from Ricinus
communis. Initial attempts to produce an affinity support using cyanogen
bromide-Sepharose 4B (Pharmacia) resulted in low capacity columns. The
best results were obtained using Affi-Gel 10 (Biorad) which contains a
10 A spacer arm. After loading 50 mg of formamide extract, the column was
eluted with approximately 70 ml of buffer to ensure elution of non-binding
material. Bound material was eluted using the staring buffer containing
0.2 M D-galactose and the antigen eluted as a sharp peak (Fig. 6). Results
obtained using ART crude FE (not shown) were identical to those depicted
in Fig. 6. The fractions under the peaks were pooled, dialyzed against
deionized water, and lyophilized. Some antigenic material passed through
the column initially without binding. This is probably due to overloading
of the column since these early fractions can be reloaded on the column
and substantial binding occurs. Initially-unbound fractions were routinely
recycled once in this manner and the bound material was pooled with the
Analysis of lectin-purified antigens. Analysis of the lectin-
purified antigens by gel diffusion indicates lines of complete identity
between the crude formamide and Lancefield extracts, and homologous
4- <; C.
o '-' Cj
-4 4 C
0 >c< J
~ Lw Cu
L4O o a0
CN e- Q *-l
W ea c -i
u so -a a)
W c e as ao
M Cu (
<0 4 0n a a
W u 3 m -4
4 C.a) r4 cJ
04 4-- 4
ud a a
u 0-1 -
.0 u oC0 (
3 4-ca *n
1-4 C Z
U-( C 0 ( )
t O CO Q
0 Cci 10
* *r CO C U
ao -i a)
Oid ao i
>. <; o ., -
< I- el -
0 --0 0
"A -- -- .
purified antigens of each strain (Fig. 7). Fig. 8 compares crude Lancefield
and formamide extracts of AHT and B13 cell walls with the homologous lectin-
purified antigen of each strain by immunoelectrophoresis. In each case a
single band with the same cathodic mobility is seen. As with the formamide
extracts, when the purified antigens are reacted with unabsorbed antiserum
by gel diffusion, partial identity between the two antigens is seen (Fig. 9a
and 9b). Absorption of the sera with the heterologous whole cells yields
specific anti-a or anti-d serum (Fig. 9c and 9d). Release of the immuno-
globulins absorbed from anti-AHT serum by B13 cells yields a serum giving
reaction of identity between the a and d antigens (Fig. 9e). Gel filtra-
tion profiles of each purified antigen are shown in Figs. 10 and 11.
A single, rather broad peak containing all of the antigen activity is seen
for each preparation. The broadness of the peaks may reflect the range of
polymer sizes which result from the rather harsh formamide extraction
procedure. The average molecular weight of the B13 antigen appears lower
than that for AHT and may reflect the increased susceptibility to hydrolysis
of B13 cell wall polysaccharides as discussed below. Chemical analyses
(Table 4) reveal that the antigens were composed primarily of galactose
and glucose in a molar ratio of 3.4:1 in AHT, and 2.1:1 in B13, with only
traces of rhamnose as shown by gas-liquid chromatography. Each antigen
contained less than 1% phosphorus. Both antigens were quite hygroscopic,
and although they were stored lyophilized and dessicated over CaSO4,
water of hydration collected over long periods of time and was not taken
into account in data shown in Table 4. Amino acid analyses did not reveal
any gross contamination with protein derived from the lectin-containing
affinity column used in the purification.
Gel immunodiffusion experiments using crude extracts and
purified antigens. In Fig. 7a, a reaction of identity is
seen between AHT Lancefield extract (aLE), AHT formamide
extract (aFE), and AHT purified antigen (a), using unabsorbed
anti-ART serum (center well). In Fig. 7b, identity is also
seen using B13 Lancefield extract (dLE), B13 formamide
extract (dFE), and B13 purified antigen (d). Center well in
Fig. 7b contains unabsorbed anti-B13 serum.
Immunoelectrophoretic patterns obtained with crude extracts
and purified antigens. Fig. 8a: comparison of AHT formamide
extract (aFE), purified antigen (a), and AHT Lancefield ex-
tract (aLE). Troughs contain unabsorbed anti-AHT serum.
Fig. 8b: comparison of B13 formamide extract (dFE), purified
antigen (d), and B13 Lancefield extract (dLE). Troughs con-
tain unabsorbed anti-B13 serum.
Gel immunodiffusion experiments using purified antigens from
AHT (a) and B13 (d). Antisera used are: unabsorbed anti-AHT
(a-AHT, Fig. 9a), unabsorbed anti-B13 serum (a-B13, Fig. 9b),
absorbed anti-ART serum (anti-a, Fig. 9c), absorbed anti-B13
serum (anti-d, Fig. 9d), and immunoglobulins released from
B13 cells used to absorb anti-AHT serum (anti-a-d, Fig. 9e).
Table 4. COMPOSITION OF LECTIN-PURIFIED ANTIGENS FROM
S. mutans AHT (a) and B13 (d)
Major Componenta AHT B13
umol/mg Pg/mg Pmol/mg Ug/mg
Rhamnose tra -- tr
Galactose 3.34 601.2 3.33 599.8
Glucose 0.97 175.4 1.58 284.8
Phosphorus 0.28 8.9 0.22 7.1
Total Recovery 785.5 891.7
a Trace amount not measured
Precipitin inhibition. Inhibition of antigen-antibody precipita-
tion was performed using absorbed sera specific for the a or d determi-
mants of each antigen, and serum specific for the cross-reactive (or a-d)
determinants of each antigen. The highest titers of anti-a-d reactivity
were obtained from anti-AHT serum and this was used in the inhibition
studies. Both absorbed a-d sera from anti-AHT or B13 showed the same
immunological reactivity (Fig. 12).
A precipitin curve was run in triplicate using varying amounts of
antigen and constant amounts of the absorbed antiserum (Fig. 13). For
the d-antigen, 1.25 ug was chosen as the approximate equivalence level
for studies using anti-d and anti-a-d sera, and 1.42 ug of the a antigen
was found optimally effective for precipitation of the anti-a and anti-
A variety of monosaccharides, methyl glycosides and disaccharides
were used as potential inhibitors. All samples and controls were run in
triplicate. Table 5 shows the results of inhibition of the anti-a serum
reaction with the purified a antigen. Galactose at the 50 umol level
gives the greatest inhibition (44%) versus 29% for glucose. The a and 3-
methyl glucopyranosides and galactopyranosides show no significant differ-
ences in inhibition from their non-methylated parent sugars. Of possible
significance is the finding that 3-methyl galactofuranoside shows greater
inhibition than the a-methyl galactofuranoside, although they are both
substantially lower than the corresponding galactopyranoside derivatives.
Among the disaccharides tested, melibiose provides the greatest inhibition
(46%) while lactose gives 39%. Both of these galactose-glucose-containing
disaccharides yield greater inhibition than the three diglucosides (cello-
biose, maltose, and gentiobiose), suggesting that the a determinant involves
Gel immunodiffusion experiments demonstrating the reaction of
identity of cross-reacting antibodies (a-d) obtained from anti-
AHT or anti-B13 sera. Top well contains the purified antigen
from AHT; bottom well, the purified antigen from B13. The
left well contains cross-reactive antibodies from anti-ART
serum, and the right well contains cross-reactive antibodies
from anti-B13 serum.
~tr 63 tT
0 0 0 0 0 0
S uO L- ro d o-
INHIBITION OF THE PRECIPITIN REACTION BETWEEN
AHT ANTIGEN AND ANTI- a SERUM
Inhibitor ymola % Inhibition
aAmount in a final volume of 17511
a galactose-glucose sequence possibly in a 1-6 linkage. (Methylation
studies to be discussed below eliminate the possibility of the 1-*4
linkage being present in the antigen). The anomeric nature of the link-
age is not clear when one considers the results obtained with the methyl
glycosides (Table 5).
Table 6 shows the results of inhibition of the anti-d serum re-
action with the purified d antigen. In this case the results are not
nearly as clear as for the a-anti-a system. At the 100 pmole level,
glucose provides 59% inhibition of precipitation while galactose yields
47% inhibition. The B-methyl glucopyranoside derivative gives nearly
twice the inhibition provided by the a-derivative, while both galactopyrano-
sides give about the same inhibition as the parent sugar, D-galactose.
The galactofuranosides are considerably less effective. All of the disac-
charides provide substantial inhibition although the two diglucosides,
maltose and gentiobiose, allow slightly higher inhibition, 53% and 50%
respectively. Cellobiose was insoluble at the 40 mole level, however,
at the 20 pmole level it gave slightly less inhibition than the other two
diglucosides at the lower level. It appears that a glucose-glucose sequence
or a glucose-galactose sequence is involved in the d antigen specificity.
However, it is not possible to establish clearer indications of the nature
of the d-specific hapten, including linkage patterns due to the relative
insensitivity of specific antibody in these inhibition studies. However, as
also the case for the a antigen, 1-4 linkages (as in maltose) are absent (as
shown by methylation studies to be presented below) and thus could not con-
tribute to the specificity of the d-reactive site.
Table 6. INHIBITION OF THE PRECIPITIN REACTION BETWEEN
B13 ANTIGEN AND ANTI- d SERUM
aAmount in a final volume of 170p1
The results of the a-d specificity studies for both antigens were
quite similar, as would be expected (Tables 7 and 8). The general trends
are the same in both cases. Galactose gives greater inhibition than
glucose. The two 1-6 linked disaccharides, melibiose and gentiobiose,
allow the greatest inhibition. No conclusions concerning anomeric link-
age can be drawn from the use of methylated derivatives of glucose or
galactose since the inhibitions provided are quite similar. From these
data (Tables 7 and 8) one can only conclude that galactose plays a major
role in the a-d specificity and a 1-6 linkage may be involved.
Some of the conclusions derived from these precipitin inhibition
studies are at variance with those recently published for strain B13 (d)
(42) and another serotype a organism, strain HS6 (46). These differences
will be presented below in the Discussion.
Methylation analysis. Results of the methylation analysis of the
purified antigens are shown in Tables 9 and 10. These data give us specific
information on the types of linkages in which all constituent sugars are
involved, and the relative frequencies of each linkage type in each
To determine the exact position of each residue in the polymer
would necessitate isolation and characterization of several disaccharides
and oligosaccharides produced by partial hydrolysis of the antigenic
polysaccharide. The large amounts of material required prevented us from
taking this classical approach: however, much information on polymeric
structure can be gained from methylation analysis alone.
In each polymer there are four residues in identical molar ratios
with respect to the branch-point. Methylation of each polymer yields
1,5-di-0-acetyl-2,3,4,6-tetra-O-methyl-D-galactitol (2,3,4,6 Gal),
INHIBITION OF THE PRECIPITIN REACTION BETWEEN
AHT ANTIGEN AND ANTI- a-d SERUM
Inhibitor pmola % Inhibition
Glucose 25 21
Galactose 25 38
c-methyl glucopyranoside 25 26
s-methyl glucopyranoside 25 23
a-methyl galactopyranoside 25 29
8-methyl galactopyranoside 25 23
a-methyl galactofuranoside 25 18
B-methyl galactofuranoside 25 4
Melibiose 10 47
Lactose 10 11
Cellobiose 10 33
Maltose 10 24
Gentiobiose 10 47
aAmount in a final volume of 1201l
INHIBITION OF THE PRECIPITIN REACTION BETWEEN
B13 ANTIGEN AND ANTI- a-d SERUM
Inhibitor umola % Inhibition
aAmount in a final volume of 120ul
p -+ c> + -+ -+ -+ -+ ;4-,
e c ,.. CM ,.I i-4 3 ,-.4 ,'- 3 -
LM ON N 01% M Lt
..4 N -4 -
c a r-o
S O O O
a si00 0C
,..44 0 0
-,-4 P4-i O 0
Q o 1 4i
H l I O C
0 Q Q u 0 0
z ai o 0 11
Sa U 3
9 0 (U
< 3 ,- 4 -4 0
ol I I I I I -
Cc I CO I a w -
41 4 0 -4 4- W CBr 00
E U U U 4- 1 .l.l (U 00 04J
> 0 4 I 4O
d M M m b M -
- 4Ij | 1 Q |4 1 0-
St i il O w II
0 CIS W m: co th S U
I 4-) 4-j 4 4-1i coi 4
a) 0 10 0 a T > u c U
W~ La .r-4 .
-i i i i e I a-I
I I '0 5 5
.0 LIn 4 i 1 L I ) 4h o o al
E (U C -4r I ^o i01 0 C u
U U l l Nl rN l O O l(C
I | >, U U U ? U (U I < ll
g CU 3 ) 4B CU t8 4j 4.1
s-a (U | I II (U | U
4.1 4-i i4 i-i -4 j i. i) (U<) a3
cr% 'o '0 i 4+i 4-i 4.1 I 4l 7-C -i-4
1 I I I *i4 I 4- U5-i -
r7 N L | L C L C ri U i-I Q)
o cM L 0
c5 I a0l C !a o
6-~~~~ t- m alillll- u a, s
-+ -C% -+ C
E--4 "I co 00
E-4 c -1 C-4r-1 4 3 i- 3 r
1 f i M < 00 0
ca r wCJ
O 0 w
0 0 0 0 ~0
--0 W Li0 00- 4 0
6 il u -1 4 c
m m cc -1 u 0 4r- Z
1-4 1 -41 0 Cu 2d
o o I
a a1 o uc
4 My .1 .4l CO 0
6j -' ao 0 00 >4
EO l4 CO1 O -O (
0 I 00 I -4 CO 0 C
0I I I I
O > cn V
E0 i 4- N C C N u
I I I I B- 0
u c u o O c a
> -41 o n -14 II
2 U 4 U .0 .0 .0 .- 0 C-
o< N >N >N 0
cr3I 1 0 00 02 00 N 1
aD a Q i u o W
- I I I I Z
I I 4 .j Q 3 > 41
4. 4. r4 -r-4 0 = a)
I I w I I I >
0 <- (i) aj 4-j 41. 4-1 W-i 4
O 0 JW J
M u e U U JO
en l- 1 U 0 0 0 0 > 1 < a4J
0 I 0- t i 1- 1- 4- o Mh~ 0
<< a- h h 4 4-h r4 aO a
z > -d WI
M1 L3I C) U- 4) 4J-
EO 0 O i- 0 i-
'"I ~~~~~ ~ c~ \0>> | 4 4 J1- l--
representing a terminal galactose residue, 1,5,6-tri-0-acetyl-2,3,4-tri-
O-methyl-D-glucitol (2,3,4 Glu), representing glucose linked at C-1 and
C-6, and 1,5,6-tri-0-acetyl-2,3,4-tri-O-methyl-D-galactitol (2,3,4 Gal)
which represents a galactose residue linked at C-1 and C-6. Finally,
both polymers yield the same branch-point derivative which has been
identified as either 1,3,5,6-tetra-0-acetyl-2,4-di-0-methyl-D-galactitol
(2,4 Gal) or 1,2,4,6-tetra-O-acetyl-3,5-di-0-methyl-D-galactitol (3,5 Gal).
These two derivatives are actually identical since they are symmetrical.
This derivative is obtained from the methylation of a galactopyranose
residue linked at C-1,3, and 6 or a galactofuranose residue linked at
C-1,2, and 6. These two particular galactose residues yield the same
derivative when methylated and reduced and can only be differentiated by
tagging C-l by reduction with sodium borodeuteride.
The four residues mentioned above may comprise a common sequence
of sugars in each polymer. It will be postulated that the terminal galac-
tose residues attached at the galactose branch-point compose part or all of
the a-d reactive site since these residues are common to both polymers, and
inhibition studies indicate that alactose may be the principal determinant of
a-d specificity (see Discussion).
There are two 1-3 linked moieties in each polymer in nearly equal
molar ratios with respect to the branch-point: a 1-3 linked galactose residue
in the AHT polymer (1.9:1), and 1-3 linked glucose in the B13 polymer (1.5:1).
The 1-3 linkage of these residues may produce pronounced "buckling" of the
polymeric chain allowing these and neighboring residues to be presented as a
sterically-accessible antigenic unit to antibody molecules. It will be pro-
posed (see Discussion) that these as well as neighboring residues constitute
the a and d determinants in the respective polymers. The AHT polymer
contains the 1-3 linked galactose residue and inhibition studies described
above show that a galactose-glucose sequence gives the greatest inhibition,
while the B13 polymer contains the 1-3 linked glucose residue and inhibition
studies suggest a glucose-glucose sequence is involved.
Another feature of both polymers that becomes evident from the
methylation studies is the presence of galactofuranose residues.
Methylation of the ART polymer yields 1,3,4-tri-0-acetyl-2,5,6-tri-0- methyl
galactitol (2,5,6 Gal) which represents a galactofuranose residue linked at
C-l and 3 to neighboring sugars. The B13 polymer yields 1,4,6-tri-O-acetyl-
2,3,5-tri-0- methyl galactitol (2,3,5 Gal) upon methylation, which represents
a galactofuranose residue linked at C-l and 6 in the native polymer.
Furanosidic linkages are known to be much more sensitive to acid hydro-
lysis than the corresponding linkages between pyranosides (3,50,52).
A study using immunoelectrophoresis was undertaken to demonstrate
the acid liability of these polymers and to compare the apparent greater
mobility of formamide extracts with products yielded by Lancefield acid
extractions of B13. Samples of B13 cell walls were hydrolyzed at pH 2.5
according to the procedure outlined in Methods, except that samples were
hydrolyzed for 10,20,40 or 60 min before neutralization. Fig. 14a and b
show the results of the hydrolysis of B13 cell walls at pH 2.5. After 10
min, essentially a single arc is seen. With increasing hydrolysis times,
however, a "gull-wing" type of precipitin band is obtained with the
faster-moving portion of the band becoming more evident and possessing
a mobility similar to that of the formamide extract shown in the center
well. To determine whether this represents the degradation of the polymer
or the extraction of a second antigen from the cell wall, the supernatant
Immunoelectrophoretic study of the degradation of Lancefield
extracts of B13 cell walls at various extraction times. In
Fig. 14a and Fig. 14b, cell walls were extracted for 10,20,40,
and 60 min and compared to a formamide extract (FE). In Fig.
14c the supernate from a 10 min Lancefield extract was hydro-
lyzed for an additional 30 min (10 + 30). In Fig. 14d the
supernate from a 10 min Lancefield extract was subjected to
formamide extraction (FE of LE). Unabsorbed anti-B13 serum
was used in all troughs.
from the 10 min extract which had been neutralized and centrifuged to
remove cell walls was readjusted to pH 2.5 and boiled for an additional
30 min. The appearance of the familiar "gull-wing" pattern upon additional
hydrolysis (Fig. 14c) suggests that this phenomenon is likely due to size
degradation or an alteration of the charge of the polymer and not due to
extraction of a second antigen from the wall. Likewise, if the neutralized
10 min supernatant is subjected to formamide extraction (Fig. 14d), further
degradation of the polymer is apparent.
The same type of experiment was repeated using AHT walls (Fig. 15).
Again, degradation can be seen with increasing hydrolysis time. When the
neutralized, cell wall-free supernatant of the 10 min acid extract is
subjected to formamide extraction, the conversion to a faster moving
component is seen as in the case of B13 (Fig. 15c).
Figure 15. Repeat of the experiment shown in Fig. 14 using AHT
Lancefield cell wall extracts (10,20,40, and 60 min)
and AHT formamide extract (FE). Time degradation of
AHT antigen showing the appearance of a faster moving
component (Fig. 15a and Fig. 15b). Supernate from a
10 min Lancefield extract was further treated by
formamide extraction (FE of LE) and compared to a
formamide extract of cell walls (FE) and a 10 min
Lancefield extract (LE, Fig. 15c). All troughs were
filled with unabsorbed anti-AHT serum.
Two antigenic polysaccharides were extracted and purified from
cell walls of the cross-reactive strains S. mutans AHT (a) and S. mutans
B13 (d). The antigens, extracted by the hot formamide method, were
purified by affinity chromatography on columns containing the galactose-
specific lectin from the castor bean (ricin) and found to be diheteroglycans
consisting primarily of galactose and glucose. Antigenic specificities
of both the specific and the cross-reactive sites on each polymer were
determined, and methylation analysis revealed structural similarities
that may reflect the nature of the cross-reactive sites and differences
that may reflect the natures of specific haptenic regions.
At the present time there are 5nly three papers dealing with the
immunochemistry of these two serotypes (41, 42, 46). Two points must be
considered when comparing the present study with the previously published
data: first, the starting materials (cell walls us whole cells) and the
methods of extraction and purification will undoubtedly affect not only
the size of the polymer extracted but also the amounts and types of
materials that may be associated with the purified products (peptidoglycans,
lipoteichoic acids, etc.) but which do not contribute to the antigenic
reactivity of the product. These differences should not, however, affect
the specificity of the combining sites with respect to antibodies when
measured by precipitin inhibition assays. Second, one is limited by the
availability of appropriate potential inhibitors particularly disaccha-
rides and higher oligosaccharides for use in precipitin inhibition studies.
The results of the inhibition of precipitation of the AHT antigen
with anti-a serum show that galactose is the best monosaccharide inhibitor
(44%), with melibiose and lactose also giving relatively high degrees of
inhibition. Since methylation analysis indicates the absence of 1-4
linkages in the purified AHT antigen, the inhibition by lactose is probably
not a function of its linkage pattern. But the presence of 1-6 linkages in
the antigen may explain the efficacy of melibiose in the inhibition (46%)
of the homologous reaction. Both disaccharides, however, contain galactose
which undoubtedly is located at the a-specific haptenic site. The use of
methylated galactopyranosides and galactofuranosides in inhibition studies
failed to reveal the probable anomeric linkages of galactose moieties at
the antibody binding site.
The a antigen isolated by Mukasa and Slade (46) from hot water
extracts of whole cells of HS6, differs in chemical composition from the
AHT antigen in the molar ratio of galactose to glucose (5.2:1 for HS6 (26)
us 3.4:1 for AHT) but appears to behave similarly in gel diffusion and
immunoelectrophoresis. Each product apparently possesses both an a and
a-d reactive site. Inhibition studies on the a site of the HS6 poly-
saccharide were carried out by Mukasa and Slade (46) with only a limited
number of inhibitors. Significant inhibition was obtained with glucose
(25%), and the diglucosides maltose (51%), and gentiobiose (49%). The
only galactose-containing inhibitor used, lactose, yielded less than 10%
inhibition. All inhibitors were compared on an equal weight rather than
equimolar basis. From these limited data they concluded that either
a-1-4 or a 0-1+6 configuration involving only glucose was responsible for
the a specificity in HS6. Their failure to employ galactose or galactosides
in their studies makes interpretation of their findings difficult.
Certainly, the present study of the AHT antigen fails to confirm the
immunochemical nature of the serotype a antigen presented by Mukasa and
Slade (46). Confidence in data presented in this dissertation is provided
by the method of antigen extraction (purified cell walls rather than whole
cells), the process of antigen purification (affinity chromatography) and
the wider range of monosaccharides and disaccharides tested in precipitin
inhibition analyses. It would appear, therefore, that further studies of
serotype a polysaccharides may be required to present a more coherent
picture of this important antigenic polysaccharide.
With respect to the specificity of the d site on the B13 antigen,
the results in Table 6 suggest that the major determinant may be glucopyrano-
side that is linked 1-6 to another glucose or galactose residue, possibly
by means of a B-linkage as indicated by the excellent inhibition obtained
using gentiobiose (50%) and B-methyl glucopyranoside (48%). Linzer and
Slade (42) meanwhile purified an antigen from cold, 5% trichloroacetic acid
(TCA) extracts of whole cells of B13. This antigenic polysaccharide had
about the same molar ratio of galactose: glucose (1.9:1) as the B13 antigen
isolated and purified in this study (2.1:1). Inhibition of the precipitation
of the Linzer-Slade antigen (42) with anti-d serum gave the following
results: galactose (31%), glucose (5%), rhamnose (0%), lactose (9%), and
melibiose (4%). Again, inhibitors were used on an equal weight (500 ug)
basis rather than an equimolar basis, despite the fact that a selection of
mono- and disaccharides with different molecular weights were being com-
pared. It was also reported that galactosamine, glucosamine, melibiose,
cellobiose, stachyose([Gal-a-l-6-] -Glu-a-1l2- fructofuranose), and
maltose inhibited less than 5%, again using 500 ug amounts of each. The
discrepancy between the published results and those of the present study
cannot readily be explained. There are differences in extraction techniques
(formamide on purified cell walls here vs 5% TCA on whole cells (42)),
purification (affinity chromatography here vs gel and ion-exchange chroma-
tography), and in the organisms used to absorb the whole B13 serum to yield
anti-d immunoglobulins (AHT cells here vs HS6 (42)). Again, it is concluded
there is much doubt as to the true nature of the d determinants, and further
studies with additional strains of the serotype are suggested to help re-
solve the discrepancies.
There is, however, agreement as to the apparent specificity of the
a-d site; namely galactose (probably in the pyranose configuration). This
specificity was found using both the AHT and B13 antigens in this study,
and in HS6 by Mukasa and Slade (46). (The specificity of the a-d site of
the B13 antigen purified by Linzer and Slade (42) was not determined in
that study). Tables 7 and 8 reveal pronounced inhibition by melibiose
(72% with B13 and 47% with AHT, each at 10 moless. These data suggest
that the galactose determinant of a-d reactivity may reside at the non-
reducing end of a branch of the main polymeric chain since melibiose con-
tains galactose linked a-1-6 to glucose.
The data provided by methylation analysis of purified polysaccha-
rides provide much useful information concerning the chemical composition
and linkage patterns of each antigen, and when considered with the inhi-
bition data, likely structures for each determinant can be postulated.
Although methylation studies do not provide information concerning the
sequential arrangement of sugar residues, nor data concerning anomeric
linkages, it is possible to derive plausible structures. Figure 16 shows
a schematic arrangement of sugar residues in the AHT and B13 antigens
taking into account the more important features suggested by the
-~ ~ ~
methylation analyses and inhibition data. All residues present in each
polymer are not shown; only the suggested common (a-d) and specific (a
and d) haptenic areas are detailed.
Both polymers possess the same branch-point: a galactose residue
linked at C-1,3, and 6 (or a galactofuranose residue linked at C-1,2, and
6). Also, both polymers probably contain a terminal galactose moiety that
is always linked to this branch-point since they are present in equimolar
amounts. Two other common residues exist: almost 2 moles (relative to the
branch) of glucose linked at C-l and C-6 plus 1.4 moles of galactose linked
at the same carbon atoms. The sharing of a terminally-linked galactose
residue, and the fact that galactose was the best inhibitor of anti-a-d
activity, makes the terminal galactose linked to the galactose branch-point
a likely arrangement for the a-d reactive site. There may be a residue
or residues intervening between the terminal galactose and the branch-
point (perhaps the common non-terminal galactose linked at C-l and C-6),
however, this cannot be determined at present.
Another important feature found in each polymer is the presence of
1-3 linked galactose (AHT), and glucose (B13). D.A.R. Simmons (55, 56)
in studying molecular models of ShigZela flexneri 0-specific side chains
found that 1-2 and 1-3 linkages on rhamnose residues caused marked buckling
of the polymers, which "broke" the primary chains into sterically discrete
repeating sequences which correspond to the antigenic sequences determined
by serological procedures. It was also noted that these 1-2 or 1-3 linkages
were not involved in determining the antigenic specificity of these deter-
minants but only served to set them apart as sterically discrete and
accessible sequences. Simmons proposed that these determinants should be
called "apodeterminants" because "by conferring a distinctive shape on
the molecule, they determine the sequence and steric accessibility of the
more distant structures that compromise the antigenic determinant without
themselves forming an integral part of it" (56). He proposed that the
determinants themselves should be designated as "non-terminal determinants."
The presence of a galactose residue linked at C-1 and C-3 in the AHT polymer
and the fact that inhibition studies show that galactose (possibly linked
1-6 to glucose) is the more inhibitory hexose, support the arrangement shown
in Fig. 16. Likewise, the B13 polymer possesses a glucose residue linked
at C-l and C-3 and inhibition studies show that glucose is the better inhib-
itor of anti-d serum activity. Whether the d determinant involves a glucose-
glucose or a glucose-galactose sequence in a 1-6 linkage is not clear,
however both 1-6 linked glucose and galactose residues are found in the
813 polmer. In Fig. 16 the glucose-glucose (1-i6) linkage is suggested in
light of the fact that glucose residues linked at C-1 and C-6 are so common
in this polysaccharide.
Finally, each polymer possesses galactofuranose residues at some
location within the chain, and these residues may account for the acid
liability of these antigens, as was discussed previously.
More precise determinations of the actual sequence of sugars in
each polymer will necessitate isolation of oligosaccharides by partial
acid or enzymatic degradation, the determination of their structures, and
use of these authentic oligosaccharides as potential inhibitors.
Detailed studies of the fine structure of the S. mutans antigens
will aid in our greater understanding of the serological relationship
among the various strains, and may lead to further subdivision of strains
on a serological basis. The precise knowledge of these relationships also
will aid in epidemiological studies, and ultimately in the selection of
appropriate strains for use in future vaccination programs.
1. Augestad, I.A., and E. Berner. 1954. Chromatographic separation of
anomeric glycosides. II. New crystalline methylfuranosides of
galactose, arabinose, and xylose. Acta Chem. Scand. 8:251-256.
2. Augestad, I.A., E. Berner, and E. Weigner. 1953. Chromatographic
separation of anomeric glycosides. Chem. and Ind. 1953:376-377.
3. Barker, S.A., C.N.D. Cruickshank, and J.H. Holden. 1963. Structure of
a galactomannan-peptide allergen from Tricophyton mentagrophytes.
Biochim. Biophys. Acta 74:239-246.
4. Bjorndal, H., C.G. Hellerqvist, B. Lindberg, and S. Svensson. 1970.
Gas-liquid chromatography and mass spectrometry in methylation
analysis of polysaccharides. Angew. Chem. Internat. Edit.
5. Bleiweis, A.S., R.A. Craig, S.E. Coleman, and I. van de Rijn. 1971.
The streptococcal cell wall: Structure, antigenic composition,
and reactivity with lysozyme. J. Dent. Res. 50:1118-1130.
6. Bleiweis, A.S., R.A. Craig, D.D. Zinner, and J.M. Jablon. 1971.
Chemical composition of purified cell walls of cariogenic
streptococci. Infect. Immun. 3:189-191.
7. Bleiweis, A.S., W.W. Karakawa, and R.M. Krause. 1964. Improved
technique for the preparation of streptococcal cell walls.
J. Bacteriol. 38:1198-1200.
8. Bleiweis, A.S., M.C. Taylor, J. Deepak, T.A. Brown, and J.R. Wetherell,
Jr. 1976. Comparative chemical composition of cell walls of
Streptococcus wutans. J. Dent. Res. 55:A103-A108.
9. Bratthall, D. 1969. Immunodiffusion studies on the serological
specificity of streptococci resembling Streptococcus mutans.
Odont. Revy 20:231-243.
10. Bratthall, D. 1970. Demonstration of five serological groups of
streptococcal strains resembling Streptococcus mutans. Odont.
11. Bratthall, D. 1970. Serological studies on Streptococcus mutans.
Odont. Revy 23:1-20.
12. Bratthall, D. 1972. Immunofluorescent identification of
Streptococcus mutans. Odont. Revy 23:181-196.
13. Carlsson, J. 1968. A numerical taxonomic study of human oral
streptococci. Odont. Revy 19:137-160.
14. Chen, P.A., Jr., T.Y. Toribara, and H. Warner. 1956. Micro-
determination of phosphorus. Anal. Chem. 28:1756-1758.
15. Colman, G., and R.E.O. Williams. 1967. Classification of non-
haemolytic streptococci. Int. J. Syst. Bacteriol. 17:306-314.
16. Coykendall, A.L. 1970. Base composition of DNA isolated from
cariogenic streptococci. Arch. Oral Biol. 15:363-368.
17. Coykendall, A.L. 1971. Genetic heterogeneity in Streptococcus
mutans. J. Bacteriol. 106:192-196.
18. Coykendall, A.L. 1974. Four types of Streptococcus mutans based on
their genetic, antigenic and biochemical characteristics.
J. Gen. Micro. 83:327-338.
19. Coykendall, A.L., O.P. Daily, M.J, Kramer, and M.E. Beath. 1971.
DNA DNA hybridization studies of Strevtococcus mutans.
J. Dent. Res. 50:1131-1139.
20. de Stoppelaar, J.D. 1971. Streptococcus mutans, Streptococcus
sanguis, and dental caries. Publication from the Department of
Preventieve Tandheelkunde, Ryksuniversiteit, Utrecht.
21. Dische, Z., and L.B. Shettles. 1948. A specific color reaction of
methyl pentoses and a spectrophotometric micromethod for their
determination. J. Biol. Chem. 175:595-603.
22. Drucker, D.B., and T.H. Melville. 1969. Computer classification of
streptococci mostly of oral origin. Nature (London) 221:664.
23. Drucker, D.B., and T.H. Melville. 1971. The classification of some
oral streptococci of human and rat origin. Arch. Oral Biol.
24. Dubois, M., K.A. Gilles, J.K. Hamilton, P.A. Rebers, and F. Smith.
1956. Colorimetric method for determination of sugars and
related substances. Anal. Chem. 28:350-356.
25. Edwardsson, S. 1968. Characteristics of caries-inducing human
streptococci resembling Streptococcus mutans. Arch. Oral Biol.
26. Fitzgerald, R.J., H.V. Jordan, and H.R. Stanley. 1960. Experimental
caries and gingival pathologic changes in the gnotobiotic rat.
J. Dent. Res. 39:923-935.
27. Fitzgerald, R.J., and P.H. Keyes. 1960. Demonstration of the
etiological role of streptococci in experimental caries in
hamsters. J. Am. Dent. Assoc. 61:9-19.
28. Fuller, A.T. 1938. Formamide method for the extraction of poly-
saccharides from hemolytic streptococci. Br. J. Exp. Pathol.
29. Gibbons, R.J. 1972. Ecology and Cariogenic Potential of Oral
Streptococci. p 373. In L.W. Wannamaker, J.M. Matsen (eds.),
Streptococci and Streptococcal Diseases. Academic Press, Inc.,
New York, New York.
30. Guggenheim, B. 1968. Streptococci of dental plaques.
Caries Res. 2:147-163.
31. Hakomori, S. 1964. A rapid permethylation of glycolipid, and poly-
saccharide catalyzed by methylsulfinyl carbanion in dimethyl
sulfoxide. J. Biochem. (Tokyo) 55:205-207.
32. Hamada, S., K. Gill, and H.D. Slade. 1976. Chemical and immuno-
chemical properties of the type f polysaccharide antigen of
Streptococcus mutans. Infect. Immun. 14:203-211.
33. Hamada, S., and H.D. Slade. 1976. Purification and Immunochemical
characterization of type e polysaccharide antigen of Streptococcus
mutans. Infect. Immun. 14:68-76.
34. lacono, V.J., M.A. Taubman, D.J. Smith, P.R. Garant, and J.R. Pollock.
1976. Structure and function of the type-specific polysaccharide
of Streptococcus mutans 6715. p 75. In W.H. Bowen, R.J. Genco,
and T.C. O'Brien (eds.), Immunologic aspects of dental caries.
Information Retrieval Inc., Washington, DC.
35. lacono, V.J., M.A. Taubman, D.J. Smith, and M.J. Levine. 1975.
Isolation and immunochemical characterization of the group-
specific antigen of Streptococcus mutans 6715. Infect. Immun.
36. Jablon, J.M., and D.D. Zinner. 1966. Differentiation of cariogenic
streptococci by fluorescent antibody. J. Bacteriol. 92:1590-1596.
37. Lancefield, E. 1933. A serological differentiation of human and
other groups of hemolytic streptococci. J. Exp. Med. 57:571-595.
38. Lindberg, B. 1972. Methylation analysis of polysaccharides.
In: S.P. Colwick and N.O. Kaplan (eds.), Methods in Enzymology.
39. Linzer, R. 1976. Serotype polysaccharide antigens of Streptococcus
mutans: Composition and serological cross reactions. p 91-99.
In W.H. Bowen, R.J. Genco, and T.C. O'Brien (eds.), Immunologic
aspects of dental caries, Information Retrieval Inc.,
40. Linzer, R., K. Gill, and H.D. Slade. 1976. Chemical composition of
Streptococcus mutans type c antigen: Comparison to type a, b, and
d antigens. J. Dent. Res. 55A:109-115.
41. Linzer, R., H. Mukasa, and H.D. Slade. 1975. Serological purification
of polysaccharide antigens for Streptococcus mutans serotypes a
and d: Characterization of multiple antigenic determinants.
Infect. Immun. 12:791-798.
42. Linzer, R., and H.D. Slade. 1974. Purification and characterization
of Streptococcus mutans group d cell wall polysaccharide antigen.
Infect. Immun 10:361-368.
43. Lowry, O.H., N.R. Roberts, K.Y. Leiner, M. Wu, and A.L. Farr. 1954.
The quantitative histochemistry of the brain. J. Biol. Chem.
44. Lowry, O.H., N.J. Rosebrough, A.L. Farr, and R.J. Randall. 1951.
Protein measurement with the Folin phenol reagent. J. Biol.
45. Mukasa, H., and H.D. Slade. 1973. Structure and immunochemical
specificity of the Streptococcus mutans group b cell wall
antigen. Infect. Immun. 7:578-585.
46. Mukasa, H., and H.D. Slade. 1973. Extraction, purification, and
chemical and immunochemical properties of the Streptococcus
mutans group "a" polysaccharide cell wall antigen. Infect.
47. Nicolson, A.L., and J. Blaustein. 1972. The interaction of Ricinus
communis agglutinin with normal and tumor cell surfaces.
48. Perch, B., E. Kjems, and T. Ravn. 1974. Biochemical and serological
properties of Streptococcus mutans from various human and animal
sources. Acta. Path. Microbiol. Scand. Sec. B. 82:357-370.
49. Pierce Chemical Co. 1972. Handbook of Silylation, p 48.
Pierce Chemical Co., Rockford, Ill.
50. Preston, J.F., III, and J.E. Gander. 1968. Isolation and partial
characterization of the extracellular polysaccharides of
Penicllium charlesii. Arch. Biochem. Biophys. 124:504-512.
51. Rantz, L.A., and E. Randall. 1955. Use of autoclaved extracts of
hemolytic streptococci for serological grouping. Standford
Med. Bull. 13:290-291.
52. Roberts, W.K., J.G. Buchanan, and J. Baddily. 1962. Galactofuranose
units in the specific substance from type 34 Pneumococcus.
Biochem. J. 82:42P.
53. Sanford, P.A., and H.E. Conrad. 1966. The structure of the Aerobacter
aerogenes A3(SL) polysaccharide. I. A reexamination using improved
procedures for methylation analysis. Biochem. 5:1508-1517.
54. Shklair, I.L., and H.J. Keene. 1974. A biochemical scheme for the
separation of the five varieties of Streptococcus mutans.
Arch. Oral Biol. 19:1079-1081.
55. Simmons, D.A.R. 1971. Immunochemistry of Shigella flexneri 0-antigens:
a study of structural and genetic aspects of the biosynthesis of
cell-surface antigens. Bact. Rev. 35:117-148.
56. Simmons, D.A.R. 1971. Sterochemical aspects of antigenic specificity
in polysaccharide determinants. Eur. J. Biochem. 18:53-58.
57. Slade, H.D., and W.C. Slamp. 1972. Peptidoglycan composition and
taxonomy of group D, E, and H streptococci, and Streptococcus
mutans. J. Bacteriol. 109:691-695.
58. Spackman, D.H., W.H. Stein, and S. Moore. 1958. Automated recording
aparatus for use in the chromatography of amino acids. Anal.
59. Van de Rijn, I., and A.S. Bleiweis. 1973. Antigens of Streptococcus
mutans. I. Characterization of a serotype-specific determinant
from Streptococcus mutans. Infect. Immun. 7:795-804.
60. Vaught, R.M., and A.S. Bleiweis. 1974. Antigens of Streptococcus
mutans. II. Characterization of an antigen resembling a
glycerol teichoic acid in walls of strain BHT. Infect. Immun.
61. Wetherell, J.R., Jr., and A.S. Bleiweis. 1974. Antigens of
Streptococcus mutans: Characterization of a polysaccharide antigen
from walls of strain GS-5. Infect. Immun. 12:1341-1348.
62. Wetherell, J.R., Jr., and A.S. Bleiweis. 1978. Antigens of
Streptococcus mutans: Isolation of a serotype-specific and a
cross-reactive antigen from walls of strain V-100 serotypee e).
Infect. Immun 19:160-169.
63. Zinner, D.D., and J.M. Jablon. 1968. Human streptococcal strains
in experimental caries. In Art and science of dental caries
research, R.H. Harris (ed.), p 87. Academic Press, Inc.,
64. Zinner, D.D., J.M. Jablon, A.P. Aran, and M.S. Saslaw. 1965.
Experimental caries induced in animals by streptococci of
human origin. Proc. Soc. Exp. Biol. Med. 118:766-770.