- Email: [email protected]
Immunochemical Characterization of Type i Carbohydrate Antigen of “Streptococcus milleri” (Streptococcus anginosus)
Zbl. Bakt. 274, 40-49 (1990) © Gustav Fischer Verlag, Stuttga rt/New York
Immunochemical Characterization of Type i Carbohydrate Antigen of "Streptococcus milleri" (Streptococcus anginosus) R YOSUKE KONAGA WA, TSUYOSHI YAKUSHIjI, and MASAKAZU INOUE"" Department of Preventive Dentistry, Kagoshima University Dental School, Kagoshima 890, Japan
With 4 Figures ' Received Februar y 9, 1990 . Accepted in revised form May 8, 1990
Summary Type-specific carbohydrate antigen of the serotype i "Streptococcus milleri" was extracted with trichloroacetic acid from purified cell walls of the type strain K39K. The extracts were then purified by chromatography on DEAE-Sephadex A-25 and Sephadex G100 columns. The purified serotype i carbohydrate antigen produced a single precipitin band against its homologous type-specific antiserum, which fused with the band produced by the autoclaved extract of the type strain cells. The serotype i antigen was a polysaccharide composed of rhamnose, galactose and glucose in a molar rat io of 1.6: 6.8 : 1.0. The quantitative precipitin inhibit ion test with variou s haptenic sugars showed that galactose as well as lactose produced the greatest inhibition, which suggested that a galactose in terminal beta linkage is the immunodeterminant of the serotype i-specific antigen. Galactose was detected in the autoclaved extracts from cells of all the 15 serotype i strains tested. Zusammenfassung Das typenspezifischen Kohlenhydrat -Anrigen des "Streptococcus milleri", Serotyp i, wurde mittels T richloressigsaure aus gereinigten Zellwanden des Typenstammes K39K extrahiert. Die Extrakte wurden dana ch durch DEAE-Sephadex A25- und Sephadex G-lOO Saulenchromarographie gereinigt. Das gereinigte, typenspezifische i-Antigen bildete eine einzige Prazipitationsbande mit dem homologen Antiserum, die mit der Bande fusionierte , die von autoklavierten Extrakten des Typenstammes gebildet wurde . Das Serotyp i-Antigen erwies sich als ein Polysaccharid aus Rhamnose , Galaktose und Glukose im Verhaltnis von 1,6: 6,8 : 1,0. 1m Prazipitationshemmtest erwiesen sich Galaktose und Laktose als die starksten Inhibitoren. Damit liegt die Annahme nahe, daB Galaktose in terminaler B-Bindung die Immunodominante des i-Antigens ist, Galaktose wurde in den Autoklaven-Extrakten sarntIicher 15 Serotyp i-Stiimme nachgewiesen.
* Corresponding author
Carbohydrate Antigen of "5. milleri"
41
Introduction The organisms referred to as "Streptococcus milleri" [Streptococcus anginosus (3)] are indigenous flora in humans. Clinical significance of the streptococcal group has been increasingly documented (7, 17). Over the past two decades, however, continuing confusion has surrounded the taxonomy and nomenclature of this group of commensal streptococci (5), and the epithet is not included in the approved list of bacterial names (20). Regardless of the confusion, the organisms known as "S. milleri" share a core of physiological traits (2). Serological classification of this clinically important commensal would be of great value not only in terms of taxonomy but also for ecological and epidemiological reasons. It has been shown that "S. milleri" strains often carry the Lancefield group antigens A, C, F, or G, (2, 13, 19) and /or the Ottens type antigens I to V (15). Chemical and immunological properties of the group and type carbohydrate antigens have been studied (14, 23-26) . However, it has not been possible to group or to type many isolates from various clinical specimens (2, 13). We have recently isolated a large number of "S. milleri" strains from human dental plaque (27) and demonstrated wide serological variation in these isolates on the basis of their cell surface carbohydrate antigens (28). Our previous study (29) revealed that the immunodeterminant of the type-specific carbohydrate antigen of serotype b "S. milleri" is a rhamnose residue . This paper describes the isolation and purification of carbohydrate antigens from th e type strain of serotype i "S. millen". The chemical composition and immunological properties of the type-specific antigen are examined. Materials and Methods
"S. milleri" strains. 15 strains of serotype i "S. milleri" previously isolated from human oral cavities (28) were used. They were grown anaerobically at 37°C for 18 h in brain heart infusion broth (BHI, Difco Laborator ies, Detroit, MI, USA) supplemented with glucose at a concentration of 1%. They were then washed three times with distilled water and lyophilized as described previously (28). Extraction of anitgen from purified cell walls. First, purified cell walls were prepared from the serotype i type strain K39K [biotype Ia (27)] by the procedures described previously (29). Briefly, whole cells were disrupted in a mechanical cell homogenizer (model MSK; B. Braun Apparatebau, Melsungen, Germany) with the aid of glass beads (0.17 to 0.18 mm in diameter). The cell walls were collected by centrifugation (26000 g for 60 min at 4°C) and then treated with trypsin (0.4 mg/ml; 1: 250, Difco) at pH 7.0 at 37°C for 2 h. The digested cell walls were extensively washed in distilled water by centrifugation and then lyophilized. The yield of the purified cell walls was 1.7 g from 6.3 g (dry weight) of whole cells. Carbohydrate antigen was extracted from the purified cell walls (1.5 g, dry weight) in 50 ml of 5% trichloroacetic acid (TCA) at 90 °C for 15 min with continuous stirring. The reaction mixture was centrifuged (15 000 g for 20 min) to collect the supernatant and the extraction was repeated three times on residual cell walls. The supernatants thus obtained were combined, neutralized with 2 N NaOH and dialyzed against distilled water. The nondializable portion was concentrated in a rotary evaporator at 50°C. Acetone was added to the concentrate to a final concentration of 80% and stored at - 20 °C for 18 h. The precipitated carbohydr ates were dissolved in a small volume of distilled water and dialyzed extensively against 0.05 M (NH4 h CO J • The dialysate was finally clarified by centrifugation (crude TCA extract).
42
R. Konagawa, T. Yakushiji, and M. Inoue
Purification of type antigen . The TCA extract (ca. 2 ml) was applied to a DEAE-Sephadex A-25 (Pharmacia Fine Chemicals, Uppsala, Sweden) column (1.5 by 45.5 ern). The column was first washed with 0.05 M (NH 4)zC03 (100 ml) and then eluted with a linear gradient of o to 1 M NaCI (150 ml + 150 ml) in the starting buffer at a flow rate of 45 ml/h, The fractions (12 ml each) reactive with the type i-specific antisera were combined, dialyzed against distilled water and finally concentrated in a rotary evaporator at 50 "C. The concentrate (5 ml) was applied to a Sephadex G-100 (Pharmacia) column (1.5 by 45 em) and eluted with distilled water at a flow rate of 8 ml/h. The immunologically active fractions (2.4 ml each) were combined and lyophilized. The yield of purified carbohydrate antigen was 40 mg from 1.5 g of cell walls. Autoclaved extraction of antigen from whole cells. Cell surface carbohydrate antigens were also extracted from whole cells by the method of Rantz and Randall (16). Whole cells (0.6 g, dry weight ) were heated at 121 °C for 20 min in 30 ml of saline and then removed by centrifugation. The supernatant was dialyzed against distilled water and lyophilized (designated Rantz-Randall antigen ). Chem ical analyses. Neutral sugars and amino sugars were quantitatively analyzed by gasliquid chromatography according to the methods of Griggs et al. (8) and Sloneker (21). Briefly, 1-5 mg portions of antigen preparations, together with inositol (0.5 mg) as an internal standard, were hydrolyzed in 2 ml of 3 N HCl at 100°C for 4 h. After neutralizing
7 .0
1.0
6 .0
i.....
c:
.!
0.5 ~
....,!. ....
E "CI
...CI
5.0
(J
o
4.0
III
Z
.....E Q)
III
o
)(
Q)
s: 1.0
0.10
III
....
'0
I-
E
0 .5
0.05 Q .....E
o Fraction
20
30
number
(12ml/tube)
40
0
Fig. 1. DEAE-Sephadex A-25 ion-exchange chromatography of the TCA extract of K39K (i) cell walls : The column (1.5 by 45.5 ern) was eluted first with 0.05 M (NH 4)zC03, followed by a linear grad ient of 0 to 1 M NaCl in the same buffer. P, phosphorus. The bar indicates the range where fractions were reactive with the type-specific antiserum.
Carbohydrate Antigen of "5. milleri"
43
with AgzC0 3 , the hydrolyzates were treated first with sodium borohydride and then with acetic anhydride. The alditol acetate derivatives thus obtained were analyzed by gas-liquid chromatography under the conditions reported previously (29). Total hexoses and N-acetyl hexosamines were estimated by the anthrone method (22) and by the method described by Ghuysen et al. (6), respectively. Phosphorus and protein were quantitated by the colorimetric methods of Lowry et al. (11) and Lowry et aI. (12), respectively. Immunological methods. The type i-specific antiserum used was the one previously prepared (28). Reactivities of the carbohydrate antigens with the type-specific antiserum were determined in capillary tube and/or by immunodiffusion in agar gels. Immunodiffusion was performed in 1% Noble agar (Difco) gels in 0.01 M phosphate buffer, pH 7.2 as described previously (28). Immunoelectrophoresis was performed at 10 rnA for 70 min on the gels in 0.05 M barbital buffer, pH 8.6 (10). Quantitative precipitin reaction and its inhibition by haptens were performed as described previously (29). To obtain the quantitative precipitin curve, the purified serotype antigen (0-5 ug) and the type-specific antiserum (5 ul) were incubated in the mixture adjusted to 50 !tl with saline at 37°C for 1 h and then at 4 °C for 18 h. The precipitate was collected by centrifugation (4000 g for 10 min) and estimated for proteins. For the hapten inhibition test, the serum (5 ul) was preincubated with 20 umol (unless otherwise described) of haptens in a total volume of 30 III at 3rC for 1 h and then the antigen was added. The reaction mixture without hapten served as the control. Results
Purification of type-specific antigen Hot TCA extracts of the purified K39~ (i) cell walls were fractionated on a DEAESephadex A-25 column (Fig. 1). Two car:bohydrate peaks were obtained, one was not adsorbed on the column and the other was eluted at 0.35 M NaC! . Only the latter peak fractions reacted with the type i-specific antiserum. The antigenic component was primarily composed of hexose moieties and small amounts of phosphorus were detec ted in the peak fractions . Fractions (No .17-21) which reacted strongly with the antiserum were combined, oncentrated and then applied to a Sephadex G-100 column (not shown) . The elution profile indicated a single major peak composed of hexose and phosphorus but no protein. The fractions (No. 11-42) reactive with the type-specific antiserum were combined, lyophilized and are hereafter designated type i antigen.
Immunochemical characterization of type-specific antigen The immunodiffusion in agar gels (Fig. 2) revealed that the purified type i antigen yielded a single precipitin line with !the type i-specific antiserum which was completely fused with the line given by the Rantz-Randall antigen or the TCA extracts of strain K39K . This indicated that all the antigen preparations contained an identical determinant. Immunoelectrophoresis (Fig. 3) showed that the purified type i antigen as well as the K39K Rantz-Rundall antigen were neutral at pH 8.6. Chemical analyses (Table 1) revealed that the purified cell walls contained significant amounts of various sugars, amino sugars, phosphorus and peptides. However, the purified type i antigen was composed of only galactose, rhamnose and glucose in a molar ratio of 6.8 : 1.6 : 1.0. A significant amount of phosphorus was also detected (2.0), but neither glucosamine, galactosamine, glycerol, ribitol nor proteins were detected.
44
R. Konagawa, T. Yakushiji, and M. Inoue
Fig. 2. Immunodiffusion in an agar gel of the serotype antigen: Antigens (25 Ill) and the type-specific antiserum (25 Ill) were allowed to react in 1.0% Noble agar gels in 0.01 M phosphate buffer, pH 7.2, overnight at room temperature . Well R, Rantz-Randall extract; well T, TeA extract; well P, purified type antigen; well A, type-specific antiserum.
Fig. 3. Immunoelectrophoresis of the serotype antigen: Immunoelectrophoresis was performed on 1.0% agar gel in 0.05M barbital buffer (pH 8.6) at 10 rnA for 70 min. The cathode was on the right. Well P, purified serotype antigen; Well R, Rantz-Randall extract ; well A, type-specific antiserum.
The maximum amount of antibody protein was precipitated when 0.75 Ilg of the type i antigen was allowed to react with 5 III of the homologous antiserum (not shown) . The quantitative precipitin inhibition test with various haptenic sugars (Table 2) showed that galactose and its related sugars inhibited the antigen-antibody reaction most effectively. Lactose induced complete inhibition. Galactose and lactose were the most effective inhibitors at the concentrations tested (Fig. 4).
Carbohydrate Antigen of "S. milleri"
45
Table 1. Chemical composition of the purified cell walls and serotype antigen of "Strep-
tococcus milleri" K39K
~g/mg
sample (molar ratio to glucose )
Components"
Purified cell walls
Serotype i antigen
Rhamnose Galactose Glucose Glucosamine* * Pept ide " * * Phosphoru s
443 (6.2) 165 (2.4) 70 (1.0)
150 633 93 0 0 33
75 (1.1) 174 (7. 1)
II (0.9)
(1.6) (6.8) (1.0) (0.0 ) (0.0) (2.0)
* Mannose, glycerol , ribitol and galactosamine were not detected. • * No N-acetyl hexosam ine was detected. *** The molar ratio in par enthes es is expressed as glycine. Table 2. Hapten inhibition of the precipit in reaction between the purified serotype i antigen and its homologous antiserum * Ha ptens
Inhib ition (%)
Glucose Galactose Rhamn ose Lacto se Melib iose Maltose Isomaltose Gluco samine Galac tosamine N-acet ylglucosamine N-acetylgalactosamine N-acetylma nnosamine a- Meth ylglucoside
36.8 82.9 56.6 100 .0 69 .7 26 .3 36. 8 23 .7 96.1 30 .3 55 .3 27.6 40 .8 59.2
~-M eth ylglucoside
a-Methy lgalactoside ~ - Meth ylgalacto side
71.1
85.5
* 5
~I portion of type-specific antiserum was incubated with 20 umo l of hapten for 1 h at 37 °C, and then 0.75 ~g of pur ified type antigen was added.
Chemical compositions of the Rantz-Randall antigen from serotype i whole cells As summarized in Table 3, the a utoclaved extracts from all the 15 serotype i strains contained rhamnose, galactose, glucose, and glucosamine in an average molar ratio of 4.0 : 0.9 : 1.0 :0.9. Significant amounts of ribitol and phosphorus were detected in the antigen extracts of all the strains but there was no glycerol. In immunodiffusion in agar gels, the extracts of all strains yielded a single band with the homologous type-specific anitserum (not sho wn) , wh ich wa s identical to that produced by the purified type antigen as shown in Fig. 2 .
46
R. Konagawa, T. Yakushiji, and M. Inoue
80
..... ~
~
60
-
40
e
.S!
:0 J:
s
20
o
1
5
10 Haptens
20 (umo t)
Fig. 4. Inhibition of the quantitative precipitin reaction by the constituent sugars of the purified serotype antigen: The experimental conditions were the same as given in Table 2, foot note, except for the concentration s of haptens. (A) rhamnose, (e) galactose, (_) glucose, (0) lactose.
Table 3. Contents of sugars and phosphorus in the Rantz-Randall extracts of the 15 serotype i "Strepto coccus miller;" strains Components
Mean molar ratio to glucose (range)
Rhamnose Galactose Glucose Glucosamine Ribitol Phosphorus
4.0 0.9 1.0 0.9 4.3
(1.2- 8.6) (0.2- 2.0)
(0.3- 2.4) (0.5- 22.1) 1.6 (0.7- 3.2)
* Mannose, glycerol and galactosamine were not detected.
Discussion As reported in the previous study (27), more than half of the oral "5. millen" isolates, like other non-haemolytic streptococci (4, 15), cannot be classified in the Lancefield serogrouping system (at least into groups A - G), but show wide serological variations in their cell surface carbohydrate type antigens (28). Recently, we have purified and immunochemically characterized serotype b-specific carbohydrate antigen . The antigen is found to be composed of rhamnose and glucose and its immunodeterminant terminal is a rhamnose residue (29). In the present study, serotype i-specific
Carbohydrate Antigen of "5. milleri"
47
antigens of "S. millen " were extracted with TCA, purified using column chromatography and characterized immunochemically. The type i antigen was composed of rhamnose, galactose and glucose (Table I). Of the three component sugars constituting the type antigen, galactose exhibited the highest inhibition at all concentrations tested in the quantitative precipitin reaction (Table 2 and Fig. 4), suggesting that the immunodeterminant terminal of the type i antigen of "S. millen" is a galactose residue. Moreover, the hapten inhibition test (Table 2) showed that I} among the disaccharides tested, lactose exhibited complete inhibition and melibiose also possessed a significantly higher effect, but maltose and isomaltose produced a weak inhibition; 2} among the hexosamines, N-acetyl hexosamines or a and ~-metyl glycosides, the galactose derivatives exhibited consistently greater inhibition than the glucose derivatives; 3} ~-methyl glycosides had higher inhibitory capacity than a-methyl glycosides. All of these findings suggest that the immunodeterminant of the serotype i-specific antigen possesses the ~-galactose- (I-4}-a-glucose- sequence (lactosyl residue). Galactosamine, which was not detected in the purified type i antigen (Table I), exhibited strong inhibition in the precipitin reaction between the type antigen and its homologous antiserum (Table 2). This inhibition, however, may have been due to a non-specific inhibition of the precipitin reaction by hexosamines, as observed with the reaction of Streptococcus sanguis b carbohydrate antigen and Lactobacillus plantarum cell wall teichoic acid with their respective antisera (1, 9). Similar non-specific inhibition by hexosamines has been known with the lectin-carbohydrate cell surface interaction between Actinomyces viscosus and S. sanguis (18). A significant amount of phosphorus was detected in the purified i antigen (Table 1). However, this does not necessarily indicate the involvement of teichoic acid in the antigenic determinant of the antigen . Neither ribitol nor glycerol was found in the purified antigen (Table 1) and the anti-i serum used in this study had no reactivity with a preparation of teichoic acid from Streptococcus sobrinus 6715-DP (not shown) . The presence of ribitol in the autoclaved extracts (Rantz-Randall antigen) from whole cells of serotype i strains may be due to the contamination with RNA although it has not been verified yet. The chemical background for the presence of phosphorus in the purified antigen preparation remains obscure at the present time. It is well known that "5. milleri " strains occasionally carry the Lancefield group antigens and/or the Ottens type antigen s (IS). Our previous study, however, has shown that the serotype i antigen can be serologically distinguished from the Lancefield group antigen s A to G (28). The serot ype i antigen (Table 1) possessed the same sugar composition as the Ottens type III antigen (24) of which the antigenic determinant is shown, however, to be ~-glucose (25, 26). Meanwhile, identification of the type i antigen and the Ottens type IV antigen appears more feasible. Willers et al. (23) have shown that this type antigen is compo sed of rhamnose, galactose and glucose as well as N-acetylglucosamine in a molar ratio of 4 : 4 : 4 : 1 and its immunodeterminant possesses a ~-galactose-glucose-galactose (or-rhamnose)- sequence. Thus, though carbohydrate compositions of these two antigens are not strictly identical, since no N-acetylglucosamine but galactose was detectable in the type i antigen (Table I), these type antigens appear to possess a lactose-like sugar sequence in a terminal position that renders them immunodeterminant. This possibility should be verified in further structural studies . All the Rantz-Randall antigens extracted from the 15 serotype i strains including the type strain possessed similar carbohydrate compositions (Tables 1 and 3). In imrnuno-
48
R. Konagawa, T. Yakushiji, and M. Inoue
diffusion in agar gel, precipitin lines with Rantz-Randall extracts from all the 15 serotype i strains were completely fused with the line formed between the typing serum and the antigen extract from the vaccine strain K39K (28). The results support the proposed chemical background of the immunodeterminant of serotype i "S. milleri" .
Acknowledgments. We thank Professor Dr. Shigeyuk i Hamada for providing a teichoic acid preparation and Dr. Richard M. Shelton for correcting the English version of the manuscript. We are also grateful to Miss Kayoko Yuge for her expert secretarial assistance.
References 1. Appelbaum, B. and B. Rosan: Antigens of Streptococcus sanguis: purification and characterization of the b antigen . Infect. Immun . 21 (1978) 896-904 2. Colman , G. and R. E. O . Williams: Taxonomy of some human viridans streptococci. In: Streptococci and Streptococcal Diseases (L. W. Wannamaker and j. M. Matsen , eds.), pp. 281-299. Academic Press, New York (1972) 3. Coykendall, A. L., P. M. Wesbecher, and K. B. Gustafson: "Streptococcus millen", Streptococcus constellatus, and Streptococcus intermedius are later synonyms of Streptococcus anginosus. Int . J. System Bact. 37 (1987) 222-228 4. Packlam, R. R.: Physiological differentiation of viridans streptococci.]. Clin. Microbiol.
5 (1977) 184-201 5. Facklam, R. R.: The major difference in the American and British streptococcus taxonomy schemes with special reference to Streptococcus milleri. Eur. J. Clin. Microbiol. 3 (1984) 91-93 6. Ghuysen, J. M., D. J. Tipper, and j. L. Strominger: Bacterial cell walls and bacteriolytic enzymes. In: Methods in Enzymology, vol. VIII (E. H. Neufeld and V. Ginsburg, eds.), pp . 685-699. Academic Press, New York (1966) 7. Gossling, J.: Occurrence and pathogenicity of the Streptococcus milleri group. Rev. Infect . Dis. 10 (1988) 257-285 8. Griggs, L. j., A . Post, E. R. White, j. A . Finkelstein, W. E. Moeckel, K. G. Holden, j. E. Zaremba, and]. A. Weisbach: Identification and quantitation of alditol acetates of neutral and amino sugars from mucins by automated gas-liquid chromatography. Anal. Biochem. 43 (1971) 369-381 9. Knox , K. W. and A. f. Wicken: Non-specific inhibition of the precipitin reaction between teichoic acids and antisera. Immunochem. 10 (1973) 93-98 10. Koga, T ., N. Okahashi, T. Yamamoto, J. Mizuno , M. Inoue, and S. Hamada : Purification and immunochemical characterization of Streptococcus sanguis ATCC 10557 serotype II carbohydrate antigen . Infect. Immun. 42 (1983) 696-700 11. Lowry, O . H., M. R. Roberts, K. Y. Leiner, M.-L. Wu, and A. L. Farr: The quantitative histochemistry of brain . I. chemical methods.]. BioI. Chern. 207 (1954) 1-17 12. Lowry, O . H., N. f. Rosebrough , A . L. Parr, and R. f. Randall: Protein measurement with the Folin phenol reagent . J. BioI. Chern. 193 (1951) 265-275 15. Liitticken, R., U. Wendorff, D. Liitticken, E. A. Johnson , and L. W. Wannamaker: Studies on streptococci resembling Streptococcus milleri and on an associated surfaceprotein antigen. J. Med . Microbiol. 11 (1978) 419-431 14. Michel, M. F. , I. van Vonn o, and R. M. Krause: Stud ies on the chemical structure and the antigenic determinant of type II antigen of group F streptococci. ]. Immunol. 102 (1969) 215-221 15. Ottens, H. and K. C. Winkler: Indifferent and haemolytic streptococci possessing group antigen F. J. Gen. Microbiol. 28 (1962) 181-191 16. Rantz, L. A . and E. Randall: Use of autoclaved extracts of hemolytic streptococci for serological grouping. Stanford Med . Bull. 13 (1955) 290-291
Carbohydrate Antigen of "S. milleri"
49
17. Ruoff, K. L. : Streptococcus anginosus ("Streptococcus millen"} : the unrecognized pathogen. Clin. Microbiol. Rev. 1 (1988) 102-108 18. Sato, S., T . Koga, and M. Inou e: A possible mechanism for the cellular coaggregation between Actinomyces viscosus ATCC 19246 and Streptococcus sanguis ATCC 10557. ]. Gen. Microbiol. 190 (1984) 1351-1357 19. Shlaes, D . M., P. I. Lerner, E. Wolinsky, and K. V. Gopalakrishna: Infections due to Lancefield group F and related streptococci (S. milleri, S. anginosus ). Medicine 60
(1981) 197-207
20 . Skerman, V. B. D., V. McGowan, and P. H. A. Sneath : Approved lists of bacterial names. Int. ]. System Bact. 30 (1980) 225-420 21. Sloneker,]. H.: Gas-liquid chromatography of alditol acetates. In: Methods in Carbohydrate Chemistry, Vol. 6 (R. L. Whistler and]. N. BeMiller, eds.), pp. 20-24. Academic Press, New York (1972) 22. Trevelyan, W. E. and]. S. Harrison : Studies on yeast metabolism . Biochem.]. 63 (1956)
23-33
23 . Willers, J. M. N ., M. F. Michel, and R. Benner: lmmunochemical studies of type IV and two group-like (Z) carbohydrate antigens of minute streptococci . Antonie v. Leeuwenhoek 39 (1973) 609-617 24. Willers,]. M. N., M. F. Mich el,]. H.]. Huis in't Veld, and G. H.]. Alderkamp: The type antigen III of group F streptococci ; separation of group and type antigens and partial characterization of type III antigen. Antonie v. Leeuwenhoek 39 (1973) 369-382 25. Willers,]. M. N., M. F. Michel , M . Sysma , and K. C. Winkler: Chemical analysis and inhibit ion reactions of the group and type antigens of group F streptococci. J. Gen. Microbiol. 36 (1964) 95-105 26. Willers,]. M. N. , H. Ottens, and M. F. Michel: Immunochemical relationship between Streptococcus MG, FIll and Streptococcus salivarius. ]. Gen. Microbiol. 37 (1964)
l-
425-431
27 . Yakushiji, T., M. Katsuki, A. Yoshimitsu, ]. Mizuno, and M. Inoue : Isolation and physiological characterization of Streptococcus milleri strains from human dental plaque. Microbios 55 (1988) 161-1 71 28 . Yakusbiji, T., R. Konagawa , M. ada, and M. Inoue : Serological variation of oral Streptococcus milleri. J. Med. Microbiol. 27 (1988) 145-151 29 . Yakusbiii, T., M. Inoue, and T. Koga: Purification and immunochemical studies of Type b carbohydrate antigen of oral Streptococcus milleri. Infect. lmmun. 56 (1988)
2264-2269
Prof. Dr. Masakazu Inoue , Department of Preventive Dentistry, Kagoshima University Dental School, 1208-1 Usuki-cho, Kagoshima 890, Japan
4 Zbl. Bakt. 274/1
Immunochemical Characterization of Type i Carbohydrate Antigen of "Streptococcus milleri" (Streptococcus anginosus) R YOSUKE KONAGA WA, TSUYOSHI YAKUSHIjI, and MASAKAZU INOUE"" Department of Preventive Dentistry, Kagoshima University Dental School, Kagoshima 890, Japan
With 4 Figures ' Received Februar y 9, 1990 . Accepted in revised form May 8, 1990
Summary Type-specific carbohydrate antigen of the serotype i "Streptococcus milleri" was extracted with trichloroacetic acid from purified cell walls of the type strain K39K. The extracts were then purified by chromatography on DEAE-Sephadex A-25 and Sephadex G100 columns. The purified serotype i carbohydrate antigen produced a single precipitin band against its homologous type-specific antiserum, which fused with the band produced by the autoclaved extract of the type strain cells. The serotype i antigen was a polysaccharide composed of rhamnose, galactose and glucose in a molar rat io of 1.6: 6.8 : 1.0. The quantitative precipitin inhibit ion test with variou s haptenic sugars showed that galactose as well as lactose produced the greatest inhibition, which suggested that a galactose in terminal beta linkage is the immunodeterminant of the serotype i-specific antigen. Galactose was detected in the autoclaved extracts from cells of all the 15 serotype i strains tested. Zusammenfassung Das typenspezifischen Kohlenhydrat -Anrigen des "Streptococcus milleri", Serotyp i, wurde mittels T richloressigsaure aus gereinigten Zellwanden des Typenstammes K39K extrahiert. Die Extrakte wurden dana ch durch DEAE-Sephadex A25- und Sephadex G-lOO Saulenchromarographie gereinigt. Das gereinigte, typenspezifische i-Antigen bildete eine einzige Prazipitationsbande mit dem homologen Antiserum, die mit der Bande fusionierte , die von autoklavierten Extrakten des Typenstammes gebildet wurde . Das Serotyp i-Antigen erwies sich als ein Polysaccharid aus Rhamnose , Galaktose und Glukose im Verhaltnis von 1,6: 6,8 : 1,0. 1m Prazipitationshemmtest erwiesen sich Galaktose und Laktose als die starksten Inhibitoren. Damit liegt die Annahme nahe, daB Galaktose in terminaler B-Bindung die Immunodominante des i-Antigens ist, Galaktose wurde in den Autoklaven-Extrakten sarntIicher 15 Serotyp i-Stiimme nachgewiesen.
* Corresponding author
Carbohydrate Antigen of "5. milleri"
41
Introduction The organisms referred to as "Streptococcus milleri" [Streptococcus anginosus (3)] are indigenous flora in humans. Clinical significance of the streptococcal group has been increasingly documented (7, 17). Over the past two decades, however, continuing confusion has surrounded the taxonomy and nomenclature of this group of commensal streptococci (5), and the epithet is not included in the approved list of bacterial names (20). Regardless of the confusion, the organisms known as "S. milleri" share a core of physiological traits (2). Serological classification of this clinically important commensal would be of great value not only in terms of taxonomy but also for ecological and epidemiological reasons. It has been shown that "S. milleri" strains often carry the Lancefield group antigens A, C, F, or G, (2, 13, 19) and /or the Ottens type antigens I to V (15). Chemical and immunological properties of the group and type carbohydrate antigens have been studied (14, 23-26) . However, it has not been possible to group or to type many isolates from various clinical specimens (2, 13). We have recently isolated a large number of "S. milleri" strains from human dental plaque (27) and demonstrated wide serological variation in these isolates on the basis of their cell surface carbohydrate antigens (28). Our previous study (29) revealed that the immunodeterminant of the type-specific carbohydrate antigen of serotype b "S. milleri" is a rhamnose residue . This paper describes the isolation and purification of carbohydrate antigens from th e type strain of serotype i "S. millen". The chemical composition and immunological properties of the type-specific antigen are examined. Materials and Methods
"S. milleri" strains. 15 strains of serotype i "S. milleri" previously isolated from human oral cavities (28) were used. They were grown anaerobically at 37°C for 18 h in brain heart infusion broth (BHI, Difco Laborator ies, Detroit, MI, USA) supplemented with glucose at a concentration of 1%. They were then washed three times with distilled water and lyophilized as described previously (28). Extraction of anitgen from purified cell walls. First, purified cell walls were prepared from the serotype i type strain K39K [biotype Ia (27)] by the procedures described previously (29). Briefly, whole cells were disrupted in a mechanical cell homogenizer (model MSK; B. Braun Apparatebau, Melsungen, Germany) with the aid of glass beads (0.17 to 0.18 mm in diameter). The cell walls were collected by centrifugation (26000 g for 60 min at 4°C) and then treated with trypsin (0.4 mg/ml; 1: 250, Difco) at pH 7.0 at 37°C for 2 h. The digested cell walls were extensively washed in distilled water by centrifugation and then lyophilized. The yield of the purified cell walls was 1.7 g from 6.3 g (dry weight) of whole cells. Carbohydrate antigen was extracted from the purified cell walls (1.5 g, dry weight) in 50 ml of 5% trichloroacetic acid (TCA) at 90 °C for 15 min with continuous stirring. The reaction mixture was centrifuged (15 000 g for 20 min) to collect the supernatant and the extraction was repeated three times on residual cell walls. The supernatants thus obtained were combined, neutralized with 2 N NaOH and dialyzed against distilled water. The nondializable portion was concentrated in a rotary evaporator at 50°C. Acetone was added to the concentrate to a final concentration of 80% and stored at - 20 °C for 18 h. The precipitated carbohydr ates were dissolved in a small volume of distilled water and dialyzed extensively against 0.05 M (NH4 h CO J • The dialysate was finally clarified by centrifugation (crude TCA extract).
42
R. Konagawa, T. Yakushiji, and M. Inoue
Purification of type antigen . The TCA extract (ca. 2 ml) was applied to a DEAE-Sephadex A-25 (Pharmacia Fine Chemicals, Uppsala, Sweden) column (1.5 by 45.5 ern). The column was first washed with 0.05 M (NH 4)zC03 (100 ml) and then eluted with a linear gradient of o to 1 M NaCI (150 ml + 150 ml) in the starting buffer at a flow rate of 45 ml/h, The fractions (12 ml each) reactive with the type i-specific antisera were combined, dialyzed against distilled water and finally concentrated in a rotary evaporator at 50 "C. The concentrate (5 ml) was applied to a Sephadex G-100 (Pharmacia) column (1.5 by 45 em) and eluted with distilled water at a flow rate of 8 ml/h. The immunologically active fractions (2.4 ml each) were combined and lyophilized. The yield of purified carbohydrate antigen was 40 mg from 1.5 g of cell walls. Autoclaved extraction of antigen from whole cells. Cell surface carbohydrate antigens were also extracted from whole cells by the method of Rantz and Randall (16). Whole cells (0.6 g, dry weight ) were heated at 121 °C for 20 min in 30 ml of saline and then removed by centrifugation. The supernatant was dialyzed against distilled water and lyophilized (designated Rantz-Randall antigen ). Chem ical analyses. Neutral sugars and amino sugars were quantitatively analyzed by gasliquid chromatography according to the methods of Griggs et al. (8) and Sloneker (21). Briefly, 1-5 mg portions of antigen preparations, together with inositol (0.5 mg) as an internal standard, were hydrolyzed in 2 ml of 3 N HCl at 100°C for 4 h. After neutralizing
7 .0
1.0
6 .0
i.....
c:
.!
0.5 ~
....,!. ....
E "CI
...CI
5.0
(J
o
4.0
III
Z
.....E Q)
III
o
)(
Q)
s: 1.0
0.10
III
....
'0
I-
E
0 .5
0.05 Q .....E
o Fraction
20
30
number
(12ml/tube)
40
0
Fig. 1. DEAE-Sephadex A-25 ion-exchange chromatography of the TCA extract of K39K (i) cell walls : The column (1.5 by 45.5 ern) was eluted first with 0.05 M (NH 4)zC03, followed by a linear grad ient of 0 to 1 M NaCl in the same buffer. P, phosphorus. The bar indicates the range where fractions were reactive with the type-specific antiserum.
Carbohydrate Antigen of "5. milleri"
43
with AgzC0 3 , the hydrolyzates were treated first with sodium borohydride and then with acetic anhydride. The alditol acetate derivatives thus obtained were analyzed by gas-liquid chromatography under the conditions reported previously (29). Total hexoses and N-acetyl hexosamines were estimated by the anthrone method (22) and by the method described by Ghuysen et al. (6), respectively. Phosphorus and protein were quantitated by the colorimetric methods of Lowry et al. (11) and Lowry et aI. (12), respectively. Immunological methods. The type i-specific antiserum used was the one previously prepared (28). Reactivities of the carbohydrate antigens with the type-specific antiserum were determined in capillary tube and/or by immunodiffusion in agar gels. Immunodiffusion was performed in 1% Noble agar (Difco) gels in 0.01 M phosphate buffer, pH 7.2 as described previously (28). Immunoelectrophoresis was performed at 10 rnA for 70 min on the gels in 0.05 M barbital buffer, pH 8.6 (10). Quantitative precipitin reaction and its inhibition by haptens were performed as described previously (29). To obtain the quantitative precipitin curve, the purified serotype antigen (0-5 ug) and the type-specific antiserum (5 ul) were incubated in the mixture adjusted to 50 !tl with saline at 37°C for 1 h and then at 4 °C for 18 h. The precipitate was collected by centrifugation (4000 g for 10 min) and estimated for proteins. For the hapten inhibition test, the serum (5 ul) was preincubated with 20 umol (unless otherwise described) of haptens in a total volume of 30 III at 3rC for 1 h and then the antigen was added. The reaction mixture without hapten served as the control. Results
Purification of type-specific antigen Hot TCA extracts of the purified K39~ (i) cell walls were fractionated on a DEAESephadex A-25 column (Fig. 1). Two car:bohydrate peaks were obtained, one was not adsorbed on the column and the other was eluted at 0.35 M NaC! . Only the latter peak fractions reacted with the type i-specific antiserum. The antigenic component was primarily composed of hexose moieties and small amounts of phosphorus were detec ted in the peak fractions . Fractions (No .17-21) which reacted strongly with the antiserum were combined, oncentrated and then applied to a Sephadex G-100 column (not shown) . The elution profile indicated a single major peak composed of hexose and phosphorus but no protein. The fractions (No. 11-42) reactive with the type-specific antiserum were combined, lyophilized and are hereafter designated type i antigen.
Immunochemical characterization of type-specific antigen The immunodiffusion in agar gels (Fig. 2) revealed that the purified type i antigen yielded a single precipitin line with !the type i-specific antiserum which was completely fused with the line given by the Rantz-Randall antigen or the TCA extracts of strain K39K . This indicated that all the antigen preparations contained an identical determinant. Immunoelectrophoresis (Fig. 3) showed that the purified type i antigen as well as the K39K Rantz-Rundall antigen were neutral at pH 8.6. Chemical analyses (Table 1) revealed that the purified cell walls contained significant amounts of various sugars, amino sugars, phosphorus and peptides. However, the purified type i antigen was composed of only galactose, rhamnose and glucose in a molar ratio of 6.8 : 1.6 : 1.0. A significant amount of phosphorus was also detected (2.0), but neither glucosamine, galactosamine, glycerol, ribitol nor proteins were detected.
44
R. Konagawa, T. Yakushiji, and M. Inoue
Fig. 2. Immunodiffusion in an agar gel of the serotype antigen: Antigens (25 Ill) and the type-specific antiserum (25 Ill) were allowed to react in 1.0% Noble agar gels in 0.01 M phosphate buffer, pH 7.2, overnight at room temperature . Well R, Rantz-Randall extract; well T, TeA extract; well P, purified type antigen; well A, type-specific antiserum.
Fig. 3. Immunoelectrophoresis of the serotype antigen: Immunoelectrophoresis was performed on 1.0% agar gel in 0.05M barbital buffer (pH 8.6) at 10 rnA for 70 min. The cathode was on the right. Well P, purified serotype antigen; Well R, Rantz-Randall extract ; well A, type-specific antiserum.
The maximum amount of antibody protein was precipitated when 0.75 Ilg of the type i antigen was allowed to react with 5 III of the homologous antiserum (not shown) . The quantitative precipitin inhibition test with various haptenic sugars (Table 2) showed that galactose and its related sugars inhibited the antigen-antibody reaction most effectively. Lactose induced complete inhibition. Galactose and lactose were the most effective inhibitors at the concentrations tested (Fig. 4).
Carbohydrate Antigen of "S. milleri"
45
Table 1. Chemical composition of the purified cell walls and serotype antigen of "Strep-
tococcus milleri" K39K
~g/mg
sample (molar ratio to glucose )
Components"
Purified cell walls
Serotype i antigen
Rhamnose Galactose Glucose Glucosamine* * Pept ide " * * Phosphoru s
443 (6.2) 165 (2.4) 70 (1.0)
150 633 93 0 0 33
75 (1.1) 174 (7. 1)
II (0.9)
(1.6) (6.8) (1.0) (0.0 ) (0.0) (2.0)
* Mannose, glycerol , ribitol and galactosamine were not detected. • * No N-acetyl hexosam ine was detected. *** The molar ratio in par enthes es is expressed as glycine. Table 2. Hapten inhibition of the precipit in reaction between the purified serotype i antigen and its homologous antiserum * Ha ptens
Inhib ition (%)
Glucose Galactose Rhamn ose Lacto se Melib iose Maltose Isomaltose Gluco samine Galac tosamine N-acet ylglucosamine N-acetylgalactosamine N-acetylma nnosamine a- Meth ylglucoside
36.8 82.9 56.6 100 .0 69 .7 26 .3 36. 8 23 .7 96.1 30 .3 55 .3 27.6 40 .8 59.2
~-M eth ylglucoside
a-Methy lgalactoside ~ - Meth ylgalacto side
71.1
85.5
* 5
~I portion of type-specific antiserum was incubated with 20 umo l of hapten for 1 h at 37 °C, and then 0.75 ~g of pur ified type antigen was added.
Chemical compositions of the Rantz-Randall antigen from serotype i whole cells As summarized in Table 3, the a utoclaved extracts from all the 15 serotype i strains contained rhamnose, galactose, glucose, and glucosamine in an average molar ratio of 4.0 : 0.9 : 1.0 :0.9. Significant amounts of ribitol and phosphorus were detected in the antigen extracts of all the strains but there was no glycerol. In immunodiffusion in agar gels, the extracts of all strains yielded a single band with the homologous type-specific anitserum (not sho wn) , wh ich wa s identical to that produced by the purified type antigen as shown in Fig. 2 .
46
R. Konagawa, T. Yakushiji, and M. Inoue
80
..... ~
~
60
-
40
e
.S!
:0 J:
s
20
o
1
5
10 Haptens
20 (umo t)
Fig. 4. Inhibition of the quantitative precipitin reaction by the constituent sugars of the purified serotype antigen: The experimental conditions were the same as given in Table 2, foot note, except for the concentration s of haptens. (A) rhamnose, (e) galactose, (_) glucose, (0) lactose.
Table 3. Contents of sugars and phosphorus in the Rantz-Randall extracts of the 15 serotype i "Strepto coccus miller;" strains Components
Mean molar ratio to glucose (range)
Rhamnose Galactose Glucose Glucosamine Ribitol Phosphorus
4.0 0.9 1.0 0.9 4.3
(1.2- 8.6) (0.2- 2.0)
(0.3- 2.4) (0.5- 22.1) 1.6 (0.7- 3.2)
* Mannose, glycerol and galactosamine were not detected.
Discussion As reported in the previous study (27), more than half of the oral "5. millen" isolates, like other non-haemolytic streptococci (4, 15), cannot be classified in the Lancefield serogrouping system (at least into groups A - G), but show wide serological variations in their cell surface carbohydrate type antigens (28). Recently, we have purified and immunochemically characterized serotype b-specific carbohydrate antigen . The antigen is found to be composed of rhamnose and glucose and its immunodeterminant terminal is a rhamnose residue (29). In the present study, serotype i-specific
Carbohydrate Antigen of "5. milleri"
47
antigens of "S. millen " were extracted with TCA, purified using column chromatography and characterized immunochemically. The type i antigen was composed of rhamnose, galactose and glucose (Table I). Of the three component sugars constituting the type antigen, galactose exhibited the highest inhibition at all concentrations tested in the quantitative precipitin reaction (Table 2 and Fig. 4), suggesting that the immunodeterminant terminal of the type i antigen of "S. millen" is a galactose residue. Moreover, the hapten inhibition test (Table 2) showed that I} among the disaccharides tested, lactose exhibited complete inhibition and melibiose also possessed a significantly higher effect, but maltose and isomaltose produced a weak inhibition; 2} among the hexosamines, N-acetyl hexosamines or a and ~-metyl glycosides, the galactose derivatives exhibited consistently greater inhibition than the glucose derivatives; 3} ~-methyl glycosides had higher inhibitory capacity than a-methyl glycosides. All of these findings suggest that the immunodeterminant of the serotype i-specific antigen possesses the ~-galactose- (I-4}-a-glucose- sequence (lactosyl residue). Galactosamine, which was not detected in the purified type i antigen (Table I), exhibited strong inhibition in the precipitin reaction between the type antigen and its homologous antiserum (Table 2). This inhibition, however, may have been due to a non-specific inhibition of the precipitin reaction by hexosamines, as observed with the reaction of Streptococcus sanguis b carbohydrate antigen and Lactobacillus plantarum cell wall teichoic acid with their respective antisera (1, 9). Similar non-specific inhibition by hexosamines has been known with the lectin-carbohydrate cell surface interaction between Actinomyces viscosus and S. sanguis (18). A significant amount of phosphorus was detected in the purified i antigen (Table 1). However, this does not necessarily indicate the involvement of teichoic acid in the antigenic determinant of the antigen . Neither ribitol nor glycerol was found in the purified antigen (Table 1) and the anti-i serum used in this study had no reactivity with a preparation of teichoic acid from Streptococcus sobrinus 6715-DP (not shown) . The presence of ribitol in the autoclaved extracts (Rantz-Randall antigen) from whole cells of serotype i strains may be due to the contamination with RNA although it has not been verified yet. The chemical background for the presence of phosphorus in the purified antigen preparation remains obscure at the present time. It is well known that "5. milleri " strains occasionally carry the Lancefield group antigens and/or the Ottens type antigen s (IS). Our previous study, however, has shown that the serotype i antigen can be serologically distinguished from the Lancefield group antigen s A to G (28). The serot ype i antigen (Table 1) possessed the same sugar composition as the Ottens type III antigen (24) of which the antigenic determinant is shown, however, to be ~-glucose (25, 26). Meanwhile, identification of the type i antigen and the Ottens type IV antigen appears more feasible. Willers et al. (23) have shown that this type antigen is compo sed of rhamnose, galactose and glucose as well as N-acetylglucosamine in a molar ratio of 4 : 4 : 4 : 1 and its immunodeterminant possesses a ~-galactose-glucose-galactose (or-rhamnose)- sequence. Thus, though carbohydrate compositions of these two antigens are not strictly identical, since no N-acetylglucosamine but galactose was detectable in the type i antigen (Table I), these type antigens appear to possess a lactose-like sugar sequence in a terminal position that renders them immunodeterminant. This possibility should be verified in further structural studies . All the Rantz-Randall antigens extracted from the 15 serotype i strains including the type strain possessed similar carbohydrate compositions (Tables 1 and 3). In imrnuno-
48
R. Konagawa, T. Yakushiji, and M. Inoue
diffusion in agar gel, precipitin lines with Rantz-Randall extracts from all the 15 serotype i strains were completely fused with the line formed between the typing serum and the antigen extract from the vaccine strain K39K (28). The results support the proposed chemical background of the immunodeterminant of serotype i "S. milleri" .
Acknowledgments. We thank Professor Dr. Shigeyuk i Hamada for providing a teichoic acid preparation and Dr. Richard M. Shelton for correcting the English version of the manuscript. We are also grateful to Miss Kayoko Yuge for her expert secretarial assistance.
References 1. Appelbaum, B. and B. Rosan: Antigens of Streptococcus sanguis: purification and characterization of the b antigen . Infect. Immun . 21 (1978) 896-904 2. Colman , G. and R. E. O . Williams: Taxonomy of some human viridans streptococci. In: Streptococci and Streptococcal Diseases (L. W. Wannamaker and j. M. Matsen , eds.), pp. 281-299. Academic Press, New York (1972) 3. Coykendall, A. L., P. M. Wesbecher, and K. B. Gustafson: "Streptococcus millen", Streptococcus constellatus, and Streptococcus intermedius are later synonyms of Streptococcus anginosus. Int . J. System Bact. 37 (1987) 222-228 4. Packlam, R. R.: Physiological differentiation of viridans streptococci.]. Clin. Microbiol.
5 (1977) 184-201 5. Facklam, R. R.: The major difference in the American and British streptococcus taxonomy schemes with special reference to Streptococcus milleri. Eur. J. Clin. Microbiol. 3 (1984) 91-93 6. Ghuysen, J. M., D. J. Tipper, and j. L. Strominger: Bacterial cell walls and bacteriolytic enzymes. In: Methods in Enzymology, vol. VIII (E. H. Neufeld and V. Ginsburg, eds.), pp . 685-699. Academic Press, New York (1966) 7. Gossling, J.: Occurrence and pathogenicity of the Streptococcus milleri group. Rev. Infect . Dis. 10 (1988) 257-285 8. Griggs, L. j., A . Post, E. R. White, j. A . Finkelstein, W. E. Moeckel, K. G. Holden, j. E. Zaremba, and]. A. Weisbach: Identification and quantitation of alditol acetates of neutral and amino sugars from mucins by automated gas-liquid chromatography. Anal. Biochem. 43 (1971) 369-381 9. Knox , K. W. and A. f. Wicken: Non-specific inhibition of the precipitin reaction between teichoic acids and antisera. Immunochem. 10 (1973) 93-98 10. Koga, T ., N. Okahashi, T. Yamamoto, J. Mizuno , M. Inoue, and S. Hamada : Purification and immunochemical characterization of Streptococcus sanguis ATCC 10557 serotype II carbohydrate antigen . Infect. Immun. 42 (1983) 696-700 11. Lowry, O . H., M. R. Roberts, K. Y. Leiner, M.-L. Wu, and A. L. Farr: The quantitative histochemistry of brain . I. chemical methods.]. BioI. Chern. 207 (1954) 1-17 12. Lowry, O . H., N. f. Rosebrough , A . L. Parr, and R. f. Randall: Protein measurement with the Folin phenol reagent . J. BioI. Chern. 193 (1951) 265-275 15. Liitticken, R., U. Wendorff, D. Liitticken, E. A. Johnson , and L. W. Wannamaker: Studies on streptococci resembling Streptococcus milleri and on an associated surfaceprotein antigen. J. Med . Microbiol. 11 (1978) 419-431 14. Michel, M. F. , I. van Vonn o, and R. M. Krause: Stud ies on the chemical structure and the antigenic determinant of type II antigen of group F streptococci. ]. Immunol. 102 (1969) 215-221 15. Ottens, H. and K. C. Winkler: Indifferent and haemolytic streptococci possessing group antigen F. J. Gen. Microbiol. 28 (1962) 181-191 16. Rantz, L. A . and E. Randall: Use of autoclaved extracts of hemolytic streptococci for serological grouping. Stanford Med . Bull. 13 (1955) 290-291
Carbohydrate Antigen of "S. milleri"
49
17. Ruoff, K. L. : Streptococcus anginosus ("Streptococcus millen"} : the unrecognized pathogen. Clin. Microbiol. Rev. 1 (1988) 102-108 18. Sato, S., T . Koga, and M. Inou e: A possible mechanism for the cellular coaggregation between Actinomyces viscosus ATCC 19246 and Streptococcus sanguis ATCC 10557. ]. Gen. Microbiol. 190 (1984) 1351-1357 19. Shlaes, D . M., P. I. Lerner, E. Wolinsky, and K. V. Gopalakrishna: Infections due to Lancefield group F and related streptococci (S. milleri, S. anginosus ). Medicine 60
(1981) 197-207
20 . Skerman, V. B. D., V. McGowan, and P. H. A. Sneath : Approved lists of bacterial names. Int. ]. System Bact. 30 (1980) 225-420 21. Sloneker,]. H.: Gas-liquid chromatography of alditol acetates. In: Methods in Carbohydrate Chemistry, Vol. 6 (R. L. Whistler and]. N. BeMiller, eds.), pp. 20-24. Academic Press, New York (1972) 22. Trevelyan, W. E. and]. S. Harrison : Studies on yeast metabolism . Biochem.]. 63 (1956)
23-33
23 . Willers, J. M. N ., M. F. Michel, and R. Benner: lmmunochemical studies of type IV and two group-like (Z) carbohydrate antigens of minute streptococci . Antonie v. Leeuwenhoek 39 (1973) 609-617 24. Willers,]. M. N., M. F. Mich el,]. H.]. Huis in't Veld, and G. H.]. Alderkamp: The type antigen III of group F streptococci ; separation of group and type antigens and partial characterization of type III antigen. Antonie v. Leeuwenhoek 39 (1973) 369-382 25. Willers,]. M. N., M. F. Michel , M . Sysma , and K. C. Winkler: Chemical analysis and inhibit ion reactions of the group and type antigens of group F streptococci. J. Gen. Microbiol. 36 (1964) 95-105 26. Willers,]. M. N. , H. Ottens, and M. F. Michel: Immunochemical relationship between Streptococcus MG, FIll and Streptococcus salivarius. ]. Gen. Microbiol. 37 (1964)
l-
425-431
27 . Yakushiji, T., M. Katsuki, A. Yoshimitsu, ]. Mizuno, and M. Inoue : Isolation and physiological characterization of Streptococcus milleri strains from human dental plaque. Microbios 55 (1988) 161-1 71 28 . Yakusbiji, T., R. Konagawa , M. ada, and M. Inoue : Serological variation of oral Streptococcus milleri. J. Med. Microbiol. 27 (1988) 145-151 29 . Yakusbiii, T., M. Inoue, and T. Koga: Purification and immunochemical studies of Type b carbohydrate antigen of oral Streptococcus milleri. Infect. lmmun. 56 (1988)
2264-2269
Prof. Dr. Masakazu Inoue , Department of Preventive Dentistry, Kagoshima University Dental School, 1208-1 Usuki-cho, Kagoshima 890, Japan
4 Zbl. Bakt. 274/1