BACrEMIOLOocAL REVIEWS, Jime 1973, p. 215-257 Vol. 37, No. 2Copyright 0 1973 American Society for Microbiology Printed in U.S.A.

Immunological Properties of Teichoic AcidsK. W. KNOX AND A. J. WICKEN

Institute of Dental Research, United Dental Hospital, Sydney, New South Wales 2010, and School ofMicrobiology, University of New South Wales, Kensington, New South Wales 2033, Australia

INTRODUCTION ............................................................ 215LOCATION OF TEICHOIC ACIDS ................ .............................. 217

Teichoic Acid-Peptidoglycan Complexes ........... ............................. 218Evidence for a covalent linkage between teichoic acid and peptidoglycan ....... 218Extraction of peptidoglyean-free teichoic acid.218Extraction of peptidoglyean-associated teichoic acid.219

Teichoic Acids as Membrane Components ........... ............................ 219Evidence for location ......................................................... 219Extraction of lipid-free teichoic acid ............ .............................. 220Extraction of lipoteichoic acids ................. .............................. 221Nature of the lipid-teichoic acid association ......... .......................... 223

Criteria of Purity of Teichoic Acid Complexes ......... .......................... 224Teichoic Acids as Surface Components ............. ............................. 225Wall teichoic acids ............................................................ 225Membrane teichoic acids...................................................... 226

Extracellular Teichoic Acid?.................................................... 227CONSTANCY OF OCCURRENCE AND STRUCTURE OF TEICHOIC ACIDS 228Wall Teichoic Acids ............................................................ 228

Effects ofgrowth conditions.................................................... 228Effects of mutation ........................................................... 229Variations in glycosidic substitution ............ .............................. 229

Membrane Teichoic Acids .............. ........................................ 230IMMUNOGENICITY OF TEICHOIC ACIDS .......... ............................ 231Wall Teichoic Acids ............................................................ 231Membrane Teichoic Acids ....................................................... 232

DETECTION OF THE ANTIGEN-ANTIBODY REACTION ...... ............... 233Reactions of Cells and Cell Walls .................. .............................. 233Precipitin Reaction ............................................................ 234Hemagglutination ............................................................ 235Effect of Ionic Concentration.................................................... 235

SPECIFICITY OF ANTIBODIES .................. .............................. 236General Considerations ......................................................... 236Specificity of Antibodies to Carbohydrate Substituents.238D-Alanine as an Antigenic Determinant ............ ............................. 239Antibodies Specific for Glycerol Phosphate .......... ............................ 240Glycerol Teichoic Acids as Heterophile Antigens ........ ........................ 241

TEICHOIC ACIDS AS GROUP ANTIGENS .......... ........................... 242Streptococci............................................................. 242Lactobacilli ............................................................ 243Staphylococci................................................................... 244Micrococci ............................................................ 244

IMMUNOBIOLOGICAL PROPERTIES OF TEICHOIC ACIDS ...... ........... 245Wall-Associated Teichoic Acid-Peptidoglyean Complexes ....... ................. 245Membrane Lipoteichoic Acids.................................................... 246

CONCLUDING REMARKS...................................................... 247LITERATURE CITED ........................................................... 248

INTRODUCTIONTeichoic acids have been known since 1959 either cell wall or cell membrane components,

to be components of the outer layers of probably the latter being the operational fractions ob-all gram-positive bacteria; they apparently are tained from disrupted bacteria from which tei-not synthesized by any gram-negative bacteria. choic acids may be extracted. Such distinctionsConventionally, teichoic acids are regarded as may imply a locational significance which is not

215

KNOX AND WICKEN

absolute, particularly in relation to the in situserological activity of these polymers. For thisreason, it is perhaps more meaningful to con-sider teichoic acids on the basis of their co-valent linkage either to peptidoglycan (wall tei-choic acids) or a membrane-associated lipid(membrane teichoic acids); the question of cyto-logical location can then be considered as aseparate issue. Although the occurrence of wallteichoic acids is variable between differentgenera, membrane teichoic acids appear to bevirtually ubiquitous in gram-positive bacteria.The detailed structures, function, and biosyn-thesis of these polymers have been the subjectsof extensive investigations since the time oftheir discovery and have been equally exten-sively reviewed (10, 13, 14, 28, 29).

Membrane-associated teichoic acids arecharacterized by their uniformity of structure.These polymers possess a linear backbone ofpoly (glycerol phosphate) in which linkage isthrough phosphodiester groups involving posi-tions one and three of adjacent glycerol resi-dues (Fig. 1). Structural variation in this groupof teichoic acids appears to be confined to thenature and extent of glycosyl substitution of thesecondary hydroxyl groups (position two) of theglycerol units. D-Alanine ester residues are usu-ally also found as substituents of either glycerolor glycosyl hydroxyl groups. It now appearslikely, as is discussed later in this review, thatmost, if not all, membrane teichoic acids arelinked covalently to membrane glycolipid, andthe term lipoteichoic acid has been given tosuch complexes (295).Peptidoglycan- or wall-associated teichoic

acids, on the other hand, are remarkable fortheir structural diversity, and it is difficult todefine what is and what is not a teichoic acid.Baddiley (29) suggests that the term mightencompass all "polymers that possess phos-phodiester groups, polyols, and/or sugar resi-dues, and usually, but not always, D-alanineester residues." The simplest and, for the pur-pose of this review, classical wall teichoic acidsare either glycerol teichoic acids similar to themembrane-associated polymers (Fig. 1) or ribi-tol teichoic acids in which ribitol replaces glyce-rol as the backbone polyol unit (Fig. 2). Thesepolymers contain only a restricted array of sugarsubstituents (Table 1).

,OHC HO O HC HO ,-

HROL P AlaOf .A

C H O ° CH 0O~ No2 2

FIG. 1. A glycerol teichoic acid. R, H or glycosyl;ala, D-alanyl.

OH C-, 2, HO

O.

CH 0~~2O R

C H 2°FIG. 2. A ribitol teichoic acid. R, H or glycosyl; ala,

D-alanyl.

Although these structures describe the major-ity of teichoic acids that have been studied,there are variations of structure among wall tei-choic acids, which are of considerable signifi-cance to the overall shape or conformation, and,in turn, the comparative serological specificityof these molecules. Glycerol teichoic acids withphosphodiester linkages between positions twoand three, rather than one and three, of the glyc-erol residues have been reported from the wallsof Bacillus stearothermophilus (290) and Actino-myces antibioticus (203). Other variations ofstructure that have been encountered mainlycenter around inclusion of sugar residues in thebackbone of the polymer, and a wide range ofpolymers have been isolated and characterized(Fig. 3). The number of sugar residues (x in Fig.3) may vary from a single hexose unit as in Bacil-lus licheniformis (96), through disaccharides asin Lactobacillus plantarum C 106 (23), to oligo-saccharides of several different sugars as occurin the type specific capsular substances of somepneumococci (231, 236, 250). Teichoic acids ofthis type (Fig. 3) may also have side-chain sub-stitution of the polyol residue as is found, for ex-ample, in the wall teichoic acids of Bacilluscoagulans (88, 89) and pneumococci (42). Thelatter pneumococcal C substance, has been par-tially characterized as having a poly (diamino-sugar-ribitol phosphate) backbone variouslysubstituted with choline phosphate.Polymers in which N-acetylglucosamine-

1-phosphate is attached through phosphodi-ester linkage to glycerol phosphate (Fig. 4) havebeen isolated from the walls of "Staphylococcuslactis" 13 (= Micrococcus lactis; 32, 33) andMicrococcus sp. 24 (18, 26). Related polysac-charide polymers, where the monomer units aresugar-i-phosphates, have also been isolatedfrom the walls of some Micrococcus species (18).While most wall teichoic acids are substitutedwith D-alanyl esters, these components may bereplaced in rare instances by succinate (Ac-tinomyces streptomycini, 201) or acetate (Ac-tinomyces violaceus, 202).

It is generally recognized, from the results ofseveral studies, that teichoic acids, free of

216 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

TABLE 1. Examples of carbohydrate components of teichoic acids

Location Polyol Carbohydrate(s)a Occurrence References

Membrane Glycerol Glc L. helveticus NCIB 7220 63, 254Glc-a-1 a 2-Glc S. faecalis 39 291Glc-j6-1 - 6-Glc S. aureus H 223Gal, Gal-a-1 -- 2-Glc L. fermenti NCTC 6991 295"

Wall Glycerol Glc B. subtilis Marburg 49L. buchneri NCIB 8007 21S. epidermidis T1 65

GalNAc S. epidermidis (albus) 81NCTC 7944

Ribitol Glc B. subtilis W-23 189L. plantarum (arabinosus) 17-5 14

GlcNAc S. aureus Copenhagen 280GalNAc S. aureus phage type 187 152

aGlc, n-glucose; Gal, D-galactose; GlcNAc, N-acetyl-D-glucosamine; GalNAc, N-acetyl-D-galactosamine." A. J. Wicken and K. W. Knox, unpublished data.

sugor )- poly - P04

FIG. 3. Repeating unit of a teichoic acid with oneor more sugar components in the "backbone." Polyolis either glycerol or ribitol.

C H20AIa -

HO -1

I N HAcCH20

FIG. 4. Teichoic acid containing N-acetylglucosa-mine-i-phosphate as a "backbone" component. Ala,D-alanyl; Ac, acetyl.

association with other wall or membrane com-ponents, do not induce antibody formation,whereas they will react with antibody formedagainst the appropriate bacterial cell or cellfraction. This distinction is well known amongother categories of serologically reactive sub-stances and led to the use of the terms "anti-gen" (inducing antibodies) and "hapten" (sero-logically reactive). Currently, the term "im-munogenic" is preferred for indicating the capa-bility of a substance to induce an immunologi-cal response whereas "antigenic" relates moreto its reactivity and specificity in a reaction.The use of the term "group antigen" preceded

this newer distinction, but would still seem anappropriate definition of a bacterial compo-nent; the definition of a group antigen hasusually preceded its identification, and in manyinstances the isolated antigen does not induceantibody formation. Similarly, the use of theterm "antigen-antibody reaction" would seem

valid in referring to the various detection sys-tems discussed in this review, in which the re-action observed depends on a particular com-ponent of an immunogenic complex and inwhich the isolated component itself is notnecessarily immunogenic. For example, and asis discussed later, the lipoteichoic acid-proteincomplex from the membrane of lactobacilli isimmunogenic; specificity depends on the tei-choic acid component but this componentalone, although reacting with antibodies incertain systems, will not induce antibody for-mation (160).

LOCATION OF TEICHOIC ACIDSIsolation and purification of bacterial cell

walls has generally been considered as the firststep in the extraction of wall teichoic acids inorder to avoid contamination with membraneteichoic acid. The assumption that cell wallscan be prepared free of membrane material is,however, not valid in every case. Cell walls ofgram-positive organisms, prepared convention-ally by mechanical disruption, treatment withdeoxyribonuclease and ribonuclease, and re-peated washing with buffers and water, fre-quently contain small quantities of lipid (85, 88,242). Recent studies (194) on cell walls similarlyisolated from strains of Staphylococcus aureushave shown that the lipid content is due toresidual membrane which may be either physi-cally joined to the wall or trapped by thecollapse of the rigid wall during disruption ofthe cells. Membrane teichoic acid was as-sociated with these membrane components, andsmall amounts of teichoic acid and phos-pholipid still persisted as wall contaminantseven after additional purification procedures,among which heating, proteolysis with trypsin,

VOL. 37, 1973 217

KNOX AND WICKEN

and treatment with detergents were the mosteffective. Membrane teichoic acid was alsofound to be still associated with cell walls of L.plantarum after treatment of the latter withtrypsin (164). Caution must therefore be exer-cised in interpreting serological activities ofextracts of cell walls. For a general and practicalaccount of the preparation, purification, andanalysis of bacterial cell walls we recommendthe recent review by Work (305).

Teichoic Acid-Peptidoglycan ComplexesEvidence for a covalent linkage between

teichoic acid and peptidoglycan. It has longbeen evident that the conditions required forthe complete extraction of teichoic acids fromcell walls involve covalent bond breakage if theproduct obtained is to be free of other wallmaterial. Teichoic acid-peptidoglycan com-plexes from enzymic digests of cell walls weredevoid of detectable phosphomonoester groups,suggesting that the terminal phosphomonoestergroup of the teichoic acid was involved in alinkage with a component of peptidoglycan(275). Muramic acid phosphate was identifiedin 1958 (5), and there is now evidence that thephosphate grouping of this component providesthe cross-link between the terminal polyol unitof teichoic acid and peptidoglycan (46, 105).A relatively acid-labile phosphoramidate

bond between the terminal phosphate group ofteichoic acid and a hexosamine amino group ofthe peptidoglycan has also been proposed fororganisms whose wall teichoic acid is fairlyreadily extracted under acid conditions (121).Although the existence of such a linkage stillhas to be demonstrated, the possibility of therebeing more than one type of linkage of teichoicacid to peptidoglycan may be indicated by theobservations of rapid release of teichoic acidfrom some organisms under alkaline conditions(130), whereas less than 10% of the wall teichoicacid of B. stearothermophilus was solubilized in48 h by the same procedure (105). The degree ofextractability of teichoic acid from Bacillussubtilis var. niger with trichloroacetic acid (79)has been noted to vary with the growth condi-tions employed. A phosphodiesterase capable ofdepolymerizing teichoic acid has been obtainedfrom B. subtilis (302), and the possibility ofsuch an enzyme occurring in some bacteria andcausing partial degradation of teichoic acidsduring the preparation of cell walls or extractionof teichoic acid should not be ignored.

In view of the structural diversity of wallteichoic acids and possible different modes oflinkage to other cell wall polymers, it is notsurprising that a variety of procedures have

been developed for their isolation in polymericform. No single procedure is likely to extract allteichoic acids without degradation, and forserological and other studies, teichoic acid prep-arations extracted by different proceduresshould be compared.Extraction of peptidoglycan-free teichoic

acid. Complete extraction of teichoic acid fromwalls free of association with peptidoglycanoccurs only slowly at 4 C in dilute trichloroaceticacid solutions (5 to 10%), but this is probablythe most widely used, and indeed classical,procedure (27). There has been some contro-versy as to whether these acidic conditionscause degradation of wall teichoic acid throughhydrolysis of inter-unit phosphodiester bonds(95). A careful re-examination of the trichloroa-cetic acid-extracted teichoic acid from the wallsof S. aureus, B. subtilis, and Lactobacillusarabinosus (122) suggested that the rate of sucha hydrolysis is low under the conditions em-ployed for extraction, and little degradationoccurs. However the product obtained withtrichloroacetic acid may not be representativeof all of the teichoic acid(s) in the wall. With B.subtilis W23, which contains two ribitol teichoicacids, trichloroacetic acid-extraction removesmost of the glucosyl-substituted polymer, butonly small amounts of the unsubstitutedpolymer (49). Similarly, with Bacillus subtilisMarburg, where the wall contains glycerol tei-choic acids, the extract contained mostly theglucosylated polymer (49). Studies on tri-chloroacetic acid extracts from B. licheniformis(135), B. subtilis (310), and B. stearothermophi-lus (105) also indicated that the extracted tei-choic acid was not representative of the wallpolymers.A serological examination of extracts of L.

plantarum NCIB 7220 provided evidence forribitol teichoic acid molecules differing in gluco-sidic substitution over at least a fourfold range(164). The molar ratio of glucose to phosphorusin the extract was higher than in unextractedwalls, suggesting that nonglucosylated ribitolteichoic acid was present in the wall, but wasnot recovered in the extract. The apparentabsence of nonglucosylated teichoic acid intrichloroacetic acid extracts may reflect differ-ences in the ease of extraction rather thandegradation; no difficulty, for instance, wasexperienced in extracting the nonglucosylatedribitol teichoic acid from L. plantarum ATCC10241/Ri where it was the only teichoic acidcomponent (73, 164).

Although exposure to trichloroacetic acid inthe cold for relatively short periods (16 h)probably avoids extensive degradation of tei-

218 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

choic acids, yields are often rather low (20 to30%). Complete extraction usually requires pro-longed exposure or higher temperatures, condi-tions which can bring about hydrolysis of glyco-sidic as well as phosphodiester bonds; treat-ment of S. aureus wall teichoic acid withtrichloroacetic acid for 15 min at 90 C resultedin almost complete loss of serological activity(110).To avoid the possibility of hydrolysis of glyco-

sidic linkages, wall preparations have beenextracted under various alkaline conditions,usually 0.1 N NaOH either at room temperatureor 35 C (20, 25, 137). Alanine ester linkageswould be hydrolysed, but, depending on thetype of teichoic acid structure, interunit phos-phodiester bonds may also be broken (20).Generally, the peptidoglycan component is notsolubilized, although S. aureus provides anexception because of the lability of the glycyllinkages in the cross-bridges (19). In specificcases the method may prove useful for thedifferential extraction of cell wall components,for it has been shown that alkali treatment of B.licheniformis wall solubilizes the teichoic acidcomponent but not teichuronic acid (137), apolymer of glucuronic acid and N-acetylgalac-tosamine (138, 145). The term "teichuronicacid," although originally applied to thispolymer is now used more generally to refer togram-positive bacterial cell wall polysaccha-rides containing uronic acids.A method which has the potential for being

generally applicable to the extraction of serolog-ically active teichoic acids is the treatment ofwalls with phenylhydrazine (12, 111) or (toavoid the formation of tar) N, N-dimethylhydra-zine (7). The method, which probably involvesfree radicals, gives high yields of both polysac-charides and teichoic acids, completely remov-ing teichoic acid from S. aureus wall withoutdetectable loss of serological activity (111).

Older methods used for the extraction ofantigenic material from bacteria such as forma-mide at 150 to 160 C (92) or HCl at pH 2 and 100C (170) have found little application in theisolation of teichoic acids, and in view of theextreme conditions involved and consequentrisk of degradation are best avoided.

Extraction of peptidoglyean-associatedteichoic acid. To obtain soluble teichoic acid-peptidoglycan complexes requires hydrolysis oflinkages within the peptidoglycan component,and a wide variety of suitable enzymes has beendescribed; these include carbohydrases andpeptidases of plant, animal, and bacterial originas well as endogenous wall autolysins (94, 276).Fractionation of the hydrolysates by ion ex-

change chromatography or gel filtration yieldsteichoic acid-peptidoglycan complexes withhigher molecular weights than free teichoicacids, and therefore they are more suitable forserological reactions (95, 105, 106, 168, 310).A novel and potentially useful approach to

the purification of teichoic acids in enzymaticdigests or other extracts depends on the affinityof the plant lectin concanavalin A for thea-anomers of D-glucose, N-acetyl-D-glucosa-mine and N-acetyl-D-galactosamine (74). Themethod employs a column of cyanogen bro-mide-activated agarose gel which will bindconcanavalin A. When an autolysate of B.subtilis 168 was passed through the column,only teichoic acid-containing fractions were re-tained, and these could be eluted with methyl-a-D-glucopyranoside. The resultant product wasless polydisperse than teichoic acid obtained byacid extraction. (Con A-Sepharose is now avail-able from Pharmacia Ltd., Uppsala, Sweden.)

Teichoic Acids as Membrane ComponentsEvidence for location. In contrast to the

variable occurrence of peptidoglycan-associatedwall teichoic acid, most gram-positive bacteriacontain a glycerol teichoic acid that can beisolated from the intracellular fraction of dis-rupted cells. Although initially referred to sim-ply as "intracellular teichoic acid" (59), it wasfound that the teichoic acid was present in theparticulate ribosome-membrane fraction (59,291). Subsequent studies based on an examina-tion of products resulting from protoplast for-mation led to the introduction of the term"membrane teichoic acid." It was found fromchemical analysis that the formation of proto-plasts from Streptococcus faecalis and Bacillusmegaterium was associated with the release ofthe major portion of the teichoic acid into thesoluble fraction, although some remained as-sociated with the cell membrane (123). Similarfindings were obtained with several strains ofstreptococci by using serological techniques toidentify teichoic acid (258, 270). A detailedinvestigation on S. faecalis ATCC 9790 (S.faecium) (263) showed that the teichoic acidwas associated with the membrane from whichlocation it could readily be removed by washingwith water or salt solutions. All of these investi-gations strongly suggested that membrane tei-choic acid was located in or on the externalsurface of the protoplast membrane, althoughthe nature of the attachment remained obscure.It now seems probable that membrane associa-tion depends on a covalent linkage betweenteichoic acid and membrane glycolipid (seebelow).

VOL. 37, 1973 219

KNOX AND WICKEN

More direct evidence for membrane associa-tion has been provided by the use of ferritin-labeled antibodies, and Fig. 5 shows the resultof reacting protoplasts of Lactobacillus fer-menti NCTC 6991 successively with antibodiesprepared against the membrane teichoic acidand ferritin conjugated to antirabbit gammaglobulin (283). The antiteichoic acid antiserumused was prepared by injection of the lipotei-choic acid-protein complex into rabbits (seeImmunogenicity of Teichoic Acid) and wasspecific for the teichoic acid moiety of thecomplex, and no serological activity towardseither the protein or lipid portions of the com-plex was detected. Teichoic acid from whichboth protein and lipid had been removed com-pletely absorbed antibodies (294) from the an-tiserum. (This restriction of antibody specificityto the teichoic acid portion of lipoteichoicacid-protein complexes appears to be a commonone and is discussed later in this review.)Extraction of lipid-free teichoic acid. Mem-

brane teichoic acids were originally obtained bytrichloroacetic acid extraction of the 100,000 gmembrane-ribosome fraction from disrupted or-ganisms after removal of cell walls (59). Crudeextracts were heavily contaminated with poly-

nucleotides, polysaccharide, and varyingamounts of wall teichoic acid; the latter con-taminant may have represented fragmentationof the wall during the disruption procedure orpartially synthesized and still membrane-associated wall teichoic acid. Trichloroaceticacid-extraction of whole organisms, as might beexpected, increases the amount of contamina-tion by wall polymers. Fractional precipitationwith alcohol or acetone followed by chromatog-raphy on columns of ion-exchange celluloses orlow porosity gels has been used to obtainmembrane teichoic acids with a high degree ofpurity and free of association with other cellularcomponents (59, 291). Such preparations hadaverage chain lengths of 17 to 22 glycerophos-phate units and molecular weights in the range3,000 to 12,000, depending on the degree ofsugar substitution. Gel chromatography of tri-chloroacetic acid-extracted material gave asmall fraction of the teichoic acid apparentlybound to polynucleotide, and such "complexes"were excluded from the low-porosity gels (Se-phadex G75) available at that time (291). In thelight of more recent studies (see below) itappears likely that the higher-molecular-weightmaterial represented partially degraded lipotei-

FIG. 5. Protoplast of L. fermenti treated with antiserum to the lipoteichoic acid and then reacted withferritin conjugated to goat antirabbit gamma globulin. [rom Journal of Ultrastructure Research (283).]

220 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

choic acid and that the association with polynu-cleotide was more apparent than real. It is,however, of some practical interest to note thatpreparations of bacterial ribosomes (291) anddeoxyribonucleic acid (DNA) (311) are invaria-bly contaminated with membrane teichoic acid,presumably as a result of cellular disruption.Teichoic acid has also been reported as both acontaminant of transfer ribonucleic acid(tRNA) of B. subtilis and an inhibitor of accept-ing activity for tRNAPhe and tRNATyr (308).

Older methods for the extraction of serologi-cally reactive teichoic acids from bacteria, in-cluding formamide at 150 to 160 C (92), diluteacid (170), and mild alkaline conditions (76) arelikely to yield degraded products. For example:(i) deacylation of lipoteichoic acids would beexpected, (ii) there is the previously mentionedpossibility of interunit phosphodiester bondsbeing hydrolyzed by acid, and (iii) glycosidicbonds may be particularly susceptible to acidhydrolysis through the influence of neighboringphosphate groups (291). Differences that proba-bly relate to the lability of glycosidic bonds havebeen found when comparing the serologicalreactions of Lancefield acid extracts and forma-mide extracts (131) and also of Lancefield acidand trichloroacetic acid extracts (292) fromgroup D streptococci with homologous an-tiserum where specificity depends on the 2-0-a-D-glucosyl-D-glucose (kojibiose) substituent (43).Extraction of lipoteichoic acids. High-

molecular-weight preparations of teichoic acidwere obtained from B. licheniformis by coldphenol extraction (44), and it was suggestedthat prolonged exposure to the acid conditionsof trichloroacetic acid-extraction might causethe hydrolysis of phosphodiester linkages. Thepossibility that the less drastic conditions ofextraction afforded by cold phenol might pro-vide a means of isolating "native" membraneteichoic acid still associated with other cellularcomponents prompted us to compare the prop-erties of phenol- and trichloroacetic acid-extracted preparations of membrane teichoicacid from L. fermenti NCTC 6991 (295).The cell contents from disrupted L. fermenti

were extracted with either 45% aqueous phenolor 10% trichloroacetic acid at 4 C and, afterfurther purification steps designed to removelipids and degrade high-molecular-weight poly-nucleotides, the crude teichoic acids were frac-tionated on gel columns of various porosities(295). The cold-phenol-extracted product waseluted from 6% agarose columns with a Kd of 0.1and was excluded from gels of lower porosity(Fig. 6a). Trichloroacetic acid-extracted mate-rial, on the other hand, yielded two fractions of

teichoic acid from columns of Sephadex G75 atKd = 0.0 and 0.50; the smaller, excluded frac-tion was eluted from 6% agarose with a Kd Of0.15 (Fig. 6b). Treatment of the phenol-extracted teichoic acid with trichloroacetic acidin the cold gave an elution profile from 6%agarose identical to that obtained for trichloroa-cetic acid-extracted teichoic acid.

Subsequent analyses and chromatographicstudies of chemical degradation products in-dicated that the phenol-extracted product was acomplex of teichoic acid with lipid and protein,the lipid association being indicated by describ-ing the complex as lipoteichoic acid (295).Trichloroacetic acid-extraction of cell contentsor treatment of the complex with trichloroaceticacid resulted in disruption of the complex andthe production of lipid- and protein-free tei-choic acid together with smaller amounts ofpartially degraded lipoteichoic acid. It wassuggested (295) that the apparent high molecu-lar weight of lipoteichoic acid, as shown by itschromatographic elution properties and sedi-mentation value of 9.5s (1% wt/vol), might beexplained by micelle formation. Negative stain-ing of preparations of lipoteichoic acid withsodium phosphotungstate has shown uniformoval bodies 7 nm in diameter (A. J. Wicken andK. W. Knox, unpublished data). Similar micel-lar formations have been reported for extractedlipopolysaccharides from gram-negative orga-nisms (239). It was also pointed out (295) thatthe existence of membrane teichoic acid as acomplex with lipid consistent with its proposedmembrane location. Lipoteichoic acid is thusregarded as native membrane teichoic acid,whereas the lower-molecular-weight lipid-freeproduct obtained by acid extraction is degradedteichoic acid.Lipoteichoic acids with similar properties

have been extracted by the cold-phenol proce-dure from the cell contents of a variety oflactobacilli, streptococci, and bacilli (160, 163,164, 296; A. J. Wicken and K. W. Knox,unpublished data) and also from S. faecalis,where the lipoteichoic acid is the group Dantigen (279). In retrospect it is of interest tonote that earlier work on cross-reactive anderythrocyte-sensitizing antigens obtained froma number of gram-positive organisms by aque-ous phenol, water, or hot saline extraction cannow be explained in terms of lipoteichoic acid asthe reactive antigen in such extracts (see Glyce-rol Teichoic Acids as Heterophile Antigens). Al-though it appears likely that all membraneteichoic acids exist as lipoteichoic acids, part ofthe membrane teichoic acid of Lactobacilluscasei NCTC 6375 is extracted as lipid-free

221VOL. 37, 1973

KNOX AND WICKEN

4 Kd:O-15

(a)~~~~~~~~~~~~~~~(b)~

102 /\2

-}U~~~~~~~~~rL /40 60 so 100 120 20 40 60 so

Froction numer

FIG. 6. Agarose gel (6%) chromatography of teichoic acid obtained by (a) phenol extraction and (b) tri-chloroacetic acid extraction of the cell contents of L. fermenti. Columns were eluted with 0.2 M-ammoniumacetate, pH 6.9; 0, Atmol of phosphorus per 0.25-ml fraction; continuous line, extinction at 260 nm. Fractionsare approximately 4 ml Kd (distribution coefficient) data from A. J. Wicken and K. W. Knox [295]).

low-molecular-weight teichoic acid by the cold-phenol procedure, and all of the membraneteichoic acid extracted by the same procedurefrom a strain of Lactobacillus fermentum ap-pears to be of low molecular weight (A. J.Wicken and K. W. Knox, unpublished data);these variations may reflect a higher lability ofthe linkage of the lipid moiety to the teichoicacid in these organisms.

In a comparative study of other methods ofextracting lipoteichoic acid from whole orga-nisms of L. fermenti NCTC 6991, instead of thesubcellular fractions, it has been shown thatsome of the procedures commonly used forextracting lipopolysaccharide from gram-nega-tive organisms are applicable to the extractionof lipoteichoic acid (294). The most frequentlyused method for obtaining lipopolysaccharide,low in protein, involves extraction with hot,aqueous phenol (288), and this method provedto be the most efficient for extracting lipotei-choic acid from L. fermenti. Crude extractswere heavily contaminated with high-molecu-lar-weight polynucleotides, but treatment ofextracts with deoxyribonuclease and ribonu-clease at pH 7, followed by chromatography on6% agarose, resulted in considerable purifica-tion, and the isolated lipoteichoic acid waslower in polynucleotide content and associatedprotein than that obtained by the cold-phenolprocedure. A lower protein content for hot-phenol-extracted lipopolysaccharide in compar-ison with the cold-phenol-extracted product hasalso been reported (287), and it has been pro-posed that under the conditions of hot-phenolextraction a protein-lipid-polysaccharide com-plex is degraded to a lipoprotein and a lipopoly-saccharide (303). There appeared to be no

difference between cold- and hot-phenol-extracted lipoteichoic acids in terms of appar-ent molecular size of the micelles and sugarsubstitution of the teichoic acid, although, asmight be expected, a greater loss of D-alanylesters was noted with extraction at the highertemperature.

Lipoteichoic acid complexes, containing highamounts of associated protein, have been ob-tained from whole organisms of L. fermenti, andother lactobacilli by a mild procedure involvingpartial removal of cell lipids from freeze-driedorganisms with chloroform-methanol (2: 1, vol/vol) at 20 C, followed by aqueous extraction ofthe lipoteichoic acid-protein complex withwater at various temperatures (294, 296).Chloroform-methanol extraction would producea lipid-depleted membrane although no changewas detected in the normal membrane profileby electron microscopy (294). In the subsequentextraction of cells with water, the lipoteichoicacid-protein complex was removed, with theconcurrent disappearance of the membrane pro-file in electron micrographs, whereas the cellwall still appeared intact. Sufficient damagecan also be imparted to the membrane byfreeze-drying to allow for a partial release oflipoteichoic acid by aqueous extraction, and asimilar partial release of the complex from L.fermenti following "cold shock" has also beenobserved (294). The high protein content ofthese aqueous-extracted lipoteichoic acidscould be considerably reduced by hot-phenolextraction or digestion with papain followed bycold-phenol extraction and gel chromatographyon 6% agarose (294).The choice of method for extracting lipotei-

choic acid relates to both specific requirements

222 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

and the organism concerned. Extraction fromwhole organisms is attractive from the point ofview of simplicity of procedure. The chloroform-methanol-water-extracted product, having ahigh associated protein content, is a good im-munogen (294), but may be variably contami-nated with cell wall polymers. Hot, aqueousphenol extraction of whole organisms yieldsa lipoteichoic acid which is low in protein and apoor immunogen but, at least in the case of L.fermenti, free from contamination with cell wallmaterial. Where the presence of wall material inlipoteichoic acid extracts is likely to interferewith subsequent operations, such as would bethe case with organisms possessing a wall tei-choic acid, extraction of lipoteichoic acid fromsubcellular fractions would be the method ofchoice. Here cold-phenol extraction shouldyield a good immunogen, whereas hot-phenolextraction yields a product low in associatedprotein.

It should be noted that lipoteichoic acids bindvery tenaciously to ion-exchange celluloses orgels and cannot be eluted from them withoutpartial degradation through deacylation (294).Further purification of extracts of whole orga-nisms, although still preserving the hydro-phobic nature of the lipid moiety of lipoteichoicacids, may therefore demand the application ofsuch techniques as preparative electrophoresisor immunoabsorption. In this latter connection,concanavalin A and similar glycosyl-specificlectins may prove to be as useful in the purifica-tion of lipoteichoic acids as they have been inthe case of wall teichoic acid (74).Nature of the lipid-teichoic acid as -

sociation. Lipoteichoic acid preparations fromL. fermenti contained glycolipid indistinguisha-ble from the free glycolipid of the cell mem-brane and small amounts of phospholipid (295).Lipoteichoic acids extracted in our laboratoriesfrom other lactobacilli similarly contain glyco-lipid, but the association with phospholipid isvariable and when present may represent amixed micelle of lipoteichoic acid and phos-pholipid (A. J. Wicken and K. W. Knox,unpublished data).The linkage of glycolipid with teichoic acid in

L. fermenti lipoteichoic acid appears to becovalent in that the glycolipid is not removed byorganic solvents or hot phenol. Deacylation ofthe polymer removes fatty acid residues anddestroys the micellar nature of the originalcomplex, but the disaccharide glycerol portionof the glycolipid is still attached to the teichoicacid (295). Deacylated lipoteichoic acid con-tains no detectable phosphomonoester groups,and the glycolipid glycoside is completely re-

leased, free of esterified phosphate, during alka-line hydrolysis of the polymer. These results areconsistent with a linkage of the terminal phos-phate group of the teichoic acid with a sugarhydroxyl group of the glycolipid glycoside. Peri-odate oxidation studies showed that the glycerolportion of the glycolipid was not involved in aphosphate linkage and that the minimumlength of the polyolphosphate chain was 29glycerophosphate residues (A. J. Wicken and K.W. Knox, unpublished data). More recently(279), the lipoteichoic acid from S. faecalis hasbeen shown to be a 28-unit, kojibiosyl-substi-tuted, glycerophosphate polymer linked tomembrane glycolipid by a presumed phospho-diester bond similar to that proposed for L.fermenti. In neither case, however, has the link-age actually been proved. In both lipoteichoicacids the presumed phosphodiester bond ismore labile to the action of trichloroacetic acidthan the interpolymer phosphodiester bonds,and the glycolipid glycoside and fatty acidesters are lost from trichloroacetic-acid-ex-tracted teichoic acid.

Phosphatidylglycolipids (glycerophospholip-ids; 261), in which a phosphatidyl residue isattached to a sugar hydroxyl group of a glycolip-id, have been reported as membrane compo-nents of a number of gram-positive bacteria (6,71, 86, 260, 261). In at least one case, S. faecalis,there is a remarkable constancy of disaccharidesubstitution (kojibiose) of glycolipid, phos-phatidylglycolipid, and lipoteichoic acid (279).In L. fermenti, galactosyl-1, 2-glucosyl substitu-tion of both glycolipid and lipoteichoic acid hasbeen shown (295; A. J. Wicken and K. W. Knox,unpublished data), and in S. aureus, gentiobi-ose is found as the disaccharide substituent ofboth membrane teichoic acid and glycolipid(223, 260). Evidence for the glycolipid moiety ofS. faecalis lipoteichoic acid being a phos-phatidylglycolipid has been afforded by a studyof the products of hydrofluoric acid hydrolysis ofthe polymer (279), but the linkages of theteichoic acid and the phosphatidyl group to theglycolipid are not known. It is suggested (279)that these involve different sugar hydroxylgroups (Fig. 7a). Alternatively, it is possiblethat the phosphatidyl group, bearing one in-stead of two fatty acid residues, is also theterminal glycerophosphate unit of the teichoicacid portion of the lipoteichoic acid (Fig. 7b). Ifthis latter possibility proved to be true thenthese three classes of membrane lipid compo-nents, glycolipid, phosphatidylglycolipid (acyl-ated glycerophosphorylglycolipid), and lipotei-choic acid (acylated polyglycerophosphoryl-glycolipid) form a graded series.

VOL. 37, 1973 223

KNOX AND WICKEN

a

amx:

He

b

FIG. 7. Postulated nature of the lipid-teichoic acid association in S. faecalis lipoteichoic acid showing (a) thephosphatidyl group as a separate substituent of the glycolipid to the teichoic acid chain (279) and (b) an alterna-tive structure in which the phosphatidyl group, bearing a single fatty acid residue, forms part of the main tei-choic acid chain. R, glycosyl or H; R', fatty acid ester residue.

Isolated lipoteichoic acids are also complexedwith protein, the amount varying with theextraction procedure and influencing the rela-tive immunogenicity of the preparation (seebelow). The possibility of the protein being ofmembrane origin and complexed with the glyco-lipid component is suggested by studies withMycoplasma pneumoniae. M. pneumoniaeglycolipids, similar in structure to those presentin the membrane of gram-positive bacteria (220,221, 229), are not immunogenic but can berendered immunogenic by aggregating withmembrane proteins from Archelloplasma (for-merly Mycoplasma) laidlawii; interaction wasachieved by dispersing the glycolipids and lipid-depleted A. laidlawii membrane in sodiumdodecyl sulphate followed by dialysis againstbuffer containing magnesium ions (230).

Criteria of Purity of Teichoic AcidComplexes

It is evident that with both wall and mem-brane teichoic acids, newer methods of isolationhave aimed at obtaining these polymers in acovalent association with other cellular compo-nents. In such circumstances, the usual criteriafor chemical purity are not relevant, for we aredealing with either "units" of the cell wallderived by random enzymic digestion or of thecell membrane obtained by disorganizing themembrane structure and therefore containinglipid and protein. That it is these complexes,rather than "pure" teichoic acids, which aremuch more likely to display immunological andother biological properties is the subject ofmuch of the remainder of this review.

224 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

Although "purity," in the sense of a singlemolecular species, cannot be applied to suchcomplexes, it is essential that their componentsbe defined both qualitatively and quantita-tively. It is regrettable that, in many instances,crude extracts of teichoic acids used in serologi-cal or other biological investigations have re-ceived no further chemical characterizationthan the chromatographic demonstration ofpolyols and polyol phosphates in acid hydroly-sates. Furthermore, the assumption of glyco-sidic substitution of a teichoic acid by thefinding of reducing sugars in acid hydrolysatesis rarely valid, for contamination of teichoicacid complexes by polysaccharides is a fairlycommon occurrence. Here, more rigid chemicaldefinition of glycosidic union of sugar withpolyol is required, which generally involvesisolation and characterization of glycosides de-rived from alkaline, enzymatic, or hydrofluoricacid hydrolysates of the polymer or complex (forreviews of methods see references 10, 13, 14,305).

Definition of a teichoic acid complex in termsof a single chromatographic peak or a singleprecipitin line in immunodiffusion or immuno-electrophoresis could be regarded as a first stepin establishing the "purity" of a complex.However, at the risk of stating the obvious, itmust be remembered that, on the one hand, gelchromatography alone establishes only a degreeof homogeneity of molecular size as opposed tokind and, on the other hand, a single precipitinline merely marks the presence of a single,reactive antigenic species in what might be acomplex mixture. Analysis of immune precipi-tates and the use of fractions derived frompartial degradations of complexes as inhibitorsin serological reactions will often yield usefulinformation as to the reactive components ofthese complexes.

Teichoic Acids as Surface ComponentsReaction of antibodies with intact cells has

usually been taken to imply a surface locationfor the antigenic components or rather, "byplacing an antigen on the cell surface, we areimplying that it behaves as though it werethere. By placing it beneath the surface wemean that it seems ... to be overshadowed bysome other bacterial component" (301).With the subsequent identification of specific

wall components, a surface location frequentlybecame synonymous with wall, so that we findreference to "the unexpected absence of group Dantigen from the cell walls of group D strep-tococci" (270), and membrane teichoic acids of

streptococci and lactobacilli are termed "in-tra-mural cementing substance" (181, 272).Wail teichoic acids. The spatial organization

of teichoic acid-peptidoglycan complexes withinthe cell wall has received relatively scant atten-tion. In B. megaterium strain M (204), the wallsare composed of approximately 50% peptidogly-can and 50% teichoic acid; hot formamide ex-traction of the teichoic acid reduced the ap-parent thickness of the walls by 50%, leavingthe rigid peptidoglycan layer, and suggestingthat the teichoic acid was located mainly as aplastic layer on the outer cell wall surface.Extraction of the teichoic acid from L.arabinosus cell walls, on the other hand, pro-duced no obvious change in their electron mi-croscopic appearance (11). Studies on B.licheniformis (136) showed a considerable heter-ogeneity of the molecular organization of thethree wall polymers of this organism, peptido-glycan, teichoic acid, and teichuronic acid.Some 50% of the peptidoglycan was free oflinkage to the other two polymers, whereassubstituted glycan strands contained either tei-choic acid or teichuronic acid substituents.These findings are-in keeping with the ideas ofattachment of teichoic acid molecules to newlysynthesized glycan strands before cross-linkingof the latter into the cell wall (190) and closeassociation of biosynthesis of wall polymersthrough a common membrane polyisoprenolphosphate carrier (8). Mauck and Glaser (190)interpret their findings of essentially randominsertion of new cell wall polymers in B. subtilisby suggesting a tangential arrangement of pep-tidoglycan strands, rather than their beinglayers parallel to the cell surface, newly syn-thesized glycan strands with attached teichoicacid molecules being intercalated with old ones.Such a model would provide for an overallexposure of at least part of the total teichoicacid at or near the cell surface.The uptake of antibody to teichoic acid by

whole cells and isolated cell walls has been usedas a probe for the location of teichoic acidwithin the wall (43). Cell walls of B. subtilis3610 bound only 19% of the total possibleantibody, but the uptake was markedly in-creased by a brief exposure to lysozyme; thiswas interpreted as "a loosening of the peptido-glycan network to allow access of antibodies toteichoic acid in a deeper layer of the wall or apreviously sterically unfavored situation." Thatwhole cells of this organism bound only 10% lessantibody than purified cell walls was suggestedto indicate that little or no teichoic acid waslocated on the inner surface of the cell wall; it is

VOL. 37, 1973 225

KNOX AND WICKEN

unfortunate for unequivocal interpretation thatthis particular strain appears to contain identi-cal wall and membrane teichoic acids (43).Similar results were claimed for B. subtilisW23, where the wall and membrane teichoicacids are different, but no experimental detailswere given (43). In contrast, B. licheniformiswalls bound 71% of the calculated amount ofantibody to teichoic acid, with only a slightincrease on lysozyme treatment, suggesting a"surface" location of the antigen (43). Themarked difference in the results with B.licheniformis and B. subtilis may reflect thedifferences in the structure of the two teichoicacids, with the former containing glycosyl resi-dues as part of the main polymer chain and thelatter being a classical glycosylated polyolphos-phate. The precise configuration of the highlycharged teichoic acid molecules in the cell wallis largely dependent on the electrostatic andhydrogen bonds formed with other wall compo-nents (144, 183, 184), and the effect with B.subtilis may be at least partly due to analteration in surface charge through adsorptionof the basic protein, lysozyme.The affinity of concanavalin A for a-D-glu-

cosylated teichoic acids has been used as amarker for the teichoic acid in the walls of B.subtilis 168 (Birdsell, Doyle, and Morgenstern,personal communication). Electron micro-graphs of this organism showed a discontinuouslayer on the surface of whole cells and the outersurface of cell walls after treatment with con-canavalin A. The layer was absent from cells nottreated with lectin or washed with methyl-a-D-glucoside solutions to remove concanavalinA. A mutant strain that does not produce a wallteichoic acid was unaffected by treatment withconcanavalin A. There also appeared to be anaccumulation of the presumed teichoic acid-concanavalin A complex at the site of septumformation. These authors suggest that the tei-choic acid in this organism is situated at theouter surface of the cell wall and that theindividual molecules are oriented perpendicu-larly to the long axis of the cell because of theirnegative charge and highly hydrophilic nature;such an orientation would be in keeping withthe Mauck and Glaser (190) model of tangentialintercalation of newly synthesized teichoic acid-peptidoglycan into the cell wall.Thus the question of the precise location of

wall teichoic acids in or on the peptidoglycannetwork is still an open one. The possibility thatteichoic acid conformation rather than depth oflocation within the wall is a more importantfactor in determining accessibility to antibodies

as well as reaction with them needs much closerexamination.Membrane teichoic acids. A membrane lo-

cation for lipoteichoic acids might suggest thatthey would not be able to react with antibodiesin situ. Yet in certain instances, antibodiesspecific for the membrane teichoic acid willagglutinate whole organisms, the best knownexample probably being the group D strep-tococci (257). Antibody reaction with the cellsurface can also be shown by the removal ofspecific antibodies from serum, and this proce-dure provided evidence for L. fermenti mem-brane teichoic acid being a surface component(128, 160).

In view of the generally recognized structureof the gram-positive bacterial cell, a reaction ofwhole organisms with antibodies specific for amembrane component requires that antibodieseither penetrate the cell wall to react with themembrane or react with membrane componentsprotruding through the cell wall or, alterna-tively, that both phenomena are occurring. Insupport of the first proposal, Burger's sugges-tion (43) that the peptidoglycan architecture ofthe wall be considered as a highly porousnetwork or as a sheet with large-size discon-tinuities has received some support from elec-tron micrographs of freeze-etched cells of S.lactis (140). These studies showed the existenceof "pores" in the wall, the calculated size beingtheoretically sufficient for the passage of mole-cules up to 3 x 106 daltons, i.e., an immuno-globulin M (IgM) molecule should be able toenter the pores. However, detailed studies of thepassive permeability of the cell wall of B.megaterium (245) showed a much lower thresh-old of polymer exclusion in the range of 3 to 5 x104 daltons. If this lower limit proved generalamong gram-positive bacteria, then even im-munoglobulin G (IgG) molecules would be un-likely to penetrate the cell wall.The absorption of membrane lipoteichoic

acid-specific antibodies by whole organisms ofL. fermenti and L. casei has been examined inmore detail (283). These two organisms differedmarkedly in their behavior towards standardserological tests. L. fermenti absorbed and wasreadily agglutinated by anti-lipoteichoic acidantiserum, and the IgM fraction was moreeffective than the IgG fraction. L. casei, on theother hand, neither appeared to absorb nor wasagglutinated by such antisera. A more sensitivemethod of detecting adsorbed antibody, involv-ing electron microscopy of organisms that hadbeen treated successively with antiserum andferritin conjugated to goat anti-rabbit-y-

226 BACTrERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

globulin, showed the apparent difference be-tween the two organisms to be quantitativerather than absolute (283). L. casei showed, bythis technique, some surface adsorption of tei-choic acid antibody, but it was irregular indistribution and significantly less than theconfluent labeling shown by L. fermenti. Awall-membrane model has been proposed toexplain these differences in surface reactivity ofa membrane antigen (283). The model (Fig. 8) isbased on current hypotheses on membranestructure (228, 267). The lipid bilayer depictedis asymmetric with glycolipid derivatives pres-ent only in the outer half of the bilayer. This isadmittedly speculative, although there is evi-dence for membranes displaying such asymme-try with respect to lipid and carbohydratecomponents (39). Lipoteichoic acid moleculesare shown as being held either in or on themembrane by hydrophobic interaction of theirglycolipid moieties with membrane lipid orprotein; the role of magnesium ions (283) inmaintaining the integrity of lipoteichoic acidassociation with the membrane suggests alsothat ionic bonds may be involved. The long,polar polyolphosphate chains are proposed toextend by intercalation into the peptidoglycan-polysaccharide net-work of the cell wall andmay come near enough to the outer surface ofthe cell wall for part of the polyolphosphatechain to act as a surface antigen. Factors onwhich this latter condition might be dependentinclude (i) thickness of the cell wall and degreeof peptidoglycan cross-linking, (ii) the length ofthe polyolphosphate chain of the lipoteichoicacid, and (iii) the conformation of the chainwithin the ionic environment of the cell wall.The walls of L. fermenti and L. casei are quitedifferent in their composition and lysosymesensitivity (157, 161), and it is suggested (283)that these different milieux may allow forgreater penetration of the wall matrix by lipo-teichoic acid in L. fermenti than in L. casei;localized differences in wall composition mayalso explain the "patchy" nature of exposure ofL. casei lipoteichoic acid at the surface asevidenced by the pattern of ferritin labeling.

This wall-membrane model suggests a highlyorientated arrangement of the polyglycerophos-phate chains of the lipoteichoic acid molecules.Enzymatic introduction of D-alanine ester resi-dues into L. casei membrane teichoic acid (234)requires a ligase that is only effective withteichoic acid bound to membrane in, as theauthors suggest, a highly orientated state; theligase was totally ineffective with membrane-free teichoic acid. The model also suggests that

FIG. 8. Diagrammatic representation of the cellwall (A) and membrane of a gram-positive organism.The membrane components shown are protein (B),phospholipid (C), glycolipid (D), phosphatidyl glyco-lipid (E), and lipoteichoic acid (F). Depending on thelength and conformation of the glycerophosphatechains and the thickness of the wall, the lipoteichoicacid molecules may function as surface antigens; fromJournal of Ultrastructure Research (283).

lipoteichoic acid forms a physical link betweenwall and membrane and is held by hydrophobicand possibly ionic forces in the membrane andby ionic interactions between teichoic acid andpeptidoglycan in the wall.

Extracellular Teichoic Acid?In several instances, serological studies on

antigens known to be teichoic acids employedthe culture fluid as a source of material. Suchproducts may have derived from viable orga-

VOL . 37, 1973 227

KNOX AND WICKEN

nisms or alternatively from autolysis of deadorganisms. Pollock (222) has defined the term"extracellular" as referring to products thathave "originated from the cell without anyalteration to cell structure greater than themaximum compatible with the cell's normalprocesses of growth and reproduction." An earlyexample of a wall teichoic acid being detected inthe growth medium is the soluble polysaccha-ride A of S. aureus (298), which has beenidentified as a teichoic acid-peptidoglycan com-plex (108). It might be assumed that suchcomplexes derive from autolysis of dead orga-nisms. However it has been demonstrated (189)in B. subtilis W23 and B. megaterium KM thatthe cell wall shows extensive turnover duringthe exponential phase of growth. Peptidoglycanand teichoic acid showed identical turnoverrates resulting in almost 50% loss of cell wallmaterial per generation in B. subtilis and about30% in the case of B. megaterium. The extentand importance of wall turnover to wall synthe-sis and growth in gram-positive bacteria gener-ally remains to be defined. The products of thewall turnover in B. subtilis (189) were found inthe growth medium as peptidoglycan-teichoicacid complexes and could, according to theabove definition, be regarded as extracellularproducts. Because it is likely that the samehydrolytic enzymes are involved in both wallturnover and autolysis of dead organisms, thedistinction between extracellular or autolyticproducts, as applied to teichoic acid-peptido-glycan complexes in the growth medium, be-comes largely philosophical and, in a practicalsense, probably relates mainly to the phase ofgrowth of the organism.There are also a number of instances where

reactive antigens present in growth media, anddetected by their ability to adsorb to red bloodcells, have the properties of membrane lipotei-choic acids. The earliest examples were strep-tococcal (90, 117, 156, 215) and staphylococcalculture filtrates (90, 215), whereas one studyshowed that the amount present in a donorbottle of blood contaminated with a Bacillusspecies was sufficient to render the recipients'red-blood cells "polyagglutinable" (50) becauseof the presence in human sera of antibodiesreacting with the sensitizing antigen (51, 52).

Other investigations concerned with commonor heterophile antigens (see later section)showed that serologically reactive materialcould be detected in the medium from 18- to24-h cultures of Listeria monocytogenes (205)and various species of the genera Streptococcus(272, 274) and Bacillus (9). A more detailedstudy on a variety of gram-positive bacteria

showed that "large amounts" were detectedafter 24 h of incubation, with a decrease oc-curring after 4 days, and no activity being de-monstrable after 9 days at 37 C (225). The de-scription of "large amounts" was not given aquantitative basis, but depended on the dilutionof culture fluid capable of sensitizing erythro-cytes, which were then agglutinated by specificantisera; hemagglutination, detectable by mac-roscopic examination, was also the procedureemployed by the other investigators. Themethod, which requires that the membrane tei-choic acid be present as a lipid complex (seebelow), is quite sensitive, for 2.5 Ag of purifiedlipoteichoic acid per ml is sufficient to sensitizeerythrocytes and cause visible agglutination onthe addition of specific antiserum (128). Asmembrane lipoteichoic acid probably consti-tutes 1 to 2% of bacterial cell mass, it can beseen that even a concentration of 2.5 ug/ml inthe culture fluid would represent a substantialproportion of the cellular teichoic acid.Whether these lipoteichoic acids should be

regarded as extracellular products or as releasedby autolysis of dead organisms has not beenresolved. Turnover of membrane components ingram-positive bacteria has not been widelyinvestigated. In B. subtilis, membrane lipidshowed turnover at approximately the samerate as cell wall (193), but in B. megateriummembrane lipid was largely conserved duringcell growth and division (195). In gram-negativebacteria, the outer membrane lipopolysaccha-ride may be released into the growth medium(162, 240), and it has been shown (214) thatSalmonella typhimurium lipopolysaccharide issynthesized in the plasma membrane beforebeing translocated through the peptidoglycanlayer into the outer membrane. By analogy, it ispossible that lipoteichoic acid in gram-positivebacteria could lose its association with themembrane and be translocated completelythrough the cell wall into the external medium.The release of lipoteichoic acid, by whatevermeans, could also be an important factor ininfluencing its immunobiological properties (seelater section).

CONSTANCY OF OCCURRENCE ANDSTRUCTURE OF TEICHOIC ACIDS

Wall Teichoic AcidsEffects of growth conditions. Studies with a

variety of bacilli have indicated that the cellwall is one of the most dynamic and phenotypi-cally variable structures of the whole cell, withlarge shifts in the chemical composition inresponse to relatively small changes in the

228 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

environment being well documented. Suchchanges in structure are of obvious importancein serological classification of bacteria, where itis required that antibodies be formed against a"surface" component and then react with thatcomponent.Marked changes in the wall composition of

several bacilli as a phenotypic response tochanges in the growth environment have beenshown in an extensive series of chemostat stud-ies (79, 80, 82, 83, 278). M. Iysodeikticus (M.luteus; 32, 33), normally considered not toproduce a wall teichoic acid, can be induced todo so when chemostat-grown under conditionsof limiting magnesium (83). Under conditions oflimiting magnesium, potassium, nitrogen, orsulphate walls of several strains of B. subtilis,B. licheniformis, B. megaterium, and S. aureusH contained teichoic acid, but under conditionsof phosphate limitation teichoic acid was re-placed by teichuronic acid or a similar acidicpolysaccharide (83). With B. subtilis var. niger,transition from fully magnesium-limited tofully phosphate-limited growth, or vice versa,involved a more rapid transition of polymertype than could be explained by dilution withnewly synthesized wall material, and it wasconcluded that the control of synthesis of bothpolymer types was the expression of a singlegenotype. In B. subtilis var. niger increasinggrowth rate increased the proportion of teichoicacid in the cell wall; the synthesis of teichoicacid could be maintained under conditions ofphosphate limitation by applying a constraintto magnesium uptake by having high concen-trations of sodium ions in the external environ-ment. Lowering the pH also increased the wallteichoic acid content in this organism as well asthe extent of D-alanyl ester substitution of thepolymer. Differences in the extent and nature ofthe glycosidic substitution of teichoic acids werealso noted under different conditions of growth;Boylen and Ensign (37) also reported that theglucose content of the teichoic acid from B.subtilis differed with the age of the cells. Theextensive turnover of the cell wall demanded bythese changes has been demonstrated by label-ing experiments with B. subtilis W23 and B.megaterium (189), as previously discussed.The biological detection of such changes in

wall composition is apparently limited to theobservation that B. subtilis W23 develops phagereceptor sites when grown under conditionswhere teichoic acid, rather than teichuronicacid, is synthesized (24). It has been suggested,however, that the inability to obtain groupingantisera reacting with strains of L. acidophilusis related to the variability in composition of the

cell wall, the substantial amounts of neutralpolysaccharide present in walls of exponentiallygrowing cells being replaced by anionic polysac-charides in the stationary phase (58).

Effects of mutation. Serological studies onmutants of enterobacteria have provided con-siderable evidence on the structure of theirantigens and the specificty of antibodies (175,176). Mutants of gram-positive bacteria havenot been exploited to the same extent, althoughtheir potential role in gaining insight into thephysiological function of wall teichoic acids hasbeen realized (306). Mutants of S. aureus wereexamined for their ability to agglutinate withantibodies specific for the wall ribitol teichoicacid component, lack of reactivity indicatingthe possibility of (i) a change in teichoic acidstructure, (ii) a lack of teichoic acid, or (iii)the presence of another surface polymer thatprevented teichoic acid from reacting with thespecific antibody (306); the mutant that wasdescribed fell into the third category, as thewall teichoic acid was masked by a teichuronicacid that prevented both antibody and phageadsorption. Other studies with mutants of S.aureus, B. subtilis, and L. plantarum haveprovided examples of the presence of teichoicacids lacking sugar substituents (48, 73, 97,260), the formation of a different teichoic acid(304), the absence of teichoic acid (57, 260), andits replacement by protein (284).

Investigations on mutants of S. aureus in-dicated that the fundamental changes in com-position and polymer type were associated withother phenotypic differences suggestive of sur-face structure changes and raised interestingquestions as to the relatedness of the mutants toother taxonomic subgroups of staphylococci(304).Variations in glycosidic substitution. Al-

though considerable changes in wall teichoicacids may occur through mutation or growth oforganisms under limiting conditions, the degreeand constancy of glycosyl substitution of tei-choic acids in organisms grown under normalconditions is of considerable importance to theserology of these polymers. This considerationapplies, in the main, to teichoic acids of theclassical type where sugars are borne as sidechains on the main polyolphosphate backbone(Table 1). Biosynthesis of wall teichoic acidsappears to require prior formation of the polyol-phosphate, with the involvement of polyiso-prenol phosphate, followed by the introductionof sugars via their uridine 5'-diphosphate de-rivatives (8, 14, 72, 116, 141). Baddiley andco-workers picture the enzymes for wall synthe-sis existing in the membrane in an ordered and

229VOL. 37, 1973

KNOX AND WICKEN

highly integrated manner as "wall synthesizingunits," each unit being a multienzyme complexfor the complete formation of a wall polymer(8). The units are proposed to be situated on theinside of the membrane with a limited amountof shared polyisoprenol phosphate, and eachunit completes a round of polymer synthesisbefore releasing the carrier intermediate withthe growing polymer chains extruding throughthe membrane into the wall where they arefinally located. This concept of a vectorialprocess of formation of wall polymers on theoutside of the membrane from nucleotide inter-mediates on the inside is an attractive one, andthe presence in the membrane of incompletelysynthesized wall polymers might explain theoften-observed contamination of membrane tei-choic acid preparations with wall teichoic acid.Evidence that introduction of the sugars

proceeds independently of formation of thepolyolphosphate is provided by the regularitywith which mutants are isolated that containunsubstituted polyolphosphate chains. Alsosome bacteria have wall teichoic acids which aremixtures of fully glycosylated and glycosyl-freepolymers (45, 49, 96, 135), and in S. aureusstrains there is serological evidence that the a-and fl-N-acetylglucosaminyl substituents ofthe wall teichoic acid occur on different polymerchains (200, 280). There are also a number ofexamples indicating that sugar substitution isboth partial and random (14, 21), a situationrather unusual in the synthesis of structuralpolymers.

Membrane Teichoic AcidsAlthough wall teichoic acids display a variety

of structures, and both the amounts present andtheir structures are susceptible to change, mem-brane teichoic acids display a much narrowerstructural diversity and are always of the classi-cal, polyglycerophosphate type. Even underconditions of phosphate starvation, whichleads to the replacement of wall teichoic acid byteichuronic acid, membrane teichoic acid is stillsynthesized (83).A report on the effect of growth conditions on

membrane teichoic acid production (191) sug-gested that teichoic acid synthesis, measuredserologically as the group D antigen, was de-pendent on the glucose concentration and finalpH of the medium, and good yields were onlyobtained from some strains of group D strep-tococci when 0.5 to 1.0% glucose was used inunbuffered media (final pH 4.0 to 4.2). How-ever, because the group D antisera used todetect the teichoic acid are glucosyl specific,

these results may reflect the influence of growthconditions on the extent of glucosyl substitutionof the teichoic acid rather than the extent ofsynthesis of the polyglycerophosphate backboneof the polymer, the latter being the importantpart of the molecule in its role of maintaining ahigh concentration of bivalent cations in theregion of the membrane (126). That variationsin the extent of sugar substitution of membraneteichoic acids may occur as a response to growthconditions is of general significance to serologi-cal classification and warrants further investi-gation.L-forms of group D streptococci (129, 270) do

not produce detectable membrane teichoic acidwhen they are grown in a liquid medium con-taining penicillin. The possibility that penicil-lin was inhibiting synthesis was suggested bythe observation that a stable L-form derivedfrom a group D streptococcus did produce thegroup-specific teichoic acid when grown in aglucose medium in the absence of penicillin(270); the teichoic acid was present in themedium, a not unexpected location because ofthe known release of membrane teichoic acidinto the medium during protoplast formation(123, 263).Magnesium ions are known to be of impor-

tance for the integrity of protoplast membranes(35, 87, 237, 297), and it has been shown thatprotoplasts of L. fermenti retain lipoteichoicacid in the presence of magnesium ions, butthat the polymer is readily lost from protoplastsprepared in their absence (283). It has beenproposed that membrane lipoteichoic acidsfunction as ion exchangers, particularly in re-gard to magnesium ions (126), which are re-quired in the synthesis of peptidoglycan andteichoic acids (8, 134). The mechanism of bio-synthesis of membrane lipoteichoic acids hasnot received the same attention as has beengiven to the wall components; polyisoprenolphosphate does participate in the synthesis of apolyglycerophosphate in B. Iicheniformis (8),but it is not known whether the final location ofthis product is wall or membrane.The continued formation of membrane tei-

choic acid under conditions where synthesis ofwall teichoic acid ceases (83) might indicate adifferent route of synthesis. Earlier suggestions(70) that glycolipids might be involved as sac-charide carriers in the synthesis of various wallpolymers in bacteria had been discounted onthe grounds of the lack of demonstration ofmetabolic turnover of glycolipid (259). How-ever, the association of glycolipid with mem-brane lipoteichoic acid reopens the question of arole for glycolipid in the synthesis of these

230 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

polymers as an acceptor for growing or com-

pleted teichoic acid chains. In this respect itmay be relevant that L-forms derived fromStreptococcus pyogenes (55) and S. aureus (286)contain twice as much free glycolipid in theirmembranes as the protoplasts derived fromthese organisms.

IMMUNOGENICITY OF TEICHOICACIDS

To determine whether a bacterial componentis immunogenic requires suitable detecting sys-tems and, as discussed in the next section, themethods most used with teichoic acids are

precipitation and agglutination, the latter em-

ploying bacterial cells, cell walls, or erythro-cytes sensitized with the antigen. However, thevarious classes of antibodies differ in theircapacitites to elicit different serological results;in general IgG antibodies are more effectiveprecipitins than IgM antibodies, whereas IgMantibodies are the more effective in agglutina-tion reactions (219). Immunoglobulin A anti-bodies generally display both properties (219).Thus, conclusions on the relative antibody re-

sponse to different immunogens can be in-fluenced by the method of detection. Theclasses of antibodies produced against teichoicacids have been determined in only a fewinstances, and in each case the precipitinmethod was used (Table 2).

Wail Teichoic Acids

The immunogenicity of wall teichoic acidshas been amply illustrated by the formation ofspecific antibodies on injection into rabbits ofwhole organisms representing a diverse range ofgram-positive bacteria, examples including theribitol teichoic acids of S. aureus (119, 243, 280),B. subtilis (28) and L. plantarum (164, 254),and the glycerol teichoic acids of S. epidermidis

(albus) (1), B. licheniformis (139), and group Elactobacilli (254). Grov and Rude (110) failed toobtain antibodies to teichoic acid on injection ofS. aureus wall preparations, and a similar resultwith S. epidermidis led to the conclusion that"structures present in cell walls have been al-tered or lost during preparation of the walls"(109). In neither of these cases did the procedurefor wall preparation include the addition ofproteolytic enzymes, but loss of protein from thewall could account for the observations, e.g.,

trypsinizing the wall of L. plantarum NCIB7220 results in the loss of immunogenicity of theribitol teichoic acid component (164). Crudecell wall preparations of L. plantarum containcontaminating immunogenic membrane tei-choic acid, and the most effective means ofremoving the contaminant while retaining theimmunogenicity of the wall component involvedtreatment with hot, sodium dodecyl sulphate(164).

Acid-extracted teichoic acids are not im-munogenic in rabbits (43, 118), although im-munogenicity may be regained by forming com-plexes with appropriate particulate acceptorssuch as methylated bovine serum albumin,cetyl pyridinium chloride (43), or chromiumchloride-treated erythrocytes (40). The lack ofantibody response to acid-extracted teichoicacid probably is related to its low molecularweight. Immunogenicity is related to molecularweight, and the antibody response of rabbits toboth dextran (149) and type III pneumococcalpolysaccharide (133), for example, decreased as

the molecular weight decreased from > 105 to104.The problem of obtaining antibodies to S.

aureus ribitol teichoic acids by injecting cellfractions has suggested the use of alternativemeans of detection of the antigen. Both theHelix pomatia A hemagglutinin (113) andconcanavalin A (233) react with a-N-acetyl-

TABLE 2. Classes of antibodies reacting with teichoic acids in precipitin tests

Predominant

Organism Prep injected Antibody source Teichoic acid antibody class ReferencessourceIgM IgG IgA

S. aureus Organisms Rabbit serum Wall + 307L. plantarum Organisms Rabbit serum Wall + (+) 164S. aureusa Organisms Human serum Wall + 186L. plantarumb Lipoteichoic acid Rabbit serum Membrane + (+) 164S. aureus Organisms Guinea pig milk Wall + 180S. aureusa Organisms Human nasal se- Wall + 63

1 cretions

a Presumed infective organism.b Lipoteichoic acids from L. casei and L. fermenti give similar results when examined by the hemagglutina-

tion reaction (283).

231VOL. 37, 1973

KNOX AND WICKEN

glucosaminyl residues of ribitol teichoic acid,although there is the disadvantage that theywill also react with other a-D-glycosyl substitu-ents (100, 114).Most human sera also contain antibodies

reacting with S. aureus teichoic acids (280), andincreases in antibody content occur in patientswith infected burns (268) and with staphylococ-cal infections in general (62, 185). The detectionof antibodies by double diffusion in agar hasproved to be a satisfactory method of distin-guishing patients with staphylococcal endocar-ditis from those with endocarditis caused byother organisms (60).The injection into humans of acid-extracted

teichoic acid from S. aureus Copenhagen gavean increase in the amount of precipitatingantibody (280). This result, which contrastswith that obtained by injecting ribitol teichoicacid into rabbits (118), may reflect differencesin the immune responses of humans and rab-bits; dextran, for instance, is immunogenic forman, but not for rabbit (150). However, a strictcomparison of these results may not be valid forit involves comparing a (presumed) secondaryresponse in humans with a primary response inrabbits.

Studies on the occurrence of human anti-bodies reacting with S. aureus teichoic acidhave used preparations from strains in whichf3-N-acetyl-glucosaminyl determinants predom-inate. However, this determinant occurs in anumber of other antigens, and cross-reactionshave been shown between such teichoic acidsand group A streptococci (147), group L strep-tococci (153), and the ubiquitous peptidoglycan(108). Without other evidence it is, therefore,difficult to decide whether the antibodies react-ing with the S. aureus teichoic acid do, in fact,represent antibodies produced, either wholly orpartly, as a result of infection by S. aureus. Theneed to qualify conclusions was realized in arecent study showing that the reaction ofhuman sera with a variety of organic dustscould relate in part to the presence of fl-N-acetylglucosaminyl-substituted teichoic acid orof "a substance with antigenic determinantsrelated to teichoic acid" (84).

Membrane Teichoic AcidsAlthough studies on the immunogenicity of

wall teichoic acids have usually started with theobservation that injection of whole organismswill yield specific antibodies, this method willfrequently fail to yield antibodies to the mem-brane teichoic acid. This lack of reactivityprobably is related to its location rather than toan inherent lack of immunogenicity. Thus Shat-tock (257) discussed the "notorious difficulty"

of obtaining antibodies to strains of group DStreptococcus bovis and, in one of the firstapplications of the Mickle shaker, showed thatinjection of disrupted bacteria yielded potentgroup D antiserum. Similarly, the injection ofintact L. plantarum organisms into rabbits onlyoccasionally yields antibodies to the membraneteichoic acid, whereas all rabbits injected withcrude cell wall produced antibodies to thecontaminating membrane teichoic acid (164).Injection of whole organisms of L. casei yieldsantibodies to the wall polysaccharide (the groupantigen), but not to the membrane teichoicacid, whereas injection of L. fermenti results inthe formation of antibodies to the membraneteichoic acid (the group antigen) and not to thewall polysaccharide (157, 160, 254). The variousfactors that may influence the degree of pene-tration of lipoteichoic acid molecules into thecell wall (see section on teichoic acids as surfacecomponents) may similarly influence the im-munogenicity of the membrane componentswhen whole organisms are injected.The influence of cellular location on immuno-

genicity has also been observed in studies on theformation of antibodies to peptidoglycan andcell wall carbohydrate polymers. In a discussionof results with group A and A variant strep-tococci, Schleifer and Krause (246) noted thatantisera to group A variant organisms are a goodsource of antibodies to peptidoglycan and thatwhole cells will absorb the antibodies, whereasgroup A streptococcal antiserum contains littleor no antibodies to peptidoglycan, and the or-ganisms do not absorb antipeptidoglycan anti-bodies. The peptidoglycan of B. lichenformisalso is only weakly antigenic, and antibodiesmainly are formed against the wall teichoic acidcomponent (139); removal of the teichoic acidgreatly enhances the immune response to thepeptidoglycan.These results on the immunogenicity of pep-

tidoglycan might also provide examples of anti-genic competition (139), a description appliedto the observation that simultaneous immuni-zation of an animal with two or more antigensmight result in a diminution of immune re-sponse to one or more of these as compared withcontrol animals receiving a single antigen. Ex-amples have been known for 70 years; recently,possible mechanisms were discussed and evi-dence was presented that the competition is atthe level of antigen "processing" or "localizing"(41). An understanding of the mechanism ofantigenic competition might also provide anexplanation of the observed differences in theimmunogenicity of the L. casei and L. fermentiwall and membrane components.Depending on the method of extraction, "pu-

232 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

rified" membrane teichoic acids may be ob-tained that retain their immunogenicity. Burger(43) showed that injection of phenol-extractedmaterial complexed with methylated bovineserum albumin or cetyltrimethyl ammoniumbromide proved to be an effective means ofobtaining antibodies, with trichloroacetic acid-extracted material being much less effective;injection of teichoic acid preparations (unspeci-fied) with Freund complete adjuvant failed toresult in antibody formation. Subsequent stud-ies (160) with L. fermenti membrane teichoicacid, with Freund complete adjuvant and theBurger injection procedure, showed that thephenol-extracted material was immunogenic,the high-molecular-weight trichloroacetic acid-extracted fraction was less effective, and thatthe low-molecular-weight trichloroacetic acid-extracted fraction was not immunogenic.The extraction of membrane teichoic acid

with trichloroacetic acid is an extension of itsuse for obtaining soluble teichoic acid fromwall, where there is the requirement to hydro-lyze the covalent linkage between teichoic acidand peptidoglycan. However, as discussed ear-lier, membrane components may be solubilizedby milder procedures that yield lipoteichoicacid-protein complexes (294). The relationshipof protein content to immunogenicity of L.fermenti products injected in Freund completeadjuvant is shown in Fig. 9. Preparation 1 wasobtained by cold, aqueous phenol extraction ofthe soluble fraction from disintegrated orga-nisms (160, 295) and preparation 2a was ob-tained by aqueous extraction of chloroform-methanol-water extracted organisms (294).Deacylation of preparation 2a and recovery ofthe teichoic acid fraction gave a product with areduced protein content and also reduced im-munogenicity; the antibody content of thesesera was measured with preparation 2a becausedeacylation, by disaggregating the complex,decreases its ability to precipitate antibody(160, 163). Products with lower protein contentand a corresponding decrease in immunoge-nicity were obtained by hot, aqueous phenolextraction (preparation 2c) and by digestionwith papain followed by cold, phenol extraction(preparation 2d). Attempts to achieve a morecomplete removal of protein by other chemicaland enzymatic methods, including digestionwith trypsin and pepsin, were unsuccessful. Thefailure of trypsin to remove all of the proteinassociated with membrane lipoteichoic acid isrelevant to observations that its use to obtain"pure" cell wall (61) may give preparations stillcontaining membrane components, as indicatedby lipid analyses (194) and detection of lipotei-choic acid (164).

3 ~~ U1 Antige 2r0fi120

Preparation

FIG. 9. Comparison of protein content of lipotei-choic acid preparations from L. fermenti and theirimmunogenicity upon injection into rabbits. Prepara-tion 1 was obtained from disintegrated organisms byextraction with cold, aqueous phenol and preparation2a was obtained from whole organisms by extractionwith chloroform-methanol followed by hot water;preparations, 2b, 2c, and 2d were obtained from 2a bytreatment with dilute alkali, hot aqueous phenol, andpapain, respectively. The amount of antibody pro-duced by each of the three rabbits injected with eachpreparation was determined by the quantitative pre-cipitin method. [Data from A. J. Wicken, J. W.Gibbens, and K. W. Knox (294).]

Freund complete adjuvant is necessary foreffective antibody production against lipotei-choic acids, and replacement by incompleteadjuvant results in a lessened response (160).Freund complete adjuvant consists of Mycobac-terium tuberculosis in an oil. The originalsuggested (91) that the oil component of theinjected emulsion serves to slow the release ofthe immunogen from the water droplets andalso to provide protection from rapid degrada-tion in the tissues, has been supported by recentstudies (309). White (289) discussed the specificeffect of the mycobacterial component of adju-vant on the immune response and concludedthat it depends on the surface-active propertiesof the peptidoglycolipid component of wax D,possibly by increasing the capture or persist-ence of the immunogen at sites on the dendriticcell surface or, to use his own descriptivephrase, the peptidoglycolipid may "form thebird-lime which holds the lymphocytes moreeffectively in the net of dendritic cells." Thepeptidoglycolipid, which consists of peptidogly-can linked to a mycolate residue of an arabino-galactan, is considered to be a product ofautolysis of the cell wall (103), and recently awater-soluble adjuvant-active fraction has beenobtained from mycobacterial cell wall by lyso-zyme digestion (3).

DETECTION OF THEANTIGEN-ANTIBODY REACTION

Reactions of Cells and Cell WallsEvidence for antibodies reacting with the cell

surface may be obtained by showing either the

233VOL. 37, 1973

KNOX AND WICKEN

reaction of antibodies with the cell or theremoval of antibodies from serum. Detection ofadsorbed antibodies, generally shown by theagglutination reaction, may also employ labeledantibody; colonies of group D streptococci reactwith fluorescein-labeled antibodies to theirgroup-specific membrane teichoic acid (68),and L. fermenti cells react with ferritin-labeledantibodies, the ferritin being detected on thecell surface (283).The reaction of cell wall teichoic acids with

antibodies would be influenced by the previ-ously discussed conformation of the teichoicacid chains and their accessibility. An exampleof the influence of accessibility of teichoic acidto antibody on serological reactivity is shown byS. aureus, for agglutination of cells seems todepend primarily on a protein component (pro-tein A) and type-specific products, and theteichoic acid probably is not an agglutinogen(132, 209). However, wall preparations containvery little, if any, of the type-specific products(209), and the agglutination of such prepara-

tions from several strains has been shown todepend on the teichoic acid component, specifi-cally on the linkage between the N-acetyl-glucosamine residues and ribitol (200, 209).Wall from S. aureus strain Copenhagen con-

tains both a- and fl-N-acetylglucosaminyl-sub-stituted teichoic acids, but antibodies couldonly be detected to the a substituent (209). Thewall was agglutinated by heterologous anti-bodies specific for fl-N-acetyl-glucosamine, so

that the results would indicate lack of formationof antibodies to the ,B substituent on injection ofstrain Copenhagen rather than the unavailabil-ity of the groupings for reaction with antibody.The agglutination reaction has been success-

fully applied in particular cases to the differen-tiation of bacterial strains, such as the typing ofgroup D streptococci (255). However, attemptsto classify lactobacilli by their agglutinationreactions have not given encouraging results(252).

Precipitin Reaction

A more satisfactory method for the serologicalclassification of lactobacilli was developed bySharpe (251) and was based on the formation ofa precipitate between antiserum and an acidextract of the bacterial cells (170). Most of thestrains could be allocated to one of six groupswith teichoic acids subsequently being shown tobe the grouping antigens for four of these (254).The qualitative precipitin method is very suita-ble for screening a large number of strains, butit depends on a subjective assessment, and a

negative reaction does not necessarily meanabsence of the component (257).Of necessity, the precipitin method requires a

solution of the cell fraction, but "solution" isnow extended to include sonically treated wallpreparations such as those from B. lichenifor-mis (139). Where the solution is represented bythe cytoplasmic fraction of cells, cross-reactionsmay occur which minimize the application ofthe procedure; for instance, Pease (218) showedcross-reactions between group D streptococci,Haemophilus species, and Mycoplasmahominis.

Quantitation of the precipitin reaction isachieved by analysis of the antigen-antibodyprecipitate for its protein content (150), gener-ally after dissolution in dilute alkali. By addingincreasing amounts of antigen to a constantamount of antiserum, the amount required formaximum antibody precipitation (the equiva-lence zone) can be determined (150). The pre-cipitate can also be analyzed for the teichoicacid component to show whether the prepara-tions contain molecules of teichoic acid differingin their determinants. This technique providesevidence of S. aureus teichoic acid chains hav-ing either a or (3 substituents (200, 280), B.subtilis 3610 wall teichoic acid having sub-stituted and unsubstituted chains (96), and L.plantarum NCIB 7220 wall teichoic acid havingchains with different degrees of glucosyl substi-tution (164).The results with L. plantarum also indicated

that only 30% of the acid-extracted teichoic acidwas precipitated by antibody, which suggeststhat many of the chains were of too low amolecular weight to precipitate antibody. Therelation between the amount of antibody pre-cipitated and the molecular weight of the anti-gen has been well established by studies ondextran (102) and has been confirmed withseveral preparations of teichoic acid. Low-molecular-weight trichloroacetic acid-extractedteichoic acid from L. fermenti precipitated only30% of the amount of antibody precipitated byhigh-molecular-weight lipoteichoic acid; afterdeacylation with consequent loss of hydro-phobic interaction between teichoic acid mole-cules, the latter product behaved like the low-molecular-weight fraction in the precipitin reac-tion (160). Deacylation of phenol-extracted li-poteichoic acid from Lactobacillus helveticus(163) and L. plantarum (164) decreased by 50%the amount of antibody precipitated.

Performing precipitin tests in agar by thedouble diffusion (Ouchterlony) method enablesthe detection of a mixture of antigenic compo-nents and of showing whether different bacteria

234 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

contain a particular antigen. This proceduredistinguishes between the S. aureus teichoicacids carrying a-and ,B-N-acetylglucosaminylsubstituents and enabled the presence of one or

both in a particular strain to be established (66,132).The ionic properties of teichoic acids enable

the application of electrophoresis and immuno-electrophoresis. Teichoic acids separated byelectrophoresis at pH 7 in agarose can bevisualized by staining with Toluidine Blue (R.Mollenhauer, B. Sc. Hons. thesis, University ofNew South Wales, Sydney, 1968). The applica-tion of immunoelectrophoresis to extracts ofgroup A streptococci showed that in addition topolyglycerophosphate, a slower-moving compo-

nent (E4) was present, and the fusing of thelines indicated a serological relationship (300).Subsequent work (178) showed that componentE4 contained D-alanine residues, thus account-ing for its lower mobility, whereas the cross-

reaction depended on the common glycerolphosphate backbone. Immunoelectrophoresishas also been used to separate teichoic acid,lipoteichoic acid, and lipoteichoic acid-proteincomplexes from lactobacilli and to show byappropriate absorptions that only antibodies tothe teichoic acid component can be detected(294).

HemagglutinationA method more sensitive than the precipitin

reaction is hemagglutination where erythro-cytes sensitized with the antigen are ag-

glutinated by antibodies specific for the anti-gen. Erythrocytes from a number of animalspecies, including humans and sheep, are suita-ble (36), although in some instances adsorptionis increased by treating the erythrocytes withtrypsin or tannic acid (127). Studies with staph-ylococcal ribitol teichoic acids showed thatthey would not sensitize erythrocytes (64, 107,132, 209) unless the erythrocytes were pre-

treated with chromium chloride (40). Staph-ylococcal "polysaccharide A," which containsribitol teichoic acid, will sensitize erythrocytes(107), but only the peptidoglycan componentunsubstituted with teichoic acid is adsorbed;the teichoic acid and teichoic acid-peptido-glycan fractions are ineffective (107, 108). Ag-glutination apparently depends on antibodiesreacting with f-N-acetylglucosaminyl residuesof peptidoglycan, and the reaction can be in-hibited by S. aureus ribitol teichoic acid car-rying this substituent (108). Latex beads havebeen used as an alternative to erythrocytes,but results with staphylococcal preparationswould indicate that in this case also peptido-

glycan, and not teichoic acid, is adsorbed (64).Studies on the sensitization of erythrocytes bymembrane teichoic acids from streptococciled to the conclusion that alanine ester residueswere essential for loss of alanine under mild al-kaline conditions resulted in loss of activity(143, 198). However, the conditions used wouldalso have removed fatty acids, and later studieswith L. fermenti membrane lipoteichoic acidindicated that fatty acids rather than alanineresidues are required for erythrocyte sensitiza-tion (128). It was shown that alanine-free lipo-teichoic acid sensitized erythrocytes, whereasa low-molecular-weight trichloroacetic acid-extracted teichoic acid, which contained alan-ine but no fatty acids, was inactive; treatmentof a preparation of lipoteichoic acid with aque-ous ammonia liberated fatty acids and markedlydecreased the ability to sensitize erythrocytes.

Adsorption of lipoteichoic acid to erythro-cytes is presumably analogous to the adsorptionof lipopolysaccharides to erythrocytes and othercell membranes, where the lipid component(lipid A) has been shown to play an essentialrole, probably by forming hydrophobic bondswith membrane lipids (53, 115). The fatty acidcontent of lipoteichoic acid is approximately 4to 5%, considerably less than that for lipopoly-saccharides (175, 277), but it is comparable tothat found to be effective in rendering polysac-charides capable of adsorbing to erythrocytes.Hammerling and Westphal (115) first demon-strated this effect when they showed that theaddition of 5% O-stearoyl groups to a number ofpolysaccharides gave optimal erythrocyte-sen-sitizing properties. Additional evidence in sup-port of the essential role of fatty acids forerythrocyte sensitization has been obtained bySlade and co-workers who showed that thehemagglutination procedure can be used fordetecting antibodies reacting with esterified cellwall polysaccharides from streptococci of groupsA and E (216, 269) and esterified glycerolteichoic acid from group A streptococci (188).

Effect of Ionic Concentration

The concentration of various salts is known toinfluence the antigen-antibody reaction, pre-sumably because of effects on the surface chargeand therefore the molecular shape of the react-ants (150). Probably the first study relevant toteichoic acids concerned the specificity of anti-bodies to the type VI pneumococcal polysaccha-ride, which contains a repeating sequence of atrisaccharide joined to ribitol phosphate (232).The phosphorylated repeating unit was four tofive times more potent as an inhibitor than was

VOL. 37, 1973 235

KNOX AND WICKEN

the nonphosphorylated repeating unit, and anumber of sugar phosphates were also effectiveinhibitors. It was correctly presumed that theresults were being influenced by the ionizationof the phosphate group for increasing the NaClconcentration from 0.9% to 11%, to minimizethe charge on the phosphate groups, showedthat the phosphorylated and nonphos-phorylated repeating units were equally effec-tive as inhibitors. It was also shown that theamount of antibody precipitated in 11% NaClwas only 66 to 77% of that precipitated in 0.9%NaCl (232).Morse (196), who investigated the specificity

of Staphylococcus epidermidis glycerol teichoicacid, obtained evidence for nonspecific inhibi-tion of the precipitin reaction by sugar phos-phates, but in other studies that suggested arole for galactose phosphate in the specificity ofgroup N streptococcal glycerol teichoic acid(78), for ribitol phosphate as a determinant inB. subtilis W23 teichoic acid (49), and formannose phosphate as a determinant of yeastphosphomannan (227), the possibility of non-specific inhibition was not examined.More recent studies (166) on the reaction of

the glucosyl-ribitol teichoic acid from L. plan-tarum NCIB 7220 with homologous antiserumhas confirmed that the serological reactivity ofteichoic acid is particularly sensitive to theionic environment. The maximum amount ofantibody was precipitated in the absence ofsalts; 0.4 M NaCl caused 40% inhibition andinhibition by other salts followed the lyotrophicseries. Divalent cations were more effective thanwere monovalent cations, probably becausethey formed complexes with the phosphategroups of teichoic acid. The binding of divalentmetal ions by teichoic acids was indicated bystudies with bacterial cell walls (126) and hasreceived additional support from the observa-tion that calcium ions form an insoluble saltwith cardiolipin by combining with the twophosphate groups in the glycerol-phosphoryl-glycerol-phosphoryl-glycerol portion of the mol-ecule (224).The hydrochlorides of glucosamine and ala-

nine methyl esters also cause nonspecific inhibi-tion of the precipitation of ribitol teichoic acidsby antibodies (166); presumably, the effect isrelated to the increase in ion concentration ingeneral, for the resultant decrease in pH was notsufficient to inhibit the reaction. Nonspecificinhibition by glucosamine hydrochloride couldaccount, at least in part, for its inhibitory effecton the precipitation of group A streptococcalglycerol teichoic acid (187) and group F strep-tococcal antigen by their specific antisera (299),

whereas the inhibition by the alanine ester hy-drochlorides is a factor to be considered indefining the role of D-alanine as an antigenicdeterminant of teichoic acids (see below).The reaction of concanavalin A with wall

teichoic acids from S. aureus (233) and B.subtilis (74) is also inhibited by high saltconcentrations, which explains why the precipi-tin reaction with some a-D-glucosylated teichoicacids in gel diffusion plates was not apparentuntil the agar had been washed with water toremove salt (22). Doyle and Birdsell (74) dem-onstrated significant lowering of the intrinsicviscosity of teichoic acid solutions when theionic strength was raised, and they suggestedthat this indicated that the molecule hadchanged from a rigid rod conformation in a lowionic concentration to a random coil, whichmasks serologically reactive glucosyl groups, athigh ionic concentrations. It is interesting thatthe inhibitory effect of salts in gel diffusionplates was only found with polymers having lessthan one glucosyl residue per polyol phosphateunit (22), whereas more highly substituted tei-choic acids still reacted; the bulky substituentsof these reactive molecules might hinder thesalt-induced formation of a random coil. Theinhibitory effect of salts was not observed whencell walls containing the teichoic acids wereagglutinated by concanavalin A, possibly be-cause of other wall components that modify theconformational changes of the teichoic acidmolecule that might otherwise occur (144).

SPECIFICITY OF ANTIBODIES

General ConsiderationsReactions between antigen and antibody thatare detectable by precipitation or agglutinationdepend on an antigen with more than onedeterminant combining with one of the specificsites on the bivalent IgG or multivalent IgMmolecule. The isolated determinant grouping ofthe antigen will combine with the antibody also,but being monovalent will not precipitate anti-body. Generally, the reaction of the determi-nant grouping with antibody has been shown byan inhibition of the antigen-antibody reaction,although more recently the technique of equilib-rium dialysis replacement has been introducedand shown to give comparable results (113).Either procedure enables conclusions to bemade concerning the portion of the antigencombining with antibody by comparing theresults obtained with components of the antigenderived by acid or enzymic hydrolysis or bycomparing the effects of other known com-pounds.

236 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

The studies by Kabat and co-workers ondextrans (148, 150) indicated that the size of theantibody combining site could be studied bycomparing the abilities of different oligosaccha-rides of glucose to inhibit the precipitin reac-tion, whereas studies on lipopolysaccharides(175, 176) have been instrumental in definingthe roles of different sugars in a polymer, andalso the linkages between them, in determiningspecificty. The extensive studies by Heidel-berger and co-workers (124) have also shown thevalue of cross-reactions between one polysac-charide and antibodies to another as a means ofidentifying specific sugar determinants andtheir linkages.

Inhibition studies will generally indicate thatone of the carbohydrate components of a hetero-polysaccharide is the most effective inhibitor ofthe precipitin reaction, and the term "immuno-dominant," suggested by Heidelberger, (176) isgenerally used to describe such a component.Numerous studies, particularly on lipopolysac-charides (175, 176, 265, 266) have shown thatthe immunodominant sugar may be located atthe nonreducing end of a carbohydrate chain, aterminal determinant, or within the chain. Ithas been proposed (265, 266) that the latterdeterminants be termed "non-terminal" ratherthan internal, for molecular models show thatthey occupy a "sterically superficial position inconsequence of the primary chain buckling."

In many instances, teichoic acids have termi-nal determinants, and specificity depends onthe carbohydrate residues attached to the ribi-tol- or glycerol-phosphate backbone. Evidence fornon-terminal carbohydrate determinants ismost likely to be obtained with teichoic acidsthat contain such residues as an integral part oftheir backbones (Fig. 3), for instance, the typeVI pneumococcal polysaccharide (232). L. plan-tarum C106 forms a teichoic acid of this type(23); the repeating sequence is shown in Fig. 10.Studies on the reaction of this and relatedteichoic acids with concanavalin A led to theconclusion that concanavalin A was reactingwith the nonterminal glucose component of therepeating sequence (22). In other studies on thereaction of teichoic acids with concanavalin A(233), it had been assumed that teichoic acidshaving carbohydrate units as an integral part ofthe backbone would not give a precipitatebecause a reaction depended on "multireactivesites." However, the anomalous results of Ar-chibald and Coapes (22) could be explained bythe shape of this teichoic acid molecule. Sim-mons (265, 266) has shown that a-1 - 2 and a-1- 3 linkages between D-sugars have a profound

effect on the configuration of a polymer, so that

FIG. 10. Proposed effect of a-i - 2 glycosidic link-ages on the shape of L. plantarum C106 wall teichoicacid showing buckling of the main chain with ex-posure of nonterminal glucose residues that canreact with concanavalin A (shown by arrows).

it may be considered as being composed of"steric repeating units," which, in the case ofthe lipopolysaccharides he examined, corre-sponded to the serological repeating units. WithL. plantarum C106 teichoic acid, the proposedeffect of the a-1 - 2 linkage on the shape is alsoindicated in Fig. 10; the teichoic acid does havethe required multireactive sites (shown by ar-rows), but these are nonterminal glucose resi-dues.Simmons (265, 266) has also shown that the 12 or 1 - 3 linkages, which are important in

defining the shape of the molecule, are notinvolved in the combination with the homolo-gous antibody combining site. The term"apodeterminant" was introduced to definesuch determinants which "by conferring a dis-tinctive shape in the molecule .... determinethe sequence and steric accessibility of the moredistant structures, that comprise the antigenicdeterminant without themselves forming anintegral part of it." In this context, we mayconsider the a-1 - 2 linkage in L. plantarumC106 teichoic acid as an apodeterminant, withthe sterically accessible glucose residue havingthe required free hydroxyl groups at positions 3,4, and 6 for interaction with concanavalin A.

Stereochemical restrictions on the shape of ateichoic acid molecule may also influencewhether antibodies will be formed against apotential determinant attached to the teichoicacid backbone. Based on studies on the car-bohydrate determinants of blood groups sub-stances (172) and the Forssman antigen (264), ithas been concluded that only those groups that"have a nonreducing terminal with great stericrigidity appear to be strongly antigenic and thushave greater immunogenetical significance"(264). Whether this generalization can be ex-tended to teichoic acids will depend on whethermore knowledge on the shapes of the polymers isobtained, but it might be expected that thebulky 2-acetamido grouping of the N-acetyl-

VOL. 37, 1973 237

KNOX AND WICKEN

glucosaminyl- substituted ribitol teichoic acidsof S. aureus and the disaccharide substituentsof the Streptococcus faecalis glycerol teichoicacid would hamper steric flexibility; in eachcase these substituents are the antigenic deter-minants.Because antibodies are generally formed

against the substituents on the polyribitol- orpolyglycerol-phosphate backbone, it has fre-quently been concluded that such substituentsmask the backbone and thus prevent the forma-tion of antibodies to the glycerol or ribitolcomponents (43, 152, 178, 179). However,McCarty had noted that certain group A strep-tococcal sera were specific for the glycerolphosphate backbone and reacted with sub-stituted glycerol teichoic acids. Kabat (148)discussed these results in connection with hisproposition that antibodies could be formedagainst one "face" or surface of an antigen, andhe suggested that such sera "would have aspecificity complementary to the -CH- aspectof carbon 2 of the glycerol teichoic acid, and notinvolving the more hydrophilic side of themolecule." Other examples of antibodies spe-cific for the glycerol phosphate sequence in reac-tion with teichoic acids carrying different sugarsubstituents have subsequently been described(163, 296), while the concept of antibodiesreacting with only one face of a polymer hasbeen well established by studies on lipopolysac-charides (175, 176, 265, 266).

Specificity of Antibodies to CarbohydrateSubstituents

Teichoic acids generally have a carbohydratesubstituent, and when this is the immunodomi-nant component, the teichoic acid may proveuseful in serological classification (Tables 3-5).Evidence for an immunodominant carbohy-drate substituent has come from inhibitionstudies and also from cross-reactions. For in-stance, antibodies to the wall ribitol teichoicacid from L. plantarum NCIB 7220 cross-reactwith the membrane glycerol teichoic acid be-cause of the common a-D-glucOsyl substituents(164). The a-D-glucosyl-specific antibodies toboth the ribitol and glycerol teichoic acids of L.plantarum also cross-react with dextrans al-though not with amylose, amylopectin, or glyco-gen (165). This reaction may be related to thepresence of single glucosyl residues joined by1 - 2 or 1 - 3 linkages to the main chains ofdextran. The antibody specificity does not re-quire such a linkage, but the results mightprovide another example of the importance ofthese linkages in defining the accessibility of anantigenic determinant (265, 266).

The specificity of antibodies may be suchthat structurally related carbohydrates can oc-cupy the combining site. This would account forthe observation that galactose inhibits the reac-tion between L. plantarum glucosyl-substitutedglycerol teichoic acid and homologous anti-bodies and also for the reaction of these anti-bodies with the galactosyl-substituted glycerolteichoic acid from L. fermenti (164).The extent to which antibody specificity is

shared between the glycosyl substituent(s) andthe backbone of polyol phosphate has beenexamined in only a few instances. Burger (43)provided evidence that glycerol contributed tothe specificty of antibodies to the glycerolteichoic acid from B. subtilis 3610 by showingthat 2-0-a-D-glucosyl-glycerol was a considera-bly more effective inhibitor than was 3-O-a-D-glucosyl-glycerol; however, glucose must be themajor contributor, for the antiserum did notprecipitate unsubstituted glycerol teichoic acid.A similar conclusion was reached with theglucosyl- ribitol teichoic acid of L. plantarumNCIB 7220, where a-D-glucosyl-ribitol was amore effective inhibitor than methyl-a-D-gluco-side, but the antiserum did not react withunsubstituted ribitol teichoic acid (164).With glycerol teichoic acids, the relative con-

tributions of glycosidic substituents and glyc-erol phosphate to specificity can be examinedby comparing inhibitions by the appropriatemethyl glycoside and glycerol-phosphoryl-glycerol-phosphoryl-glycerol, which is obtainedfrom cardiolipin and contains 1 - 3 linkedglycerol phosphate units. By this means it wasshown (164) that antibodies to the membraneteichoic acid of L. plantarum NCIB 7220 dif-fered in their specificity, depending on thepreparation injected-whole "cell wall" gaveantibodies specific for the glucosyl substituents,trypsinized "cell wall" that still contained re-sidual membrane gave antibodies specific forthe glycerol phosphate backbone of the mem-brane teichoic acid, whereas antibodies to theisolated lipoteichoic acid were specific for bothglucose and glycerol phosphate.

Individual rabbit variations also lead to dif-ferences in antibody specificity. With L. fer-menti membrane teichoic acid, where the sub-stituents are D-galactose and a-D-galactosyl-1 -> 2-D-glucose, a comparison of 11 rabbit serashowed that 100 Imol of D-galactose gave 5 to47% inhibition (median 32%), whereas 100,jmolof D-glucose gave 0 to 22% inhibition (median7%); those sera with low sugar specificity wereshown to be primarily specific for the glycerolphosphate backbone (160, 296). Differences inthe specificity of sera could also account for

238 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

TABLE 3. Glycerol teichoic acids as grouping antigens of streptococci

Teichoic acidOrganism References

Location Substituent(s)a Determinant

Group D Membrane Glc-a-1 - 2-GIc Glc-a-1 - 2-Glc 291, 292Group N Membrane Gal-phosphate ? ? 78, 125S. mutansGroup I (a) Membrane Glc, Gal #-Gal 281, 282Group 1 (b) Wall Gal ,-Gal 34b

a See Table 1 for abbreviations.A. S. Bleiweis, personal communication.

TABLE 4. Group antigens of lactobacilli

Serological ~~~~~~~~~~~~Immuno-Seroogicalr Group antigen Location Componentsa dominant Referencescomponent

A Teichoic acid Wall? Membrane Glycerol, a-Glc a-Glc 163, 192, 254B Polysaccharide Wall Rha, Glc, Gal, GlcNAc, a-Rha 98, 157, 158

GalNAcC Polysaccharide Wall Glc, Gal, GlcNAc, P-Glc 98, 157, 158

GalNAc 159D Teichoic acid Wall Ribitol, a-Glc a-Glc 164, 254E Teichoic acid Wall Glycerol, a-Glc ? 21, 254F Teichoic acid Membrane Glycerol, Gal, a-Gal 1601

Gal-a-1 -_ 2-Glca Rha, L-rhamnose; see Table 1 for other abbreviations.b A. J. Wicken and K. W. Knox, unpublished data.

TABLE 5. Teichoic acid antigens of staphylococci and micrococci

Staphylococcus Polyol Substituent° Polysaccharide Occurrence in Referencessynonyms micrococci

aureusstrain Wood 46 Ribitol P-GlcNAc A = A Biotypes 2-5, 7 66, 119, 211strain 263 Ribitol a-GlcNAc 263 = Aa _ 132, 211

epidermidisstrain T1 Glycerol a-Glc B = Ba Biotypes 1, 6 1,65,66, 211, 212strain T2 Glycerol ,-Glc BO 65, 66, 211

a See Table 1 for abbreviations.

conflicting conclusions on whether group Dstreptococcal teichoic acid cross-reacts withantisera to group A streptococci (78, 179).

D-Alanine as an Antigenic DeterminantIn contrast to the observation that antibodies

are frequently formed against the carbohydratecomponents of teichoic acids, there are only afew instances of antibodies with specificity forthe D-alanine substituents, namely, certain rab-bit antisera to group A streptococci (178, 188)and S. aureus phage type 187 (152). This lack ofimmunogenicity may be related to the markedlability of the alanine ester linkage, for D-ala-nine components of peptidoglycan (246) andpoly-D, L-alanyl-proteins (244) have been shownto be immunodominant.

Evidence for D-alanine being an immunodom-inant component of teichoic acids was providedby the loss of serological reactivity concomitantwith the hydrolysis of ester linkages and byD-alanine methyl ester hydrochloride being abetter inhibitor of the precipitin reaction thanL-alanine methyl ester hydrochloride; D-ala-nine, itself, did not inhibit (152, 178, 188).However, the alanine ester hydrochlorides cancause nonspecific inhibition similar to thatcaused by other ionized compounds (166).Treatment of L. plantarum NCIB 7220 ribitolteichoic acid with ammonia gave an alanine-free product retaining 95% of its serologicalactivity with homologous antiserum, yet theprecipitin reaction was inhibited 32% by 100Mmol of D, L-alanine methyl ester hydrochloride

VOL. 37, 1973 239

KNOX AND WICKEN

and 35% by 100 ,UmoI of L-alanine methyl esterhydrochloride. Injection of L. plantarum cellwall, from which alanine esters were removedby hydroxylamine, gave antiserum whose reac-tion with homologous teichoic acid was simi-larly inhibited by the ester hydrochlorides; foreach system 100 Amol of sodium chloride gave33% inhibition. From these results, it may beconcluded that the reported inhibition by theL-alanine methyl ester is nonspecific and thatthe specific effect of the D-alanine component ofthe ester is less than has been generally con-cluded.

Antibodies Specific for Glycerol PhosphateMcCarty (177) showed that antibodies spe-

cific for the glycerol phosphate backbone ofteichoic acids could be obtained from certainrabbits injected with group A streptococci;evidence was provided by reaction of the serawith unsubstituted glycerol teichoic acid and byinhibition by synthetic polyglycerophosphate(average chain length = six units). Glycosylatedteichoic acids from lactobacilli also cross-reactwith group A streptococcal antibodies, and thecross-reaction is inhibited by prior absorption ofthe serum with polyglycerophosphate (163).Analysis of the precipitate formed with L.fermenti lipoteichoic acid confirmed that glyco-syl-substituted teichoic acid was being precipi-tated.

Additional evidence for a sequence of glycerolphosphate units being immunogenic determi-nants in glycosyl-substituted teichoic acids hascome from studies with glycerol-phosphoryl-glycerol-phosphoryl-glycerol (G3P2) as an in-hibitor (296). In the reaction between L.helveticus lipoteichoic acid and homologousantiserum, the antibodies are primarily specificfor the glucosyl substituents, and 2 umol ofGP2gave only 5% inhibition, possibly because ofnonspecific ionic inhibition. However, for lipo-teichoic acids from L. plantarum, L. fermenti,and L. casei, where there is a much lower degreeof substitution, 2 ,umol of G3P2 was a veryeffective inhibitor and varied from 35 (L.fermenti) to 75% (L. casei NCTC 6375). Thesize of the immunogenic determinant has notbeen established, although specificity does re-quire that the glycerol units be joined byphosphodiester bonds involving positions 1 and3 of glycerol, for the wall teichoic acid from B.stearothermophilus, which has phosphodiesterbonds between positions 2 and 3 of glycerol(290), did not react.The degree to which specificity of sera related

to the glycerol phosphate backbone was re-flected in the cross-reactions between the lac-

tobacillus teichoic acids (296) and the resultsfor L. helveticus, L. fermenti, and L. casei arecompared in Fig. 11. Also included are resultsshowing the extent to which each of the reac-tions that precipitated sufficient antibody wasinhibited by 2 ,umol of GP2 (296). A comparisonof the two sets of results shows the importanceof glycerol phosphate determinants in the cross-reactions.The inhibition of the precipitation of glycerol

teichoic acid by the glycerol phosphate compo-nent of cardiolipin suggested the use of the serato detect cardiolipin; cardiolipin is employed inseveral complement fixation tests for syphilis,and antibodies to cardiolipin are specific for thephosphoryl glycerol phosphate component(112). It was shown (293) that rabbit antibodiesspecific for the glycerol phosphate component oflipoteichoic acids reacted as reagin in theKolmer complement fixation test, although notin the more specific tests for treponemal infec-tions; reagin produced in the few cases ofhuman syphilitic infection that were tested didnot react with lipoteichoic acid. These resultscould explain some of the biological false-posi-tive reactions for syphilis obtained in caseswhere there was a recent history of gram-posi-tive bacterial infection (93) and a consequentpossibility of a high titer of antibody to teichoicacid.That human sera may contain a significant

titer of antibody to membrane teichoic acidshas been shown by a number of studies. In mostof these, as is discussed in the next section, the

100

L.heliveticusAL.fermenti

L.casei75 U GP, Inhib.

0~

L. helveticus L.fermenti L. casel R094Antisera

FIG. 11. Cross-reactions between lipoteichoic acidsfrom L. helveticus, L. fermenti, and L. casei R094 andrabbit antisera to the lipoteichoic acid preparations.Results are expressed as a percentage of the amount ofantibody precipitated in the homologous reaction.Where sufficient antibody was precipitated, the per-cent inhibition of the reaction by 2 Mmol of glycerol-phosphoryl-glycerol-phosphoryl-glycerol is alsoshown. [Data from A. J. Wicken and K. W. Knox(296).]

BACTERIOL. REV.240

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

structure of the reactive teichoic acid was notknown. However, a comparison of the reactivityof 53 human sera with erythrocytes sensitizedwith lipoteichoic acids from L. plantarum, L.fermenti, and L. casei showed that 32 haddetectable antibodies, and that 17 reacted witheach of the teichoic acids and contained cross-reacting antibodies (182).The presence of cross-reactive antibodies

specific for the polyglycerophosphate backbonecould account, at least in part, for the observa-tions that human sera contain antibodies ag-glutinating erythrocytes sensitized with ex-tracts of oral streptococci (171) and teichoicacids from two Bacillus species (67). In a studyon oral streptococci, where a positive correlationwas obtained between the amount of dentalcaries and antibody reacting with extracts ofcertain cariogenic streptococci, it was assumedthat the antibodies detected were specific forthe particular strain, and a causal relationshipwas invoked (171). In another study (67), on"naturally occurring antibodies to bacillary tei-choic acids," the presence of cross-reactingantibodies was even more certain, for the au-thors show that specificity depends on thepolyglycerophosphate backbone.

Glycerol Teichoic Acids as HeterophileAntigens

The conclusion that glycerol teichoic acidsdiffering in glycosyl substituents can react withantibodies specific for the glycerol phosphatesequence would account for a number of obser-vations on cross-reactions between gram-posi-tive bacteria from diverse groups and has led tothe use of such terms as nonspecies specific(225), Rantz (104), heterophile (179), or hetero-genetic antigens (51).The first definitive evidence was provided by

McCarty's study (177) with the group A strep-tococcal antiserum that reacted with poly-glycerophosphate, and a precipitate was ob-tained with acid extracts from a variety ofgram-positive, but not gram-negative, bacteria.Antiserum to a strain of L. acidophilus, whichreacts with acid extracts of various gram-posi-tive bacteria (252, 253), has also been shown tocontain antibodies specific for the polyglycero-phosphate backbone (253).Most studies on heterophile antigens of gram-

positive bacteria have used the more sensitivehemagglutination method for detection. Thereaction of human sera with erythrocytes sensi-tized with culture filtrates from streptococciand staphylococci (90, 215) was interpreted asindicating a common antigen, and evidence fora more wide-spread occurrence of such antigens

among gram-positive bacteria was obtained byRantz and co-workers (225, 226); they alsoshowed that hot-water extraction of cells gaveactive material, introduced phenol extraction,and obtained suitable rabbit antisera.The identification of intracellular teichoic

acid as an antigenic component (177) led to thesuggestion by Salton that the erythrocyte-sen-sitizing antigen(s) known to be wide-spreadamong gram-positive bacteria might also beteichoic acid (see 104). This was shown to be thecase, for McCarty's polyglycerophosphate in-hibited the hemagglutination of erythrocytessensitized with lysates of S. aureus (104). Con-firmatory evidence was obtained by Stewart(271), who inhibited the agglutination of eryth-rocytes sensitized with streptococcal culturefluids containing "Hickey antigen" by addingsynthetic polyglycerophosphate.

Neter and co-workers conducted a detailedstudy of erythrocyte-sensitizing antigens ofgram-positive bacteria and improved themethod of detection (206), provided evidencefor placental transfer of antibodies reactingwith Rantz antigen (207), and extended the listof reactive organisms to include L.monocytogenes (205); this observation provideda basis for the known serological cross-reactionbetween Listeria sp. and S. aureus. Not allgram-positive bacteria examined gave an activeproduct, and negative results were obtained, forinstance, with Bacillus brevis (9) and "M.lysodeikticus" (M. luteus) (104). Other inves-tigators had also noted that not all gram-posi-tive organisms produced a reactive component,but in many cases a negative reaction couldrelate to difficulties in extracting and detectingthe antigen. For instance, the conclusion for B.brevis was based on the lack of reaction of theculture filtrate and the failure of cells to absorbantibodies (9), and a later examination of a hotsaline extract of another strain did give apositive result (51). However the study on"Micrococcus lysodeikticus" (M. luteus) usedlysozyme digests (104) and indicated that moreresearch is warranted to determine whether thisorganism does contain the expected membraneglycerol teichoic acid.

All of the above studies indicating the occur-rence of cross-reactive teichoic acids in a varietyof gram-positive bacteria depended on havingantibodies reacting with a common structuralfeature, presumably the glycerol phosphate se-quence. However, antibodies to teichoic acidsmay also be specific for the glycosyl substitu-ents and would account for the specificities ofthose human sera that differed in their ability toagglutinate erythrocytes sensitized with lipotei-

VOL. 37, 1973 241

KNOX AND WICKEN

choic acids from different lactobacilli (182).Such differences in specificity would also ex-plain the differences observed in the ability ofindividual human and rabbit sera to aggluti-nate erythrocytes sensitized with culture fil-trates (272, 274) or hot saline extracts (51, 181,273) of different gram-positive bacteria. Theerythrocyte-sensitizing antigens had the proper-ties of lipoteichoic acids, and evidence wasobtained in two instances for the presence ofteichoic acids (67, 271). In some cases, extractsfrom an individual strain reacted with sera ofdiffering specificity, and it was assumed thatthis indicated that a corresponding number ofantigens were present. Alternatively, the seracould be detecting different antigenic determi-nants on the same teichoic acid molecule.To establish the presence in different bacteria

of antigens with the same determinant group-ing, investigators have generally used the classi-cal technique of absorption of antisera withintact organisms. Where absorption occurs,there is certainly evidence for the antigen beingpresent (225). However, conclusions based on alack of absorption of antibodies, the maincriterion used by Stewart and co-workers fordelineating the antigens of streptococci andlactobacilli, do not take into account the situa-tion that membrane teichoic acid may not beavailable to react with antibodies. A moresatisfactory method was employed by Chorpen-ning and Dodd (51) who used erythrocytessensitized with antigen for absorbing antibodiesfrom human and rabbit sera.

TEICHOIC ACIDS AS GROUPANTIGENS

Teichoic acids have proved to be of onlylimited usefulness as antigenic components inthe classification of bacteria, probably becauseof the restricted array of sugar substituents onthe polyol phosphate backbone and the varia-ble, and apparently uncontrolled, degree ofsugar substitution that can occur; dissacharidesubstituents which would extend the possibili-ties of structural diversity appear to occur onlyin membrane teichoic acids (Table 1). Amongthe streptococci, where serological identifica-tion is frequently the preferred method, teichoicacids are the group antigens in only a minorityof instances (Table 3). They do, however, have amajor role in the serological classification oflactobacilli (Table 4) and may also aid in theclassification of staphylococci and micrococci(Table 5).

StreptococciThe reactivity of acid extracts of streptococci

with specific antisera (170) has enabled approx-

imately 20 alphabetically designated serologicalgroups to be identified. Group A streptococci,because of their pathogenicity, have probablyreceived the most attention, and McCarty'sstudies defined the cell wall polysaccharide asthe group antigen (179). Other studies by nu-merous investigators indicate that wall polysac-charides are the specific components for most ofthe serological groups, the principle exceptionsbeing groups D and N (101).The classification of streptococci into group D

(Table 3) depends on a specific membranelipoteichoic acid (43, 77, 279, 291, 292), andanother division into serological types dependson specific cell wall polysaccharides (69, 75, 76,217, 255). Because of its location, the presenceof the group antigen may prove to be difficult toestablish, both with regard to antibody produc-tion and the detection of the antigen in anextract (257, 270). Certain strains of strep-tococci originally classified into group Q on thebasis of their specific cell wall polysaccharide(270) are now known to contain the group Dantigen (208, 270). These strains, designatedStreptococcus avium, differ in their physiologi-cal characteristics from the established group Denterococci, S. faecalis and S. faecium, and ithas been proposed that "the group Q strep-tococci constitute a valid species which shouldbe included in the serological group D strep-tococci" (208). Nowlan and Deibel (208) foundit "difficult to reconcile" the physiological dif-ferences between the group Q and D strains "onthe basis of a variation in type-specificity."However, there is no a priori reason for anti-genic properties correlating with physiologicalcharacteristics, and it would seem logical toconsider the group Q antigen as a type-specificpolysaccharide present in certain group D strep-tococci.The specificity of antibodies to the group N

antigen (Table 3) has not been clarified; galac-tose did not inhibit the precipitin reaction, andthe inhibition by galactose-1-phosphate andgalactose-6-phosphate, rather than indicatingthat galactose phosphate is the serological de-terminant (78), could be another example ofnonspecific inhibition by a sugar phosphate(196, 232). Further, group N antiserum cross-reacted with only one galactose-containingpneumococcal polysaccharide, type XVI, whichalso contained glycerol phosphate (125). Glyce-rol phosphate cannot be a major contributor toserological specificity, for group N antibodiesdid not precipitate group A streptococcal poly-glycerophosphate (78).The oral streptococci have evoked considera-

ble attention because of their role in dentalcaries. Although certain of the oral streptococci

242 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

can be classified into one or other of theLancefield groups (158), the cariogenic strains,usually classified as Streptococcus mutans or S.sanguis, do not belong to any of the knownserological groups. Schemes for the serologicalclassification of S. mutans have been proposedon the basis of the reactivity of cell extracts withrabbit antisera (38, 142). An examination ofrepresentative strains by Bleiweis and co-work-ers has shown (Table 3) that those in groups Iand II (142), also called groups a and b (38),contain specific teichoic acids (34, 281, 282, A.S. Bleiweis, personal communication). There isevidence that S. sanguis strains also have a

group-specific teichoic acid, one with the prop-

erties of a membrane component, but not ex-tractable with trichloroacetic acid or phenol(238a; B. Rosan, Abstr. Int. Ass. Dent. Res.1972, p. 265).

LactobacilliAttempts to develop a serological classifica-

tion of lactobacilli have used both cell aggluti-nation and the precipitin reaction. The resultsof a number of investigations on agglutinatingantigens indicate a complexity of surface com-

ponents, generally type-specific and with only a

limited applicability to classification, for theyoften cut across species as defined on the basisof physiological properties (252).

Studies by Sharpe and co-workers showedthat the precipitin reaction, by using acidextracts and specific antisera, provides a muchmore satisfactory method for the serologicalclassification of lactobacilli. An examination of442 strains, representing all species then recog-nized, showed that 70% could be classified as

belonging to one of six groups and one sub-group (251) designated by the letters A to F(256); subsequently group G was defined (238).Broadly, the classification is in agreement withone based on the physiological characteristics ofthe strains, although strains of L. casei belongto one of two groups, B and C. The lactobacilliwere one of the first genera to be surveyed forthe occurrence of teichoic acids (31), and subse-quent work has shown that they are the antigensdefining four of the six groups examined (Table4); the group G antigen has not been identified.

In group A, there have been conflicting con-

clusions on whether the wall or membraneteichoic acid is the group antigen, and studiesusing the precipitin method implicated themembrane component (254), whereas aggluti-nation reactions suggested that the wall compo-nent was responsible (192). However, it is nowapparent from studies on L. helveticus NCIB8025 that the wall and membrane glycerol

teichoic acids are of similar structure, eachhaving a-D-glucosyl substituents (163). The iso-lated membrane lipoteichoic acid is immunoge-nic, with the antibodies being primarily specificfor the a-D-glucosyl substituents and, there-fore, cross-reacting with wall teichoic acid (163).In terms of the reactivity of Lancefield acid ex-tracts of group A organisms (251) or theiragglutinability (192), it is to be expected thatthe wall teichoic acid would be the majorcontributor to the serological reaction becauseof the greater amount of teichoic acid in the wall(163). However, none of the studies on the groupA lactobacilli has resolved whether the anti-bodies formed on injection of whole organismsare elicited by the wall teichoic acid, the mem-brane teichoic acid, or both. Thus the classifica-tion of a strain as group A could depend on theproduction of antibodies to the membrane tei-choic acid and their reaction primarily with thewall component.The group E antigen has also been identified

as a glucosyl-substituted glycerol teichoic acidwhich is present in the wall (254); an examina-tion of one strain, L. buchneri NCIB 8007,showed that there were a-D-glucosyl substitu-ents (21). The main difference between thisteichoic acid and the group A antigen is thedegree of sugar substitution, and the ratio ofglucose to phosphorus is 0.64: 1.00 and0.45:1.00, respectively, for the wall and mem-brane teichoic acids from L. helveticus NCIB8025 (group A) (163) as compared with0.26: 1.00 for Lactobacillus buchneri (21). TheL. buchneri teichoic acid is randomly sub-stituted with glucosyl residues, and D-alanylester residues are attached to most of theremaining glycerol units (21); antibody specific-ity may be related to the high degree of D-ala-nine substitution, for specificity based entirelyon a-D-glucOsyl substituents would be insuffi-cient to distinguish the group E antigen fromthe group A antigen.The group F antigen is also a glycerol teichoic

acid; it is present in the membrane (159, 254)and differs from the A and E antigens bycontaining both glucose and galactose, withgalactose being primarily responsible for sero-logical specificity (Table 4).Group D lactobacilli, represented by most

strains of L. plantarum (including L.arabinosus 17-5), contain a wall ribitol teichoicacid (15) and a membrane glycerol teichoic acid(59), each of which bears immunodominanta-D-glucOsyl substituents (164); antibodies tothe wall teichoic acid cross-react with the mem-brane component (164). The conclusion that thewall teichoic acid is the group antigen has beenconfirmed by observations that strains lacking

243VOL. 37, 1973

KNOX AND WICKEN

the glucosyl substituents on the ribitol teichoicacid (73, 164) or containing a different teichoicacid (4) do not react with group-specific an-tiserum.

In general, injection of group D lactobacillidoes not yield antibodies to the membranecomponent, although these can be obtained byinjecting crude cell wall or lipoteichoic acid inFreund adjuvant (164). A similar situation alsoexists with strains of L. casei where injection ofwhole organisms gives antibodies specific forthe cell wall polysaccharides (Table 4), andantibodies to the membrane component canonly be obtained by injecting disintegratedorganisms (Knox and Wicken, unpublisheddata) or lipoteichoic acid in Freund adjuvant(296). The L. casei lipoteichoic acids containminor amounts of glucose and galactose, al-though earlier studies on one of the strains(ATCC 7469 = NCTC 6375) described a tri-chloroacetic acid-extracted teichoic acid pre-sumed to be devoid of carbohydrate substitu-ents (155).As discussed previously, antisera to a number

of the purified lipoteichoic acids will cross-reactwith lipoteichoic acids from serologically un-related lactobacilli because of immunodomi-nant a-D-glucosyl substituents or glycerol phos-phate units (160, 163, 296). These observationsseem to militate against an efficacious serologi-cal classification of lactobacilli. However, basedon the reactivity of acid extracts of organismswith antiserum to whole organisms (251), such aclassification has been developed. These twoparameters may well be crucial, for injection ofwhole organisms will frequently fail to elicit theproduction of antibodies to the membrane lipo-teichoic acid, and acid-extracted teichoic acidsreact only poorly in the precipitin reaction incomparison with the high-molecular-weight li-poteichoic acids (160, 163).

StaphylococciA serological distinction between virulent and

avirulent staphylococci was first indicated bythe studies of Wieghard and Julianelle (298)who isolated fractions containing serologicallyactive phosphorylated polysaccharides fromtype A (virulent) and type B (avirulent) strains.The type A strains would now be classified as S.aureus and are generally distinguishable fromtype B strains by the production of coagulaseand ability to ferment mannitol. The currentlypreferred name for type B strains is Staph-ylococcus epidermidis, although they havealso been designated S. albus and S.saprophyticus (179).The relationships of polysaccharides A and B

to teichoic acids and their current synonyms aresummarized in Table 5. The rather confusingliterature on the nomenclature of staphylococ-cal antigens also contains references to Jensen'santigen A, a component originally thought to bea polysaccharide, but now known to be a protein(54). Antigen A appears to be the major ag-glutinogen within S. aureus strains; a largenumber of type-specific agglutinogens areknown, most of which are presumed to beproteins, whereas the wall teichoic acids do notseem to contribute to the agglutination reactionof whole organisms (209). The serological detec-tion of teichoic acid in extracts of S. aureus isinfluenced by whether the antiserum used de-tects only a- or fl-N-acetyl-glucosaminyl sub-stituents, and it has therefore been concludedthat tests for the detection of wall teichoic acidwill not replace the coagulase test in identifyingthe majority of S. aureus strains (197).A number of types of S. epidermidis can also

be distinguished by cell agglutination, althoughthe specific agglutinogens have not been identi-fied (1). Double-diffusion precipitin reactionsshowed, however, that extracts from two-thirdsof the strains examined contained an a-D-gluco-syl-substituted glycerol teichoic acid (1). Be-cause this precipitin method readily distin-guishes between the S. epidermidis and S.aureus teichoic acids and because some strainsof S. aureus are coagulase negative, the serologi-cal detection of teichoic acid has been proposedas a means of differentiating coagulase-negativestaphylococci (1).

MicrococciMicrococci are similar to staphylococci in

their physiological characteristics and, as aresult, certain strains of micrococci that havebeen examined for teichoic acid componentshave been incorrectly classified as S.epidermidis (173, 211) and S. lactis (16, 17, 32).The naming of species has also led to confusion,and it has been proposed that there should onlybe two, Micrococcus roseus and M. luteus, thelatter including strains of M. lysodeikticus (32,33).

Results of chemical (30, 109) and serologicalstudies (174, 211) indicated that some strainscontain a cell wall ribitol teichoic acid, whereasin others a cell wall glycerol teichoic acid maybe present. The ribitol teichoic acid is indistin-guishable, upon double diffusion in agar, fromS. aureus ,B-N-acetylglucosaminyl-substitutedribitol teichoic acid (174) and has been shown(Table 5) to be present in strains representingfive of the eight biotypes (211); these strainsalso contain an additional unidentified compo-

244 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

nent, C, which is detectable upon doublediffusion in agar, and the extract containingboth components is called polysaccharide A#C(174). Other strains (Table 5) contained anantigen indistinguishable from a-D-glucosyl-substituted glycerol teichoic acid upon agar geldouble diffusion, whereas biotype 8 strainslacked detectable antigenic components. Theresults of these studies led to the conclusionthat the type of teichoic acid present "should begiven more weight" in the classification ofmicrococci (211).IMMUNOBIOLOGICAL PROPERTIES OF

TEICHOIC ACIDSWall-Associated Teichoic Acid-Peptido-

glycan ComplexesCell walls of various gram-positive bacteria

are known to be toxic to animal tissues. Wallfragments of group A streptococci producechronic dermal lesions in rabbits (248), injury tothe joints of rabbits (249), and rheumatic car-diac-like lesions in mice (213). Dermal lesions inrabbits, similar to those produced by group Astreptococci, are produced at the site of injec-tion of cell walls of a number of gram-positivebacteria, including S. aureus, streptococci ofgroups B, C, E, F, and K, L. casei, Actinomycesisraelii, and Actinomyces (Odontomyces)viscosus (247). Peritoneal abscess formation inmice has also been achieved with S. aureus cellwalls (169).The rate and pattern of degradation of differ-

ent species of bacteria after phagocytosis is veryvariable (56), and it is known, for instance, thatgroup A streptococcal cell walls may persist inmigrating human phagocytes for sufficient peri-ods of time to be deposited in body tissues andmay thus be an important factor in the patho-genesis of post-streptococcal sequelae in hu-mans (99). Although the toxicity of many bac-terial cell walls has been established from theirability to either produce or perpetuate chronicinflammatory lesions in animal tissue, questionsas to whether these are due to direct toxic ef-fects or hypersensitivity reactions as well as therole(s) of the various antigenic components ofthe cell walls remain, in most instances, to beanswered.

Studies on the dermal toxicity in rabbits ofcell walls of several gram-positive bacteria (2,247) have indicated clearly that the peptidogly-can component of the walls is responsible for thelesions produced. A large particle size is alsoimportant in determining the toxic activity;acid hydrolysis progressively decreases the ac-tivity. Similar conclusions concerning peptido-glycan being the toxic principle of S. aureus cell

walls were reached in a series of skin test studies(167) on guinea pigs with a range of carefullyfractionated wall antigens. The latter includedviable and heat-killed whole organisms, cellwalls, peptidoglycan complexes, teichoic acid,teichoic acid-peptidoglycan complexes, andlow-molecular-weight peptidoglycan fragments.In nonsensitized animals, all antigens but tei-choic acid elicited acute inflammatory reactionswhich decreased in size after 10 h. In animalssensitized with whole organisms, the reactionsto all antigens but teichoic acid and peptidogly-can fragments remained erythematous and in-durated for at least 30 h and were interpreted asdelayed hypersensitivity type reactions. Al-though teichoic acid was apparently not directlyinvolved in the formation of these dermal le-sions, it was suggested to have an indirect role;teichoic acid-peptidoglycan fragments, but notpeptidoglycan fragments alone, are capable ofevoking the hypersensitivity reactions in sensi-tized animals. This, as the authors suggest, maybe due to the greater particle size of the teichoicacid-peptidoglycan fragments or their greaterpersistence in animal tissue through inhibitionof peptidoglycan digestion by associated tei-choic acid. Of the two strains of S. aureus usedin this study, it is interesting to note that strainCopenhagen had a greater sensitizing capacitythan did strain 263; the former, in common withpeptidoglycan, has a predominantly ,8-linkedN-acetylglucosaminyl-substituted teichoicacid, whereas in the latter strain the glucosa-mine substitution is essentially a in configura-tion.

Cutaneous hypersensitivity in humans tostaphylococcal wall teichoic acid has beenclaimed (185), but the test antigen used waspurified by electrophoresis from a crude extractprepared by sonic treatment of whole orga-nisms and was likely, therefore, to have been ateichoic acid-peptidoglycan complex. Earlier re-ports (47, 199) that S. aureus cell wall teichoicacid was responsible for removing phagocytosis-promoting and killing factors for this organismin human sera have subsequently been disproved(262), and once again peptidoglycan of largemolecular size appears to be the responsible cellwall component. Studies on the toxicity ofsurface antigens of S. aureus, as measured bytheir effects on the respiration of mouse livertissue (154), suggested that wall-associated tei-choic acid and peptidoglycan were the toxicfactors involved; the various fractions and ex-tracts tested were not, however, characterizedchemically. The cellular toxicity of streptococ-cal peptidoglycan, noted earlier, is paralleled bya general inhibition of phagocytosis by rabbit

VOL. 37, 1973 245

KNOX AND WICKEN

polymorphs, and this appears to be due to adirect toxic effect of the peptidoglycan on thephagocytic cells rather than to absorption andneutralization of opsonins and other factors(146).Thus, on the basis of the evidence presently

available, it seems unlikely that wall-associatedteichoic acid has a direct role in the toxicity ofcell walls.

Membrane Lipoteichoic AcidsRheumatic fever and glomerulonephritis are

well-known sequelae to f3-hemolytic group Astreptococcal infections, and in both cases animmunological etiology, involving humoral an-tibody to a streptococcal component, has beengenerally accepted. Controversy exists as towhether the mechanism of pathogenesis in-volves antibody combination with antigenfixed at the site of tissue damage, or whethercombination is intravascular, involving cir-culating antigen followed by deposition of theimmune complex at the site of subsequenttissue damage. The identity of the antigen(s)involved also has not been resolved nor whetherit is truly of streptococcal origin or is instead ahost tissue component capable of cross-reactingwith antibody to a streptococcal antigen (241,312).The ready attachment of membrane lipotei-

choic acids to erythrocytes by a presumedhydrophobic interaction of their glycolipidmoiety with the erythrocyte membrane hasbeen discussed earlier in this review. Sensitiza-tion of erythrocytes with streptococcal mem-brane teichoic acid (143, 198) has been shown tobe reversible, for antigen is able to be trans-ferred from erythrocytes to tissues and viceversa. It has been suggested (198) that this mayindicate a mechanism whereby streptococcalantigen can be distributed to host tissues, andperhaps this plays a role in the pathogenesis ofrheumatic fever (and glomerulonephritis).Whether or not the binding of lipoteichoic acidto animal cell membranes is nonspecific orspecific in involving a finite number of "recep-tor-sites" remains to be determined. If the latterproves to be true lipoteichoic acid, carried byerythrocytes from the site of streptococcal infec-tion, could become concentrated at host mem-brane surfaces where there are a large number ofreceptor sites or sites of high affinity for lipotei-choic acid. This hypothesis is attractive and,certainly, among the wide range of streptococcalantigens which have thus far been endowedwith the role of causative or reactive antigen inthese two diseases, lipoteichoic acid has re-ceived the least attention. A major objection to

this role for streptococcal lipoteichoic acid is itspotential lack of specificity in contrast to thehigh degree of strain specificity of group Astreptococci in causing these diseases, particu-larly streptococcal glomerulonephritis. How-ever, lipoteichoic acids are invariably obtainedin vitro, and are possibly released in vivo, ascomplexes with protein. It is possible that theassociated protein provides the necessary speci-ficity of the immune reaction, whereas thelipoteichoic acid acts as a carrier that is selec-tive for particular host tissues.Complement fixation by antibody reaction

with erythrocytes sensitized with lipoteichoicacid and subsequent hemolysis has been dem-onstrated in a sheep erythrocyte-rabbit anti-body system (293). Antibody agglutination ofvarious animal erythrocytes sensitized with theRantz antigen from a variety of bacterialsources has been referred to earlier (see GlycerolTeichoic Acids as Heterophile Antigens sec-tion), and may in the presence of complementlead to lysis. The clinical importance is wellillustrated by at least one recorded case oftransfusion of contaminated blood leading tosevere hemolysis and death (50).Repeated injection into rabbits of "the mem-

brane-associated teichoic acid" from S.pyogenes (285) has been reported to result inmarked calcification and tubular necrosis of thekidneys without any significant increase inhumoral antiteichoic acid antibodies. Themethod of preparation of the teichoic acid wasnot described in this preliminary report (285),but it seems likely that the lipoteichoic acidused in earlier studies (143, 198) was involved.Removal of D-alanyl esters from the teichoicacid with alkali rendered it inactive and, as hasbeen discussed earlier, such a procedure wouldalso effectively deacylate the glycolipid moietyof a lipoteichoic acid.

Lipoteichoic acids are amphipathic moleculesthat possess both hydrophilic and hydrophobicgroupings, and it would be expected that thevarious biological properties that have beendescribed above, particularly those requiringadsorption to cell membranes, relate to thisproperty. Lipopolysaccharides of gram-negativebacteria are also amphipathic molecules, andthey resemble lipoteichoic acids in their physi-co-chemical properties. Injection of lipopolysac-charides into higher animals causes a number ofcharacteristic and well-known reactions, in-cluding fever, shock, and death (endotoxic reac-tions), and they also stimulate bone resorptionin tissue culture (120). Preliminary observationsindicate that lipoteichoic acid is not pyrogenicfor rabbits nor lethal for mice (A. J. Wicken and

246 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

K. W. Knox, unpublished data). It does, how-ever, give a positive localized as well as ageneralized Schwartzman reaction in rabbits,the latter being accompanied by the bilateralcortical necrosis of the kidneys characteristic ofa lipopolysaccharide-induced reaction; higherdoses of lipoteichoic acid (100 to 500,gg) than oflipopolysaccharide are required for positive re-action, and lipid-free teichoic acid is withouteffect (A. J. Wicken and K. W. Knox, unpub-lished data). Lipoteichoic acid, but not lipid-free teichoic acid, also stimulates bone resorp-tion in tissue culture, although the amountsrequired are approximately 10-fold greater thanfor lipopolysaccharide (E. Hausmann, A. J.Wicken, and K. W. Knox, unpublished data).Lipopolysaccharide has been implicated in theresorption of bone in human periodontal disease(120), and it seems likely that lipoteichoic acidfrom gram-positive organisms in plaque andgingival pockets could play a similar role. The"odontopathic" potential of at least one gram-positive human isolate, resembling Streptococ-cus salivarius, has been indicated by its abilityto stimulate alveolar bone loss in rats (151).The observations that the biological proper-

ties of lipoteichoic acids are less pronouncedthan those of lipopolysaccharides are not sur-prising. Although both contain lipid, the lipidmoiety (lipid A) of lipopolysaccharides is un-usual, and their characteristic components arehexosamine, generally glucosamine, and a 3-hydroxy fatty acid, generally 3-D-hydroxymyris-tic acid (235). The lipid A moiety is responsiblefor endotoxicity, and it seems likely that "thepresence or absence of acylated hydroxy fattyacid esters may determine the degree of en-dotoxic activity of lipopolysaccharides" (235).

CONCLUDING REMARKSTeichoic acids derive their name from the

Greek work for "wall," and the earliest prepara-tions were derived from this source and wereconsidered "pure" when they could be obtainedfree from other cell components. Similarly, theproducts we now know as membrane teichoicacids were originally isolated by the same proce-dures and with the same intention. We nowrealize that preservation of covalent associationof teichoic acids with peptidoglycan or mem-brane glycolipid and protein is important inso-far as the biological activity of these polymers isconcerned. Historically, the distinction betweenwall and membrane teichoic acid was an opera-tional one based on disrupted organisms, i.e.,the teichoic acid was either in the cell wall or itwas not. However, as has been discussed in thisreview, membrane teichoic acids may be quite

intimately associated with the cell wall matrixin the intact cell, and the former locationaldistinction is less clear-cut. Also, a subsurfacelocation for membrane teichoic acid does notnecessarily mean that it is "hidden" where insitu serological activity is concerned. Ionic in-teractions with wall components and the conse-quent effects of these on conformation anddepth of penetration of teichoic acid into thewall may be major determining factors in theactivity of the polymer as a classical, surfaceantigen. The known variability of wall composi-tion as a response to environmental changesmay, too, be paralleled in some organisms byvariations in the extent of surface activity ofmembrane teichoic acid. Whether or not a cleardistinction between wall and membrane tei-choic acids is warranted would seem to dependon the validity of certain assumptions, viz., (i)the apparently essential requirement for amembrane teichoic acid as compared with thevariable occurrence of wall teichoic acids, (ii)the membrane polymer is always a polyglycero-phosphate, whereas the only requirement for awall teichoic acid, by definition, is that itcontain a polyolphosphate, and (iii) the con-stancy of covalent association of wall teichoicacids with peptidoglycan and of membraneteichoic acids with membrane components.Compared with cell wall polysaccharides,

teichoic acids have generally found limitedapplication in the serological classification ofgram-positive bacteria. However, the more re-cent studies with Staphylococci, micrococci,and oral streptococci indicate that teichoicacids can be a valuable aid to classification,provided that the strains being examined arewell characterized and that there are methodsfor extracting serologically reactive teichoicacid and of obtaining specific antiserum. Fur-thermore, the diversity of structure found inwall teichoic acids has potential in the study ofthe effects of shape of polar antigens on anti-body specificity that has been little used todate.

Finally, comparisons have been made in thisreview between the physical and biologicalproperties of lipopolysaccharides of gram-nega-tive bacteria and lipoteichoic acids of gram-positive bacteria. Possibly, the teichoic acidportion of lipoteichoic acid can be regarded asanalogous in many of its properties to theheptose-phosphate-containing common core ofthe outer membrane lipopolysaccharide of thegram-negative organism, whereas the cell wallpolysaccharides, and to a lesser extent wallteichoic acids, frequently found in gram-posi-tive organisms provide structures of greater

247VOL. 37, 1973

KNOX AND WICKEN

serological specificity that are comparable tothe 0-specific side chains of lipopolysaccha-rides.

ACKNOWLEDGMENTSWe thank M. Heidelberger, H. J. Rogers, M. R. J.

Salton, M. E. Sharpe, and G. D. Shockman for theirhelpful suggestions and criticisms, S. Lowe and J.Markham for their assistance, and the NationalHealth and Medical Research Council for their finan-cial support of this work.

LITERATURE CITED

1. Aasen, J., and P. Oeding. 1971. Antigenic stud-ies on Staphylococcus epidermidis. Acta Path-ol. Microbiol. Scand. Sect. B. 79:827-834.

2. Abdulla, E. M., and J. H. Schwab. 1965. Im-munological properties of bacterial wall muco-peptides. Proc. Soc. Exp. Biol. Med.118:359-362.

3. Adam, A., R. Ciorbaru, J.-F. Petit, and E.Lederer. 1972. Isolation and properties of a

macromolecular water-soluble immuno-adju-vant fraction from the cell wall ofMycobacter-ium smegmatis. Proc. Nat. Acad. Sci. U.S.A.69:851-854.

4. Adams, J. B., A. R. Archibald, J. Baddiley, H.E. Coapes, and A. L. Davison. 1969. Teichoicacids possessing phosphate-sugar linkages instrains of Lactobacillus plantarum. Biochem.J. 113:191-193.

5. Agren, G., and C.-H. de Verdier. 1958. Theisolation of protein-bound phosphoryl mu-

ramic acid from Lactobacillus casei. ActaChem. Scand. 12:1927-1936.

6. Ambron, R. T., and R. A. Pieringer. 1971. Themetabolism of glyceride glycolipids. V. Iden-tification of the membrane lipid formed fromdiglucosyl diglyceride in Streptococcusfaecalis ATCC 9790 as an acylated derivativeof glyceryl phosphoryl diglucosyl glycerol. J.Biol. Chem. 246:4216-4225.

7. Anderson, J. C., A. R. Archibald, J. Baddiley,M. J. Curtis, and N. B. Davey. 1969. Theaction of dilute aqueous NN-dimethyl-hydrazine on bacterial cell walls. Biochem. J.113:183-189.

8. Anderson, R. G., H. Hussey, and J. Baddiley.1972. Mechanisms of wall synthesis in bacte-ria. The organisation of enzymes and iso-prenoid phosphates in the membrane. Bio-chem. J. 127:11-25.

9. Anzai, H., E. Neter, and E. A. Gorzynski. 1960.Demonstration of Rantz antigen in membersof the genus Bacillus. J. Bacteriol. 80:142-143.

10. Archibald, A. R. 1972. Teichoic acids, p.162-172. In R. L. Whistler (ed.), CarbohydrateChemistry, vol. 6. Academic Press, Londonand New York.

11. Archibald, A. R., J. J. Armstrong, J. Baddiley,and J. B. Hay. 1961. Teichoic acids and thestructure of bacterial walls. Nature (London).191:570-572.

12. Archibald, A. R., and J. Baddiley. 1965. Theextraction of teichoic acid from bacterial wallswith phenylhydrazine and its significance forthe nature of their linkages with glycosamino-peptides. Biochem. J. 95:19c-20c.

13. Archibald, A. R., and J. Baddiley. 1966. Theteichoic acids. Advan. Carbohyd. Chem.21:323-375.

14. Archibald, A. R., J. Baddiley, and N. L. Blum-son. 1968. The teichoic acids. Advan. En-zymol. 30:223-253.

15. Archibald, A. R., J. Baddiley, and J. G. Bucha-nan. 1961. The ribitol teichoic acid fromLactobacillus arabinosus walls: isolation andstructure of ribitol glycosides. Biochem. J.81:124-134.

16. Archibald, A. R., J. Baddiley, and D. Button.1965. A new kind of teichoic acid from the cellwalls of a strain of Staphylococcus lactis.Biochem. J. 95:8c-11c.

17. Archibald, A. R., J. Baddiley, and D. Button.1968. The glycerol teichoic acid of walls ofStaphylococcus lactis I3. Biochem J.110:543-558.

18. Archibald, A. R., J. Baddiley, D. Button, S.Heptinstall, and G. H. Stafford. 1968. Occur-rence of polymers containing N-acetylglucosa-mine-i-phosphate in bacterial walls. Nature(London). 219:855-856.

19. Archibald, A. R., J. Baddiley, and J. Goundry.1970. The action of dilute alkali on the glycylresidues of staphylococcal peptidoglycan. Bio-chem. J. 116:313-315.

20. Archibald, A. R., J. Baddiley, E. Heckels, and S.Heptinstall. 1971. Further studies on theglycerol teichoic acid of walls of Staphylococ-cus lactis 13. Biochem. J. 125:353-359.

21. Archibald, A. R., J. Baddiley, and S. Heptin-stall. 1969. Distribution of the glycosyl sub-stituents along the chain of the teichoic acid inwalls of Lactobacillus buchneri NCIB 8007.Biochem. J. 111:245-246.

22. Archibald, A. R., and H. E. Coapes. 1971. Theinteraction of concanavalin A with teichoicacids and bacterial walls. Biochem. J.123:665-667.

23. Archibald, A. R., and H. E. Coapes. 1971. Thewall teichoic acids of Lactobacillus plantarumC106. Location of the phosphodiester groupsand separation of the chains. Biochem. J.124:449-460.

24. Archibald, A. R., and H. E. Coapes. 1971.Influence of growth conditions on the presenceof bacteriophage receptor sites in walls ofBacillus subtilis W23. Biochem. J.125:667-669.

25. Archibald, A. R., H. E. Coapes, and G. H.Stafford. 1969. The action of dilute alkali onbacterial cell walls. Biochem. J. 113:899-900.

26. Archibald, A. R., and S. Heptinstall. 1971. Theteichoic acids of Micrococcus sp24. Biochem.J. 125:361-363.

27. Armstrong, J. J., J. Baddiley, J. G. Buchanan,B. Carss, and G. R. Greenberg. 1958. Isolation

248 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

and structure of ribitol phosphate derivatives(teichoic acids) from bacterial cell walls. J.Chem. Soc. p. 4344-4354.

28. Baddiley, J. 1968. Teichoic acids and the molec-ular structure of bacterial walls. Proc. Roy.Soc. Ser. B. 170:331-348.

29. Baddiley, J. 1970. Structure, biosynthesis, andfunction of teichoic acids. Account. Chem.Res. 3:98-105.

30. Baddiley, J., J. H. Brock, A. L. Davison, and M.D. Partridge. 1968. The wall composition ofmicrococci. J. Gen. Microbiol. 54:393-396.

31. Baddiley, J., and A. L. Davison. 1961. Theoccurrence and location of teichoic acids inlactobacilli. J. Gen. Microbiol. 24:295-299.

32. Baird-Parker, A. C. 1965. Staphylococci andtheir classification. Ann. N. Y. Acad. Sci. 128:(Art. 1) 4-19.

33. Baird-Parker, A. C. 1970. The relationship of cellwall composition to the current classificationof staphylococci and micrococci. Int. J. Syst.Bacteriol. 20:483-490.

34. Bleiweis, A. S., R. A. Craig, D. D. Zinner, and J.M. Jablon. 1971. Chemical composition ofpurified cell walls of cariogenic streptococci.Infect. Immunity 3:189-191.

35. Bodman, H., and N. E. Welker. 1969. Isolationof spheroplast membranes and stability ofspheroplasts of Bacillus stearothermophilus.J. Bacteriol. 97:924-935.

36. Bokkenheuser, V., E. A. Gorzynski, E. Cohen,and E. Neter. 1962. Immune agglutinationand lysis of antigen modified erythrocytesfrom various animal species. Proc. Soc. Exp.Biol. Med. 110:94-98.

37. Boylen, C. W., and J. C. Ensign. 1968. Ratio ofteichoic acid and peptidoglycan in cell walls ofBacillus subtilis following spore germinationand during vegetative growth. J. Bacteriol.96:421-427.

38. Bratthall, D. 1969. Immunodiffusion studies on

the serological specificity of streptococci re-

sembling Streptococcus mutans. Odont. Rev.20:231-243.

39. Bretscher, M. S. 1972. Asymmetrical lipid bi-layer structure for biological membranes. Na-ture N. Biol. 236:11-12.

40. Brock, J. H., and B. Reiter. 1971. Sensitizationof sheep erythrocytes by cell wall tecihoic acidof Staphylococcus aureus. Immunochemistry8:933-938.

41. Brody, N. I., and G. W. Siskind. 1972. Studies onantigenic competition. II. Evidence for effectat level of antigen "processing." Immunology22:75-85.

42. Brundish, D. E., and J. Baddiley. 1968. Pneumo-coccal C substance, a ribitol teichoic acidcontaining choline phosphate. Biochem. J.110:573-582.

43. Burger, M. M. 1966. Teichoic acids: antigenicdeterminants, chain separation, and their lo-cation in the cell wall. Proc. Nat. Acad. Sci.U.S.A. 56:910-917.

44. Burger, M. M., and L. Glaser. 1964. The synthe-sis of teichoic acids. I. Polyglycerophosphate.J. Biol. Chem. 239:3168-3177.

45. Burger, M. M., and L. Glaser. 1966. The synthe-sis of teichoic acids. V. Polyglucosylglycerolphosphate and polygalactosyl glycerol phos-phate. J. Biol. Chem. 241:494-506.

46. Button, D., A. R. Archibald, and J. Baddiley.1966. The linkage between teichoic acid andglycosaminopeptide in the walls of a strain ofStaphylococcus lactis. Biochem. J.99:11c-14c.

47. Cameron, C. M. 1969. Antiphagocytic activity ofStaphylococcus aureus antigens. Onder-stepoort J. Vet. Res. 36:199-206.

48. Chatterjee, A. N. 1969. Use of bacteriophage-resistant mutants to study the nature of thebacteriophage receptor site of Staphylococcusaureus. J. Bacteriol. 98:519-527.

49. Chin, T., M. M. Burger, and L. Glaser. 1966.Synthesis of teichoic acids. VI. The formationof multiple wall polymers in Bacillus subtilisW-23. Arch. Biochem. Biophys. 116:358-367.

50. Chorpenning, F. W., and M. C. Dodd. 1965.Polyagglutinable erythrocytes associated withbacteriogenic transfusion reactions. Vox Sang.10:460-471.

51. Chorpenning, F. W., and M. C. Dodd. 1966.Heterogenetic antigens of gram-positive bac-teria. J. Bacteriol. 91:1440-1445.

52. Chorpenning, F. W., and M. C. Dodd. 1967."Normal" antibodies which react with hetero-genetic bacillary antigen. Nature (London)213:305.

53. Ciznar, I., and J. W. Shands, Jr. 1971. Effect ofalkali treated lipopolysaccharide on erythro-cyte membrane stability. Infect. Immunity.4:362-367.

54. Cohen, J. 0. 1969. Analytical serology of theMicrococcaceae, p. 379-424. In J. B. G. Kwa-pinski (ed.), Analytical serology of microor-ganisms, vol. 2. Interscience Publishers, Inc.,New York.

55. Cohen, M., and C. Panos. 1966. Membrane lipidcomposition of Streptococcus pyogenes andderived L-forms. Biochemistry 5:2385-2392.

56. Cohn, Z. A. 1963. The fate of bacteria withinphagocytic cells. 1. The degradation of isotopi-cally labelled bacteria by polymorphonuclearleukocytes and macrophages. J. Exp. Med.117:27-42.

57. Cole, R. M., T. J. Popkin, R. J. Boylan, and N.H. Mendelson. 1970. Ultrastructure of a tem-perature-sensitive rod mutant of Bacillus sub-tilis. J. Bacteriol. 103:793-810.

58. Coyette, J., and J.-M. Ghuysen. 1970. Structureof the walls of Lactobacillus acidophilus strain63AM Gaser. Biochemistry 9:2935-2943.

59. Critchley, P., A. R. Archibald, and J. Baddiley.1962. The intracellular teichoic acid fromLactobacillus arabinosus 17-5. Biochem. J.85:420-431.

60. Crowder, J. C., and A. White. 1972. Teichoic

249VOL. 37, 1973

KNOX AND WICKEN

acid antibodies in staphylococcal and nonsta-phylococcal endocarditis. Ann. Int. Med.77:87-90.

61. Cummins, G. S., and H. Harris. 1956. Thechemical composition of the cell wall in somegram-positive bacteria and its possible valueas a taxonomic character. J. Gen. Microbiol.14:583-600.

62. Daugharty, H., R. R. Martin, and A. White.1967. Antibodies against staphylococcal tei-choic acids and type-specific antigens in man.J. Immunol. 98:1123-1129.

63. Daugharty, H. D., R. R. Martin, and A. White.1969. Reaction of sera and nasal secretionswith staphylococcal antigens. J. Lab. Clin.Med. 73:1011-1018.

64. Daugharty, H., M. Shriver, and A. White. 1967.Serological activity of staphylococcal polysac-charide. J. Bacteriol. 93:131-134.

65. Davison, A. L., and J. Baddiley. 1964. Glycerolteichoic acids in walls of Staphylococcus epi-dermidis. Nature (London) 202:875.

66. Davison, A. L., J. Baddiley, T. Hofstad, N.Losnegard, and P. Oeding. 1964. Serologicalstudies on teichoic acids from the walls ofstaphylococci. Nature (London) 202:872-874.

67. Decker, G. P., F. W. Chorpenning, and G. T.Frederick. 1972. Naturally-occurring anti-bodies to bacillary teichoic acids. J. Immunol.108:214-222.

68. Deddish, P. A., and H. D. Slade. 1969. Detectionof polysaccharide, teichoic acid, and proteinantigens in bacterial colonies on an agarsurface. J. Bacteriol. 97:1352-1356.

69. Deibel, R. H. 1964. The group D streptococci.Bacteriol. Rev. 3:330-366.

70. Distler, J., and S. Roseman. 1964. Polysaccha-ride and glycolipid synthesis by cell-free prep-arations from Type XIV Pneumococcus. Proc.Nat. Acad. Sci. U.S.A. 51:897-905.

71. Dos Santos Mota, J.. M., J. A. F. Op den Kamp,H. M. Verheij, and L. L. M. van Deenen. 1970.Phospholipids of Streptococcus faecalis. J.Bacteriol. 104:611-619.

72. Douglas, L. J., and J. Baddiley. 1968. A lipidintermediate in the biosynthesis of a teichoicacid. F.E.B.S. Lett. 1:114-116.

73. Douglas, L. J., and M. J. Wolin. 1971. Cell wallpolymers and phage lysis of Lactobacillusplantarum. Biochemistry 10:1551-1555.

74. Doyle, R. J., and D. C. Birdsell. 1972. Interac-tion of concanavalin A with the cell wall ofBacillus subtilis. J. Bacteriol. 109:652-658.

75. Elliott, S. D. 1959. Group and type-specificpolysaccharides of group D streptococci. Na-ture (London) 184:1342.

76. Elliott, S. D. 1960. Type and group polysaccha-rides of group D streptococci. J. Exp. Med.111:621-630.

77. Elliott, S. D. 1962. Teichoic acid and the groupantigen of group D streptococci. Nature (Lon-don) 193:1105-1106.

78. Elliott, S. D. 1963. Teichoic acid and the groupantigen of lactic streptococci (group N). Na-ture (London) 200:1184-1185.

79. Ellwood, D. C. 1970. The wall content andcomposition of Bacillus subtilis var. nigergrown in a chemostat. Biochem. J.118:367-373.

80. Ellwood, D. C. 1971. The anionic polymers in thecell wall of Bacillus subtilis var. niger grown inphosphorus-limiting environments supple-mented with increasing concentrations of so-dium chloride. Biochem. J. 121:349-351.

81. Ellwood, D. C., M. V. Kelemen, and J. Baddiley.1963. The glycerol teichoic acid from the wallsof Staphylococcus albus NTCC 7944. Bio-chem. J. 86:213-225.

82. Ellwood, D. C., and D. W. Tempest. 1969.Control of teichoic acid and teichuronic acidbiosynthesis in chemostat cultures of Bacillussubtilis var. niger. Biochem. J. 111:1-5.

83. Ellwood, D. C., and D. W. Tempest. 1972.Effects of environment on bacterial wall con-tent and composition. Adv. Microbiol. Phys-iol. 7:83-117.

84. Faux, J. A., V. Holford-Strevens, I. D. Wells,and J. Pepys. 1970. "False positive" precipita-tion reactions to extracts of organic dusts dueto a teichoic acid from S. aureus. Clin. Exp.Immunol. 7:897-902.

85. Few, A. V. 1955. The interaction of polymyxin Ewith bacterial and other lipids. Biochim. Bio-phys. Acta 16:137-145.

86. Fischer, V. W., and W. Seyforth. 1968. 1- [O-a-D-Glucopyranosyl-(1-2)-O-a-D-glucopyranosylJ-glycerin aus den Glykolipiden von Streptococ-cus faecalis und Streptococcus lactis. Hoppe-Seyler Z. Physiol Chem. 349:1662-1672.

87. Fitz-James, P. C. 1968. The collection of meso-some vesicles extruded during protoplasting,p. 124-143. In C. B. Guze (ed.), Microbialprotoplasts, spheroplasts and L-forms. TheWilliams & Wilkins Co., New York.

88. Forrester, I. T., and A. J. Wicken. 1966. Thechemical composition of the cell walls of somethermophilic bacilli. J. Gen. Microbiol.42:147-154.

89. Forrester, I. T., and A. J. Wicken. 1966. A novelteichoic acid-like polymer from the cell wallsof Bacillus coagulans NRS T2007. Biochem.Biophys. Res. Commun. 25:23-27.

90. Fraser, K. B. 1954. Antibody in glandular feversera to an antigen common to streptococci andstaphylococci. J. Pathol. Bacteriol.67:301-309.

91. Freund, J., and M. V. Bonanto. 1944. The effectof paraffin oil, lanolin-like substances andkilled tubercle bacilli on immunization withdiphtheria toxoid and Bact. typhosum. J.Immunol. 48:325-334.

92. Fuller, A. T. 1936. The formamide method forthe extraction of polysaccharides from hemo-lytic streptococci. Brit. J. Exp. Pathol.19:130-139.

93. Garner, M. F. 1970. The biological false positivereaction to serological tests for syphilis. J.Clin. Pathol. 23:31-34.

94. Ghuysen, J.-M. 1968. Use of bacteriolytic en-zymes in determination of wall structure and

BACTERIOL. REV.250

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

their role in cell metabolism. Bacteriol. Rev.32:425-464.

95. Ghuysen, J.-M., D. J. Tipper, and J. L. Strom-inger. 1965. Structure of the cell wall ofStaphylococcus aureus strain Copenhagen.IV. The teichoic acid-glycopeptide complex.Biochemistry 4:474-485.

96. Glaser, L., and M. M. Burger. 1964. The synthe-sis of teichoic acids. III. Glucosylation ofpolyglycerophosphate. J. Biol. Chem.239:3187-3191.

97. Glaser, L., H. Ionesco, and P. Schaeffer. 1966.Teichoic acids as components of a specificphage receptor in Bacillus subtilis. Biochim.Biophys. Acta 124:415-417.

98. Glastonbury, J., and K. W. Knox. 1963. Theserological classification of varieties of Lac-tobacillus casei. J. Gen. Microbiol. 31:73-77.

99. Glick, A. L., R. A. Getnick, and R. M. Cole.1971. Electron microscopy of group A strep-tococci after phagocytosis by human mono-cytes. Infec. Immunity 4:772-779.

100. Goldstein, I. J., C. E. Hollerman, and E. E.Smith. 1965. Protein-carbohydrate interac-tion. II. Inhibition studies on the interaction ofconcanavalin A with polysaccharides. Bio-chemistry 4:876-883.

101. Gooder, H. 1970. Cell wall composition in theclassification of streptococci. Int. J. Syst.Bacteriol. 20:475-482.

102. Goodman, J. W., and E. A. Kabat. 1960. Theeffect of variation in molecular weight on thecross-reactivity of dextran with type II anti-pneumococcal serum. J. Immunol. 85:342-346.

103. Goren, M. B. 1972. Mycobacterial lipids: Se-lected topics. Bacteriol. Rev. 36:33-64.

104. Gorzynski, E. A., E. Neter, and E. Cohen. 1960.Effect of lysozyme on the release of erythro-cyte-modifying antigen from staphylococciand Microsoccus lysodeikticus. J. Bacteriol.80:207-211.

105. Grant, W. D., and A. J. Wicken. 1968. Muramicacid phosphate and the linkage of teichoicacid to peptidoglycan in Bacillus stearother-mophilus cell walls. Biochem. Biophys. Res.Commun. 32:122-128.

106. Grant, W. D., and A. J. Wicken. 1970. Autolysisof cell walls of Bacillus stearothermophilusB65 and the chemical structure of the peptido-glycan. Biochem. J. 118:859-868.

107. Grov, A. 1965. Studies on antigen preparationsfrom Staphylococcus aureus. 2. The influenceof modification of functional groups and en-zymic digestion on the serological activity ofpolysaccharide A. Acta Pathol. Microbiol.Scand. 65:607-612.

108. Grov, A. 1969. Studies on antigen preparationsfrom Staphylococcus aureus. 7. The compo-nent of polysaccharide A sensitizing tannederythrocytes. Acta Pathol. Microbiol. Scand.76:621-628.

109. Grov, A., and S. M. Helgeland. 1971. Immuno-chemical characterization of Staphylococcusepidermidis and micrococcus cell walls. ActaPathol. Microbiol. Scand. 79:812-818.

110. Grov, A., and S. Rude. 1967. Immunochemicalcharacterization of Staphylococcus aureus cellwalls. Acta Pathol. Microbiol. Scand.71:409-416.

111. Grov, A., and S. Rude. 1967. Immunochemicalexamination of phenylhydrazine-treated Staph-ylococcus aureus cell walls. Acta Pathol.Microbiol. Scand. 71:417-421.

112. Guarnieri, M., B. Stechmiller, and A. L. Lehn-inger. 1971. Use of an antibody to study thelocation of cardiolipin in mitochondrial mem-branes. J. Biol. Chem. 246:7526-7532.

113. Hammarstr6m, S., and E. A. Kabat. 1971.Studies on specificity and binding propertiesof the blood group A reactive hemagglutininfrom Helix pomatia. Biochemistry10:1684-1693.

114. Hammarstr6m, S., A. A. Lindberg, and E. S.Robertson. 1972. Precipitation of lipopolysac-charides from rough mutants of Salmonellatyphimurium by an A-hemagglutinin fromHelix pomatia. Eur. J. Biochem. 25:274-282.

115. Hammerling, U., and 0. Westphal. 1967. Syn-thesis and use of 0-stearoyl polysaccharides inpassive hemagglutination and hemolysis. Eur.J. Biochem. 1:46-50.

116. Hancock, I. C., and J. Baddiley. 1972. Biosyn-thesis of the wall teichoic acid in Bacilluslicheniformis. Biochem. J. 127:27-37.

117. Harris, T. N., and S. Harris. 1953. Agglutinationby human sera of erythrocytes incubated withstreptococcal culture concentrates. J. Bacteri-ol. 66:159-165.

118. Haukenes, G. 1962. Immunochemical studies onpolysaccharide A of Staphylococcus aureus. 6.Antigenic properties. Acta Pathol. Microbiol.Scand. 55:450-462.

119. Haukenes, G., D. C. Ellwood, J. Baddiley, andP. Oeding. 1961. Serological cross-reactivitybetween polysaccharide A and teichoic acid ofStaphylococcus aureus. Biochim. Biophys.Acta 53:425-426.

120. Hausmann, E., N. Weinfeld, and W. A. Miller.1972. Effects of lipopolysaccharide on boneresorption in tissue culture. Calcified TissueRes. 9:272-282.

121. Hay, J. B., A. R. Archibald, and J. Baddiley.1965. The molecular structure of bacterialwalls. The size of ribitol teichoic acids and thenature of their linkage to glycosaminopep-tides. Biochem. J. 97:723-730.

122. Hay, J. B., N. B. Davey, A. R. Archibald, and J.Baddiley. 1965. The chain length of ribitolteichoic acids and the nature of their associa-tion with bacterial cell walls. Biochem. J.94:7c-9c.

123. Hay, J. B., A. J. Wicken, and J. Baddiley. 1963.The location of intracellular teichoic acids.Biochim. Biophys. Acta 71:188-190.

124. Heidelberger, M. 1956. Chemical constitutionand immunological specificity. Annu. Rev.Microbiol. 25:641-658.

125. Heidelberger, M., and S. Elliott. 1966. Cross-reactions of streptococcal group N teichoicacid in antipneumococcal horse sera of types

251VOL. 37, 1973

KNOX AND WICKEN

VI, XIV, XVI, and XXVII. J. Bacteriol.92:281-283.

126. Heptinstall, S., A. R. Archibald and J. Baddiley.1970. Teichoic acids and membrane functionin bacteria. Nature (London) 225:519-521.

127. Herbert, W. J. 1967. Passive haemagglutination,chap. 21. In D. M. Weir (ed.), Handbook ofexperimental immunology. Blackwell Scien-tific Publications, Oxford.

128. Hewett, M. J., K. W. Knox, and A. J. Wicken.1970. Studies on the group F antigen oflactobacilli: detection of antibodies by ha-emagglutination. J. Gen. Microbial.60:315-322.

129. Hijmans, W. 1962. Absence of the group-specificand the cell wall polysaccharide antigen inL-phase variants of group D streptococci. J.Gen. Microbiol. 28:177-179.

130. Hill, M. J. 1967. Action of bile salts on bacterialcell walls. Nature (London) 214:1152-1154.

131. Hobson, P. N., and M. J. MacPherson. 1954.Some serological and chemical studies onmaterials extracted from an amylolytic strep-tococcus from the rumen of the sheep. Bio-chem. J. 57:145-151.

132. Hofstad, T. 1965. Studies on the antigenic struc-ture of the 80/81 complex of Staphylococcusaureus. 4. Antigenic properties and chemicalbasis of serological reactivity of a purifiedpolysaccharide. Acta Pathol. Microbiol.Scand. 63:422-434.

133. Howard, J. G., H. Zola, G. H. Christie, and B.M. Courtenay. 1971. Studies on immunologi-cal paralysis. V. The influence of molecularweight on the immunogenicity, tolerogenicityand antibody-neutralizing activity of type IIpneumococcal polysaccharide. Immunology21:535-546.

134. Hughes, A. H., M. Stow, I. C. Hancock, and J.Baddiley. 1971. Functions of teichoic acidsand effect of novobiocin on control of Mg2+ atthe bacterial membrane. Nature N. Biol.229:53-55.

135. Hughes, R. C. 1965. The isolation of structuralcomponents present in the cell wall of Bacilluslicheniformis NCTC 6346. Biochem. J.96:700-709.

136. Hughes, R. C., J. G. Pavlik, H. J. Rogers, and P.J. Tanner. 1968. Organization of polymers inthe cell walls of some bacilli. Nature (London)219:642-644.

137. Hughes, R. C., and P. J. Tanner. 1968. Theaction of dilute alkali on some bacterial cellwalls. Biochem. Biophys. Res. Commun.33:22-28.

138. Highes, R. C., and P. E. Thurman. 1970. Somestructural features of the teichuronic acid ofBacillus licheniformis NCTC 6346 cell wall.Biochem. J. 117:441-449.

139. Hughes, R. C., P. F. Thurman, and M. R.Salaman. 1971. Antigenic properties of Bacil-lus licheniformis cell wall components. Eur. J.Biochem. 19:1-8.

140. Hurst, A., and J. M. Stubbs. 1969. Electronmicroscopic study of membranes and walls of

bacteria and changes occurring during growthinitiation. J. Bacteriol. 97:1466-1479.

141. Hussey, H., and J. Baddiley. 1972. Lipid inter-mediates in the biosynthesis of the wall tei-choic acid in Staphylococcus lactis I3. Bio-chem. J. 127:39-50.

142. Jablon, J. M., and D. D. Zinner. 1966. Differen-tiation of cariogenic streptococci by fluores-cent antibody. J. Bacteriol. 92:1590-1596.

143. Jackson, R. W., .and M. Moskowitz. 1966. Na-ture of a red cell sensitizing substance fromstreptococci. J. Bacteriol. 91:2205-2209.

144. James, A. M., and J. E. Brewer. 1968. Non-protein components of the cell surface ofStaphylococcus aureus. Biochem. J.107:817-821.

145. Janczura, E., H. R. Perkins, and H. J. Rogers.1961. Teichuronic acid: a mucopolysaccharidepresent in wall preparations from vegetativecells of Bacillus subtilis. Biochem. J.80:82-93.

146. Jones, J. M., and J. H. Schwab. 1970. Effects ofstreptococcal cell wall fragments on phagocy-tosis and tissue culture cells. Infect. Immunity1:232-242.

147. Juergens, W. G., A. R. Sanderson, and J. L.Strominger. 1963. Chemical basis for an im-munological specificity of a strain of Staph-ylococcus aureus. J. Exp. Med. 117:925-935.

148. Kabat, E. A. 1966. The nature of an antigenicdeterminant. J. Immunol. 97:1-11.

149. Kabat, E. A., and A. E. Bezer. 1958. The effectof variation in molecular weight on the antige-nicity of dextran in man. Arch. Biochem.Biophys. 78:306-318.

150. Kabat, E. A., and M. Mayer. 1961. Experimen-tal immunochemistry. Charles C Thomas,Publisher, Springfield, Ill.

151. Kalstrup, J., and R. J. Gibbons. 1970. Inductionof dental caries and alveolar bone loss by ahuman isolate resembling Streptococcus sali-varius. Caries Res. 4:366-377.

152. Karakawa, W. W., and J. A. Kane. 1971. Im-munochemical analysis of a galactosamine-rich teichoic acid of Staphylococcus aureus,phage type 187. J. Immunol. 106:900-906.

153. Karakawa, W. W., J. E. Wagner, and J. H.Pazur. 1971. Immunochemistry of the cell wallcarbohydrate of group L hemolytic strepto-cocci. J. Immunol. 107:554-562.

154. Kedzia, B. 1967. The influence of surface anti-gens of Staphylococcus aureus on intensity ofrespiration of liver tissue in mice. Arch. Im-munol. Ther. Exp. 15:42-49.

155. Kelemen, M. V., and J. Baddiley. 1961. Struc-ture of the intracellular glycerol teichoic acidfrom Lactobacillus casei ATCC 7469. Bio-chem. J. 80:246-254.

156. Kirby, W. M. M. 1951. Hemagglutination reac-tion in streptococcal infections and acuterheumatic fever. Proc. Soc. Exp. Biol. Med.78:519-522.

157. Knox, K. W. 1963. Isolation of group specificproducts from Lactobacillus casei and L. caseivar. rhamnosus. J. Gen. Microbiol. 31:59-72.

252 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

158. Knox, K. W. 1970. Antigens of oral bacteria.Adv. Oral Biol. 4:91-130.

159. Knox, K. W., and E. A. Hall. 1965. The isolationof oligosaccharides from the cell wall polysac-charide of Lactobaccillus casei serologicalgroup C. Biochem. J. 94:525-533.

160. Knox, K. W., M. J. Hewett, and A. J. Wicken.1970. Studies on the group F antigen oflactobacilli: antigenicity and serological speci-ficity of teichoic acid preparations. J. Gen.Microbiol. 60:303-313.

161. Knox, K. W., and K. J. Holmwood. 1968. Struc-ture of the cell wall of lactobacilli. Role ofmuramic acid phosphate in Lactobacillus fer-menti. Biochem. J. 108:363-368.

162. Knox, K. W., M. Vesk, and E. Work. 1966.Relation between excreted lipopolysaccharidecomplexes and surface structures of a lysine-limited culture of Escherichia coli. J. Bacte-riol. 92:1206-1217.

163. Knox, K. W., and A. J. Wicken. 1970. Serologi-cal properties of the wall and membraneteichoic acids from Lactobacillus helveticusNCIB 8025. J. Gen. Microbiol. 63:237-248.

164. Knox, K. W., and A. J. Wicken. 1972. Serologi-cal studies on the teichoic acids of Lactobacil-lus plantarum. Infect. Immunity. 6:43-49.

165. Knox, K. W., and A. J. Wicken. 1972. Reactionof dextrans with antisera to teichoic acids.Arch. Oral Biol. 17:1491-1494.

166. Knox, K. W., and A. J. Wicken. 1973. Non-specific inhibition of the precipitin reactionbetween teichoic acids and antisera. Immuno-chemistry 10:93-98.

167. Kowalski, J. J., and D. T. Berman. 1971. Im-munobiological activity of cell wall antigens ofStaphylococcus aureus. Infect. Immunity.4:205-211.

168. Kowalski, J. J., D. J. Tipper, and D. T. Berman.1970. Preparation of cell wall antigens ofStaphylococcus aureus. Infect. Immunity.2:54-59.

169. Lam, G. T., F. J. Sweeney Jr., C. H. Wilmer,and R. I. Wise. 1963. Abscess-forming factor(s) produced by Staphylococcus aureus. J.Bacteriol. 86:611-615.

170. Lancefield, R. 1933. A serological differentiationof human and other groups of hemolytic strep-tococci. J. Exp. Med. 57:571-595.

171. Lehner, T., J. M. A. Wilson, and R. G. Ward.1970. Serum antibodies in dental caries inman. Arch. Oral Biol. 15:481-490.

172. Lloyd, K. O., E. A. Kabat, and R. E. Rosenfield.1966. Immunochemical studies on bloodgroups. XXXV. The activity of fucose-con-taining oligosaccharides isolated from bloodgroup A, B and H substances by alkalinedegradation. Biochemistry 5:1502-1507.

173. Losnegard, N., and P. Oeding. 1963. Immuno-chemical studies on polysaccharides fromStaphylococcus epidermidis. 1. Isolation andchemical characterization. Acta Pathol. Mi-crobiol. Scand. 58:482-492.

174. Losnegard, N., and P. Oeding. 1963. Immuno-chemical studies on polysaccharides from

Staphylococcus epidermidis. 2. Antigenicproperties. Acta Pathol. Microbiol. Scand.58:493-500.

175. Luideritz, O., K. Jann, and R. Wheat. 1968.Somatic and capsular antigens of gram-nega-tive bacteria, p. 105-228. In M. Florkin and E.H. Stotz (ed.), Comprehensive biochemistry,vol. 26A. Elsevier Publishing Co., Amsterdam.

176. Luideritz, O., A.-M. Staub, and 0. Westphal.1966. Immunochemistry of 0 and R antigensof Salmonella and related Enterobacteriaceae.Bacteriol. Rev. 30:192-255.

177. McCarty, M. 1959. The occurrence of poly-glycerophosphate as an antigenic componentof various gram-positive bacterial species. J.Exp. Med. 109:361-378.

178. McCarty, M. 1964. The role of D-alanine in theserological specificity of group A streptococcalglycerol teichoic acid. Proc. Nat. Acad. Sci.U.S.A 52:259-265.

179. McCarty, M., and S. I. Morse. 1965. Cell wallantigens of gram-positive bacteria. Advan.Immunol. 4:249-286.

180. McDowell, G., A. Grov, and P. Oeding. 1971.Local immunization of guinea pig mammarygland with staphylococcal antigens. ActaPathol. Microbiol. Scand. 79:805-811.

181. McLoughlin, M., and F. S. Stewart. 1971. Redcell-sensitizing antigens of lactobacilli. J.Gen. Microbiol. 66:229-238.

182. Markam, J. L., K. W. Knox, R. G. Schamschula,and A. J. Wicken. 1973. Antibodies to teichoicacids in humans. Arch. Oral Biol. 18:313-318.

183. Marshall, N. J., J. H. Hewitt, and A. M. James,1971. Changes in the surface properties ofthree strains of Staphylococcus aureus on invitro training to methicillin resistance. Mi-crobios. 4:241-251.

184. Marshall, N. J., and A. M. James. 1971. Surfaceproperties of methicillin-resistant cells ofStaphylcoccus aureus. Microbios. 4:217-225.

185. Martin, R. H., H. Daugharty, and A. White.1966. Staphylococcal antibodies and hyper-sensitivity to teichoic acids in man. Antimi-crob. Ag. Chemother. 1965, p. 91-96.

186. Martin, R. R., H. Daugharty, and A. White.1968. Immunoglobulins of specific staphylo-coccal antibodies. Pediatrics 41:1092-1096.

187. Matsuno, T., and H. D. Slade. 1970. Composi-tion and properties of a group A streptococcalteichoic acid. J. Bacteriol. 102:747-752.

188. Matsuno, T., and H. D. Slade. 1971. Group Astreptococcal polysaccharide antigens. Infect.Immunity. 3:385-389.

189. Mauck, J., L. Chan, and L. Glaser. 1971. Turn-over of the cell-wall of gram-positive bacteria.J. Biol. Chem. 246:1820-1827.

190. Mauck, J., and L. Glaser. 1972. On the mode ofin vivo assembly of the cell wall of Bacillussubtilis. J. Biol. Chem. 247:1180-1187.

191. Medrek, T. F., and E. M. Barnes. 1962. Theinfluence of the growth medium on the dem-onstration of a group D antigen in faecalstreptococci. J. Gen. Microbiol. 28:701-709.

192. Mills, C. K. 1969. Agglutinating antigens of

253VOL. 37, 1973

KNOX AND WICKEN

Lactobacillus jugurti ATCC 521 J. Gen. Mi-crobiol. 57:105-114.

193. Mindich, L. 1970. Membrane synthesis in Bacil-lus subtilis. 1. Isolation and properties ofstrains bearing mutations in glycerol metabo-lism. J. Mol. Biol. 49:415-432.

194. Mirelman, D., D. R. D. Shaw, and J. T. Park.1971. Nature and origin of phosphorus com-pounds in isolated cell walls of Staphylococcusaureus. J. Bacteriol. 107:239-244.

195. Morrison, D. C., and H. J. Morowitz. 1970.Studies on membrane synthesis in Bacillusmegaterium KM. J. Mol. Biol. 49:441-459.

196. Morse, S. I. 1963. Isolation and properties of agroup antigen of Staphylococcus albus. J.Exp. Med. 117:19-26.

197. Morse, S. I. 1965. Biological attributes of staph-ylococcal cell walls. Ann. N. Y. Acad. Sci.128:191-210.

198. Moskowitz, M. 1966. Separation and propertiesof a red cell sensitizing substance from strep-tococci. J. Bacteriol. 91:2200-2204.

199. Mudd, S., A. Yoshida, I. W. Li, and N. A.Lenhart. 1963. Identification of a somaticantigen of Staphylococcus aureus critical forphagocytosis by human blood leucocytes. Na-ture (London) 199:1200-1201.

200. Nathenson, S. G., N. Ishimoto, J. S. Anderson,and J. L. Strominger. 1966. Enzymic synthesisand immunochemistry of a- and fl-N-acetyl-glucosaminyl-ribitol linkages in teichoic acidsfrom several strains of Staphylococcus aureus.J. Biol. Chem. 241:651-658.

201. Naumova, I. B., A. Belozerskii, and F. A. Shafi-kova. 1962. Isolation and some properties ofteichoic acid of Actinomyces streptomyciniKrass. Dokl. Akad. Nauk. SSSR 143:730-733.

202. Naumova, I. B., and M. Sh. Zaretskaya. 1964.Some properties of ribitol teichoic acid fromActinomyces violaceus. Dokl. Akad. Nauk.SSSR 156:1464-1467.

203. Naumova, I. B., and M. Sh. Zaretskaya. 1964.Some properties of glycerine-teichoic acidsfrom Actinomyces rimosus T-118 and Ac-tinomyces antibioticus 39. Dokl. Akad. Nauk.SSSR. 157:207-210.

204. Nermut, M. V. 1967. The ultrastructure of thecell wall of Bacillus megaterium. J. Gen.Microbiol. 49:503-512.

205. Neter, E., H. Anzai, and E. A. Gorzynski. 1960.Identification of an antigen common to List-eria monocytogenes and other bacteria. Proc.Soc. Exp. Biol. Med. 105:131-134.

206. Neter, E., E. A. Gorzynski, A. M. Drislane, A. H.Harris, and E. Rajnovich. 1959. Detection ofstaphylococcal antibodies in human gammaglobulin and serum by hemagglutinationtests. Proc. Soc. Exp. Biol. Med. 101:484-487.

207. Neter, E., E. Rajnovich, and E. A. Gorzynski.1960. Study of staphylococcal antibodies ofthe Rantz type: placental transfer and titers insera of children of various ages. Pediatrics25:21-26.

208. Nowlan, S. S., and R. H. Deibel. 1967. Group Qstreptococci. 1. Ecology serology, physiology,

and relation to established enterococci. J.Bacteriol. 94:291-296.

209. Oeding, P. 1965. Antigenic properties of staph-ylcocci. Ann. N. Y. Acad. Sci. 128:183-189.

210. Oeding, P. 1967. Antigenic studies on Micrococ-cus strains. Acta Pathol. Microbiol. Scand.70: 120-128.

211. Oeding, P., and I. L. Hasselgren. 1972. Antigenicstudies of genus Micrococcus. 2. Double diffu-sion in agar gel with particular emphasis onteichoic acids. Acta Pathol. Microbiol. Scand.80:265-269.

212. Oeding, P., B. Myklestad, and A. L. Davison.1967. Serological investigations on teichoicacids from the walls of Staphylococcus epider-mis and Micrococcus. Acta Pathol. Microbiol.Scand. 69:458-464.

213. Ohanian, S. H., J. H. Schwab, and W. J.Cromartie. 1969. Relation of rheumatic-likecardiac lesions of the mouse to localisation ofgroup A streptococcal cell walls. J. Exp. Med.129:37-49.

214. Osborn, M. J., J. E. Gander, and E. Parisi. 1972.Mechanism of assembly of the outer mem-brane of Salmonella typhimurium. Site ofsynthesis of lipopolysaccharide. J. Biol.Chem. 247:3973-3986.

215. Pakula, R., and W. Walczak. 1955. An eryth-rocyte-sensitizing factor common to staph-ylococci and haemolytic streptococci. ActaMicrobiol. Pol. 4:235-243.

216. Pavlovskis, O., and H. D. Slade. 1969. Absorp-tion of 3H fatty acid esters of streptococcalgroups A and E cell wall polysaccharide anti-gens by red blood cells and their effect onhemagglutination. J. Bacteriol. 100:641-646.

217. Pazur, J. H., J. S. Anderson, and W. W. Kara-kawa. 1971. Glycans from streptocaccal cellwalls. J. Biol. Chem. 246:1793-1798.

218. Pease, P. 1967. Antigenic relationships betweenoral group D streptococci, and some haemoph-ilus species and Mycoplasma hominis. J. Gen.Microbiol. 49:433-441.

219. Pike, R. M. 1967. Antibody heterogeneity andserological reactions. Bacteriol. Rev.31:157-174.

220. Plackett, P., B. P. Marmion, E. J. Shaw, and R.M. Lemcke. 1969. Immunochemical analysisof Mycoplasma pneumoniae. 3. Separationand chemical identification of serologicallyactive lipids. Aust. J. Exp. Biol. Med. Sci.47:171-195.

221. Plackett, P., and E. J. Shaw. 1967. Glucolipidsfrom Mycoplasma laidlawii and StreptococcusMG. Biochem. J. 104:61c-62c.

222. Pollock, M. R. 1962. Exoenzymes, p. 121-178. InI. C. Gunsalus and R. Y. Stanier (ed.), Thebacteria, vol. 4. Academic Press, Inc., London.

223. RajBhandary, U. L., and J. Baddiley. 1963. Theintracellular teichoic acid from Staphylococ-cus aureus H. Biochem. J. 87:429-435.

224. Rand, R. P., and S. Sengupta. 1972. Cardiolipinforms hexagonal structures with divalent cat-ions. Biochim. Biophys. Acta 255:484-492.

225. Rantz, L. A., E. Randall, and A. Zuckerman.

254 BACTERIOL. REV.

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

1956. Hemolysis and hemagglutination bynormal serums of erythrocytes treated with a

nonspecies specific bacterial substance. J. In-fect. Dis. 98:211-222.

226. Rantz, L. A., A. Zuckerman, and E. Randall.1952. Hemolysis of red blood cells treated bybacterial filtrates in the presence of serum andcomplement. J. Lab. Clin. Med. 39:443-448.

227. Raschke, W. C., and C. E. Ballou. 1971. Im-munochemistry of the phosphomannah ofyeast Kloeckera brevis. Biochemistry10:4130-4135.

228. Razin, S. 1972. Reconstitution of biologicalmembranes. Biochim. Biophys. Acta265:241-296.

229. Razin, S., B. Prescott, G. Caldes, W. D. James,and R. M. Chanock. 1970. Role of glycolipidsand phosphatidylglycerol in the serologicalactivity of Mycoplasma pneumoniae. Infect.Immunity. 4:408-416.

230. Razin, S., B. Prescott, and R. M. Chanock. 1970.Immunogenicity of Mycoplasma pneumoniaeglycolipids: a novel approach to the produc-tion of antisera to membrane lipids. Proc.Nat. Acad. Sci. U.S.A. 67:590-597.

231. Rebers, P. A., and M. Heidelberger. 1961. Thespecific polysaccharide of type VI Pneumococ-cus. II. The repeating unit. J. Amer. Chem.Soc. 83:3056-3059.

232. Rebers, P. A., E. Hurwitz, and M. Heidelberger.1961. Immunochemistry of pneumococcaltypes II, V, and VI. J. Bacteriol. 82:920-926.

233. Reeder, W. J., and R. D. Ekstedt. 1971. Study ofthe interaction of concanavalin A with staph-ylococcal teichoic acids. J. Immunol. 106:334-339.

234. Reusch, V. M., and F. C. Neuhaus. 1971. D-ala-nine: membrane acceptor ligase from Lac-tobacillus casei. J. Biol. Chem. 246:6136-6143.

235. Rietschel, E. Th., H. Gottert, 0. Lideritz, and0. Westphal. 1972. Nature and linkage of thefatty acids present in the lipid-A component ofSalmonella lipopolysaccharide. Eur. J. Bio-chem. 28:166-173.

236. Roberts, W. K., J. G. Buchanan, and J. Baddi-ley. 1963. The specific substance fromPneumococcus type 34 (41). The structure of aphosphorus-free repeating unit. Biochem. J.88:1-7.

237. Rogers, H. J., D. A. Reaveley, and I. D. J.Burdett. 1967. The membrane systems ofBacillus licheniformis, p. 303-313. In H. Peet-ers (ed.), Protides of the biological fluids, vol.15. Elsevier Publishing Co., Amsterdam.

238. Rogosa, M., and M. E. Sharpe. 1959. An ap-proach to the classification of lactobacilli. J.Appl. Bacteriol. 22:329-340.

238a. Rosan, B. 1973. Antigens of Streptococcus san-

guis. Infect. Immunity 7:205-211.239. Rothfield, L., and R. W. Horne. 1967. Reassocia-

tion of purified lipopolysaccharide and phos-pholipid of the bacterial cell envelope: elec-tron microscopic and monolayer studies. J.Bacteriol. 93:1705-1713.

240. Rothfield, L., and M. Pearlman-Kothencz. 1969.

Synthesis and assembly of bacterial mem-brane components. A lipopolysaccharide-phospholipid-protein complex excreted by liv-ing bacteria. J. Mol. Biol. 44:477-492.

241. Rotta, J. 1969. Biological activity of cellularcomponents of group A streptococci in vivo, p.63-101. In J. A. Brody et al. (ed.), CurrentTopics in Microbiology and immunology, vol.48. Spring-Verlag, Berlin.

242. Salton, M. R. J. 1953. Studies on the bacterialcell wall. IV. The composition of the cell wallof some gram-positive and gram-negative bac-teria. Biochim. Biophys. Acta 10:512-523.

243. Sanderson, A. R., J. L. Strominger, and S. G.Nathenson. 1962. Chemical structure of tei-choic acid from Staphylococcus aureus strainCopenhagen. J. Biol. Chem. 237:3603-3613.

244. Schechter, I., and M. Sela. 1967. Preferentialformation of antibodies specific toward D-amino acid residues upon immunization withpoly-D, L-peptidyl proteins. Biochemistry6:897-905.

245. Scherrer, R., and P. Gerhardt. 1971. Molecularsieving by the Bacillus megaterium cell walland protoplast. J. Bacteriol. 107:718-735.

246. Schleifer, K. H., and R. M. Krause. 1971. Theimmunochemistry of peptidoglycan. Separa-tion and characterization of antibodies to theglycan and to the peptide subunit. Eur. J.Biochem. 19:471-478.

247. Schuster, G. S., J. A. Hayashi, and A. N. Bahn.1967. Toxic properties of the cell wall ofgram-positive bacteria. J. Bacteriol. 93:47-52.

248. Schwab, J. H., and W. J. Cromartie. 1957.Studies on a toxic cellular component of groupA streptococci. J. Bacteriol. 74:673-679.

249. Schwab, J. H., W. J. Cromartie. S. H. Ohanian,and J. G. Craddock. 1967. Association ofexperimental chronic arthritis with the per-sistence of group A streptococcal cell walls inthe articular tissue. J. Bacteriol. 94:1728-1735.

250. Shabarova, Z. A., J. G. Buchanan, and J.Baddiley. 1962. The composition ofPneumococcus type-specific substances con-taining phosphorus. Biochim. Biophys. Acta57:146-148.

251. Sharpe, M. E. 1955. A serological classificationof lactobacilli. J. Gen. Microbiol. 12:107-122.

252. Sharpe, M. E. 1970. Cell wall and cell membraneantigens used in the classification of lac-tobacilli. Int. J. Syst. Bacteriol. 20:509-518.

253. Sharpe, M. E., J. H. Brock, K. W. Knox, and A.J. Wicken. 1973. Glycerol teichoic acid as acommon antigenic factor in lactobacilli andsome other gram-positive organisms. J. Gen.Microbiol. 74:119-126.

254. Sharpe, M. E., A. L. Davison, and J. Baddiley.1964. Teichoic acids and group antigens inlactobacilli. J. Gen. Microbiol. 34:333-340.

255. Sharpe, M. E., and P. M. F. Shattock. 1952. Theserological typing of group D streptococciassociated with outbreaks of neonatal diar-rhoea. J. Gen. Microbiol. 6:150-165.

256. 9harpe, M. E., and D. M. Wheater. 1957.Lactobacillus helveticus. J. Gen. Microbiol.

VOL. 37, 1973 255

256 KNOX AND WICKEN

16:676-679.257. Shattock, P. M. F. 1949. The streptococci of

group D: the serological grouping of Strep-tococcus bovis and observations on serologi-cally refractory group D strains. J. Gen. Mi-crobiol. 3:80-92.

258. Shattock, P. M. F., and D. G. Smith. 1963. Thelocation of the group D antigen in a strain ofStreptococcus faecalis var. liquifaciens. J.Gen. Microbiol. 31:iv.

259. Shaw, N. 1970. Bacterial glycolipids. Bacteriol.Rev. 34:365-377.

260. Shaw, D. R. D., D. Mirelman, A. N. Chatterjee,and J. T. Park. 1970. Ribitol teichoic acidsynthesis in bacteriophage-resistant mutantsof Staphylococcus aureus H. J. Biol. Chem.245:5101-5106.

261. Shaw, N., and A. Stead. 1972. Bacterial glyco-phospholipids. F.E.B.S. Lett. 21:249-253.

262. Shayegani, M., K. Hisatsune, and S. Mudd.1970. Cell wall component which affects theability of serum to promote phagocytosis andkilling of Staphylococcus aureus. Infect. Im-munity 2:750-756.

263. Shockman, G. D., and H. D. Slade. 1964. Thecellular location of the streptococcal group Dantigen. J. Gen. Microbiol. 37:297-305.

264. Siddiqui, B., and S. Hakomori. 1971. Revisedstructure for the Forssman glycolipid hapten.J. Biol. Chem. 246:5766-5769.

265. Simmons, D. A. R. 1971. Stereochemical aspectsof antigenic specificity in polysaccharide de-terminants. Eur. J. Biochem. 18:53-58.

266. Simmons, D. A. R. 1971. Immunochemistry ofShigella flexneri O-antigens: a study of struc-tural and genetic aspects of the biosynthesis ofcell-surface antigens. Bacteriol. Rev.35:117-148.

267. Singer, S. J., and G. L. Nicolson. 1972. The fluidmosaic model of the structure of cell mem-branes. Science 175:720-731.

268. Singleton, L., G. W. Ross, and J. Kohn. 1964.Staphylococcal teichoic acid antibody in thesera of patients with burns. Nature (London)204:1173-1174.

269. Slade, H. D., and U. Hammerling. 1968. Detec-tion by hemagglutination of antibodies togroup A and group E streptococci by the use ofO-stearoyl derivatives of their cell wall car-bohydrate-grouping antigens. J. Bacteriol.95:1572-1579.

270. Smith, D. G., and P. M. F. Shattock. 1964. Thecellular location of antigens in streptococci ofgroups D, N and Q. J. Gen. Microbiol.34: 165-175.

271. Stewart, F. S. 1961. Bacterial polyglycerophos-phate. Nature (London) 190:464.

272. Stewart, F. S. 1963. Identification of a newstreptococcal red cell sensitizing antigen. Im-munology 6:403-411.

273. Stewart, F. S., and M. McLoughlin. 1965. Ared-cell sensitizing antigen of group D strep-tococci. Immunology 9:319-326.

274. Stewart, F. S., T. W. Steele, and W. T. Martin.

BACTERIOL. REV.

1959. The mechanisms involved in the produc-tion of red cell panagglutinability by strep-tococcal cultures. Immunology 2:285-294.

275. Stominger, J. L., and J.-M. Ghuysen. 1963. Onthe linkage between teichoic acid and theglycopeptide in the cell wall of Staphylococcusaureus. Biochem. Biophys. Res. Commun.12:418-424.

276. Strominger, J. L., and J.-M. Ghuysen. 1967.Mechanisms of enzymatic bacteriolysis.Science 156:213-221.

277. Taylor, A., K. W. Knox, and E. Work. 1966.Chemical and biological properties of an ex-tracellular lipopolysaccharide from Esche-richia coli grown under lysine-limiting condi-tions. Biochem. J. 99:53-61.

278. Tempest, D. W., J. W. Dicks, and D. C. Ellwood.1968. Influence of growth conditions on theconcentration of potassium in Bacillus subtilisvar. niger and its possible relationship tocellular ribonucleic acid, teichoic acid andteichuronic acid. Biochem. J. 106:237-243.

279. Toon, P., P. E. Brown, and J. Baddiley. 1972.The lipid-teichoic acid complex in the cyto-plasmic membrane of Streptococcus faecalisNCIB 8191. Biochem. J. 127:399-409.

280. Torii, M., E. A. Kabat, and A. E. Bezer. 1964.Separation of teichoic acid of Staphylococcusaureus into two immunologically distinctspecific polysaccharides with a and ,B-N-acetylglucosaminyl linkages respectively. J.Exp. Med. 120:13-29.

281. van de Rijn, I. 1972. Ph.D. dissertation, Univer-sity of Florida, Gainesville.

282. van de Rijn, I., and A. S. Bleiweis. Biochemicaland immunochemical characterisation of agroup I cariogenic streptococcal antigenic de-terminant. Bacteriol. Proc., p. 59, 1972.

283. van Driel, D., A. J. Wicken, M. R. Dickson, andK. W. Knox. 1973. Cellular location of thelipoteichoic acids of Lactobacillus fermentiNCTC 6991 and Lactobacillus casei NCTC6375. J. Ultrastruct. Res. In press.

284. van Heijenoort, J., D. Menjon, B. Flouret, J.Szulmajster, J. Laporte, and G. Batelier.1971. Cell walls of a teichoic acid deficientmutant of Bacillus subtilis. Eur. J. Biochem.20:442-450.

285. Waltersdorff, R. L., and R. W. Jackson. Induc-tion of kidney disease in rabbits with strep-tococcal teichoic acid. Bacteriol. Proc., p. 117,1970.

286. Ward, J. B., and H. R. Perkins. 1968. Thechemical composition of the membranes ofprotoplasts and L-forms of Staphylococcusaureus. Biochem. J. 106:391-400.

287. Westphal, O., and K. Jann. 1965. Bacteriallipopolysaccharides. Extraction with phenol-water and further applications of the proce-dure, p. 83-91. In R. L. Whistler (ed.), Meth-ods in carbohydrate chemistry, vol. 5. Aca-demic Press, Inc., New York.

288. Westphal, O., 0. Luderitz, and F. Bister. 1952.Ober die Extraktion von Bakterien mit Phe-

IMMUNOLOGICAL PROPERTIES OF TEICHOIC ACIDS

nol/Wasser. Naturforsch. 7b:148-155.289. White, R. G. 1968. Antigens and adjuvants.

Proc. Roy. Soc. Med. 61:1-6.290. Wicken, A. J. 1966. The glycerol teichoic acid

from the cell wall of Bacillusstearothermophilus B65. Biochem. J.99:108-116.

291. Wicken, A. J., and J. Baddiley. 1963. Structureof intracellular teichoic acid from group Dstreptococci. Biochem. J. 87:54-62.

292. Wicken, A. J., S. D. Elliott, and J. Baddiley.1963. The identity of streptococcal group Dantigen with teichoic acid. J. Gen. Microbiol.31:231-239.

293. Wicken, A. J., J. W. Gibbens, and K. W. Knox.1972. Antiteichoic acid antibodies and nontre-ponemal serological tests for syphilis. Infect.Immunity 5:982-984.

294. Wicken, A. J., J. W. Gibbens, and K. W. Knox.1973. Comparative studies on the isolation ofmembrane lipoteichoic acid from Lactobacil-lus fermenti 6991. J. Bacteriol. 113:365-372.

295. Wicken, A. J., and K. W. Knox. 1970. Studies onthe group F antigen of lactobacilli: isolation ofa teichoic acid-lipid complex from Lactobacil-lus fermenti. J. Gen. Microbiol. 60:293-301.

296. Wicken, A. J., and K. W. Knox. 1971. A serologi-cal comparison of the membrane teichoicacids from lactobacilli of different serologicalgroups. J. Gen. Microbiol. 67:251-254.

297. Wiebull, C. 1956. The nature of the "ghosts"obtained by lysozyme lysis of Bacillusmegaterium. Exp. Cell Res. 10:214-221.

298. Wieghard, C. W., and L. A. Julianelle. 1935. Theimmunological specificity of staphylococci. II.The chemical nature of the soluble specificsubstances. J. Exp. Med. 62:23-30.

299. Willers, J. M. N., M. F. Michel, M. J. Sysma,and K. C. Winkler. 1964. Chemical analysisand inhibition reactions of the group and typeantigens of group F streptococci. J. Gen.Microbiol. 36:95-105.

300. Wilson, A. T., and G. G. Wiley. 1963. Thecellular antigens of group A streptococci. J.Exp. Med. 118:527-556.

301. Wilson, G. S., and A. A. Miles. 1948. Theantigenic structure of bacteria, p. 283. In G. S.

Wilson and A. A. Miles (ed.), Topley andWilson's principles of bacteriology and im-munity, vol. 1. Arnold, London.

302. Wise, E. M. Jr., R. S. Glickman, and E. Teimer.1972. Teichoic acid hydrolase activity in soilbacteria. Proc. Nat. Acad. Sci. U.S.A.69:233-237.

303. Wober, W., and P. Alaupovic. 1971. Studies onthe protein moiety of endotoxin from Gram-negative bacteria. Eur. J. Biochem.19:340-356.

304. Wolin, M. J., A. R. Archibald, and J. Baddiley.1966. Changes in wall teichoic acid resultingfrom mutations of Staphylococcus aureus.Nature (London) 209:484-486.

305. Work, E. 1971. Cell walls, p. 361-418. In J. R.Norris (ed.), Methods in microbiology, vol.5A, Academic Press, Inc., London and NewYork.

306. Wu, T. C. M., and J. T. Park. 1971. Chemicalcharacterization of a new surface antigenicpolysaccharide from a mutant of Staphylococ-cus aureus. J. Bacteriol. 108:874-884.

307. Yoshida, K., and R. D. Ekstedt. 1968. Antibodyresponse to Staphylococcus aureus in rabbits:sequence of immunoglobulin synthesis and itscorrelation with passive protection in mice. J.Bacteriol. 96:1540-1545.

308. Youatt, J., A. Montoya, and E. W. Nester. 1970.Possible inhibitory effect of teichoic acid onBacillus subtilis transfer ribonucleic acid. Bio-chem. J. 119:317-321.

309. Youmans, A. S., and G. P. Youmans. 1967.Preparation and effect of different adjuvantson the immunogenic activity of mycobacterialribosomal fraction. J. Bacteriol. 94:836-843.

310. Young, F. E. 1966. Fractionation and partialcharacterization of the products of autolysis ofcell walls of Bacillus subtilis. J. Bacteriol.92:839-846.

311. Young, F. E., and A. P. Jackson. 1966. Extentand significance of contamination of DNA byteichoic acid in Bacillus subtilis. Biochem.Biophys. Res. Commun. 23:490-495.

312. Zabriskie, J. B. 1967. Mimetic relationshipsbetween group A streptococci and mammaliantissues. Advan. Immunol. 7:147-189.

2-57VOL. 37, 1973