JOURNAL OF BACTERIOLOGY, Nov. 1974, p. 672-678Copyright 0 1974 American Society for Microbiology

Vol. 120, No. 2Printed in U.S.A.

Lipopolysaccharide Containing L-Acofriose in the FilamentousBlue-Green Alga Anabaena variabilis

J. WECKESSER, A. KATZ, G. DREWS, H. MAYER, AND I. FROMMEInstitut far Biologie II, Lehrstuhl far Mikrobiologie der Universittt, and Max-Planck-Institut far

Immunbiologie, D-78 Freiburg i.Br., GermanyReceived for publication 1 June 1974

For the first time, an 0-antigenic lipopolysaccharide (LPS) has been isolatedfrom a filamentous blue-green alga (Anabaena variabilis). It was extractablewith phenol-water, resulting in extraction of the bulk of the LPS into the phenolphase. The polysaccharide moiety of this LPS consists of L-rhamnose, its3-0-methyl ether L-acofriose, D-mannose, D-glucose, and D-galactose. L-Glycero-D-mannoheptose and 2-keto-3-deoxyoctonate, the two characteristic sugar com-ponents of enteric LPS, and phosphate groups are absent from the A. variabilisO antigen. The only amino sugar present is D-glucosamine. Three hydroxy fattyacids were identified, namely, ,-hydroxymyristic, ,-hydroxypalmitic and B-hy-droxystearic acids, in addition to palmitic and unidentified fatty acid. The LPSof A. variabilis is localized in the outermost cell wall layer and behaves like a bac-terial 0 antigen in serological tests. The passive hemagglutination yielded hightiters with isolated LPS (pretreated by heat or by alkali) and rabbit antisera pre-pared against living or heat-killed cells. The position of the precipitation arcsafter immunoelectrophoresis of the 0 antigen indicates the lack of chargedgroups. The water phase of the phenol-water extract contains, in high yield, aglucose polymer. It is serologically inactive as shown by the passive hemagglu-tination test and by agar-gel precipitation.

It has been shown previously (for review seereference 4) that cell walls of Cyanophyta (blue-green algae) are more closely related to bacterialcell walls than had been assumed previously(24). There are significant similarities in thechemical composition and morphological orga-nization of the cell walls of these differentorganisms. Frank et al. (6) demonstrated thepresence of murein in isolated cell walls ofCyanophyta, a component that formerly wasthought to be restricted to bacterial cell walls.This finding, verified by several other investiga-tors (4), explains the susceptibility of someCyanophyta towards penicillin and lysozyme.Like cell walls of gram-negative bacteria, the

cell wall of blue-green algae consists of multiplelayers. Recent observations on ultrathin sec-tions of different Oscillatoriaceae andChroococcaceae (2, 6, 14) and on freeze-etchingsof Anabaena variabilis (9) show a farreachingsimilarity between their cell walls and thecomplex cell walls of gram-negative bacteria.Finally, Weise et al. (40) were able to extractfrom isolated cell walls of Anacystis nidulanshigh-molecular-weight material which, asjudged by its extraction properties and chemi-cal composition, could be considered to be acounterpart of the 0 antigens (lipopolysaccha-

rides [LPS]) of gram-negative bacteria. This 0antigen of A. nidulans, a unicellular blue-greenalga, has been, to date, the only reported LPSisolated from Cyanophyta. Whether other spe-cies, especially filamentous strains, also pos-sess similar LPS is of considerable interestfrom a taxonomic and a phylogenetic point ofview.The present paper describes the isolation and

the chemical and serological characterization ofthe 0 antigen of the filamentous blue-green algaA. variabilis (Hormogonales). It will be shownthat this LPS carries 0-specific groups similarto bacterial 0 antigens, but that it differsconsiderably in its chemical composition andextraction properties from the well-known en-terobacterial 0 antigens.

MATERIALS AND METHODSOrganism. The strain used was obtained from N.

G. Carr, Liverpool, and was derived from the A.variabilis (KUltzing) strain described by Kratz andMyers (17). It was cultivated in a modified JUtttnermedium (15) containing: NaNOs, 1.6 g; K.HPO4, 80mg; MgSO4.-7 H20, 80 mg; CaCl,-2 H,O, 60 mg;ethylenediaminetetraacetic acid-iron (III)-sodiumsalt (Fluka A.G. Buchs, Switzerland), 6 mg; K2COs,10 mg; H,BO4, 0.6 mg; MnCl2.4H,0, 0.6 mg;

672

on Decem

ber 26, 2019 by guesthttp://jb.asm

.org/D

ownloaded from

LIPOPOLYSACCHARIDE OF ANABAENA VARIABILIS

ZnSO4 * 7H2O, 0.01 mg; Na2MoO4* 2H2O, 0.02 mg;CuSO4 5H,O, 0.02 mg; Co(NO,)2 6H,0, 0.01 mg;and glass-distilled water, 1,000 ml. The pH wasadjusted to 7.5. Mass cultures were prepared in a10-liter Microferm MF-114 (New Brunswick Sci.Corp. Inc.). A light unit of fluorescent tubes was setup in front of the culture vessel (ca. 5,000 lx on thesurface of the Microferm vessel). A mixture of air andCO2 (98:2) was bubbled continuously through theculture. The temperature was adjusted to 32 C. Thestrain does not fix nitrogen, and grows in turbulent,stirred cultures as very short filaments (four to eightcells). In agar cultures, it forms long filaments of 30to 50 cells. Heterocysts were not observed.

Isolation of LPS. Lyophilized mass cultures wereextracted by the hot phenol-water method of West-phal et al. (41). The water and phenol phases werefreed from phenol by extensive dialysis against tapwater and finally against distilled water. After re-moval of insoluble material by centrifugation (2,500 xg, 30 min), the extracts were freed from low-molecu-lar-weight material by ultracentrifugation (threetimes at 105,000 x g for each run). Extraction of A.variabilis with liquid phenol-chloroform-petroleumether was performed as described by Galanos et al.(7).Sugar analysis. Neutral sugars were liberated by

hydrolysis with 1 N sulfuric acid at 100 C for 4 h. Thehydrolysates were neutralized with a saturated aque-ous solution of Ba(OH),. Sugars were separated bydescending paper chromatography (Whatman no. 1)with the solvent system pyridine-butan-1-ol-water(4:6:3, voVvol/vol) and were stained with silver ni-trate-NaOH (34) or with anilinium hydrogen phthal-ate (30).

For gas-liquid chromatography, the LPS was hy-drolyzed in 0.1 N HCl at 100 C for 48 h. The aldoses ofthe neutralized hydrolysates (ion-exchanger IRA 410,HCO-) were converted into alditol acetates by themethod of Sawardeker et al. (31). A Varian aerograph(model 1520 B) fitted with a glass column (0.32 by 152cm), filled with ECNSS-M (3% on Gas-Chrom Q, 100to 120 mesh) at a column temperature of 165 C, and ata nitrogen flow rate of 28 ml/min was used. Massspectra of the isolated 0-methyl sugar were deter-mined, at 70 eV, in a Perkin-Elmer mass spectrometer(model 270B) fitted with a Honeywell Visicorder(model 3508).

Conditions employed in determining the presenceor absence of 2-keto-3-deoxyoctonate (KDO) werehydrolysis of the LPS with 0.1 N sulfuric acid at 100 Cfor 20 min and reaction with thiobarbituric acidreagent by the method of Heath and Ghalambor (10).Amino sugars were liberated by hydrolysis in 6 N HClat 100 C for 16 h. After removal of HCl by repeatedevaporation in vacuo in the presence of KOH in adesiccator, amino sugars were identified by high-volt-age paper electrophoresis in the vertical chambersystem described by Kickhofen and Warth (16), usinga pyridine-acetic acid-water (10:4:86, vol/vol/vol)buffer (pH 5.3). An amino acid analyzer (Beckman,Type 120 B) packed with Custom Research Resin(type AA-15) was used for the quantitation of glucosa-mine.

Amino acids and phosphorus. Amino acids, liber-ated by hydrolysis in 6 N HCl at 100 C for 16 h, werefreed from HCI and identified in an amino acidanalyzer as described above. Phosphorus was deter-mined by the method of Lowry et al. (19).

Fatty acid analysis. Fatty acids were liberated byhydrolysis in 4 N HCl at 100 C for 6 h, extracted fromthe hydrolysate with petroleum ether (40 to 60 C),weighed, and esterified with 2 N methanolic HCl at100 C for 2 h. For identification, a Perkin-Elmer gaschromatograph (model F20) fitted with a Castorwax(2.5% on Chromosorb G, 80 to 100 mesh) stainless-steel column (0.32 by 152 cm) at a column tempera-ture of 170 C and a nitrogen flow of 25 ml/min, oralternatively an EGSS-X (15% on Gaschrom P, 100 to120 mesh) stainless-steel column (0.32 by 150 cm) at acolumn temperature of 165 C and a nitrogen flow of 30ml/min, was used. Mass spectrometry was carried outas described above, using a coupled gas chromatogra-phy-mass spectrometry system.

Hydroxylaminolysis. Ester-bound fatty acids weretransformed into ethanol-soluble hydroxamic acidsby the method of Snyder and Stephens (33). The LPSresidue with its presumable amide-bound fatty acidsbecame insoluble in ethanol and was precipitated bycentrifugation (1,450 x g, 30 min). After the superna-tant fluid was decanted, hydroxamic acids werehydrolyzed by addition of 6 N HCl to the supernatantfluid (pH 2.0) and heating at 100 C for 2 h. Theamide-bound fatty acids in the sediment were split offfrom glucosamine in alkali (4 N NaOH, 100 C, 5 h).The identification of amide- and ester-bound fattyacids was achieved by gas chromatographic analysisof the fatty acid methyl esters (see above).

Serological tests. Antisera were prepared by intra-venously immunizing albino rabbits, three times at4-day intervals, with increasing amounts (0.25, 0.5,and 1.0 ml) ofA. variabilis cell suspensions containing10' cells/ml in 0.9% saline. The immunization wascarried out either with living or with heat-killed cells(100 C, 2.5 h). The rabbits were bled 5 days after thelast injection. For passive hemagglutination tests,human erythrocytes (blood group A) were sensitizedwith untreated, heat-treated (100 C, 60 min), oralkali-treated (0.25 N NaOH, 56 C, 60 min [27]) LPSsamples. Serial dilutions of antisera were preparedwith phosphate-buffered saline (pH 7.2) by using theMetrimpex microtitrator system (Metrimpex typeOX-603). Fifty-microliter amounts of the serum dilu-tions were mixed with equal amounts of LPS-sensi-tized erythrocyte suspensions (25 Mliters of packedcells, suspended in 5 ml of saline). The titers wereread after a 60-min incubation of the plates at 37 C.The reciprocal serum dilution of the last dilution cupsshowing complete agglutination were recorded ashemagglutination titers. Agar-gel immunoelectro-phoresis was carried out, by the method of Scheideg-ger (32), using an electrophoresis chamber from theGelman Instrument Company (model 51170-1) and adiethyl-barbituric acid-sodium Veronal-calcium lac-tate buffer (12) at 10 V/cm for 1 h. Agar-gel precipita-tion was performed as described by Ouchterlony (29).Concanavalin A was obtained from Miles Yeda Ltd.,Rehovoth, Israel.

673VOL. 120, 1974

on Decem

ber 26, 2019 by guesthttp://jb.asm

.org/D

ownloaded from

WECKESSER ET AL.

Incubation with a- and fB-amylase. Incubation ofthe water-phase material with a-amylase from bac-teria or with ,B-amylase from barley (both from FlukaA.G., Buchs, Switzerland) was carried out as de-scribed by Mayer et al. (23). The incubation mixturewas examined directly by paper chromatography todetect any release of maltose (staining with AgNO3-NaOH [34]). Bacterial glycogen from Escherichia coliB was used as a positive control.

RESULTSHigh-molecular-weight material (LPS)

from the phenol phase of phenol-water ex-tracts: (i) isolation. The hot phenol-waterprocedure is generally adopted for isolation of 0antigens of gram-negative bacteria. By thismethod, a high-molecular-weight material wasextracted from A. variabilis cells into the phenolphase. The yield was about 1% of the cell dryweight. To remove possible contamination byphospholipids and other lipophilic materialsthat would also be extracted into the phenolphase, the high-molecular-weight fraction waswashed extensively with chloroform-methanol(2:1, vol/vol). In spite of its lipophilic charac-ter, we were not able to extract this materialwith phenol-chloroform-petroleum ether, amethod recently found to be highly suitablefor extracting lipophilic glycolipids (R-LPS)of enterobacterial R-mutant strains (7). Thehigh-molecular-weight material isolated fromA. variabilis represents a LPS by chemicalcriteria and an 0 antigen by its serological ac-tivity (see below).

(ii) Analyses for neutral sugars. The neu-tral sugar fraction obtained by acid hydrolysis(1 N H2SO4, 100 C, 4 h) of the LPS wasanalyzed by paper chromatography and bygas-liquid chromatography after the sugarswere converted into their respective alditolacetates. Rhamnose and glucose were present inlarge amounts; mannose and galactose, how-ever, were present in lower percentages (Table1). The optical rotation values of the isolatedsugars, separated by preparative paper chroma-tography, revealed a D-configuration for man-nose, glucose, and galactose, and an L-configu-ration for rhamnose. It is noteworthy thatL-glycero-D-mannoheptose, an ubiquitous con-stituent of enterobacterial 0 antigens, is absentfrom the LPS of A. variabilis.An additional neutral sugar, characterized by

its high R, value in pyridine-butan-1-ol-water(Rrhamnose = 1.24), amounted to 18.4% of theLPS dry weight. It was recognized by its inten-sive brown staining with anilinium hydrogenphthalate and by its weak reactivity withAgNO3-NaOH. The sugar could easily be ob-

TABLE 1. Composition of the LPS from Anabaenavariabilis

Component Amt of component(mg/100 mg of LPS)

Total neutral sugars 78.2L-Acofriose 18.4L-Rhamnose 19.3D-Mannose 3.9D-Glucose 34.3D-Galactose 2.3L-Glycero-D-mannoheptose 0

2-Keto-3-deoxy-octonate 0

D-Glucosamine 2.1

Totalfatty acids 10.7fl-C,40H 3.1,-C160H 1.7,-C18OH 1.7C 16 1.5

1.7

Phosphorus 0.03

tained in a pure state due to its high chromato-graphic mobility. For mass spectrometry, itsdeuterium-labeled alditol acetate was preparedby using NaB2H, in deuterium oxide for thereduction. Fragmentation at 70 eV gave thefollowing main fragments: m/e 43, 88, 101, 143,190, and 203, and a fragment with lower inten-sity (about 1% of the base peak at m/e 43) atm/e 290. These fragments are typical for thefragmentation of 3-0-methyl-ethers of 6-deoxy-hexoses (38). The parental 6-deoxy-hexose, ob-tained by demethylation of the 0-methyl sugarunder investigation, was identified as rhamnoseby paper and gas-liquid chromatography. Au-thentic 3-O-methyl-L-rhamnose showed thesame properties in mass spectrometry and inchromatography. It gave a single peak in gas-liquid chromatography after its alditol acetatewas mixed with that of the unknown sugar on anECNSS column. The optical rotation of L-aco-friose was reported by Muhr et al. (26) to be[a]D + 39.10, whereas that of D-acofriose wasdescribed by Morrison et al. (25) to be [aD-27°. The definitive positive optical rotation ofthe sugar under investigation proved its L-con-figuration. The 0-methyl sugar of the LPS of A.variabilis was, therefore, identified as the 3-0-methyl-ether of L-rhamnose (L-acofriose).

(iii) Analyses for amino sugars and KDO.Separation of acid hydrolysates (6 N HCl,100 C, 16 h) of the A. variabilis LPS on anamino acid analyzer gave only a glucosaminepeak, and no further amino sugars could be

674 J. BACTERIOL.

on Decem

ber 26, 2019 by guesthttp://jb.asm

.org/D

ownloaded from

LIPOPOLYSACCHARIDE OF ANABAENA VARIABILIS

traced. The same result was obtained when thehydrolysate was separated by high-voltagepaper electrophoresis at pH 5.3.The assay for KDO in a mild hydrolysate (0.1

N H2SO4, 100 C, 20 min) of the LPS was

negative: no color was obtained with the thi-obarbituric acid reagent, indicating that KDOwas absent from the LPS of A. variabilis.Preliminary observations concerning the isola-tion of the lipid A moiety of A. variabilis bymild acid hydrolysis showed that it could not besplit off from the LPS by hydrolytic conditions,usually applied for the isolation of lipid A ofgram-negative bacteria (21). This is very proba-bly due to the lack of KDO in the A. variabilis 0antigen, which forms the acid-labile linkagebetween lipid A and the polysaccharide moietyof enterobacterial 0 antigens.

(iv) Analyses for fatty acids. The materialfor fatty acid analysis, obtained from acidhydrolysates (4 N HCl, 100 C, 6 h) by extractionwith petroleum ether, corresponded to 10.7%(dry weight) of the LPS. Gas-liquid chromatog-raphy and mass spectrometry of the fatty acidsconverted into their methyl ester derivativesallowed identification of three hydroxy fattyacids: ,B-hydroxy myristic acid (as the mainfatty acid), f,-hydroxypalmitic acid, and ,B-hydroxystearic acid. Also detected were pal-mitic acid and an unidentified fatty acid withTR = 5.6 (palmitic acid = 1) on an EGSS-X-column. The type of binding by which individ-ual fatty acids are bound to glucosamine was

estimated by hydroxylaminolysis. The threedifferent ,B-hydroxy fatty acids are bound to theamino group of the glucosamine, whereas pal-mitic acid is ester bound.

(v) Analyses for phosphorus and aminoacids. The LPS of A. variabilis is almostcompletely free from phosphorus (0.03% [dryweight] of the LPS) (Table 1). The total amountof amino acids did not exceed 8.4%, and analy-sis on an amino acid analyzer showed none to bepredominate. No diaminopimelic acid was de-tected. These results indicate that the proteincontained in the LPS sample represents a

contaminant in the phenol-soluble LPS, ratherthan a specific peptide moiety.

(vi) Serological studies. Antisera againstheat-killed or living A. variabilis cells reactedequally strongly in the passive hemagglutina-tion test with erythrocytes coated with theisolated A. variabilis LPS (Table 2). It wasnecessary, however, to pretreat the 0 antigeneither by alkali (0.25 N NaOH, 56 C, 60 min) or

by heat (100 C, 60 min) to obtain an effectivecoating of human erythrocytes. Both methods ofsensitization gave essentially the same results.

TABLE 2. Hemagglutination titers of humanerythrocytes, sensitized with LPS of Anabaena

variabilis, by rabbit antisera prepared with wholecells of Anabaena variabilis

Reciprocal hemagglutination titer

LP,ha-LPS, alkali-Serum no. LPSS,theat-d treated

LPS,ate treated (0.25 Nuntreated< (00miCn) NaOH, 56 C,

60mm) ~60min)1lga < 10 2,560 2,560120a < 10 5,120 2,560737b < 40 1,280 2,560784b <80 2,560 2,560

a Serum prepared with heat-killed cells.b Serum prepared with living cells.

(It should be mentioned that most of the enteric0 antigens show properties identical to thosementioned above.) Agar-gel precipitation stud-ies showed that no differences exist betweenantisera obtained with heat-killed and livingcells, indicating the heat stability of this 0antigen. In immunoelectrophoresis with thealkali-treated LPS (pH 8.6, 10 V/cm, 60 min),the precipitation arcs were formed around theantigen well, demonstrating that chargedgroups are virtually absent in the A. variabilis 0antigen.

L-Acofriose, one of the major sugar constitu-ents of the A. variabilis 0 antigen, was detectedearlier in the LPS of the photosynthetic bacte-rium Rhodopseudomonas capsulata 37b4 (38).The 0 antigens of both organisms were testedfor serological cross-reactivity, since it is knownthat lipophilic sugars, such as dideoxy-hexosesand 0-methyl sugars, very often occupy thenonreducing terminal positions in 0 antigensand, therefore, are of special importance to theserological specificity of the respective 0 anti-gen. No cross-reactions were observed, how-ever, in the passive hemagglutination test, nei-ther by exposing erythrocytes coated with A.variabilis LPS to R. capsulata antiserum, norvice versa.Macromolecular material (glycogen) from

the water phase of phenol-water extracts.Macromolecular material different from that ofthe phenol phase could be isolated from thewater phase after extraction of the cells withphenol-water. It amounted to about 6% of thecell dry mass. Qualitative and quantitativegas-liquid chromatography of acid hydrolysatesand optical rotation measurements showed thatthe pol,ymer consists exclusively of D-glucose(about 100% [dry weight] of the material). Fatty

675VOL. 120, 1974

on Decem

ber 26, 2019 by guesthttp://jb.asm

.org/D

ownloaded from

WECKESSER ET AL.

acids were present only in very small amounts(<0.3%). The polymer was inactive in passivehemagglutination and in agar-gel precipitationtests against antisera obtained with heat-killedand living A. variabilis cells. A very strongreaction was observed, however, in agar-gelprecipitation with concanavalin A (22), a lectinthat is known to react with polysaccharidescarrying a-glycoside-linked terminal units ofglucose, mannose, or N-acetylglucosamine (8).With immunoelectrophoresis with concanavalinA, no migration of the water-phase material wasobserved, indicating the absence of chargedgroups in the glucan. Treatment of the polymerwith a- or f-amylase resulted in a high yield ofmaltose in addition to a number of oligosaccha-rides. The same result was obtained with bacte-rial glycogen from E. coli B.

DISCUSSIONThe present paper described the detection of

a LPS with 0 antigenic activity in the filamen-tous blue-green alga A. variabilis. In spite ofseveral striking similarities to enterobacterialLPS (i.e., extractability with phenol-water,presence of amide-linked fl-hydroxy fattyacids), the sugar and fatty acid compositiondeviates remarkably from that of enterobacte-rial 0 antigens. L-Rhamnose, its 3-0-methylether (L-acofriose), D-mannose, D-glucose, andD-galactose are the neutral sugars of the A.variabilis 0 antigen, and D-glucosamin its onlyamino sugar.

It is interesting that the two sugars L-glycero-D-mannoheptose and KDO, both components ofthe enterobacterial R core region, are absent inthe A. variabilis LPS. Either this LPS is devoidof an R core or it is built up by other constitu-ents. It would be particularly interesting toknow which sugar forms the linkage betweenthe lipid moiety and the polysaccharide part,since KDO is absent here. The only hithertodescribed LPS of Cyanophyta, originating fromA. nidulans, is also free from heptose, but itdoes still contain KDO (40). The same is truefor the 0 antigens of Xanthomonas campestris(11, 35), Rhodopseudomonas capsulata 37b4 (36)and R. viridis F (39). In contrast the LPS ofNeisseria catarrhalis (1), Bacteroides fragilisand B. melaninogenicus (13) are free from bothheptose and KDO. 0-methyl sugars, like L-aco-friose, are rare constituents in 0 antigens ofgram-negative bacteria. In photosynthetic bac-teria of the family Rhodospirillaceae, 0-methylsugars were however encountered frequently(37-39). It is noteworthy, that A. variabilisrepresents another photosynthetic prokaryote

in which the 0 antigen contains an 0-methylsugar. Since this sugar represents 18.4% of theLPS dry weight, it seems to be present in everyrepeating unit of the polysaccharide. Other LPSusually possess 0-methyl sugar constituents inonly minor amounts. R. capsulata 37b4, for ex-ample, contains only 1% L-acofriose and 15% L-rhamnose, indicating that not all repeating unitsof the 0-specific chain contain the 0-methylsugar (36). Similar observations were made byBjorndal et al. (3) with the 0 antigen of Kleb-sielkl K73:010. A thorough investigation of thisLPS by a methylation study showed that Lracofriose is restricted to the terminal repeat-ing unit.

Phosphate is not present in the LPS of A.variabilis, indicating that the lipid A moiety ofA. variabilis differs considerably from entericlipid A. Phosphate, as an integral part of theenteric LPS provides a bridging between twoglucosamine-disaccharide units in the lipid A(20). :-Hydroxymyristic acid, the main fattyacid of enteric 0 antigens, also represents themain fatty acid of the A. variabilis LPS, but theother fatty acids do not correspond to thoseusually found in lipid A.The occurrence of three amide-linked #-

hydroxy fatty acids in a procaryotic 0 antigen isunusual. In addition, this composition is ofspecial interest, since the fatty acid analyses ofblue-green algae that are presently availablewere usually carried out with whole cells, andthey predominantly demonstrate the presenceof saturated, mono-, di- or tri-unsaturated long-chain fatty acids, but not the presence ofhydroxy fatty acids (for summary see reference28).

All of these data demonstrate that the 0antigens of Cyanaphyta, as well as of somegram-negative bacteria, do not strictly followthe general structural concept of enterobacterialO antigens.The LPS of A. variabilis carries 0 specificity.

It behaves like a bacterial 0 antigen in passivehemagglutination and agar-gel-precipitation,but does not migrate upon immunoelectro-phoresis, substantiating the absence of phos-phate and KDO observed by chemical analysis.The LPS seems to be localized in the outermostlayer of the multiply layered cell wall, and it isobviously not covered by any capsular or slimeantigens, since similar titers were observedwhen using antisera against either heat-killed orliving cells.

Like other blue-green algae, the investigatedstrain ofA. variabilis is gram positive. However,the fine structure and the chemical composition

676 J. BACTERIOL.

on Decem

ber 26, 2019 by guesthttp://jb.asm

.org/D

ownloaded from

LIPOPOLYSACCHARIDE OF ANABAENA VARIABILIS

of the cell walls of blue-green algae so farinvestigated clearly demonstrate the gram-neg-

ative character of these cell walls.The high-molecular-weight material from the

water phase of the phenol-water extracts isclosely related to bacterial glycogen. This mate-rial has been described as a typical reserve

component in some Cyanophyta cells (for sum-mary see reference 5). The polymer is serologi-cally inactive and does not seem to be part ofthe cell wall.The present paper shows that LPS compara-

ble to the 0 antigens of gram-negative bacteriaexist not only in unicellular species (A.nidulans), but also in the filamentousCyanophyta. It should be pointed out, how-ever, that up to now investigations have beenrestricted to only one representative of eachgroup, so that the results cannot be generalized.It would, therefore, be highly desirable to obtainmore precise knowledge on the 0 antigens andother cell wall constituents from other strains ofthese organisms.

ACKNOWLEDGMENTS

We wish to thank R. Warth for the amino acid analysesand S. Gogowska, B. Straub, and D. Borowiak for experttechnical assistance.

This work was supported by the Deutsche Forschungs-gemeinschaft.

LITERATURE CITED

1. Adams, G. A., T. G. Tonabene, and M. Yaguchi. 1969.Cell wall lipopolysaccharides from Neisseria ca-

tarrhalis. Can. J. Microbiol. 15:365-374.2. Allen, M. M. 1968. Ultrastructure of the cell wall and cell

division of unicellular blue-green algae. J. Bacteriol.96:842-852.

3. Bjmrndal, H., B. Lindberg, and W. Nimmich. 1970.Structural- studies on the lipopolysaccharide fromKlebsiella. Acta Chem. Scand. 24:3414-3415.

4. Drews, G. 1973. Fine structure and chemical compositionof the cell envelopes, p. 99-116. In N. G. Carr, and B. A.Whitton (ed.), The biology of blue-green algae. Black-well Scientific Publications, Oxford.

5. Drews, G. 1973. Protophyta (Bakterien, Blaualgen), p.

413-440. In G. C. Hirsch, H. Ruska, and P. Sitte (ed.),Grundlagen der Cytologie. VEB Gustav Fischer Verlag,Jena.

6. Frank, H., M. Lefort, and H. H. Martin. 1962. Elek-tronenoptische und chemische Untersuchungen an

Zellwilnden der Blaualgen, Phormidium uncinatum. Z.Naturforsch. 17:262-268.

7. Galanos, C., 0. LUderitz, and 0. Westphal. 1969. A new

method for extraction of R lipopolysaccharides. Eur. J.Biochem. 9:245-249.

8. Goldstein, I. J., C. E. Hollermann, and E. E. Smith. 1965.Protein-carbohydrate interaction. H. Inhibition studieson the interaction of Concanavalin A with polysaccha-rides. Biochemistry 4:876-883.

9. Golecki, J. R., and G. Drews. 1974. The structure of theblue-green algal cell wall. Cell walls of Anabaenavariabilis revealed by freeze etchings before and afterextraction. Cytobiologie 8:213-227.

10. Heath, E. C, and M. A. Ghalambor. 1963. 2-Keto-3-

deoxy-octonate, a constituent of the cell wall lipo-polysaccharide preparations obtained from Esche-richia coli. Biochem. Biophys. Res. Commun.10:340-345.

11. Hickmann, J., and G. Ashwell. 1966. Isolation of bacteriallipopolysaccharide from Xanthomonas campestriscontaining 3-acetamido-3,6-dideoxy-Dgalactose andD-rhamnose. J. Biol. Chem. 241:1424-1428.

12. Hirschfeld, J. 1960. Immunolelectrophoresis procedureand application to the study of group-specific varia-tions in sera. Science Tools (LKB, Stockholm). 7:18-25.

13. Hofstad, T., and T. Kristoffersen. 1971. Lipopolysaccha-ride from Bacteroides melaninogenicus isolated fromthe supernatant fluid after ultracentrifugation of thewater phase following phenol-water extraction. ActaPathol. Microbiol. Scand. 79:12-18.

14. Jost, M. 1965. Die Ultrastruktur von Oscillatoria ru-

bescens D.C. Arch. Mikrobiol. 50:275-280.15. Jfittner, F., H. Victor, and H. Metzner. 1971. Massenan-

zucht phototropher Organismen in einer automa-tischen Kulturanlage. Arch. Mikrobiol. 77:275-280.

16. Kickhofen, B., and R. Warth. 1968. Eine TrennkammerfUir die Hochspannungselektrophorese nach demMichl'schen Prinzip. J. Chromatogr. 33:558-560.

17. Kratz, W. A., and J. Myers. 1955. Nutrition and growth ofseveral blue-green algae. Amer. J. Bot. 42:282-287.

18. Lamont, H. C. 1969. Sacrificial cell death and trichomebreakage in an Oscillatoriacean blue-green alga-thfrole of murein. Arch. Mikrobiol. 69:237-259.

19. Lowry, 0. H., N. R. Roberts, K. Y. Leiner, M. L. Wu, andA. L. Farr. 1954. The quantitative histochemistry ofbrain. I. Chemical methods. J. Biol. Chem. 207:1-17.

20. Lfideritz, O., C. Galanos, V. Lehmann, M. Nurminen, E.Th. Rietschel, G. Rosenfelder, M. Simon, and 0.

Westphal. 1973. Lipid A, chemical structure and bio-logical activity. J. Infect. Dis. 128(Suppl.):17-29.

21. Ltderitz, O., 0. Westphal, A. M. Staub, and H. Nikaido.1971. Isolation and chemical and immunochemicalcharacterization of bacterial lipopolysaccharides, p.145-233. In G. Weinbaum, S. Kadis, and S. J. Ajl (ed.),Microbial toxins. A comprehensive treatise, vol. 4.Academic Press Inc., London.

22. Mayer, H. 1972. Reaktivitat und Differenzierung derkompletten enterobakteriellen R-Basaltypen mit Con-canavalin A. Zbl. Bakt. Hyg., I. Abt. Orig. A 220:477-483.

23. Mayer, H., A. M. C. Rapin, and H. M. Kalckar. 1965.The selectivity of biosynthesis of glucosyl compoundsas illustrated by an E. coli mutant defective in UDPGsynthetase. Proc. Nat. Acad. Sci. U.S.A. 53:459-466.

24. Metzner, I. 1955. Zur Chemie und zum submikroskopi-schen Aufbau' der Zellwande, Scheiden und Gallertenvon Cyanophyceen. Arch. Mikrobiol. 22:45-77.

25. Morrison, I. M., R. Young, M. B. Perry, and G. A.Adams. 1967. The exocellular glycan produced byMyxobacterium 402. Can. J. Chem. 45:1987-1990.

26. Muhr, H., A. Hungerer, and T. Reichstein. 1955. Dievermutliche Struktur der L-Acofriose. Helv. Chim.Acta 38:499-504.

27. Neter, E. 1956. Bacterial hemagglutination and hemol-ysis. Bacteriol. Rev. 20:166-188.

28. Nichols, B. W. 1973. Lipid composition and metabolism,p. 144-161. In N. G. Carr, and B. A. Whitton (ed.), Thebiology of blue-green algae. Blackwell Scientific Publi-cations, Oxford.

29. Ouchterlony, 0. 1964. Gel-diffusion techniques. In J. F.Ackroyd (ed.), Immunological methods. BlackwellScientific Publications, Oxford.

30. Partridge, S. M. 1949. Aniline hydrogen phthalate as a

spraying reagent for chromatography of sugars. Nature(London) 164:443.

VOL. 120, 1974 677

on Decem

ber 26, 2019 by guesthttp://jb.asm

.org/D

ownloaded from

WECKESSER ET AL.

31. Sawardeker, J. S., J. H. Sloneker, and A. Jeanes. 1967.Quantitative determination of monosaccharides as

their alditol acetates by gas-liquid chromatography.Anal. Chem. 37:1602-1604.

32. Scheidegger, J. J. 1955. Une micro-methode de l'immuno-electrophorese. Int. Arch. Allergy 7:103-110.

33. Snyder, F., and N. Stephens. 1959. A simplified spectro-photometric determination of ester groups in lipids.Arch. Biochem. Biophys. Acta 34:244-245.

34. Trevelyan, W. E., D. P. Procter, and J. S. Harrison. 1950.Detection of sugars on paper chromatograms. Nature(London) 166:444-445.

35. Volk, W. A. 1968. Quantitative assay of polysaccharidecomponents obtained from cell wall lipopolysaccha-rides ofXanthomonas species. J. Bacteriol. 95:980-982.

36. Weckesser, J., G. Drews, and I. Fromme. 1972. Chemicalanalysis of and degradation studies on the cell walllipopolysaccharide of Rhodopseudomonas capsulata. J.Bacteriol. 109:1106-1113.

37. Weckesser, J., G. Drews, I. Fromme, and H. Mayer. 1973.

Isolation and chemical composition of the lipopolysac-charides of Rhodopseudomonas palustris strains. Arch.Mikrobiol. 92:123-138.

38. Weckesser, J., H. Mayer, and G. Drews. 1970. Theidentification of 3-O-methyl-L-rhamnose (L-acofriose)as constituent of the lipopolysaccharide of Rhodopseu-domonas capsulata. Eur. J. Biochem. 16:158-160.

39. Weckesser, J., G. Rosenfelder, H. Mayer, and 0. Lfider-itz. 1971. The identification of 3-O-methyl-D-xyloseand 3-O-methyl-L-xylose as constituents of the lipo-polysaccharides of Myxococcus fulvus and Rhodopseu-domonas viridis, respectively. Eur. J. Biochem.24:112-115.

40. Weise, G., G. Drews, B. Jann, and K. Jann. 1970.Identification and analysis of a lipopolysaccharide incell walls of the blue-green alga Anacystis nidulans.Arch. Mikrobiol. 71:89-98.

41. Westphal, O., 0. LUderitz, and F. Bister. 1952. Uber dieExtraktion von Bakterien mit Phenol/Wasser. Z. Na-turforsch. 7b:148-155.

678 J. BACTERIOL.

on Decem

ber 26, 2019 by guesthttp://jb.asm

.org/D

ownloaded from