Adhesion of Mycoplasma pneumoniae to Sulfated Glycolipids and ...

Val. 264, No. 16, Issue of June 5, PP.9283-9288. 1989 Printed in U.S. A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY

Adhesion of Mycoplasma pneumoniaeto SulfatedGlycolipids and Inhibition by DextranSulfate* (Received for publication, August 25, 1988)

Howard C. Krivan$$, Lyn D. Olsonll, Michael F. BarileT, Victor Ginsburg$, and David D. Roberts$ From the $.Laboratory of Structural Biology, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health and theTMycopkxma Laboratory, Center for Drugs and Biologics, Food and Drug Administration,Bethesda, . . Maryland 20892

A virulent strain of Mycoplaama pneumoniae was lung fibroblasts (MRC5), HeLa cells, hamster tracheal epimetabolically labeled with [‘Hlpalmitate and studied thelial cells, spermatozoa, and erythrocytes (6-12). Some of for binding to glycolipids and to WiDr human colon these studiessuggest that sialylglycoproteins maybe receptors adenocarcinoma cells. The organism binds strongly to for M. pneumoniae,as treatment of the cells with neuraminsulfatide and other sulfated glycolipids, such as semi- idase decreases binding (10, 14, 15). nolipid and lactosylsulfatide which all contain terminal Recent studies suggest thatthe organism recognizes Gal(3SOa)/31-residuesand weakly to some neolactoser- NeuAca2-3Gal~l-4GlcNAcsequences onerythrocytes(16), ies neutral glycolipids. M. pneumoniae do not bind as both glycolipids and glycoproteins containing this structure gangliosides including the sialylneolacto-series and inhibit adhesion of these microorganisms (17). Other studies, other neutral glycolipids that were tested. Only meta- however, suggest that glycolipids arenotreceptors for M. bolically active M. pneumoniae cells bind to sulfatide, pneumoniae (18, 19). To further examine the role of carboas binding is maximal in RPMI medium at 37 “C and almost completely abolished in nutrient-deficient me- hydrates as adhesion receptors, the binding of 3H-labeled M . dium or by keeping the cells at 4 O C . Dextran sulfate pneumoniae to glycolipids and, in the accompanying paper but not other sulfatedor anionic polysaccharides at 10 (20), glycoproteins was studied. We reporthere that inoverlay pg/ml completely inhibits bindingof M. pneumoniae to assays on thin layer chromatograms M. pneumoniae binds seminolipid, both of which contain purifiedsulfatide.Dextransulfate does not inhibit specifically to sulfatide and binding to the neolacto-series neutral glycolipids. Dex- a terminal Gal(3S04)pl-residue, and does not bind to other tran sulfate partially inhibitsadhesion of M. pneumo- acidic glycolipids includinggangliosides and the sialylated niae to cultured human colon adenocarcinoma cells linear or branchedneolacto-series glycolipids which were pro(WiDr). The biological relevance of these data is sug- posed to be receptorsfor M . pneumoniae on erythrocytes (16, gested by our finding that sulfatide occurs in large 17). amounts in human trachea, lung, and WiDr cells.Thus, EXPERIMENTALPROCEDURES there are at least two distinct receptors that mediate binding of M. pneumoniae to cells: glycolipids containMaterials-Dextran sulfate (Mr500,000, lot 44F-0408 and M, 5,000, ing terminal Gal(3S04)/31-residuesas reported here, lot 77F-06343, fucoidin, colominic acid(Escherichiacoli), hyaluronate, and glycoproteins containing terminal NeuAca2- dipalmitoyl-phosphatidylcholine(synthetic), cholesterol (grade I, 3GalB1-4GlcNAcsequences (Roberts, D. D., Olson, L. 99%), cholesterol 3-sulfate, and bovine serum albumin (A7030 fatty D., Barile, M. F., Ginsburg, V., and Krivan, H. C. acid and globulin free), were from Sigma. Bovine lung heparin (160 units/mg) was from the UpJohn Co. RPMI 1640 medium was pur(1989) J. Biol. Chem. 264,9289-9293).

Mycoplasma pneumoniae is a small procaryotic parasite of the human respiratory tract and etiologic the agent of primary atypical pneumonia. The pathogen has no cell wall and requires exogenous cholesterol forthe synthesisof plasma membrane and glucose as a carbon and energysource. In tracheal organ cultures the adhesion of viablemycoplasmas to the respiratory epithelium is essential for the initiation of infection (1-3). Once bound, M. pneumoniae does not penetrate theepithelialsurfacebutcausesextensive damage to the tracheal epithelium, leading to ciliostasis, loss of cilia, and finally cell death (4, 5 ) . M. pneumoniaealso bindsin vitro to many othereucaryotic cells, including human colon carcinoma cells (WiDr), human * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. To whom correspondenceshould be addressedLaboratory of Structural Biology, Bldg. 8, Rm. 2A27, NIDDK, NIH, Bethesda, MD 20892.

chased from Biofluids. Glycolipids-Bovine brainsulfatide (galactosylceramide-13-sulfate), ceramide monohexoside, ceramide trihexoside, globoside, and gangliosides GM1 and GDla were obtained from Supelco. Lactosylceramide andglucosylceramide were from Behring Diagnostics. Other reference gangliosides were from Bachem, Inc. Seminolipid (p-galactosylalkylacylglycerol-13-sulfate)was isolatedfrombovinetestes (Pel-Freez Biologicals) as previously described (21). Galactosyl ceramide-16-sulfate was prepared as previously described by sulfation of galactosyl ceramide (22). Sulfated glucuronosylparagloboside (IV3[3’S03GlcA]- nLcOse4Cer) was purifiedfromhumanperipheral nerve (23). Lactosylceramide-113-sulfate,GM3, and sialyllactofucopentaosyl-(111)-ceramidewere purified from human kidney (24-26). a-Galactosylparagloboside (IV3GalnLcOserCer) and the I-active aGalJactoisooctaosylceramide were purified from rabbit erythrocytes (Pel-Freez) (27). Lactoisooctaosylceramide was prepared from the latter lipidby treatmentwith coffee beana-galactosidase.a2-3Sialylparagloboside (NeuGc), a2-3-sialyllactoneohexaosylceramide, GM3 (NeuGc), and an I-active ganglioside were prepared frombovine erythrocytes (28). a2-3-Sialylparagloboside (NeuAc) was isolated from type 0 human erythrocytes (29). Paragloboside and lactoneohexaosylceramide were prepared by desialylation of the respective gangliosides with 1 M formic acid for 60 min at 100 ’C. Asialo-GM1 and asialo-GM2 were prepared as previously described (30). LactoN-triaosylceramide was prepared by digestion of paragloboside with bovine testes @-galactosidase(Boehringer Mannheim). The identities of the neolacto-series glycolipids were confirmed by immunostaining

9283

9284

Mycoplasma pneumoniae Adhesion

to Sulfatide

Mycophma Adhesion to Cultured Cells-Adhesion of [3H]-labeled with monoclonal antibody My-28before andafterneuraminidase digestion (31). Concentrations of galactosyl ceramide I'-sulfate, ga- M.pneumoniae to cells on glass coverslips was measured by a modilactosyl ceramide 13-sulfate,cholesterol sulfate, glucosylceramide, ga- fication of a method previouslydescribed (8). WiDr human colon adenocarcinoma(ATCCCCL 218) was grown in Eagle's minimal lactosylceramide,lactosylceramide,asialo-GM1, andasialo-GM2 were determined by dry weight. Other sulfated glycolipids were deter- essential medium with 10% fetal calf serum (Biofluids) in a 5% COZ mined by the dye-binding assayof Kean (32) asmodified by Tadano- atmosphere a t 37 "C. The cells were removed with trypsin and plated Aritomi and Ishizuka (33). The concentrations of the other neutral on 12-mm round glass coverslips in 24-well tissue culture plates and and acidic glycolipids listed in Table I were determined by densitom- grown for 3 days. Control coverslips were preincubated in medium medium then etry (Quickscan, Helena Laboratories) of orcinol-stained thin layer without cells. The coverslips werewashed in serum-free chromatograms compared with authentic standards. The purity of all incubated in RPMI-BSA for 15 min. Themedium was removed, and labeled M. pneumoniuesuspendedin 0.5 ml of RPMI-BSA were lipids were confirmed by thin layer chromatography in neutral and added to eachwell. The plates were incubated on a rocking table for acidic solvent systems. Lipids were extracted from normal human lung, trachea, and WiDr 60 min, a t 37 "C. The coverslips were washed by dipping in salinesix by scintillation cells (30,34) and separated into neutral and acidic fractions by anion times and the bound radioactive bacteria determined counting.Forinhibitionstudies,theinhibitors were added to M. exchange chromatography on DEAE-Sepharose in the bicarbonate pneumoniae prior to adding the bacteria to thecoverslips. form. For some experiments, WiDr cells were metabolically labeled with [S5S]sulfate (ICN Radiochemicals). Labeling was done for 48 h in Hams F-12 medium with 10% fetal calf serum, 10% RPMI 1640, RESULTS pM). and 100pCi/ml [35S]sulfate(totalsulfateconcentration80 Equilibration of [35S]sulfate withthe intracellular pool in WiDr cells Binding of M . Pneumoniae to Glycolipids on Thin Layer is complete within 4 h (35). The carrier sulfate concentration was Chromutograms-Incubation of 3H-labeled M . pneumoniae selected to minimize dilution of theintracellularsulfate pool by with various glycolipids resolved on thin layer chromatograms metabolism of sulfur-containing amino acids and under sulfation due to low carriersulfateconcentrations (36, 37).Thus,the specific was used to determine the carbohydrate binding specificity of the organism. As shown by an autoradiogram (Fig. L4) comactivity of the incorporated sulfate under these conditions should equal that in the medium. Cells were removed fromthe tissue culture pared with a similar thin layer plate visualized with orcinol flasks by removing the medium and adding 2.5 mM EDTA in 10mM reagent (Fig. lB), M . pneumoniue bound avidly to authentic phosphate-buffered saline, pH 7.3. After 60 min at 37 "C, the cells sulfatide, detecting 100 ng of this glycolipid ( l a n e cg), and to were collected by centrifugation and extracted as described above. a glycolipid with the same mobility as sulfatide in the acidic Desalted lipid extracts were analyzed by high performancethin lipid fraction of human trachea ( l a n e f). This trachealglycolayer chromatography developed in chloroform/methanol/0.25%KC1 in water (5:4:1) or chloroform/methanol/acetone/acetic acid/water lipid was confirmed to be sulfatide by its specific staining ( 8 2 4 2 1 ) . T h e labeled sulfated glycolipids were visualized by auto- with lZ5I-labeledvon Willebrand factor (38) (data notshown). radiography and quantified by scraping the bands and scintillation Sulfatide was also detected in human lung lipids but atlower counting. Sulfated glycolipids in the tissue extractswere detected by levels than in trachea. M . pneumoniae also bound to other staining of the lipids separated by high performance thin layer chro- sulfated glycolipids including lactosyl sulfatide and seminomatography with '"I-von Willebrand factor (38). Growth and Labeling of Organisms-Virulent M. pneumoniae strain lipid, which contain the same terminal Gal(BSO&?l-residue M129, passage 4-6, were grown and metabolically labeled with [3H] as sulfatide, and anisomer of sulfatide in which the terminal sulfate is linked to the6-position of galactose (see Table I for palmitic acid (12-17 Ci/mmol, Du Pont-New England Nuclear) as previouslydescribed (8). The organisms were passedfour times structures). Interestingly, M . pneumoniue also binds to high through a 26-gauge needle and suspended to approximately lo7 cpm/ amounts of lactosylceramide and to a lesser extent glucosylml of degassed RPMI 1640 medium containing 1% bovine serum ceramide, paragloboside, lactotriaosylceramide, and cu-galacalbumin (Sigma, fatty acid free) and 25 mM Hepes,' pH 7.3 (RPMIBSA). _ ~ ~ " -~ . - . Mycophma Overlay Assay-M. pneumoniae were bound to glycolipids separated on thin layer chromatograms as described in detail for other bacteria (30,39). Briefly, glycolipids were separated by thin CMHlayer chromatography on aluminum-backed silica gel high performanceplates(Merck,WestGermany) developedwith chloroform/ methanol/0.25% CaCIz in water (6035:8). After chromatography, the CDHplates were coated with 0.1% polyisobutylmethacrylate, soaked in CTH0.05 M Tris-HCI, pH 7.6, containing 110 mM sodium chloride, 5 mM GL4CaCI2,0.2 mM phenylmethanesulfonyl fluoride, and 1%bovine serum GM3albumin (TBS-BSA) and incubated for 3 h at 25 "C with 60 pl/cm2 GM2GM 1of 3H-labeled M. pneumoniae (approximately lo7 cpm/ml of RPMIGD laBSA). The plates were gently washed five times in 0.01 M sodium GO 1 bJ 1 phosphate, p H 7.2, containing 0.15 M sodiumchloride (PBS)to a D c ~ c t c g de f g h i a b c d e f g h I remove unbound organisms, dried, andexposed for 24 h to Ultrofilm FIG.1. Binding of M. pneumoniae to glycolipids separated 3H (2208-190) high speed film (LKB). by thin layerchromatography. Glycolipids were chromatographed Solid-phase Binding Assay-The binding of M. pneumoniae to purified glycolipids immobilized in microtiter plates (Falcon 3912, on aluminum-backed silica gel HPTLC plates developed in chlorowere coated Becton Dickinson) was measured as previously described (39). Puri- form/methanol/0.25% CaCIz in water, 6035% The plates fied glycolipids were serially diluted in 25 pl of methanol containing with plastic, soaked in Tris-BSA, and incubatedfor 3 h a t 25 "C with 0.1 pg each of the auxiliarylipids cholesterol and phosphatidylcholine. [3H]palmitate-labeledM. pneumoniae suspended in RPMI 1640 conAfter the solutions were dried by evaporation, the wells were filled taining 1% BSA and 25 mM Hepes,pH 7.3, as described under 1 h, rinsedwithRPMI-BSA,and orcinol reagent "Experimental Procedures" (panel A ), or sprayed with with TBS-BSA,emptiedafter incubated with 25 pl of 3H-labeled M. pneumoniae (approximately to identify glycolipids (panel B). Lane a, acidic glycolipid standards lo7 cpm/ml RPMI-BSA). After incubation for 2 h a t 37 "C (unless sulfatide (0.5 pg). GM3 (2 pg), GM2 (2 pg),GDla (2 pg), GDlb pg), (2 otherwise stated), the wells were washed five times with saline, and GTlb (2 pg); lane b, neutral standards galactosyl ceramide (4 pg), lactosylceramide (4 pg), globotriaosylceramide (2 pg), and globotebound M. pneumoniae was quantified by scintillation counting in Aquasol. For inhibition studies, variouspolysaccharides were serially traosylceramide (2 pg); lunes c and cl, sulfatide (2 pg), c2, (0.5 pg), diluted in 25 p1 of RPMI-BSA in microtiter wells followed by the and c3 (0.1 pg); lane d, seminolipid (2pg); lane e, cholesterol 3-sulfate addition of 25 pl of 3H-labeled M. pneumoniae. (2pg); lane f, human trachea acidic glycolipids from100 mg wet weight of tissue; lane g, monosialoganglioside from 100 mg wet weight ' The abbreviations used are: Hepes,4-(2-hydroxyethyl)-l-pipera- of bovine erythrocytes; lune h, a2-3-sialylparagloboside(2 pg); lane i, zineethanesulfonic acid; BSA, bovine serum albumin; HPTLC, high I-active monosialylganglioside from bovine erythrocytes (2 pg). For performance thin layer chromatography. abbreviations see Footnote 1 and Table I. ,

to Sulfatide


9285

TABLEI Glycolipids tested for ability to bind M. pneumoniae on thin-layer chromatograms Name"

Structure

Bindingb

Sulfatide Sulfatide Lactosylsulfatide Seminolipid Glucosylcer (CMH) Lactosylcer (CDH) Lacto-N-triaosylcer Paragloboside a-Galactosylparagloboside Galactosylcer (CMH) S04-Glucuronosylparagloboside Trihexosylcer (CTH) Asialo GM2 Globoside (GL4) Asialo GM1 GM3 GM3 (NeuGc) GM2 GM1 Sialylparagloboside Sialylparagloboside (NeuGc) Sialylneolactofucopentaosylcer GDla GDlb GTlb Sialylneolactohexaosylcer acI-active sialyllactoisooctaosylcer acI-active lactoisooctaosylcer acI-active Galz-lactoisooctaosylcer

Trivial names and structures and represented according to recommendations in Ref. 52 and references cited therein; cer, ceramide; CMH, ceramide monohexoside; CDH, ceramide dihexoside; CTH, ceramide trihexoside; GL4, globoside. 'Negative binding (-) indicates no binding to 4 pg of lipid and positive binding to less than 0.5 pg (+++), 0.52 pg (++), and 2-4 pg (+).

tosylparagloboside, but not to other neutral glycolipids (Table quires energy and physiological temperatures for maximal I). No binding was detected to otheracidic glycolipids includ- binding tooccur. Inhibition of M. pneumoniae Binding to Immobilized Suling a2-3-sialylparagloboside,I-active monosialylganglioside, or to thegangliosides GM3, GM2, GM1,GDla, GDlb, GTlb. fatide, Lactosylceramide, and WiDr Monolayers by Dextran In addition, sulfate itself is not sufficient for binding as M . Sulfate-Various anionic polysaccharides were tested for inpneumoniae does not bind to high amounts of cholesterol hibition of M . pneumoniae binding to sulfatide immobilized sulfate or to sulfated glucuronosylparagloboside, which has a in microtiter plates (Table 11, Fig. 4). High molecular weight terminal sulfate linked to the 3-position of glucuronic acid. dextran sulfatewas the most potent inhibitor of M. pneumoQuantitative Bindingof M . pneumoniae to ImmobilizedGly- niae binding to1pg of sulfatide with50% inhibition obtained colipids i n Microtiter Plates-Binding of M. pneumoniae to at 0.4 pg/ml,whereas dextran had no effect (Fig. 4). The purified glycolipids adsorbed on microtiter plates was exam- sulfated fucan, fucoidin, and low molecular weight dextran ined to further define binding specificity. Binding to sulfatide sulfate were weak inhibitors andseveral other polysaccharides was sensitive and dose dependent (Fig. 2). M. pneumoniae tested were inactive (Table 11). High molecularweight dextran bound weakly to lactosylceramide and paragloboside, whereas sulfate did notinhibit M . pneumoniae bindingto high no binding was detected to cholesterol sulfate or other glyco- amounts of lactosylceramide and other neolacto-series glycolipids tested at 10 pg/well, consistent with the data obtained lipids suggesting that the inhibitory activity of this polysacfrom theoverlay assay.Binding of M. pneumoniae to sulfatide charide is specific for sulfatide (data not shown). is both energy and temperature dependent (Fig. 3). At 37 "C Because virulent strains of M . pneumoniae adhere to mamabout 0.25 pg of sulfatide was requiredfor half-maximum malian cells, monolayersof WiDr cells were used to determine binding. The binding activity was about 5 times lower at 25 "C if dextransulfateinhibits adhesion of the organism. 3Hand was minimal at 4 "C. M. pneumoniae also bound poorly Labeled M . pneumoniae was incubated with WiDr cell monoat 37 "C innutrient-deficient medium (Tris-BSAwithout layers attached to coverslips in triplicate with and without RPMI) with binding activities comparable to that obtaineddextran at sulfate (Table 111).Dextran sulfate inhibited adhesion 4 "C (Fig. 3). These results suggest that M . pneumoniae re- in all three experiments, but the degree of inhibition varied

Mycoplasma pneurnoniae Adhesion to Sulfatide

9286

I

o

I

0.1

0

1

10

INHIBITOR (pg/ml)

GLYCOLIPID C p g )

FIG. 2. Binding of M. pneumonia to purified glycolipids. FIG. 4. Inhibition of M. pneumoniae binding to sulfatide by Lipids in 25 pl of methanol containing 0.1 pg each of the auxillary dextran sulfate. Polysaccharides were serially diluted in 25 plof lipids cholesterol and phosphatidylcholine were evaporated in flat RPMI-BSA in microtiter wells previouslycoated with1pg of purified bottom wells of polyvinylchloride microtiter plates. The wells were sulfatide. Binding was determined after incubation for 2 h at 37 "C blocked with 1%albumin for1h, washed twice withRPMI-BSA,and with 25 pl of 3H-labeledM. pneumoniae with the indicated concenincubated at 25 "C with 25 pl of 3H-labeledM.pneumoniae (approx- tration of dextran (0)or dextran sulfate (A). imately 10' cpm). After2 h, the wells were washed5 times with saline, cut from the plate, and bound radioactivity quantified in a scintillaTABLEI11 tion counter. In control experiments organismswere incubated with auxiliary lipids only to correct for nonspecific binding (typically ~ 1 % Inhibition of M. pneumoniae adherenceto adenocarcinoma cell monolayers (WiDr) by dextran sulfate of the total radioactivity added).M.pneumoniae binding was determined in RPMI-BSA for sulfatide ,).( lactosylceramide (m),para3H-labeledM.pneumoniae attached" Dextran sulfate globoside (e), and cholesterolsulfate,ceramide trihexoside, globoside, Exp. 1 Exp. 2 Exp. 3 GM1, GM2, or GM3 (0). dm1 1 89

19

% of control

69

83 77

10 74 31 ( p < 0.002) 78 ( p C 0.05) 72 ( p C 0.1) 100 Results are the average of triplicate determinations normalized to control binding in the absence of inhibitor: 7, 21, and 11%of the added mycoplasma, respectively, for the three experiments. Nonspecific binding to medium-treated coverslips without cellswas 2-3% of the total added. The significance of the inhibition at 100 pg/ml of dextran sulfate, relative to the control adhesion to WiDr cells in the absence of inhibitor, was determined using a two-sided t test.

SULFATIDE (pg)

FIG. 3. Energy and temperature-dependent binding of M. pneumoniae to sulfatide. Microtiter wells were coated with sulfa-

tide and blocked with albumin as described in the legend of Fig. 2. Binding of 3H-labeledM. pneumoniae was determined in RPMI-BSA for 2 h at 4 "C (m),25 "C (O),37 "C (A),and at 37 "C in BSA without RPMI (A).

TABLEI1 Inhibition of M. pneumoniae binding to sulfatide by anionic polysaccharides Inhibitor

160"

between experiments. In general, greater inhibition was obtained with subconfluent WiDr cells than with confluent cells (data not shown). In each experiment, 10 pg/ml of dextran sulfate inhibited more than1pg/ml but 100 pg/ml of dextran sulfate caused no further inhibition. Thus, maximal inhibition was obtained with approximately 10 pg/ml of dextran sulfate. Similar results were obtained with MRC5 lung fibroblasts where in three experiments a mean of 47% of M . pneumoniue adhesion was inhibited by dextran sulfate (data not shown). Metabolic labeling with [35S]sulfate confirmed that WiDr cells make a large amount of sulfatide (Fig. 5). An orcinolpositive resorcinol-negative glycolipid that comigrates with authentic brain sulfatide is detected acidic in lipids from WiDr cells (Fig. 5A, lane c). This lipid contains35S(Fig. 5A, lane a) and was verified to be sulfatide by comigration with authentic sulfatide in an acidic solvent system that resolves sulfatide, seminolipid, and cholesterol 3-sulfate(Fig. 5B, lane a). Based on incorporationof 35Sunder steady statelabeling, WiDr cells contain approximately 90 nmol of sulfatide/gram wet weight of cells representing 41%of the sulfate incorporated into lipid. This agrees with the densitometric estimationof sulfatide concentration of orcinol-stainedthin layer chromatograms of 50 nmol of sulfatide/gram wet weight of cells. Significant label was also incorporated into cholesterol 3-sulfate (39%) and small amounts into a lipid that comigrates with seminolipid (Fig. 5B, lane a).

Dextran sulfate (Mr500,000) 0.4 Dextran sulfate (Mr5,000) 10 Fucoidin 50 Heparin >200 sulfate Chrondroitin >200 acid Colominic >200 DISCUSSION phosphomannan Yeast >zoo Recently, Loomes et al. (17) reported that adhesion of M . Concentration giving 50% inhibition of M.pneumoniae binding pneumoniae to human erythrocyteswas inhibited by ganglioto 1 pg of sulfatide/well inthe solid-phase binding assay.

Mycoplasma pneumoniae Adhesion to Sulfatide

SM4s-

GM3GM2GM 1GD 1 a-

a

b c

a

,

.-.

,

II

.

,

b c d e f

FIG. 5. Identification of sulfatide synthesized by WiDr adenocarcinoma cells. WiDr cells were metabolically labeled with [35S]sulfateas described under “Experimental Procedures.” Neutral and acidic lipids were chromatographed on silica gel high performance thin layer plates developed in chloroform/methanol/0.25% KC1 in water, 5:4:1 (panel A ) or chloroform/methanol/acetone/acetic acid/ water, 8:2:421 (panel B ) . The lipids were detected by autoradiography (lane a ) or orcinol reagent (lanes b-f).Panel A, 35S-labeledacidic lipids from lo6 WiDr cells (lane a ) , neutral (lane b ) , and acidic (lane c) lipids from 30 mg wet weight of WiDr Cells. The orcinol positive Migration of reference sulfatide band is indicated by the arrow (c). glycolipids is indicated in the left margin: sulfatide, GM3, GM2, GM1, GDla, GDlb, and GTlb. Panel B, 35S-labeledacidic lipids from lo6 WiDr cells (lanes a and c), bovine brain sulfatide (lane b ) , seminolipid (lone d ) , cholesterol 3-sulfate (lane e), and neutral glycolipid standards from top to bottom CMH, CDH, CTH, and GL4 (lane f). For abbreviations see footnote 1 and Table I.

9287

lacto-series core structure, inhibitadhesion of M.pneumoniae (16, 17), and our finding that the organism binds to asparagine-linked oligosaccharides bearing this terminal structure (20). However, as demonstrated for laminin-mediated hemagglutination and laminin binding to sulfatide (40), inhibition by gangliosides may be indirect in that they mask sulfatide receptors. Why M. pneumoniae does not bind to sialylneolacto-series glycolipids is unclear. Possibly, mannose which occurs in asparagine-linked glycoproteins but is absent in glycolipids may be required for tight binding (20). The biological relevance for sulfatide for adhesion of M. pneumoniae is suggested by three findings. First, only metabolically active M . pneumoniae cells bind to sulfatide (Fig. 3). At physiological temperatures, binding was maximal in RPMI medium and almost completely abolished in nutrient-deficient medium or by keeping the cells at 4 “C. These results are consistent with the findings of others that adhesion of M. pneumoniue is decreased by metabolic poisons, low temperatures and by using nonviable organisms (3, 41-43), and that adhesion may be influenced by an energized membrane (3, 44). Second, sulfatide occurs in high amountsinhuman trachea and is present in human lung and cultured WiDr human colon adenocarcinoma cells (Figs. 1and 5). The latter have approximately 50 million sulfatide molecules/cell. Additional sulfated glycoconjugates that are recognized by M. pneumoniae may bepresent onglycoproteins or proteoglycans in these tissues. It remains to be determined whether sulfatide or other sulfated glycoconjugates recognized by the organism are expressed on the apical membrane of the tracheal epithelium where M . pneumoniae adheres to initiateinfection. And third, dextran sulfate, which specifically inhibits sulfatidedependent binding, partially inhibits M . pneumoniae adhesion to WiDr cells. The existence of receptors other than sialylglycoproteins would explain why inhibition of M. pneumoniae binding to cultured cell lines by sialylglycoconjugates (16, 17, 19, 42,46) or following neuraminidase treatment (7,18,41,45)is usually incomplete. This is also the case with our inhibition studies in which 10 pg/ml of dextran sulfate completely abolished binding to purified sulfatide immobilized on plastic (Fig. 4), but only partially inhibitedM. pneumoniae adhesion to WiDr cells (Table 111). Thus, there are probably a t least two distinct receptors that mediate binding of M. pneumoniae to cells: glycolipidscontaining terminal Gal(3S04)@1-residues and glycoproteinscontaining terminalNeuAca2-3Gal~l-4GlcNAc sequences, both of which must be blocked for complete inhibition of M. pneumoniae binding to cultured cells (20). Neutral neolacto-series glycolipids are possibly a third distinct receptor for M. pneumoniae, but binding to these lipids is probably of too low avidity to be physiologically important. Sulfatide-mediated adhesion, however, may be important in mycoplasma pathogenicity to guarantee intimate contact of the parasite with the hostmembrane to satisfy itsstrict nutritional requirements. Many functions have been suggested for sulfated glycolipids in biology (reviewed in Ref. 47), including a role in cell adhesion in eucaryotes (reviewed in Ref. 48). Sulfatides immobilized on plastic promote adhesion of several human tumor cell lines (21), and sulfated glycolipids participate in laminin and thrombospondin-mediated cell adhesion (21,49). Interestingly, dextran sulfate which inhibits eucaryotic cell adhesion (48) also inhibits human immunodeficiency virus binding to T lymphocytes (50,51).

sides and that sialylated linear or branched neolacto-series glycolipids containing NeuAca2-3Gal@l-4GlcNAcsequences were possible receptors for adhesion. Other investigators, however, have proposed that only glycoproteins are receptors, since gangliosides from both human and bovine tissues did not bind M. pneumoniae, and did not inhibitadhesion of the organism to humanlung fibroblasts(18,19). Because of these conflicting results, we have reexamined the role of glycolipids and, in the accompanying paper (20), glycoproteins in adhesion of M . pneumoniae. The glycolipid binding specificity of M. pneumoniue was established by the thin layer overlay assay. Of the many glycolipids present on thechromatogram (Table I), the organism bound only to sulfated glycolipids and weakly to lactosylceramide, glucosylceramide, lactotrihexaosylceramide,paragloboside, and a-galactosylparagloboside(Fig. 1). In solidphase binding assays, however, only sulfatide exhibited good a dose response, whereas lactosylceramide and paragloboside bound M. pneumoniae weakly and the other glycolipids not a t all (Fig. 2). Thus, M. pneumoniae differs from many other lung pathogens which bind specifically to glycolipids containing unsubstituted GalNAcB1-4Gal sequences, such as Streptococcus pneumoniae, Pseudomonas aeruginosa, and Haemophilus influenzae (30). Interestingly, M. pneumoniae does not discriminate between galactosyl ceramide 13-sulfate and its unnatural 6-sulfate isomer (Table I), yet the organism does not bind to cholesterol sulfate or to sulfated glucuronosylparagloboside, which has aterminalsulfate linked to the 3position of glucuronic acid. These results indicatethat sulfate alone is not sufficient for M . pneumoniae binding and that at least Gal(3S04)@1-residuesin glycolipids are required. M . pneumoniae did not bind to a2-3-sialylneolacto-series glycolipids either on thin layer chromatograms or adsorbed in cholesterol-phosphatidylcholine onmicrotiterplates. This Acknowledgments-We thank Dr. R. C. Frates, Jr., Department of finding appears to be in variance withreports that these Pediatrics, University of Texas Medical Branch, Galveston, for proglycolipids as well as brain gangliosides, which lack the neo- viding human trachea and lung specimens, Dr. G. w. Jourdian,

9288

Department of Biological Chemistry, University of Michigan, Ann Arbor for yeast phosphomannan, and Dr. C. I. Civin,Division of Pediatric Oncology, The Johns Hopkins University School of Medicine, Baltimore, for My-28 monoclonal antibody. 1. 2. 3. 4. 5. 6.

7. 8. 9.

10. 11. 12. 13. 14. 15. 16.

17. 18. 19. 20. 21. 22. 23.

to Sulfatide


REFERENCES Collier, A. M., and Clyde, W.A. Jr. (1971) Infect. Immun. 3 , 694701 Hu, P. C., Collier, A. M., and Baseman, J. B. (1975) Infect. Immun. 11,701-710 Hu, P. C., Collier, A.M., and Baseman, J. B. (1976) Infect. Immun. 14,217-224 Collier, A. M. (1972) Ciba Found. Symp. 6 , 307-320 Carson, J. L., Collier, A. M., and Hu, S. S. (1980) Infect. Immun. 29,1117-1124 Krause, D. C., and Chen, Y. Y. (1988) Infect. Immun. 56, 20542059 Gabridge, M. G., Taylor-Robinson, D., Davies, H. A., and Dourmashkin, R. R. (1979) Infect Immun. 2 5 , 446-454 Chandler, D. K. F., Collier, A. M., and Barile, M. F. (1982) Infect. Immun. 35,937-942 Gabridge, M. G., Barden-Stahl, Y. D., Polisky, R. B., and Engelhardt, J. A. (1977) Infect. Inmun. 16, 766-772 Manchee, R. J., and Taylor-Robinson, D. (1969) Br. J. Exp. Pathol. 50,66-75 Manchee, R. J., and Taylor-Robinson, D. (1969) J. Bacteriol. 9 8 , 914-919 Banai, M., Razin, S., Bredt, W., and Kahane, I. (1980) Infect. Immun. 30,628-634 Larin, N.M., Saxby, N. V., and Buggey,D. (1969) J. Hyg. (Cambridge) 67,375-385 Sobeslavsky, O., Prescott, B., and Chanock, R. M. (1968) J. Bacteriol. 9 6 , 695-705 Barile, M. F. (1979) in The Mycophmas, Vol. I I (Tully, J. G., and Whitcomb, R. F., eds) pp. 425-464, Academic Press, New York Loomes, L. M., Uemura, K., Childs, R. A., Paulson, J. C., Rogers, G. N., Scudder, P. R., Michalski, J., Hounsell, E. F., TaylorRobinson, D., and Feizi, T. (1984) Nature 307, 560-563 Loomes, L. M., Uemura, K., and Feizi, T. (1985) Ifect. Immun. 47, 15-20 Gabridge, M. G., and Taylor-Robinson, D. (1979) Infect. Immun. 25,455-459 Geary, S. J., and Gabridge, M. G. (1987) Isr. J. Med. Sci. 2 3 , 462-468 Roberts, D. D., Olson. L.D. Barile, M.F., Ginsburg, V., and Krivan, H. C. (1989) J. Bid. Chem. 264,9289-9293 Roberts, D. D., Wewer, U. M., Liotta, L. A., and Ginsburg, V. (1988) Cancer Res. 48,3367-3373 Roberts, D. D., Rao, C. N., Liotta, L. A., Gralnick, H. R., and Ginsburg, V. (1986) J. Biol. Chem. 2 6 1 , 6872-6877 Chou, D. K. H., Ilyas, A. A., Evans, J. E., Costello, C., Quarles,

24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40.

41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52.

R. H., and Jungalwala, F. B. (1986) J. Biol. Chem. 261,1171711725 Martensson, E., (1966) Biochim. Biophys. Acta 1 1 6 , 521-531 Rauvala, H. (1976) J . Bid. Chem. 251,7517-7520 Hanfland. P... Eaae,. H.. Dabrowski, U.. Kuhn, S., Roelche,. D.,. and Dabrowski, J. (1981) Biochemist& 2 0 , 5310-5319 Watanabe, K., Hakomori,, S., Childs, R. A., and Feizi, T. (1979) J. Biol. Chem. 254,3221-3228 Watanabe, K., Powell, M.E., and Hakomori, S. (1979) J. Bid. Chem. 254,8223-8229 Ando, S., Kon, K., Isobe, M., Nagai, Y., and Yamakawa, T. (1976) J. Biochem. (Tokyo)79,625-632 Krivan, H. C., Roberts, D. D., and Ginsburg, V. (1988) Proc. Natl. Acad. Sci. U.S. A. 8 5 , 6157-6161 Spitalnik, S. L., Schwartz, J. F., Magnani, J. L., Roberts, D. D., Spitalnik, P. F., Civin, C. I., and Ginsburg, V. (1985) Blood 66, 319-326 Kean, E. L. (1968) J. Lipid Res. 9 , 319-327 Tadano-Aritomi, K., and Ishizuka, I. (1983) J. Lipid Res. 2 4 , 1368-1375 Svennerholm, L., and Fredman, P. (1980) Biochim. Biophys. Acta 617,97-109 Iozzo, R. V. (1984) J. Cell Bwl. 99,403-417 Iozzo, R. V. (1987) J. Biol. Chem. 2 6 2 , 1888-1900 Humphries, D. E., Silbert, C.K., and Silbert, J. E. (1988) Biochem. J. 2 5 2 , 305-308 Roberts, D.D., Williams, S. B., Gralnick, H. R., and Ginsburg, V. (1986) J . Biol. Chem. 2 6 1 , 3306-3309 Krivan, H. C., Ginsburg, V., and Roberts, D. D. (1988) Arch. Biochem. Biophys. 2 6 0 , 493-496 Roberts, D.D., Liotta, L.A., and Ginsburg, V. (1986) Arch. Biochem. Biophys. 250,498-504 Powell, D. A,, Hu, P. C., Wilson, M., Collier, A. M., and Baseman, J. B. (1976) Infect. Immun. 13,959-966 Gorski, F., and Bredt, W. (1977) FEBS Lett. 1 , 265-267 Banai. M.. Kahane. I.. Razin., S.., and Bredt, W. (1978). Infect. . Immun. '21, 365-372 Razin, S., Kahane, I., Banai, M., and Bredt, W. (1981) Ciba Found. Symp. 80,98-118 Izumikawa, K., Chandler, D. K. F., and Barile, M. F. (1986) Proc. SOC. EXP. BWl. 181, 507-511 Chandler, D. K. F., Grabowski, M. W., and Barile, M. F. (1982) Infect. Immun. 38,598-603 Farooqui, A. A., and Horrocks, L. A. (1985) Mol. Cell. Biochem. 66,87-95 Roberts, D. D., and Ginsburg, V. (1988) Arch. Biochem. Biophys. (1989) 267, 405-415 Roberts, D. D., Shenvood, J. A., and Ginsburg, V. (1986) J . Cell Bid. 104, 131-139 Ueno. R.. and Kuno. S. (1987) Lancet 2,796-797 Mitsuya,'H., Looney, D.'J., Kuno, S., Ueno, R., Wong-Stahl, F., and Broder, S. (1988) Science 240,646-649 IUPAC-IUB Joint Commission on Biochemical Nomenclature (1986) Eur J. Biochem. 159, 1-6 "