Binding of Pseudomonas cepacia to normal human intestinal mucin and respiratory mucin from patients with cystic fibrosis

Binding of Pseudomonas cepacia to normal

human intestinal mucin and respiratory mucin

from patients with cystic fibrosis.

U S Sajjan, … , M A Karmali, J F Forstner

J Clin Invest. 1992;89(2):648-656. https://doi.org/10.1172/JCI115631.

Although not as prevalent as Pseudomonas aeruginosa, Pseudomonas cepacia is another opportunistic pathogen which colonizes the lungs of at least some patients with cystic fibrosis. A subgroup of these patients exhibits the "cepacia syndrome", i.e., a rapid clinical deterioration and death within one year. To investigate potential early sites of bacterial attachment, we have measured the specific binding of P. cepacia isolates from cystic fibrosis (CF) sputa to both CF and non-CF mucins purified from respiratory and intestinal secretions, respectively. As shown in microtiter binding assays, clinical isolates from 19/22 patients were found to bind to both mucins, with the highest specific binding exhibited by isolates from eight patients, seven of whom later died with the cepacia syndrome. No differences were observed in the binding capacity of the two (CF versus non-CF) mucins. Binding was specific, saturable, and not influenced by tetramethylurea, a disruptor of hydrophobic associations. Individual sugars were ineffective as hapten inhibitors, as were several lectins. Mucins treated by reduction/alkylation or chloroform/methanol extraction showed enhanced bacterial binding, findings which were attributed to exposure of

underlying binding sites. Deglycosylation procedures indicated that mucin receptors for P. cepacia include N-acetylglucosamine and N-acetylgalactosamine, probably linked together as part of core oligosaccharide structures. P. cepacia isolates also bound to buccal

epithelial cells, and mucin partially inhibited the binding of those isolates of […]

Research Article

Find the latest version:

Binding of Pseudomonas cepacia

to

Normal

Human Intestinal Mucin

and Respiratory Mucin from Patients with Cystic Fibrosis

UmadeviS.Sajjan,MaryCorey,Mohamed A.Karmali, and JanetF.Forstner

Research Institute,HospitalforSickChildren, Universityof Toronto, Toronto, Ontario, M5G 1X8, Canada

Abstract

Although not as prevalent as Pseudomonasaeruginosa, Pseu-domonascepaciais another opportunistic pathogen which col-onizes the lungsofatleastsomepatients with cystic fibrosis.A

subgroup of these patients exhibits the "cepacia syndrome", i.e.,arapid clinical deteriorationanddeath within one year. To

investigate potential early sites of bacterial attachment, we havemeasuredthespecific binding ofP.cepacia isolatesfrom

cysticfibrosis(CF)sputa tobothCF andnon-CFmucins

puri-fied from respiratoryandintestinal secretions, respectively. As shown in microtiterbindingassays,clinical isolates from19/22

patients werefound tobindto both mucins, with the highest

specificbinding exhibited by isolates from eight patients,seven

of whomlaterdied withthecepacia syndrome.Nodifferences

wereobserved in the binding capacityof the two (CF versus

non-CF)mucins. Bindingwas specific, saturable, and not in-fluenced by tetramethylurea,adisruptor of hydrophobic associ-ations. Individualsugars wereineffectiveashapteninhibitors,

as wereseverallectins. Mucins treated byreduction/alkylation

orchloroform/methanolextraction showed enhancedbacterial binding, findings whichwereattributedtoexposureof underly-ing binding sites. Deglycosylation procedures indicated that mucin receptors for P. cepacia include N-acetylglucosamine andN-acetylgalactosamine,probably linked togetheraspartof

coreoligosaccharidestructures. P.cepaciaisolates also bound

tobuccalepithelial cells,andmucinpartiallyinhibited the bind-ing of those isolates ofP.cepaciathatalso had the abilityto

bindtomucin. We speculate thatspecific bindingof P. cepacia

tosecretedmucinsmaybeanearlystepin thepathogenesisof thecepaciasyndrome.(J.Clin.Invest.1992.89:648-656.)Key words: mucin adhesion*carbohydratereceptors *buccal

epithe-lial cells *bindingassays *mucin competition

Introduction

Pseudomonascepacia, awell knownplantpathogen,wasfirst describedasanopportunistic pathogenincysticfibrosis

(CF)'

Preliminaryreportsofthiswork haveappearedinabstract form(1990. Pediatr. Pulmonology. Suppl. 5:196A; and 1991. Glyconjugate J. 8:209).

Address correspondence to J. F. Forstner, M. D., Ph.D., The Hospi-talfor Sick Children, 555 University Avenue, Room 3423, Toronto, OntarioM5G 1X8, Canada.

Receivedforpublication7 June 1991and in revisedform 11 October 1991.

1.Abbreviationsused in this paper: BEC, buccal epithelial cells; CF, cystic fibrosis;dpm,disintegrations per minute; FEV1,forced

expira-patients in the early 1970s (1). Unlike Pseudomonas

aerugi-nosa,thisorganisminfectsasmall percentage of CFpatients.

Over the last 20years, however, theprevalence of P. cepacia infection in CF has graduallyincreased in some centers(2, 3).

A recent survey at the Hospital for Sick Children, Toronto, suggested that approximately 40% ofatotal of 500 CFpatients

are nowcolonized with this organism.Theinfection isbelieved

tospread bynosocomialtransmission(4, 5)butonceacquired,

the clinical course is quite variable. Some patients show no

particular alterationin thecourseoftheirdisease,someshowa

mildly accelerated decline, while others succumb to arapid declineinpulmonaryfunction withfever,elevationof

erythro-cytesedimentationrate,leukocytosis, and resistanceto

antibi-otictherapy, culminatingindeath withinseveral monthsto a year.Thisrapidfatal deterioration has been called the "cepacia syndrome"(2,6). Very little is known about thepathogenesis

of thissyndromeorthe virulence factorsresponsible,although

severalpotential virulence factors have been described forP.

cepacia, including proteolytic, hemolytic, lipolytic, and

sac-charolytic activities (7, 8). Unfortunately,evidencethat

pulmo-narypathology isadirectresultofthese factorsisnonexistent

or atbestveryweak.

Kuehnetal.(9) reportedthat both CF and non-CF P.

cepa-cia isolates bind to respiratory epithelial cells, and suggested

thatpolarpilion P. cepacia mayplaya role inadhesion and

initiation of infection.In contrast,Saimanetal.(10) suggested

thatoutermembraneproteinsofP. cepacia mediateadhesion

torespiratorycells. Krivanetal.(I1)havedemonstrated that selectedCF and non-CFP.cepacia andP.aeruginosa isolates

also bind invitrotoasialoglycolipids,with thespecificreceptor sequencein eachcasebeing N-acetylgalactosamine (GalNAc)(

1,4galactose (GAL).

Autopsyspecimens ofCFlung (12)and ferretlung (usedin

aCF model system) (13),haveshown thatrespiratory

patho-gens ingeneral tend to localize withinluminalmucus

secre-tionsrather than adheretocells. Toourknowledgenoreports

ofP.cepacialocalization in CF lungs have beenpublished.

In anefforttoexplorepotentially important earlyeventsof

P. cepacia infection in CFpatients at the Hospital for Sick ChildreninToronto,weundertookastudyofthebindingofP. cepacia isolatestopurifiedhumanmucins. Ourgoalswere to

discover whether therewas specific binding ofP. cepacia to

mucins, whether binding wasdifferent for non-CF intestinal mucinthanfor CF tracheobronchial mucin, and whetherwe

could obtain information about thespecificreceptorstructures on mucins involved in P. cepacia adhesion. In addition we

explored thebindingofP.cepacia isolatestobuccalepithelial

cellsand theabilityof mucinstointerfere with thisbinding.

tory volumein 1 _s; Gal, galactose; GalNAc, N-acetylgalactosamine;

GlcNAc, N-acetylglucosamine; SIM,smallintestinalmucin; TBM,

tra-cheobronchialmucin;TFMS,trifluoromethanesulfonicacid.

J.Clin.Invest.

©TheAmericanSocietyforClinical Investigation, Inc. 0021-9738/92/02/0648/09 $2.00

Methods

Bacterial

strains

and culture conditions

Fromsputumcultures of 22patientswithcystic fibrosis,whowere seen

in theout-patientclinicorwhowerehospitalizedfor briefperiods

dur-ing 1986to 1989, 53 clinical isolates ofP. cepaciawere identified.

Although theseisolatesweresent tothe Center for DiseaseControl, Atlanta, Georgia,forribotype analyses,the resultsarenotyet available.

Therefore precisestrain differences(if any)amongst the isolatesare

unknown. Identification ofP. cepaciainthe Clinical Microbiology Laboratorywasmadeonthe basis of identification schema outlinedby

Gilardi(14), includingcolonialmorphologies, odor,and

Gram-nega-tivestainingof motilerod-shapedbacteria.Inaddition,isolateswere

catalasepositive, weaklyoxidasepositive,didnotfermentglucose,gave

good growthonMcConkey medium,wereresistant topolymyxinB

sulphate (10 mg/liter),gavepositive glucoseoxidation reaction(Hugh

and Leifson medium) andpositive growthinSimmon's citrate

me-dium. Results ofspecific tests used todistinguish P. cepaciafrom

closelyrelatedspeciesinPseudomonas RNAgroupII included

nega-tivearginine dihydrolase activity, positive hydrolysisof O-nitrophenyl-fl-D-galactopyranoside,andpositiveesculinhydrolysis (formostof the

isolates).Each culturewassuspendedinsterilewater,transferred onto

brain heart infusionagarplates,grownfor 36-48 h at370C,andthen

maintainedat-70'Cinvialscontaining5% trisodium citrate in 40%

glycerol.

In preparationforbinding studies,bacteria weresubculturedon

brain heartinfusionagarplates,grownfor 36-48h,andsinglecolonies

inoculated into 10mloftrypticasesoybroth(Difco Laboratories, Inc.,

Detroit, MI) supplementedwith 0.1mCiof[3H]sodiumacetate(ICN

Biomedicals Inc., CostaMesa, CA) (sp act 2.5 Ci/mmol)or

[35SJL-methionine(New EnglandNuclear ResearchProducts, Boston, MA) (spact800Ci/mmol).Bacteriaweregrownunder aerobic conditionsin

anOrbit shaker(CanLab, Mississauga, Ontario, Canada)at 150rpm

for 16-18h at37°C.The brothwascentrifugedat 10,000g for10min at4°C,and the bacterialpelletwashed three times with 20 vol of PBS

pH 7.0 containing0.1% BSA (fraction V, essentially globulin free; SigmaChemicalCo.,St.Louis, MO).Thepelletwas_resuspendedinthe

samePBS-BSAsolution togiveaconcentration suitableforsubsequent bindingassays(0.1ODat650nmwasequalto 2.52x

IO'

CFU/mlas

determinedby plating seriallydilutedbacteria).

Patient

data

Selected information on each patient from whom isolateswere ob-tainedwasenteredroutinelyinto the CFpatientdataregistryat the

Hospitalfor Sick Children. On the basis of severalclinicalparameters,

including pulmonaryfunctiontests(FEV1,forced vitalcapacity,chest

xray,andweightandheightforage,patientswereplacedintooneof

threegroupsatthe timeeach isolatewasobtained:i.e., advanced, mod-erate, ormild severity ofillness. Followup mortalitydataon each

patientwere maintainedaspartof theongoingdataregistryatthis

hospital.

Purification of

mucins

Tracheobronchial mucin(TBM)fromcombinedsputaofCF patients,

and small intestinal mucin (SIM) from ileal secretions ofa

non-CFpatientwerepurified,andcompositionalassaysforcarbohydrate

and aminoacidswere performed,asdescribedintheaccompanying

paper(15).

Modifications ofintestinal mucinstructure

Reduction andalkylation.Mucin(377 ,ugasproteinor3.7mgasdry

wt)wasmixedwithDTT(10 mM)in0.05 MTrisHCl buffer pH 8.6 containing1%SDS,boiled for 5min,cooledtoroomtemperature,and

alkylatedwithiodoacetamide(20 mM)for1hatroomtemperaturein

the dark. Excess reagentswereremovedby dialysisfor 24 hagainstsix

changesofdistilledH20.

Delipidation.Fivesequentialchloroform:methanolextractions of

mucin(20mgdry wt)werecarried outbythemethodofSlomianyet al.

(16) to remove potential lipid contaminants. Extracts were pooled, concentrated by rotary flash evaporation, and redissolved in 100 y1 of chloroform:methanol (2:1 vol/vol) for subsequent use in thin layer chromatography. Theresidual mucin was used in bacterial binding or hapteninhibition assays.

Desulfation. Mucin (3.77 mg) was treated with 4 ml of 0.05 MHCl in drymethanol at370Cfor4h asdescribed bySlomiany et al. (17), neutralized with 1 M NaOH,dialyzed againstthreechanges of distilled H20,and concentrated byfreezedrying.Mucin sulfatewasmeasured by the method of Silvestrietal.(18) before and after methanolic-HCl

treatment.

Deglycosylation.(a)Partialdeglycosylation was accomplished by subjectingmucin(27 mg)to twosequentialtreatmentswith trifluoro-methanesulfonic acid (TFMS):anisole (2:1) for 2 hat00Casdescribed by Edgeetal.( 19).

(b)Totaldeglycosylation wasachieved bysubjectingthepartially deglycosylated mucinto"reverse"f,-eliminationasoutlinedby Ger-ken etal. (20).Briefly, this entailedtheadditionof0.1 Maceticacidtoa

solution ofpartly deglycosylated mucin(6.4mg in 1 mlof distilled H20)tobringthe pH to 4.5,incubationwith sodiummetaperiodate (200 mM) at 4°C overnight in the dark,destruction of excess periodate by the addition of sodiumiodide/sodium thiosulphatereagent(1:1vol/

vol200 mMNal and 800 mM Na2S203), and incubation for 12 hat

4°C.Thedeglycosylated mucinwasthendialyzedat4°C against dis-tilledH20for 3 d and freeze dried.

Proteolytic digestion.Mucin (3.7 mg)wasincubated with 754_tLgof protease(typeXlV,Sigma Chemical Co.) in PBS pH 7.2 for 24 hat

37°Casdescribed earlier(21). PMSF(2mM)wasthen added and the mixture dialyzed for 24 hagainst distilledH20.No attempt wasmade

to removeresidualpronasefromtheretentate.

Periodate treatment. Sodium metaperiodate (0-100 mM) was addedtomucin-coatedmicrotiterwells andincubated for1 h at 37°C. After washingthe wells, standardbacterial binding assays were per-formed. Moredetails ofthe procedureare provided in Results (Ta-ble IV).

Bacterial

binding to mucin

Microtiterplateassays.Binding ofP. cepaciatoeither SIMorTBM wasperformedin wells ofpolyvinyl microdilutionplatesasdescribed

intheaccompanying paper (15) with a minormodification, namely, thatthroughout the assay PBScontaining 0.1%BSAwasusedinstead of PBS alone.This modification didnotaffectspecific binding,it sim-ply reducedbackgroundbindingtoBSA.Specific bindingtomucinwas

calculated insomeexperiments by subtractingthedisintegrationsper minute (dpm)ofBSA-bound bacteria from thedpmofmucin-bound bacteria,and theresultingvalue then convertedtoCFU/,ggmucin. In otherexperiments specificityofbindingtomucinwasexpressedas a

bindingratio R, which is the dpm boundtomucindividedby dpm boundtoBSA. In hapteninhibition experimentsthe R valuefor bind-ing of controls (no hapten) was used as a reference (0% inhibition) to facilitatecomparisonswith samples exposedtohaptens.

Slot-blotbindingassays.Native or modified mucins (50-500 ng as protein) were slot-blotted onto _0.45-,gm nitrocellulose membranes (Bio-Rad Laboratories, Richmond, CA) in a slot-blot system manifold II(Schleicherand Schuell Inc., Keene, NH), washed, and blocked with 3%gelatinfor 1 hat37°C. The membrane was then incubated with

"S-labelledP.cepacia(2-3 x 109CFU/ml) in PBS-BSA pH 7.0 for 1h at 37°C. Afterwashing,themembranesweredried,incubated for 5 min inrapidautofluor enhancer (New England Nuclear,Boston,MA)

andexposedtoX-OMATARfilm (EastmanKodak,Rochester,NY) for 4-5 d.Intensityof slotswasjudged visuallyandbyscanning densi-tometryby theuseofalaser(UltrascanXL; PharmaciaLKB,Uppsala, Sweden).

Thin layer chromatography overlayassay. Chloroform:methanol

extractsof mucin(10gl)aswellaspurifiedreferenceglycolipids (1-2

Aig)wereseparated bythin layer chromatographyonsilicagelusing

Theglycolipidstandards were kindly supplied by Dr. C. Lingwood, Division ofMicrobiology, Hospital for Sick Children, and included asialoandmonosialo gangliotetraosyl ceramides, asialoandmonosialo gangliotriosylceramides, globotetraosyl-, globotriosyl-,and lactosyl cer-amides(asialo GMl, GMl, asialo GM2, GM2, Gb4, Gb3, and LC, respectively).Thesilicaplateswereeither stainedwith Orcinol reagent (SigmaChemicalCo.,St. Louis, MO) (22)andvisualized, orblocked with1% gelatin overnightat370C,incubatedwith 10 ml of35S-labelled P. cepacia (3 X 109 CFU/ml) for 1 h at 370C,washedsix timeswith PBS-BSA, airdried,andprocessedforautoradiography.The method wasessentiallythe same asthatdescribed by Hansson et al. (23).

Hapten inhibitionassay. Potential haptens were preincubated with 3H-labelled bacteria for1 h at37°C, followedbymeasurement of bacte-rialbindingtomucinby themicrotiterbinding assay protocol. Further details forspecificexperimentsaregiveninfigure legendsortables in Results.Individualcarbohydrates used as haptens were obtained from

SigmaChemical Co. Lectins (Helixpomatia, wheat germ agglutinin, peanutagglutinin,andUlexeuropaeusI)werepurchasedfrom Biocarb Chemicals, Lund,Sweden.

Buccalepithelialbindingassay. Binding of3H-P.cepacia tobuccal

epithelialcells(BEC)wasdetermined byfollowingthe methodof Doig

etal.(24). To testtheeffect ofmucin on P. cepacia binding to BEC, 3H-bacteriawerefirst incubatedwith mucin (25-200

Ag/ml)

for1hat

37°C,the mixturewasthen incubated with BEC,and the assay was continuedasusual.

Statistical

analysis

Statistical comparisons ofbinding results, Rvalues, andpulmonary functiontests(FEV1),wereperformed byStudent's t testusing Stat-works, Cricket Software, Philadelphia,PA.

LinearregressionsofLangmuirisotherm plots were computed by Cricket graph,(CricketSoftware)and theslopes foreachisolatewere thencompared witheach otherby multiplelinearregression analysis

usingSASStatisticalSoftware, Cary,NC.

Results

Binding

ofP.

cepacia

to

intestinal

orrespiratory mucin Initialbinding experiments involvedtheincubation ofafixed concentration (1.0-1.5 X 109 CFU/ml) of3H-labelled P.

cepa-ciafor 1 hat37°Cinmicrotiterwells coatedwith either

puri-fied normal human SIMorCF TBM.Radioactivity (dpm) of

bacteria boundtomucinwasdividedby dpmofbacteria bound

toBSA-coated wellstogeneratea

ratio,

R,whichreflects

speci-ficity

ofbindingtomucin.Rvalues>2.0weretakento

indi-cate a

preference

of

binding

for mucin (25). Incontrastto P.

aeruginosastrains(15)sevenof the

eight

isolates ofP.cepacia exhibited

preferential

bindingtomucin.Nomajor differences

in

binding

between the normaland CF mucinswereobserved

(Table I).

Using

twoisolatesthat showed thegreatestmucinbinding

values(PC 5 and 7),experimentswereconductedtodiscover theoptimum conditions of time(0.5-3 h), temperature

(22-37°C),and pH(5.8-7.5),forbinding.The results (not shown) indicated that 1 h at 37°C and pH 7.0gavethe bestbinding, and theseconditionsweretherefore used in subsequentbinding

orbinding/inhibition studies. Forthreeisolates thatwere

ex-ploredindetail (PC 5, 7, 24),specificbinding dependedupon

bacterial concentration and exhibited saturationat a

concen-tration of- _{4-5 x 109 CFU/ml. Fig. 1ashowsarepresentative}

plot forone isolate (PC 7)andFig. 1 bthederivedLangmuir isotherm, whichgave a linearregression line suggestive ofa

singletype ofnoninteracting mucin receptors. Fig.2shows that thebindingofPC7 tointestinal mucinwas80-90%inhibitable

TableLBinding ofP. cepacia Isolateto Mucins

dpm bound dpm bound R value

Isolates Mucin toBSA(a) tomucin(b) (b/a)

PC 1 SIM 70(69-73) 218 (170-254) 3.1

TBM 70 (69-73) 238 (197-273) 3.4

PC 2 SIM 785 (714-872) 2849 (2780-2902) 3.6 TBM 785 (714-872) 2506(2234-2644) 3.2

PC 3 SIM 71(61-78) 285 (264-297) 4.0

TBM 71(61-78) 256 (247-275) 3.6

PC 4 SIM 41(21-61) 169(133-194) 4.1

TBM 41(21-61) 167(152-185) 4.1

PC 5 SIM 96(94-99) 801(770-850) 8.3

TBM 96(94-99) 833 (805-886) 8.6

PC7 SIM 58(52-63) 2820(2515-3035) 48.3

TBM 70(66-75) 3515 (3116-3710) 50.2

PC18* SIM 35 (29-42) 43 (35-51) 1.2

PC 24* SIM 27(22-36) 640(557-703) 23.7

3H-P.cepacia isolates (1.0-1.5X 109CFU/ml) were tested for binding

tomicrotiterwells coatedwithBSA(control), SIM, or TBM. R values

>2.0 wereconsidered toreflect specific bindingto mucin. Values of dpmarethe average (± total rangeobtained) ofthreeindependent bindingexperiments. * BindingtoTBMwas notdetermined.

by preincubatingthe bacteria witheither SIM orTBM solu-tions before carrying outstandard microtiter bindingassays.

Forthe threeisolates, Langmuir isotherms permitted

calcula-tion of values for N(saturation of mucin sites expressed as

CFUf/,g

mucin)andK(associationconstantexpressedas

ml/

tg

mucin)(Table II).Multiplelinearregression analyses of the

isothermsindicated thatdifferencesinbindingvaluesamongst

the isolateswerestatistically significant. However,the Kand N

values wereall within one order ofmagnitude, which likely

reflectssimilarity ofbacterial adhesins andcommon,but

per-haps not identical, mucin receptors. In later experiments, whichweredesignedtoprobe the identityof mucinreceptors, wefrequently used both isolates PC 5 and 7(and in selected

casesboth mucins)to ensure thatour findings wouldnotbe

restrictedto onestrain-specificmechanismof mucin binding. Hydrophobic interactions,which playanimportantrolein

P.aeruginosa bindingtomucin or BSA(15),were notfound to

influenceP. cepaciabindingtomucin,asjudged bythetotal lackof effectoftetramethylurea(0.05-0.4 M)uponthe

bind-ing of eitherPC 5orPC7 tomucins (notshown). This result in

additiontotheunequivocal specificityofP.cepacia bindingto

mucin(versus BSA),madeitunnecessary to resort to

solution-phase assaysas wasdoneearlier forP.aeruginosa ( 15). Survey

of

clinical

isolates

From 22patientswithcysticfibrosis whovisited theoutpatient

clinicor werehospitalized during 1986to1989,therewere53 P.cepacia isolates obtained. The clinicaldatarelevanttoeach

patientatthetime the isolatewasobtained and thesubsequent

clinicalcourseof eachpatient (up untilDecember 31, 1990)

weredocumentedandstoredin theCF PatientDataRegistryat

theHospitalfor Sick Children.Eachisolate(atafixed bacterial concentration of 1.2x I09CFU/ml)wastested for itscapacity

5r

a

0

x

Im

4

3

2

1

1 2 3

U

xlO 9

4 5

90r

b

80

70 60

50

40

30

0 1 2 3 4 5

U

x10

l

9

Figure 1. (a)Binding ofP. cepacia to tracheobronchialmucin.An

aliquot of150

_;A

of3H-P.cepacia(PC7)(2X 108-4X 109 CFU/ml) was added tomicrotiterwells coated with tracheobronchial mucin

orBSA (0.52Agprotein/well), incubated for 1 h at37°C,andafter washing four times to remove nonbound bacteria, wells were counted for boundradioactivity. Mucinbindingvalues were correctedfor nonspecific BSA binding. Each value represents the mean (±SEM) of sixindependent bindingexperiments. U, unbound bacteria(CFU/ml x 109); B,boundbacteria(CFU/,ugmucin dry wt).(b) Langmuir iso-therm plotof the data of Fig. 1 a.

study period. Therefore a total of 19 out of the 22 patients produced38isolateswhichdid bindtomucin,andthese were segregated and grouped accordingtothe severity of illnessof thepatientatthe time ofP. cepaciaisolation(TableIII).

R values varied within each group, but the average was highest in the isolates of those patients judgedto haveadvanced

disease(group a).Sixof the eight patients in groupadiedwith severeillness within oneyear(three withinonemonth) ofthe time the last isolate was obtained. Three of these six had R

values> 20.0,whilethe otherthreehadvaluesof1 1.8, 5.7,and 2.5, respectively. Therewere twoadditional patientswith ad-vanceddisease whoseRvalues are notincluded in Table III because none of their isolates boundto mucin. Oneofthose patients died with severe illness, but the lastisolate available from that patient was obtained10months before death,and we

therefore donot knowwhetherthe Rvaluemight have become positive had it beentestedduring that time. Theother patient with advanced disease(but R < 2.0) is still alivedespitealow FEV1(i.e.,22% ofpredicted). Despitethetestingofmanyofhis

isolates,nomucin-binding activityhas ever beenobserved.In

summary, therefore, most patients with P. cepacia and

ad-80 ;; j

z b

20

0 20 40 60 80 100 120 140 160

MUCIN_(jigofprotein/ml)

Figure2. Inhibition of 3H-P. cepacia bindingtoimmobilized intes-tinal mucinby preincubationwith intestinalortracheobronchial mucin.3H-labellel P.cepacia (2.55 X 108CFU/ml)(isolatePC7)

were_preincubateawith intestinal(SIM)orCF tracheobronchial

mu-cin_(TBM)0-150,ug/mlfor 1 hat370CinPBS,and then 150Mlof each mixturewereaddedtomicrotiterwellspreviouslycoated with intestinal mucinorBSA. Standard_bindingassayswereperformed

asdescribedinMethods and corrections made fornonspecificBSA

binding.Inhibitionwascalculatedasa_percentageof control(0%

in-hibition)containingnomucin inpreincubationmixtures. Each value is themeanofthree_{independent binding}assays and barsrepresent the total range of values obtained. o,TBM;*, SIM.

vanceddiseasewerecolonizedwithmucin-binding

organisms,

butatleastone

patient

carriedanon-mucin-bindingstrain and is still

alive,

andonewithalow R value(< 2.0)died.

In sixmoderatelyillpatients (group b),theaverageR value

wasloweralthoughnotstatisticallydifferentfromgroupa

_(P

=

0.077).

There was onedeath in group b, and it occurred

within threemonthsofthelast isolate obtained. The R value

was23.7andthe FEV1 was63,butwithinaveryshortperiod

thereafter the clinical condition of the patient deteriorated

sharply. Nofurther isolatesorFEV1 measurements were ob-tained

during

the three monthsbefore death.Ifweassumethat

this

_patient

wasincorrectlyclassifiedclinicallyas

"moderate",

and add his R value to the data for group a, the difference betweengroupaand bbecomes statistically significant (i.e.,P

=

0.024).

Oneothermoderatelyillpatientwascolonizedwith

TableII. ValuesofNandKforP.cepacia BindingtoMucins

Isolate Mucin N K

CFU/Ilgmucin ml/lgmucin

PC 5 SIM 1.46(+0.1 1)X 107 1.34(±0.19)x 10-9 PC 24 SIM 3.01(±0.16)x 107 3.29 (±0.23) x 10-10 PC 7 SIM 6.08(±0.98)x 107 3.33 (±0.70) x 10-1° PC 7 TBM 1.66(±0.29)x 108 1.23 (±0.32) x 10-'

Langmuirisothermswereconstructed from the results of binding

ex-perimentsof 3H-labelled P. cepacia(1 x 107-5 x 109CFU/ml)to

immobilized SIMorTBM in microtiter wells. For each concentration of bacteria sixbinding experimentswerecarriedoutper isolate and thebindingtomucin corrected for nonspecific binding to BSA. The average number ofbindingsitesonthemucin,N(±SEM),andthe association constant, K(±SEM),werederived from the intercept('/ KN)andslope('IN),respectively. Figures in brackets represent SEM.

Table III. Correlation of P. cepacia Binding to Mucin and Clinical Severity ofDisease

Number of Number of Binding

Severity of positive patients parameter

disease isolates who died R±SEM FEV1±SEM

(a) Advanced 13 6 15.98±3.49 33.6±2.19

(b) Moderate 12 1 8.49±1.84 64.7±3.27

(c) Mild 13 0 6.85±0.82 82.0±3.19

P. cepacia isolates were labelledwith3H-acetate,concentrated to 1.2 X 109 CFU/ml;and testedfor their binding to intestinal or respiratory mucins inmicrotiter bindingassays. Specific binding to mucin, R refersto aratioof mucin bindingdivided by BSA binding, and was determined fromthreeindependent bindingassays.Patients who died didsowithinthe three year studyperiod(1986 to 1989). Statis-ticaldifferencesweresignificant for comparisons ofR valuesof group a versus c(P=0.0 18),but notfora versus b(P=0.077) or b versus c(P=0.41).If datafromoneclinicallyanomalouspatient(R

=23.7) in group b (see text) are placed in group a the P valuefora versus b would be<0.05. Comparisons ofFEV1 values gave P =

<0.001 foraversusb,a versusc, and b versus c.

P. cepaciawhich did not bind to mucin, and this patient

re-mains clinicallyunchangedfouryears later.

Themild group (c) contained nine patients with isolates which bound to mucin. The average R value was significantly lower than that ofgroup a (P= 0.018) but not group b (P

=0.41).There were two otherpatientsinthe mild group whose isolates did not bind to mucin, and those patientswere still

aliveasofDecember31, 1990.

Mucin receptors

for

P. cepacia

Todiscoverthe natureofthe receptorsitesonmucins which

areresponsible for the adhesion ofmucin-binding isolates,a number ofexperimental approaches were taken. These in-cluded the useofselectedhapten inhibitors ormucins of al-teredcomposition orstructure. In mostexperiments purified

smallintestinal mucinwasused rather than tracheobronchial

mucindue to greateravailability oftheformer, and most

ex-perimentswereperformed withisolates PC 5and/orPC 7.

Treatment with sodium

metaperiodate

MucinandBSA-coated wells were incubated in the darkfor1 h

at37°Cwithincreasing concentrations of NaIO4,washedfour

times withPBS,and the residual boundproteinmeasuredby

the methodofSorensonetal. (26).No loss ofboundprotein (0.52 ,g protein)occurredfor either BSAormucin (not shown)

asresult ofperiodate treatment. IsolatesPC5and PC 7were then added andthe usualbinding assayperformed.Asshown in Table IV for SIM, periodate treatment caused a gradual decreaseinPC 5 and PC 7binding to mucin, which suggests thatmucin carbohydratesareinvolvedasreceptors. No

signifi-cance couldbeattached to thepercentagedifferences between thetwo isolates since the concentration of bacteria added in eachcase wasdifferent.

Treatment

with

carbohydrates or lectins as inhibitors

Hapten inhibition studies were carried out by preincubating PC5 orPC7(1 x I09 CFU/ml)for1 hat 37°C with individual monosaccharides which included D-fucose, D-glucose,

D-galac-Table I V.Effect ofSodium Metaperiodate on P. cepacia Binding to Intestinal Mucin

Binding (percent of control)

NaIO4 PC 5 PC 7

mM

0 100 100

6.2 72(68-75) 63(59-70)

12.5 30(20-36) 49 (47-53)

25.0 20(16-25) 41 (4046)

50.0 5(4.1-5.7) 27(23-30)

100.0 3(2.8-3.6) 13(12.7-13.4)

NaIO4(0-100mM) was added tomucin-coated microtiterwells and incubatedfor 1 hat 37°C in the dark. The wells were then washed four timeswithPBS-BSA.3H-labelled bacteria(1.34 x I09 CFU/ml for PC5 and 5X I0O CFU/ml forPC7)wereaddedintriplicateto wellsandbindingmeasured asdescribedin Methods.Similarly treated BSA-coated wellswereused to correctfornonspecific binding. Control(noperiodate)bindingdatawerenormalizedto100%.Values represent the mean and total rangeofvaluesobtained(brackets).

tose, D-mannose, N-acetylglucosamine, N-acetylgalactosa-mine, andN-acetylneuraminicacid (20 mM). The mixture of

3H-bacteriaandmonosaccharide was then added to mucin- (or

BSA-)coated wells andbindingmeasured in the usual fashion. Noneofthe sugars inhibitedbacterialbinding to mucin. Simi-larly there was no inhibition by the disaccharides Gal

_(l

1,3

GalNAc,Gal 3 1,6 GlcNAc,Gal

f3

1,4 GlcNAc,orGlcNAc

_f3

1,4 GlcNAc used at concentrationsof2.5 and5.2 mM (not shown).

Inseparateexperimentsthefollowing lectins(5-25 ,ug/ml) were individually incubated for 1 h at 37°C in mucin- (or BSA-) coated wells:helixpomatia,peanutagglutinin, Ulex Eu-ropaeus I, and wheat germ agglutinin. After four washes to removenonboundlectins,3H-P. cepacia(PC5 or PC7)were

added andbindingassays carriedout.Onceagainnoinhibition

wasobserved (notshown),suggestingthat thesimple

monosac-charidesor disaccharides suchasGalNAc,Gal ,B1,3 GalNAc, fucose,and,B-linked GlcNAc, respectively,didnotprovide spe-cific mucin receptor sitesforthe bacteria.

Chloroform:methanol

extracts

of

mucin

Since Krivanetal.(11)havedemonstratedbinding ofselected

strains ofP. cepacia toasialoglycolipidsGM1 andGM2, we

wishedtoconfirm thisfinding fortheisolatesPC 5 andPC7.In

addition,sequentialchloroform:methanol extractions of intes-tinal mucin(20 mg)wereperformedtodiscoverwhether the

mucinpreparationcontained anylipid contaminants capable

ofbindingP.cepacia.Glycolipidstandards(1-2 ug)aswellas 10Ml ofmucinextract(fromatotalof 100Ml)wereseparated by

thin layerchromatographyonsilicagelandsprayedwith Or-cinol reagent.Aminor band of free sugars(?)wasdetectednear

the solventfront in the mucintract(Fig.3a) butno

recogniz-ableglycolipidcomponentswerefound. Sincethesensitivityof detectionwas - 0.5

Mg

ofglycolipid,themucinglycolipid

con-tamination was judged to be negligible (i.e., <0.05% by

weight).35S-labelledP.cepaciawereoverlayedontoduplicate (Orcinol-free)TLCplates,washed,andsubjectedto

con-a -;, bindingtomucin bypeanutagglutininorby Gal( 1,3GalNAc disaccharidemadeit unlikely thatcoreGal

#

1,3 GalNAc units of mucinweresufficient to account for the observed bacterial

bindingtomucin.

.

.,, :.. *.

..<22_....4e

.: :

.s '' ...., @

S

1 2 3 4 5 6

b

K'-

'77

8xAa~jS

-

'

,.:~~~~~~~~~~~~~~~~~~~~~~~~~~~7

r.~~-,F, S~

'''as s s

1 2 3 4 5 6

Figure 3. Thin layer chromatography of lipids and bacterial overlay with"S-P.cepacia. (a) Standardglycolipids ( 1 -2Mg)were subjected

toTLC and detected by Orcinol spray 1 =lactosylceramide(LC),2

=Gb3, 3=

_GM,,

4=asialoGM1,5=GM2 and asialo GM2. Lane 6,lipidextractfrom 2 mg of intestinal mucin. (b) Duplicate TLC plateswerenotstained but were blocked with1%gelatin overnight at 37°C,washed, and then overlaid with"S-labelledP. cepacia(PC 5, 2 x 109CFU/ml)andincubated for 1 h at 37°C.After further wash-ing to remove nonbound bacteria, TLC plates were subjected to au-toradiography(5-7 d exposure to Kodak x-ray film). Bacteria bound

to standard asialo GM1 (lane 4) but not to mucin lipid extract (lane 6).

cluded thatourpreviouslyobserved bacterial binding to mucin wasunlikely to have been due to contaminating glycolipids.

Theonly positiveband was that of theasialo-GMl standard.

Asialo GM2 didnotbindeither PC 5 or PC 7. A consideration

of the differencesbetween the two glycolipids suggested that the receptor structurespecific forPC5 and PC 7 might be the

disaccharide Gal (3 1,3GalNAConasialoGM1.This

disaccha-ride isalso commonin mucin core 1 or core 2 oligosaccharides

(30). However, the previous lack of inhibition of P. cepacia

Chemical and enzymatic treatmentofmucin

Both hapten inhibition and directP. cepaciabinding studies

wereperformedon intestinalmucinsamplesbeforeandafter delipidation, desulfation, partialortotaldeglycosylation,thiol

reductionandalkylation,ordigestion withpronase. Composi-tional analyses of the mucins were carried out to monitor

changes in the mucinas aresult of eachtreatment(Table V)

and to assist in theinterpretationofbindingresults. No

signifi-cant changesinamino acid profileswerenoted(exceptfora 25% decreasein aspartic plus glutamic acid content after

pro-nase treatment). Carbohydrate composition, however, was al-teredin severalcases:delipidationcaused anunexplained24%

decrease infucose(21.8 to 16.6%),desulphation with metha-nolicHC1causedamajor decrease in thetwoacid labile resi-dues fucose (62%) and sialic acid (59%), aswell as a loss of

sulphate(98%); TFMS (partial deglycosylation) causedatotal

lossof fucose, sialic acid,andgalactose,aswellas an83%

de-creasein GlcNAc anda57%decreaseinGalNAc. Reduction/ alkylation andpronase treatmentscausednomajor changesin

carbohydrate. TFMS plus 'reverse'(-elimination(total degly-cosylation) released all ofthe mucin carbohydrate. Thetreated

mucinswerethen usedaspotential hapteninhibitors in

stan-dardmucin (SIM) bindingassaysafterpreincubationwith 3H-P.cepacia (PC 5) at370C for1 h. As shownin TableVI,totally deglycosylated mucinwas noteffectiveas ahapten inhibitor,

whereasdelipidated,and reduced/alkylatedmucinswere more

effective inhibitors than untreated (control) mucin. Desulfated mucin (whichwasalso reduced insialic acid andfucose)was

thesame asuntreatedmucin. Themosteffective inhibitorwas

thepartiallydeglycosylated(TFMS-treated) mucin which

con-tained only GlcNAcandGalNAcresidues.Pronasedigestion

gaveanomalouslyhighbackgroundbindingtoBSA-containing

wells,whichrenderedmucin bindingdatainvalid.

Someofthe treated mucinswere nextused indirect binding

experiments performedasslot-blotbindingassayswith 35S-la-belledP.cepaciaoverlay andautoradiographicdetection. BSA providedanegative control. Fig.4 wasconsistent withthe

re-sults of Table IV in showing thatbindingwas strongest for TFMS-treated (partially deglycosylated)mucin, moderately in-creasedabove controlfordelipidated mucin,andnegligiblefor totally deglycosylated mucin.Pronase-digested mucinwasthe

same asuntreatedcontrol mucin. Desulfated mucinwas not

tested duetolackof material.

Taken together, both the hapten inhibition and direct bind-ing experiments suggest that core oligosaccharide structures

containingGalNAc and/or GlcNAcprovidedthe bestreceptor structuresforP. cepacia. Theenhancementofbinding by

re-ducingagentsand lipidextractionwerejudgedto have been

causedby conformational modifications(ineither thepeptide

ortheoligosaccharidemoieties ofthemucin)which "unmask"

orexposethemoredeeplyburiedcorecarbohydrate(GlcNAc

andGalNAc-containing)receptors.

Use

of other

mucinsin

binding studies

Asialoporcinesubmaxillarymucin,which containscore

car-bohydrate chains having the structuresGalNAc-(ser/thr) or

Gal (31,3

_{GalNAc-[ser/thr]}

withorwithouta1,2linked fucose

TableV. CompositionofHuman IntestinalMucinBefore and After Various Chemical and EnzymaticTreatments

Reduced Pronase Partially Totally

Mol% Notreatment Delipidated Desulfated andalkylated treated deglycosylated deglycosylated

Asp+Glu 12.8 14.7 12.4 13.6 9.7 12.1 13.2

Ser+Thr+Pro 52.8 48.0 49.8 50.2 51.7 49.4 49.0

Gly+Val+Ala 16.9 16.5 18.0 17.1 16.2 17.4 18.4

Ile+Leu+Phe 11.3 13.3 12.6 10.0 13.9 13.3 11.3

Rest 6.2 7.5 7.2 9.1 8.5 7.8 8.1

Fucose 21.8 16.6 8.3 19.8 18.5 ND ND

Galactose 23.0 26.1 36.3 25.3 25.3 ND ND

N-acetylglucosamine 45.1 46.6 43.3 42.5 45.2 68.3 ND

N-acetylgalactosamine 8.4 7.6 11.4 8.3 8.8 31.7 ND

Sialic acid 1.7 3.1 0.7 4.1 2.2 ND ND

Sulphatenmol/jgprotein 0.45 NT 0.007 NT NT NT NT

Each value represents the averageof twodeterminations.ND means notdetected. NT means not tested.Amino acids were analyzed by the Pi-cotag system(WatersAssociates,Toronto,Canada)after hydrolysisin 6 M HC1for24 h at 110C. Carbohydrates wereanalyzedbytheBio LC-HPLC(DionexCorp., Sunnyvale, CA) systemafterhydrolysis in 2 Mtrifluoroaceticacidfor2 h at0C.Pronase-treated samples probably contained residualpronase,therefore amino aciddata may not bereliable.Sulphate was assayed by thebarium-rhodizonatemethodof Silvestri

etal.(18).

residues (27), as well as asialo ovine submaxillary mucin, which containsmainly a-linked GalNAcasthemajor carbohydrate (28), were compared with normal SIM in microtiter binding studiesof 3H-P. cepacia (2.5X I09 CFU/ml of PC 5). Porcine submaxillary mucin bound P. cepacia only 1/7 as much as

SIM, and ovine submaxillary mucin did not exhibit any spe-cificbinding (not shown).Theseexperimentssuggested, by

ex-clusion, that neither a-linked core GalNAc-norGal

_fl

1,3 Gal-NAc inmucinsaresufficientto act aseffectivereceptors. Since O-glycans are the dominant chains in mucincarbohydrates, the presence of GlcNAc linked to core GalNAc residues in

humanSIM or TBM would thus seem to be necessary for opti-mumbinding.

The

effect of mucin on P. cepacia binding to buccal

epithelial

cells

Intheprevious paper (15) mucinswerefoundnotto interfere with the binding ofP.aeruginosa to buccalepithelial cells. In

Table VI.Inhibition ofBindingofP. cepaciatoHumanIntestinal Mucin inthePresenceofTreatedMucinsUsed

asHaptenInhibitors

Hapten Percentinhibition

Nohapten 0

Untreatedmucin (control) 28.5±2.57

Delipidatedmucin 57.9±3.75

Reduced and alkylatedmucin 57.3±6.26

Desulfatedmucin 26.6±4.20

TFMS-treated mucin 75.3±2.60

TFMS plus reverse ,B-eliminated mucin 3.3±0.68

3H-labelledP.cepacia(PC 5isolate) (2.65 X 109CFU/ml)were preincubatedwith each hapten (5-10 ug/well) for 1 h at 37°C, and themixture then addedtomucin-coatedwells as in standard mucin-bindingassays. All valueswerecorrectedfor nonspecific bind-ingtoBSA. In the absenceofhapten, the R value was 15.5. Each value represents the mean(±SEM) for six independent experiments.

thecase ofP.cepaciawe obtained adifferentresult,butonlyfor

isolates whichwere known tobecapable of bindingtomucin.

Isolates PC 5, PC7(whichcanbind tomucin) and PC 45,PC

61 (whichdo not exhibitspecific bindingtomucin),werefirst

testedfor their abilitytobind tobuccalepithelial cells, (2-3

E

r- _0.6

CO c0

0

0.4

0

L..

0 _0.2

0

.

.

I

1

2

3

4

5

6

7

Figure 4. Slot blot direct assay ofP.cepacia bindingtochemicallyor

enzymaticallymodified mucinsamples. Samples containing 50, 100, and500 ng of mucinproteinwereslot-blottedontonitrocellulose membranes, washed, and then blocked with 3%globulin-freeBSA in TBSfor1 hat37°C. Afterfurtherwashing, sampleswereincubated overnightwithI'S-labelledP.cepacia (PC 5, 3.6 x 109CFU/ml), washedsix timesto removenonboundbacteria, and thensubjected

tofluorography.Overlaysfor500-ng samplesareshownonthe bot-tom;densitometryonthe top. 1,BSA(control);2, untreatedmucin (TBM); 3,chloroform:methanol-treated mucin;4, TFMS-treated mucin, 5,pronase-treatedmucin; 6,TFMS +reverse,B-eliminated

mucin;7, untreated mucin(SIM).

0.0

X 105BEC/ml)underconditionssimilarto those used

previ-ouslyfor P. aeruginosa(I15,24).Allfour PC isolatesboundina

saturablefashion, reachingvalues of52,64, 16,and 36CFU/

BEC at bacterialconcentrations of 1.0 x 109,1.0X 109,2X 108,

and 5X 108CFU/ml for PC 5, PC 7, PC 45, and PC61, respec-tively.

At similarconcentrations ofbacteria andBEC, the reported binding of P. aeruginosa (PAK) is - 181 CFU/BEC (29),

which suggeststhat there may be more receptor sites on BEC for P. aeruginosa than forP. cepacia.

Table VII presents the results of hapteninhibition

experi-ments in which P. cepaciaisolates werepreincubated with

dif-ferent concentrations of mucin beforetestingtheirbindingto

BEC. Only PC 5and PC 7bindingwereinhibitedsignificantly

by mucin, and the maximum inhibition obtained wasabout 60%.PC 45 and PC 61 (the non-mucinbinding isolates)were

not inhibited. All four strains therefore haveadhesins capable of bindingtoBEC,butit islikelythat theseadhesins, (aswellas

the specific receptorsites forthem on BEC), differ from the specific adhesins (andreceptors)involved inmucinbinding. Discussion

Clinical isolatesofP. cepaciaobtained fromthe sputum of 19 out of 22colonizedpatientswithcysticfibrosiswereobserved to adhere to non-CF and CF mucinsofthe intestinal and

respi-ratory tracts, respectively. There was nopreference ofany of

theisolates tested for CF mucin versusnon-CF mucin. Thus the specificreceptorsonmucinsare

essentially

normalmucin

components present in the respiratory and gastrointestinal

tracts. The overallquantityand accessibility of mucin within

theairways ofpatients with CF may therefore be

important

determinants ofP. cepaciacolonization. Isolatesvaried

some-what in theextent (saturation) andaffinity of their adherence to mucin, which suggests that the number(and possibly the nature) of the bacterial adhesin(s) for mucins vary amongst

strains.

TableVII. EffectofMucinon P. cepaciaBinding

toBuccalEpithelialCells

Percentbinding ofcontrol

Mucin PC 5* PC 7* PC45$ PC61t

gg/ml

0 100 100 100 100

25 74(72-76) 68 (66-70) 113(104-128) 104(102-107) 50 59(54-51) 48(47-51) 110(105-115) 103(87-109) 100 58(52-66) 60(53-66) 113(106-119) 103(98-106)

150 ND 52(47-56) ND 95(87-100)

200 ND 40(36-47) ND 87(81-98)

3H-labelled PC 5, PC 7, PC 45,orPC 61(2X 108or1 X 109CFU/ml)

werepreincubated with mucin (0-200,g/ml)for 1 hat

370C,

then addedtobuccalepithelialcells(1.5or5X I05cells/ml)andincubated forafurther 1h at370C.Thereaction mixturewasfilteredona

polycarbonatefilter and theretentatecounted.Bindingin the absence of mucin (control)(100%)was23, 17, 21,and 10CFU/BECfor PC 5, PC 7, PC 45, and PC61,respectively.ND,notdetermined. * Iso-lates knowntobind; $ isolates knownnot tobindspecificallyto

mucins.

SelectedstrainsofP.

_cepacia

_{(particularly}

PC5 and PC

₇₎

were chosen for a moredetailed examination of the

binding

process. In eachcase

binding

to mucinwas

specific

(several-fold

higher

than to BSA

controls),

saturable with

_increasing

bacterial

concentration,

and

_(for

atleastoneisolate

tested)

in-hibited

by

preincubation

with mucins

_(SIM

or

TBM)

before

binding

assays.Thesecharacteristicswerenotobserved in

ear-lier studies withP.

aeruginosa,

which showedno

preferential

binding

tomucin

compared

toBSA

(15).

Hydrophobic

interactions didnot

_participate

in the

recep-tor-mediated

binding

ofP.

cepacia

to

mucin,

since

tetrameth-ylurea

had no effect on

binding.

In contrast, P.

_aeruginosa

binding

to mucin

(and

BSA)

was found to be almost

com-pletely

dependent

uponweak

nonspecific

hydrophobic

interac-tions

(15).

The 'naked'

peptide

regions

of mucin

_appeared

to

play

nopartin

mediating

P.

cepacia

attachment,

sincepronase

digestion (which

removes

_{nonglycosylated}

or'naked'areasof

mucins) (21)

didnotreduce

binding.

Several

experiments

suggested

that

specific

mucin_receptors

forP.

cepacia

were

_carbohydrate

innature.Sodium

metaperio-datetreatmentof mucininhibited

_binding,

asdid removal of

all

_carbohydrate

_by

TFMS

plus

'reverse' 13-elimination.

Al-though

various

_lectins,

individualmonosaccharides

(including

GlcNAc and

GalNAc)

andselecteddisaccharidesdidnotactas

hapten

inhibitors,

mucin

_samples

which were almost

com-pletely

(88%)

deglycosylated

were_potentinhibitors. The

only

detectablesugars

remaining

onthe TFMS-treatedmucinwere

GlcNAc and

GalNAc,

presumably

linked in the

configuration

of GlcNAca

1,6

(-

or(

_1,3)

-GalNAc

(or

both) (i.e.,

inacore3

or

_partial

core4mucin

oligosaccharide configuration [30]).

As

shown in

binding

assaysof PC 5 toasialo ovine

submaxillary

mucin andasialo

porcine

submaxillary mucin,

the

configura-tions

_{GalNAc-peptide (Ser/Thr) or-Gal}

(3

_1,3

_{GalNAc-peptide}

(Ser/Thr)

were

_inadequate

forP.

_cepacia

_binding.

Thereforeit

appearstobenecessarytohave GlcNAclinkedtoGalNAcfor

optimum

bacterial adhesion.

_Synthetic

_{oligosaccharides}

will be used infuture

_experiments

to

_identify

the best

_{configurations.}

Othertreatmentsof

_mucin,

_{including reduction/alkylation}

and

_{chloroform/methanol extraction,}

werefoundtoenhance

P.

cepacia

binding,

as

_{judged by hapten}

inhibitionanddirect

binding

studies. Since

_reducing

agents act to break disulfide

bonds andtorelease the

11

8-kD "link"

glycopeptide

from

na-tiveintestinal

mucin, (but

leave the

_{remaining carbohydrates}

intact

[21,

31])

reduction

probably

caused

_unfolding

_and/or

separation

of

peptide

chains inafashionthat

_exposed

underly-ing

or

normally

masked

_carbohydrate

receptors.

Chloroform:methanol extraction may also have

exposed

receptor

sites,

either

by

denaturing

(unfolding?)

mucin

pep-tides,

or

_by

removing

minor

(nondetectable)

lipid

contami-nants. Since

"delipidated"

mucin exhibited greaterbacterial

binding

than intact

mucin,

it is

_highly

_unlikely

that mucin

receptors forP.

_cepacia

resideonminor

_lipid

contaminants in

themucin

_preparation.

The isolates of P.

cepacia

that bound to mucin were

grouped

according

tothe

_severity

of illness ofeach

_patient

at

the time the isolate was obtained.

Although

the number of

samples

was

_small,

itwas

interesting

thatisolates

_exhibiting

the

highest

average mucin

_binding

R values tended to _segregate

withsevere illness

(group

a),

whilethosewiththe lowest

aver-ageRvaluessegregatedwith mild illness_(groupc).Thesedata

between the extent ofmucin binding and progressive lung dam-age. Itwill be necessary in the future to carry out a prospective study to discover whether colonization withmucin-bindingP.

cepacia is a poorprognostic signinpatients with cystic fibrosis.

Experiments comparing the effects of mucin on P. cepacia

bindingto buccalepithelialcells suggestedthat thoseisolates

that are capableof bindingtomucin (i.e., PC 5 and PC 7) were inhibited (in part) from binding to BEC by added mucin,

whereas thoseisolates (i.e.,PC 45 and PC 61) with no specific affinity for mucin, wereinsensitivetoaddedmucin.One might expect apriorithatbindingtomucinin vivo would be protec-tive, since it would retard accessof bacteriatorespiratory

epi-thelialcells and subsequent lung damage. On the other hand,

withinareasof stasisin the lung, secreted mucin may simply

provide a reservoir for mucin-binding P. cepacia strains, therebyfacilitating colonization. Ourfuture effortswill be

di-rected toisolating andcharacterizingthemucin-specific

adhe-sin(s) ofP. cepacia.

Acknowledgments

Financial assistance for this work was provided by the Canadian Cystic FibrosisFoundation. Dr. U. Sajjanwastherecipient ofaCanadian Cystic Fibrosis Foundation postdoctoralfellowshipaward. The authors areindebted to Ms. AnneRobson, Division ofMicrobiology, Hospital

for Sick Children, for identification andprovision ofPseudomonas cepacia isolates.

References

1.Ederer,G.M., and J. M. Matsen. 1972.Colonizationandinfection with Pseudomonascepacia. J.Infect.Dis. 125:613-618.

2. Isles, A., I.Maclusky,M.Corey,R.Gold,C.Prober,P.Fleming,and H.

Levison. 1984.Pseudomonascepaciainfection in cystic fibrosis:anemerging problem. J.Pediatr. 104:206-210..

3. Tablan, 0.C., T. L.Chorba,D. V.Schidlow,J. W.White,K. A.Hardy, P. H.Gilligan,W. M. Morgan, L. A.Carson,W. J. Jason, and W. R.Jarvis. 1985. Pseudomonascepaciacolonization in patients with cystic fibrosis: risk factors and clinicaloutcome.J.Pediatr. 107:382-387.

4.Sobel,J.D.,N.Hashman, G. Reinherz,and D. Merzbach. 1982.

Nosoco-mialPseudomonascepaciainfection associated with chlorohexidin contamina-tion.Am. J.Med. 73:183-186.

5.Tablan, 0.C.,W. J.Martone,and C. F.Doershuk. 1987. Colonization of

therespiratorytractwithPseudomonas cepaciaincystic fibrosis:riskfactorsand outcome.Chest. 91:527-532.

6.Thomassen, M.J.,C. A.Demko,J. D.Klinger,and R.C. Stern. 1985. Pseudomonascepaciacolonizationamongpatients with cystic fibrosis.A new

opportunist.Am.Rev.Respir.Dis. 131:791-796.

7.Gessner,A.R.,and J. E. Mortensen. 1990.Pathogenicfactors of

Pseudo-monas cepacia isolates frompatients withcysticfibrosis.J. Med. Microbiol.

33:115-120.

8.Lonon,M.K.,D. E.Woods,and D. C.Straus. 1988.Production oflipase by clinical isolates ofPseudomonas cepacia.J.Clin. Microbiol. 26:979-984.

9. Kuehn, M.,K. Lent, J. Hass,J.Hagenqieker,and A. L. Smith. 1988. Binding of Pseudomonas cepaciato respiratorycells. Pediatr. Pulmonology.

(Suppl. 2):77. (Abstr.)

10.Saiman,L., G.Cacalano, and A.Prince. 1990.Pseudomonascepacia adherencetorespiratory epithelialcellsis enhancedbyPseudomonasaeruginosa.

Infect.Immun.58:2578-2584.

11. Krivan, H. C., V.Ginsburg,and D. D. Roberts. 1988.Pseudomonas

aeruginosaandPseudomonas cepaciaisolatedfromcysticfibrosis patients bind

specificallytogangliotetraosylceramide(asialoGMI)andgangliotriosyl

cera-mide(asialo GM2).Arch. Biochem.Biophys.260:493-496.

12.Baltimore,R.S.,C. D. C.Christie,andG.J. W.Smith.1989.

Immunohis-topathogeniclocalizationofPseudomonas aeruginosainlungsfrompatientswith

cysticfibrosis.Am.Rev.Respir.Dis. 140:1650-1661.

13.Sanford,B.A.,V. L.Thomas, andM. A. Ramsay. 1989. Bindingof

Staphylococcito mucusin vivo and in vitro.Infect.Immun.57:3735-3742.

14.Gilardi,G. L. 1985. Pseudomonas.InManualof ClinicalMicrobiology.

4th ed. E. H.Lennette,A.Balows,W. J.Hausler,Jr.,andH. J.Shadomy,editors. ASMPublications,Washington,DC350-372.

15.Sajan,U., J.Reisman,P.Doig, R. T. Irvin, G. Forstner, and J. Forstner.

1992.Binding of nonmucoid Pseudomonasaeruginosa to normal human

intes-tinal mucinandrespiratory mucinandrespiratory mucin frompatientswith

cysticfibrosis. J. Clin.Invest.89:657-665.

16.Slomiany,A., B. L.Slomiany,A.Witas,M. Aono,and L. J.Newman.

1983.Isolation of fatty acidscovalently bound to the gastric mucus glycoprotein

of normal and cystic fibrosis patients. Biochem. Biophys. Res. Commun. 113:286-293.

17.Slomiany,B.L., and A. Slomiany. Isolation and characterization of the

sulfated neolactotetraosylceramide from hog gastric mucosa. J. Biol. Chem. 253:3517-3520.

18.Silvestri,L. J., R. E. Hurst, L.Simpson,and J. M. Settine. 1982.Analysis of sulfateincomplex carbohydrates.Anal.Biochem. 123:303-309.

19.Edge,A.S. B., C. R.Faltynek,L.Hof,L.E.Reichart, and P. Weber.1981.

Deglycosylationofglycoproteins bytrifluoromethanesulfonicacid. Anal.

Bio-chem. 118:131-137.

20.Gerken,T. A., N.Jentoft, and R.Gupta. 1989. A new approachfor chemically deglycosylating mucous glycoproteins. Glycoconjugate J. 6:155.

(Abstr.)

21.Mantle, M.,G.G. Forstner,and J. F. Forstner. 1984. Biochemical

charac-terizationof the component parts ofintestinalmucinfrom patients withcystic fibrosis. Biochem.J.224:345-354.

22.Laine,R. A., K.Stallner,andS.Hakomori. 1974. Isolation and

character-izationofmembraneglycospingolipids.InMethods in MembraneBiology.2nd ed. E. D. Korn, editor. Plenum PublishingCorp.,New York. 205-244.

23.Hansson, G.C.,K. A.Karlsson,G. Larson, N. Stromberg, and J. Thurin. 1985. Carbohydrate specific adhesion ofbacteria to thin-layerchromatograms:a

rationalized approachtothestudyof host cellglycolipidreceptors.Anal. Bio-chem. 146:158-163.

24. Doig, P., T. Todd, P. A. Sastry, K. K. Lee, R. S. Hodges, W. Paranchych, and R. T. Irvin. 1988. Role of pili in the adhesion of Pseudomonas aeruginosa to humanrespiratory epithelial cells. Infect. Immun. 56:1641-1646.

25. Laux, D. C., E. F. McSweegan,and P. S. Cohen. 1984. Adhesionof

enterotoxigenicEschierichia coli to immobilized intestinal mucosal preparations: amodel for adhesion to mucosal surface components. J. Microbiol. Methods.

2:27-39.

26.Sorenson, K., and U. Brodbeck. 1986. Assessment ofcoating-efficiencyin ELISA plates by direct protein determination. J. Immunol. Methods. 95:291-293.

27. Carlson, D. M.1968. Structure and immunochemical properties of oligo-saccharides isolated from pigsubmaxillarymucins. J.Biol. Chem. 243:616-626. 28.Sadler, J.E., J. L. Rearick, J. C.Paulson,and R. L.Hill. 1979. Purification tohomogeneityof a,B-galactosidasea2-3sialyltransferaseandpartial

purifica-tion of ana-N-acetylgalactosaminidea2-6sialyltransferasefromporcine sub-maxillary glands. J. Biol. Chem. 254:4434-4443.

29.Irvin, R. T., P. Doig, P. A. Sastry, B.Heller,and W.Paranchych. 1989. Competition for bacterial receptor sites onrespiratory epithelialcellsby Pseudo-monasaeruginosa strains of heterologous pilus type: usefulness ofkinetic parame-tersof adherence in predicting the outcome. Microbiol. Ecol. Health Dis. 3:39-47.

30.Schachter,H., and I. Brockhausen. 1991. The biosynthesis ofserine

(threo-nine)-N-Acetylgalactosamine-linkedcarbohydratemoieties. InGlycoconjugates:

Composition, Structure and Function. H. Allen and E. C.Visailus,editors. Mar-cel Dekker, Inc., NY. In press.