• No results found

Rapid method to detect shiga toxin and shiga like toxin I based on binding to globotriosyl ceramide (Gb3), their natural receptor


Academic year: 2020

Share "Rapid method to detect shiga toxin and shiga like toxin I based on binding to globotriosyl ceramide (Gb3), their natural receptor"


Loading.... (view fulltext now)

Full text



Copyright C 1989,American Societyfor Microbiology

Rapid Method



Shiga Toxin


Shiga-Like Toxin






Globotriosyl Ceramide

(Gb3), Their Natural Receptor


Department of Pediatrics andProgram inInfectiousDiseases and Clinical Microbiology, UniversityofTexas MedicalSchoolat Houston, Houston, Texas 77030

Received19October 1988/Accepted 25January 1989

Shigatoxinand thecloselyrelatedShiga-liketoxins produced by Escherichia colirepresentagroupofvery similar cytotoxinsthatmayplayanimportant role in diarrhealdiseaseandhemolytic uremic syndrome. These toxinshave thesamebiologic activitiesand accordingto recentstudies also share thesamebindingreceptor,

globotriosyl ceramide(Gb3). They arecurrently detected, onthebasisof their abilitytodamage several cell

lines, by using expensive and tediousassaysthat requirefacilities for andexperiencewithtissue cultures and arethereforemostsuitable for researchlaboratories. We have developedarapidmethodtodetect Shiga toxin and Shiga-like toxin I based on specific binding to their Gb3 natural receptor, which was coated onto

microdilution plates. Bound toxin was then detected by enzyme-linked immunosorbent assay (ELISA) with monoclonalantibodies. The sensitivity of the Gb3 ELISAwas0.2ng(2ng/ml) of purified toxin.Theassay was positive with sonic extracts ofShigella dysenteriae serotype 1 strain60R (a Shiga toxin producer), E. coli serotype026:Hll1strain1H30, andE.coli serotype0157:H7(bothShiga-liketoxinIproducers). Theassaywas veryspecific inthatnocross-reactivitywasnoted withpurified cholera toxin,E.coli heat-labile andheat-stable enterotoxins,and Clostridiumdifficile cytotoxin,orsonicextractsof othercytotoxin-producing organisms,such asothershigellae, pathogenicandnonpathogenicE.coli,Salmonellaspp.,Campylobacterspp.,and Aeromonas spp.These resultswereincompleteagreementwitha [3H]thymidine-labeledHeLacellcytotoxicityassayand with detection of the structural genes by DNA hybridization studies with a Shiga-like toxin I probe. Quantitative analysis showed a high correlation between Gb3 ELISA and HeLa cell assay when fractions obtained at various stages of toxinpurificationwereexaminedbyboth methods(r=0.99,P<0.01).Thisrapid Gb3 ELISAis sensitive andspecific andmaybediagnostically useful incytotoxin-related infections.

Cytotoxin production by enteric organisms has been

in-creasinglyinvestigated inrecentyears. Information regard-ingthe role of cytotoxinsin the intestinalmanifestations of

Shigella spp. (26), Escherichia coli (6), and Clostridium

difficile (3) is gradually accumulating. Hemorrhagic colitis

syndrome has been recently definedandrelatedtohigh-level cytotoxin-producingE. coli,suchasserotypes0157:H7 and 026:H11,which have been calledenterohemorrhagicE. coli (19, 24, 27, 29). Inaddition, cytotoxin production probably playsa major pathogenic role inthe development of hemo-lyticuremicsyndrome (HUS) (5, 12). Forexample, E. coli 0157:H7 has been observedin one-half of the patientswith HUSduringaprospective studyin thenorthwesternUnited

States(20),andthe syndromedeveloped in 22%ofpatients involvedinanoutbreakofhemorrhagic colitisrelatedtothis

organism (4). Evidence of infection with cytotoxin-pro-ducingE. colihas beendemonstratedin75%(12)to88%(5)

ofpatientswithidiopathic HUS. It iscurrently thoughtthat the cytotoxins damage vascular endothelial cells, thereby initiatingtheabnormalities ofHUS(5).

The cytotoxins are defined by their ability to damage

mammaliancells. Amongthemost important cytotoxinsare Shiga toxin, produced by Shigella strains (mainly Shigella dysenteriae serotype 1) (2, 21), and the closely related

Shiga-liketoxins(SLTs), produced byE. coli (23).The latter are also referred to as verotoxins by those who assay cytotoxic activity in Verocells (15). At least two

immuno-logically distinct toxins, SLT-I and SLT-II, have been de-fined (19, 25,29). SLT-I is neutralized byantibodies raised

* Correspondingauthor.

againstShiga toxin,towhichit isvirtually identical;thereis only a three-nucleotide difference, resulting in a

single-amino-acid difference, between the two toxins (11). Al-though immunologically distinct, SLT-II is very similar to

Shigatoxin and SLT-I withregardto structureandbiologic

activities. Thesetoxinsactbyinhibiting protein synthesis by catalytic inactivation of the 60S ribosomal subunits. Recent studies have shown that these toxins also share the same glycosphingolipid receptor, globotriosyl ceramide (Gb3) (9, 10, 16, 17, 31).Other less-defined variants of SLTsalso exist (23).

The cytotoxins arecurrentlydeterminedby theirdamage

to severalcell lines, which isusually evidencedby morpho-logic changes.Theseassaysareexpensiveandtedious.They require facilitiesfor andexperiencewith tissue cultures and are therefore not suitable for routine usage and

less-well-equipped laboratories. Inlightofrecentdataindicatingthat the well-defined cytotoxins share thesamebindingreceptor, wehavedevelopedarapid-detection methodforShigatoxin and SLT-I based on their binding to Gb3, the natural receptor, with detection of bound toxin by enzyme-linked

immunosorbentassay(Gb3ELISA).Resultswerecompared with cytotoxicity determinedby aquantitative radiolabeled HeLacellassaythatwehavedescribedpreviously(2, 6, 26)

and with the structural genes fortoxin productiondetected

by DNAhybridization assay(21, 22).

(Thispaperwaspresentedinpart atthe annualmeetingsof the Southern SocietyforPediatricResearch, NewOrleans, La., February 1989;and the Societyfor PediatricResearch, Washington, D.C., May 1989.)


on April 11, 2020 by guest



TABLE 1. Procedure forGb3 ELISA

Concnor DletVol Incubationconditions

Stepandcompound dilution Diluent (j./well) Temp Time

1. Coat plates with Gb3 10ikg/ml Chloroform-methanol (2:1) 100 Room 4h

2. Wash Tween-PBSa 200 Room 1 min(5x)

3. Block with BSA 50mg/ml PBS 200 370C 1 h

4. Wash Tween-PBS 200 Room 1 min(3x)

5. Add test sample 0.1% BSAinTween-PBS 100 370C 1 h

6. Wash Tween-PBS 200 Room 1min(5x)

7. Addmonoclonoal antibodies toSLT-I 1:1,000 0.1% BSAinTween-PBS 100 370C 1h

8. Wash Tween-PBS 200 Room 1min(5X)

9. Addperoxidase-conjugated rabbit anti- 1:1,000 0.1% BSA in Tween-PBS 100 370C 1h mouseIgG"

10. Wash Tween-PBS 200 Room 1 min(5x)

PBS 200 Room 1min (2x)

11. Addsubstrate (orthophenylene diamine) 0.4mg/ml Citrate-phosphate buffer 100 Room 15 min

12. Stop reaction with2.5 N H2SO4 100 Room

aTween-PBS,0.05%Tween80in PBS (0.2Msodium phosphate, 0.2Mpotassium phosphate).

bIgG, Immunoglobulin G.


Bacterial strains. Bacteria which characteristically

pro-duce Shiga toxin andSLT-I wereinitially used for

develop-ment ofthe assay. They included S. dysenteriae serotype 1 strain 60R(aShiga toxin producer); E. coli serotype 026: Hl strainH30 (a SLT-I producer), originally described by J.Konowalchuk(15);andE.coli serotype0157:H7 (a SLT-I

producer), provided by John Mathewson, who


obtained the strain from Richard Wilson at the E. coli

ReferenceCenter,Pennsylvania StateUniversity. For

spec-ificity studies,otherbacterial strainswereused:E.coliK-12 strain C600(anon-cytotoxin-producing strain)andlysogenic

E. coli K-12 strain C600 (933w) (a SLT-II producer),

pro-vided by A. D. O'Brien, Uniformed Services University of the HealthSciences, Bethesda, Md. Other bacterialisolates examined were Shigellaspp. (2),pathogenic and

nonpatho-genic E. coli strains (6), Salmonella spp. (1), Aeromonas spp. (13), Campylobacter spp., andCitrobacterspp.

Preparation of bacterial sonic extracts. Bacteria were grownin iron-depletedsyncase


for48 hwithshakingat

200 rpm,harvested, lysedbysonication, andfilter sterilized, as previously described (2, 6). Protein concentration was determined by protein assay (Bio-Rad Laboratories, Rich-mond, Calif.) with bovine serum albumin (BSA) as a stan-dard.

Preparation of purified Shiga toxin. Shiga toxinwas puri-fied fromasonic extractofS. dysenteriae serotype 1strain 60R by Affi-Gel Blue column, chromatofocusing, and hy-droxylapatite column, as wehavedescribed elsewhere (26). The finalcytotoxic fraction was electrophoresedon sodium dodecyl sulfate-polyacrylamide gel (15%) under denaturing conditions (,-mercaptoethanol), and purity was confirmed withsilverstaining.Thepurified toxin had a specificactivity

of107350%cytotoxic doses


per mgofprotein.

Preparation of anti-Shiga toxin serum. Rabbit anti-Shiga

toxinserum,whichalso neutralized SLT-I,wasobtainedby repeated injections offormaldehyde-treated purified Shiga toxin (100 Ftg per dose) in complete Freund adjuvant as

previously described (2).

Gb3ELISA. The detailed procedure of theGb3 ELISA is described in Table 1. Gb3 (Supelco, Bellefonte, Pa.) was

obtained and diluted for use in chloroform-methanol (2:1). Afterinoculation,thepolyvinyl chloride microdilution plates (Dynatech Laboratories, Inc., Alexandria, Va.) were left

uncovered, sothat thedil ientevaporated and Gb3 attached to the plates. Monoclonal antibodies (13C4, mouse ascitic fluid) to SLT-I (B subunit) were obtained from N. A. Strockbine, Centers for Disease Control, Atlanta, Ga. (28), and used at a 1:1,000 dilution.Peroxidase-conjugated rabbit anti-mouse immunoglobulin G (Bionetics Laboratory Prod-ucts, Charleston, S.C.) was then added at a dilution of 1:1,000. Afterbeingwashed,the substrate, orthophenylene diamine (0.4 mg/ml)-0.1% H202 in citrate-phosphate buffer

(25 mMcitrate,50 mM phosphate [pH 5]), was added.The

reaction was stoppedwith 2.5 N H2SO4, and then samples were transferred to new plates and the A490 wasmeasured (Dynatech microELISAplatereader MR580).Controlwells, towhich0.1%BSA in0.05%Tween80-phosphate-buffered

saline (PBS) (sample diluent) instead of the sample was

added and inwhich all steps wereperformed,gavereadings of near zero (0.000 to 0.003). Positivity was defined as

absorbance of 0.1 ormore.

Reproducibility of the assaywas determinedbyduplicate testing. Sensitivity was determined by titration curve with twofold dilutions ofaknownamountofpurifiedShigatoxin and sonic extracts of SLT-I-producing bacteria. Specificity

wasdeterminedby examining cross-reactivity with (i)other purified toxins (including cholera toxin and E. coli heat-stableenterotoxin[SigmaChemicalCo.,St. Louis,Mo.];E. coli heat-labile enterotoxin [B subunit], provided by John

Mathewson,whooriginallyobtained thetoxinfrom J. Clem-ents, Tulane Medical School, New Orleans, La.; and C.


cytotoxin [Bartels Immunodiagnostic Supplies,

Bellevue, Wash.]), (ii) sonic extracts of well-characterized

bacterial strainsthatproduce SLT-II (E. coli C600 [933w])or are non-cytotoxin producing (E. coli C600), and



extracts ofa varietyof other cytotoxin-producingbacterial isolates (Table 2).

Correlation with a quantitative HeLa cell assay was de-termined by simultaneousexamination ofpurifiedtoxin and

sonic extracts by both methods. In addition, we examined

samplesobtained during various stagesofShigatoxin puri-fication for cytotoxicity


per milligram of bacterial

protein) and by Gb3 ELISA (micrograms oftoxin) (blind format). The fractions were a crude sonic extract of S.

dysenteriae serotype 1 strain 60R, a cytotoxin fraction eluted from an Affi-Gel Blue column, a cytotoxin fraction eluted bychromatofocusing, andpurified Shigatoxin.

on April 11, 2020 by guest



TABLE 2. Detection of Shiga toxin andSLT-I in sonic extracts of cytotoxin-producing bacteria by Gb3 ELISA and comparison

with HeLa cell assay and DNA hybridization

Result by:

Bacterium (toxin) Gb HeLacell DNA ELISAEIA cytotoxinassay probe forSLT-I Referencestrains

S.dysenteriaeserotype1 strain + + + 60R(Shiga toxin)

E.coli 0157:H7(SLT-I) + + +

E.coli026:H11 strain H30 + + +


E. coli C600 (933w)(SLT-II) - -

-E. coliC600 - -

-Other bacteria(no. ofstrains)

S.dysenteriae serotype 1(10) + + NDb Shigellasonnei(13)

-Shigella flexneri (13) -

-Shigella boydii (4) -

-Nonpathogenic E.coli (11) - - ND

Pathogenic E. colic

EAEC (11) - - ND

EIEC (30) - - ND

EPEC (6) - - ND

ETEC (10) - - ND

E. coli026:H11 (SLT-I) (3) + + ND E. coli0157:H7(SLT-I) (4) + + ND E. coli0157:H7(SLT-II) (4) - - ND Salmonella enteritidis (30) -

-Salmonella typhi (12) -

-Salmonella choleraesuis (15) -

-Aeromonashydrophila(30) - - ND

Campylobacterjejuni(25) - - ND

Citrobacter spp. (5) - - ND

a Allsonicextracts wereexaminedwith aprotein concentration of1mg/ml. Positivity with Gb3 ELISAwasdefinedas anA490of0.1or more.Positivity with thecytotoxicity assaywas determined whenneutralizationwith anti-Shigatoxinserum wasdemonstrated.DNAhybridizationwasperformedwith aprobe forSLT-I (22),and positivitywasdeterminedby comparison with

resultsforcontrol strains. bND, Notdetermined.

' EAEC, Enteroadherent E. coli; EIEC, enteroinvasive E. coli; EPEC, enteropathogenicE. coli;ETEC, enterotoxigenicE.coli.

Cytotoxicity and neutralization assays. Cytotoxicity was quantitatively determined by overnight incubation of cell-free toxin preparations with



cells as described previously (2, 6). Attached viable cells

were countedfor radioactivity, and results were expressed

aspercentageofHeLacellsurvival relatedtocellsurvivalin control wells with no toxin. CD50was calculated from the linear regression of the percentage of HeLa cell survival

versus the logarithm of toxin dilution (2). Neutralization studies were


by preincubation of the toxin (4h,


with rabbit


serum or rabbit nonim-mune serum(2, 26) followedby determination of cytotoxic-itywith radiolabeled HeLacellsasdescribed above.

DNAhybridizationstudies. The DNAprobe forSLT-Iwas provided by J. W. Newland and R. J. Neill (Walter Reed Army Institute ofResearch, Washington, D.C.). The probe

is a 1,142-base-pair BamHI fragment of the recombinant

plasmid pJN37-19, cloned from E. coli


933J (22). As Shiga toxin and SLT-I differ in only three

nucleo-tides, the SLT-I probe also hybridizes with Shiga

toxin-producing organisms. The


was extracted from

low-temperature-gelling agarose and labeled bynick translation



o 0




0 50 100 1 50 200 250 300

Purified toxin(ng/ml)

FIG. 1. Detection ofpurified Shigatoxinby Gb3 ELISA. Serial twofold dilutions ofpurifiedtoxin were inoculated into wells(100ptl perwell)ofamicrodilution platecoated withGb3;bound toxin was detected by ELISA. Each dilution was assayedin triplicate, and resultsarethe means of three determinations. Background readings were near zero (0.000to0.003).

with ax-32P-labeled 5'-dCTP (Dupont, NEN Research Prod-ucts,Boston, Mass.). Bacterial coloniesweretransferredto

Whatman 541 filter papers, lysed as described previously (18), andhybridized under stringentconditions


form-amide, 42°C,2x SSC [lx SSC is 0.5 MNaCI plus0.015 M

sodium citrate], 18 h). On each filter the following positive

and negative control strains were placed: S. dysenteriae

serotype 1 strain 60R, E. coli 026:H11 strain H30, and E.


Statistical analysis. Neutralization studies were done in

triplicate, and the Student t test was used to examine the




frac-tions preincubated with


toxin serum or with

nonimmune serum.


Gb3 ELISA. We found that the


concentration for

coatingthe plates which gave highest



repro-ducibilitywas 10 to30




was then used in all

further assays and for evaluation of the method.


was achievedafter4h, but coated







2 weeks without


the results. Because


assays (optical density at 490 nm of 0.1 or more) had an





while the



clear with a


near zero, visualdetection of


results waspossible. The


ofthe assaywas

deter-mined by titration curve with different concentrations of


Shiga toxin


1). The assay was



detectingaslittleas0.2ngof toxin(100,ulof

2-ng/ml purified

toxin). The assay was positive with sonic extracts of S.

dysenteriae serotype 1 strain 60R (Shiga toxin), E. coli

026:H11 strain H30 (SLT-I), and E. coli 0157:H7 (SLT-I)

untiladilution ofatleast 1:256(initialconcentration,1mgof

bacterial protein per ml) was achieved. In


culture supernatants were alsoexamined and gave


results, although optic densities suggest that the amount of toxin detectedwasabout fivefold lower than that in sonicextracts.

on April 11, 2020 by guest



TABLE 3. Gb3 ELISA readingsforpurified toxins

Toxin Concn A4(,

Shiga toxin 0.5 1.394

Choleratoxin 1 0.041

E. coliheat-labile enterotoxin 1 0.040

E.coliheat-stableenterotoxin 1 0.013

C.difficile cytotoxin 1 0.014

Specificity of theassay was determined by several

exper-iments. Examination of other purified toxins (at a

concen-tration of1,ug/ml) showed nocross-reactivity with cholera

toxin,E.coliheat-labile and heat-stable enterotoxins, and C.

difficile cytotoxin (Table 3). Examination ofother bacterial

sonic extracts (1 mg of bacterial protein perml) showed no

cross-reactivity with sonic extracts ofE. coli C600 (933w)

(which produces SLT-II) or with the parent strain E. coli C600 (non-cytotoxin producing) (Table 2). In addition, we

examined sonic extracts (at a concentration of 1 mg of

bacterial protein per ml) of a variety of other

cytotoxin-producing bacteria, includingothershigellae,pathogenic and

nonpathogenic E. coli, and strainsof Salmonella,

Aeromo-nas, Campylobacter, and Citrobacter. None of these sonic

extracts showed positive results with the Gb3 ELISA.

Correlationof Gb3 ELISA with HeLa cell cytotoxicityassay

and with DNA hybridization studies. Quantitative correlation

with the cytotoxicity assay wasdetermined by examination

offractions obtained atvarious stagesofShiga toxin

purifi-cation with the two methods simultaneously. Increased

specific activity


per milligram of bacterial protein)









10~~~0 10 1

jgtoxin (Gb3-ELISA)

FIG. 2. Correlation between specific activity and Gb3 ELISA readings of fractions obtained at various stages of Shiga toxin purification. Specific activity was determined with radiolabeled

HeLacell assay and expressed as


per milligram ofprotein.

Readings of Gb3 ELISA


were expressed as micrograms of

purified Shiga toxin, obtained from the titration curve of purified

toxin. Thefractions examined were(dots from leftto right) crude

sonicate ofS. dysenteriae 1 strain 60R, cytotoxic fraction eluted fromAffi-Gel Blue column, cytotoxic fractioneluted by chromato-focusing, and purified Shiga toxin. A high correlation was found

between specific activity and Gb3 ELISA readings (r = 0.99, P<


during purification of the toxin resulted in simultaneous increased readings in the Gb3 ELISA and higher amountsof toxindetected (Fig. 2). Regression analysis indicated a high correlation between the two methods (r = 0.99, P < 0.01). Theeffect of heat treatment(100°C,30min) onpurifiedShiga toxin and sonic extracts of E. coli 026:H11 strain H30 (SLT-I) was examined. Heating caused inactivation of most cytotoxicactivityin the HeLacell assay (Fig. 3B) and atthe

same time reduced the absorbance in the Gb3 ELISA

method to near zero (Fig. 3A). In addition, when various bacterial sonic extracts were examined, all extracts positive for Shiga toxin or SLT-I in the cytotoxicity assay (neutral-ized by rabbit antiserum to Shiga toxin) or hybridized with DNA probe for SLT-I were also positive in the Gb3 ELISA. Bacterial sonic extracts negative in the cytotoxicity assay (no neutralization with an anti-Shiga toxin serum) and in DNA hybridization assays were always negative in the Gb3 ELISA (Table 2).


Shiga toxin and SLTs are currently diagnosed by biologic assays. Threebiologic activities have been ascribed to these cytotoxins: lethality in mice, fluid accumulation in ligated rabbit ileal loops, and toxic activity in several cell lines (8). Of these, the assays most frequently used to detect specific activity are based on cytotoxicity in HeLa or Vero cells. These assays are complicated, cumbersome, and time-con-suming, and they require facilities for and experience with tissue cultures. They are usually based on cytotoxin-related morphologic changes of the cells or on radiolabeled assays. Identification of the specific cytotoxin also requires neutral-ization studies with the appropriate antiserum. In addition, the sensitivities of various celllinesto cytotoxic effects vary significantly (23). ELISA methods that use monoclonal antibodies to detect these cytotoxins have been described elsewhere (7, 14).

Recent data have showed that Shiga toxin, SLT-I, and SLT-II share the same binding receptor, Gb3 (10, 16, 31). Because these well-defined toxins are emerging as the most important and clinically relevant cytotoxins produced by members of the family Enterobacteriaceae, we have devel-oped a rapid diagnostic assay based on specific binding to this receptor; bound Shiga toxin orSLT-I was then detected immunologically. Gb3 was obtained from the manufacturer inchloroform-methanol (2:1) solution. The Gb3 was difficult to getinto aqueous solution. Organic solvents were therefore used, and a special method for coating onto microdilution plates was necessary. A similar principle has been used to detectheat-labile enterotoxins by using their specific binding to the GM1 receptor (30). However, since GM1 is water soluble, it has been diluted in PBS, and a standard procedure hasbeen used forcoating.

Thepresent study shows that the Gb3 ELISA is sensitive andspecific, detecting nanogramquantities of purified Shiga toxin. Sonic extracts ofSLT-I-producing organisms (E. coli 026:H11 and E. coli 0157:H7) were also positive in this assay. Production ofSLT-I by these organisms was verified both byneutralization studies with specific antiserum and by DNAhybridization. The specificity of the assay was high, in that nocross-reactivity was found with SLT-II or cytotoxins produced by a variety of other organisms, including other Shigella and E. coli strains andSalmonella, Campylobacter, and Aeromonas strains. Lack of SLT-I production was

demonstrated in all of these strains by neutralization studies withspecific antiserum in a HeLacell assay, and in most of

on April 11, 2020 by guest



c 1.2-o>




.0 o .0





0M Shiga


1 100° C


80-> 60


0 40 o

0> 2J

0 20 B



|- SLT-I





FIG. 3. Effect of heat treatment (100°C, 30min) of purified Shiga toxin (0.5,ug/ml) and sonicextractofE.coli026:H11 strain H30(1 mg ofbacterial protein per ml) on specific activity and on Gb3 ELISA readings. (A) Gb3 ELISA readings(absorbance). Readingswerereduced to nearzero by heat treatment. (B) HeLa cell survival. Cell survival (as determinedinradiolabeled assay)indicatedhigh cytotoxic activities of bothpreparationsthat were nearlycompletely inactivated by heattreatment.Resultsarepresentedas means+standarddeviationsof three


them, SLT-I production was also demonstrated by DNA

hybridization studies. Quantitative correlation with the

ra-diolabeled HeLacell assay was demonstrated by examina-tionoffractions at the various stages of toxinpurificationby

both methods. Because Gb3 is also thebinding receptor of SLT-II (31), the latter toxin might also be detected by the samemethodbyappropriate monoclonal antibodiesto SLT-II. It would be very helpful if plates precoated with Gb3

could be used to detect the main cytotoxins, with specific identification of bound toxins by useofdistinct antibodies.

The presentGb3ELISA hasobviousadvantages over the HeLa or Verocell assays: (i)itis inexpensive, simple, and

rapid,canbeperformedin oneworkingday, and canbeused

by relatively inexperienced personnel; (ii) there is no need forexperience with orfacilitiesfor tissuecultures; (iii)Gb3, themonoclonal antibodies (American Type Culture

Collec-tion, cell line 13C4), and the second antibodies (rabbit anti-mouse immunoglobulin G) arecommercially available;

and(iv) objective results areobtained (opticaldensityat490

nm) rather than subjective interpretation of morphologic changesin cell lines.

The Gb3 ELISAseemscomparable to the ELISA (7, 14). The methods described used anti-Shiga toxin monoclonal

antibodies (7) or antibodies purified by immunoaffinity col-umn chromatography (14) and then detection by enzyme

immunoassay. The sensitivities of the assays were 0.06

ng/ml (7) and about 10 ng/ml (14), and the assays were specificwhen tested with othertoxins. However, the entire

Gb3 ELISA canbeperformedin oneworkingday, theassay is based oncommercially availablematerials, and it has the

potential advantage of detecting additional cytotoxins that

bind toGb3 astheir naturalreceptor.

Any rapid and simple method to detect this group of

cytotoxins is ofprimeimportance,asitmay be also used by

less-well-equipped laboratories, for example, in developing

countries. This may significantly simplify epidemiologic

studies of cytotoxin-related infections and increase our

understandingof the role of these toxins in diarrheal disease and HUS.


We thank David Newburg for helpful suggestions, Barbara E. Murray for help with DNA hybridization, and Anne Wright for preparation of themanuscript.

This work was supported by Public Health Service grant HD-13021 from the National Institute of Child Health and Human Development andbyagrantfrom the Thrasher ResearchFund.


1. Ashkenazi, S., T. G. Cleary, B. E. Murray, A. Wanger, and L. K.Pickering. 1988. Quantitative analysis andpartial charac-terizationofcytotoxinproductionby Salmonella strains. Infect. Immun.56:3089-3094.

2. Bartlett, A. V.,HI,D.Prado, T. G.Cleary,and L. K.Pickering. 1986. ProductionofShiga toxin and other cytotoxins by sero-groupsofShigella. J. Infect. Dis. 154:996-1002.

3. Bartlett,T. G. 1979. Antibiotic-associated pseudomembranous colitis. Rev.Infect.Dis. 1:530-539.

4. Carter, A. O., A. A. Borczyk, and J. A. K. Carlson. 1987.A severeoutbreak of Escherichia coli0157:H7-associated hemor-rhagic colitis in anursing home. N. Engl. J. Med. 317:1496-1500.

5. Cleary, T. G. 1988. Cytotoxin-producing Escherichia coliand hemolytic uremic syndrome. Pediatr. Clin. North Am. 35: 485-501.

6. Cleary, T.G., J. J. Mathewson, E. Faris, and L. K. Pickering. 1985. Shiga-like cytotoxin production by enteropathogenic Escherichia coli serogroups. Infect. Immun. 47:335-337. 7. Donohue-Rolfe,A., M. A. Kelley, M.Bennish, and G. T. Keusch.

1986.Enzyme-linked immunosorbentassayforshigella toxin.J. Clin. Microbiol. 24:65-68.

8. Eiklid, K., and S. Olsnes. 1983. Animal toxicity ofShigella dysenteriae cytotoxin: evidence that the neurotoxic, entero-toxic,andcytotoxicactivities are due to onetoxin.J. Immunol. 130:380-384.

9. Fuchs, G.,M. Mobassaleh, A.Donohue-Rolfe,R. K. Montgom-ery, R. J. Grand, and G. T. Keusch. 1986. Pathogenesis of shigella diarrhea: rabbit intestinal cell microvillus membrane binding site forshigellatoxin. Infect. Immun.53:372-377. 10. Jacewicz, M., H. Clausen, E.Nudelman,A.Donohue-Rolfe,and

G. T. Keusch. 1986. Pathogenesis of Shigella diarrhea. XI. Isolation of a Shigella toxin-binding glycolipid from rabbit jejunumand HeLacells and its identification as



on April 11, 2020 by guest



ceramide. J.Exp. Med. 163:1391-1404.

11. Jackson, M. P., J. W. Newland, R. K. Holmes, and A. D. O'Brien. 1987. Nucleotide sequenceanalysis of the structural geneforShiga-like toxinIby bacteriophage933J from

Esche-richiacoli. Microb.Pathog. 2:147-153.

12. Karmali,M.A., M.Petric,C. Lim, P. C.Fleming, G. S. Arbus, and H. Lior. 1985. The association between idiopathic hemo-lytic uremic syndrome and infection by Verotoxin-producing Escherichiacoli. J.Infect. Dis. 151:775-782.

13. Kindschuh, M., L. K. Pickering, T. G. Cleary, and G. Ruiz. Palacios. 1987. Clinical and biochemical significance of toxin production byAeromonas hydrophila. J. Clin. Microbiol. 25: 916-921.

14. Kongmuang, U., T. Honda, and T. Miwatani. 1987. Enzyme-linked immunosorbentassayto detectShiga toxin of Shigella dysenteriae1and related toxins.J.Clin. Microbiol.25:115-118. 15. Konowalchuk, J., J. I. Speirs, and S. Stavric. 1977. Vero responsetoacytotoxin ofE. coli. Infect. Immun. 18:775-779.

16. Lindberg, A. A.,J. E. Brown,N.Stromberg,M.Westling-Ryd, J. E. Schultz, and K.-A. Karlsson. 1987. Identification of the carbohydrate receptor for Shiga toxin produced by Shigella dysenteriaetype1. J. Biol. Chem.262:1779-1785.

17. Lingwood, C., H.Law,S. Richardson, M.Petric,J. Brunton, S. deGrandis, and M.Karmali. 1987.Glycolipid binding of purified andrecombinant Escherichia coliproduced verotoxinin vitro. J. Biol.Chem. 262:8834-8839.

18. Maas, R.1983. Animproved colony hybridization method with significantly increased sensitivityfordetectionof singlegenes.


19. Marques, L. R. M., M. A. Moore, J. G. Wells,I.K.Wachsmuth,

and A. D. O'Brien. 1986. Production of Shiga-like toxin by Escherichia coli. J. Infect. Dis. 154:338-341.

20. Neill, M. A., P. I. Tarr,andC. R. Clausen. 1987. Escherichia

coli0157:H7asthepredominant pathogen associated with the

hemolytic uremic syndrome: aprospective studyin thePacific Northwest.Pediatrics 80:37-40.

21. Newland, J. W., and R. J. Neill. 1988. DNAprobes for Shiga-like toxinsIand Iland fortoxin-converting bacteriophages. J.

Clin. Microbiol.26:1292-1297.

22. Newland,J. W.,N.A.Strockbine, S. F.Miller,A. D. O'Brien,

andR.K.Holmes. 1985. Cloning of Shiga-like toxinstructural genesfrom atoxin converting phage of Escherichia coli. Sci-ence230:179-181.

23. O'Brien,A. D., and R. K. Holmes. 1987. Shiga and Shiga-like

toxins.Microbiol. Rev. 51:206-220.

24. O'Brien, A. D., J.W.Newland, S. F. Miller, R. K.Holmes,H. W.Smith, and S. B. Formal. 1984. Shiga-liketoxinconverting

phages from Escherichia coli strains that cause hemorrhagic colitisorinfantile diarrhea. Science 226:694-696.

25. Perera, L. P., L. R. M. Marques, and A. D. O'Brien. 1988. Isolation and characterization of monoclonal antibodies to Shiga-like toxin Il of enterohemorrhagic Escherichia coli and use of the monoclonal antibodies in acolony enzyme-linked immunosorbent assay. J. Clin. Microbiol. 26:2127-2131. 26. Prado, D.,T. G.Cleary,L. K.Pickering, C. D. Ericsson,A. V.

BartlettIII,H. L.DuPont, and P. C.Johnson. 1986. Therelation betweenproduction of cytotoxin andclinical features of shigel-losis.J. Infect. Dis. 154:149-155.

27. Riley,L.W.,R.S. Remis,S. D.Helgerson,H. B.McGee, J.G. Wells,B. R.Davis,R.J. Herbert,E. S.Olcott,L. M.Johnsons, N. T. Hargrett, P. A.Blake, and M. L. Cohen. 1983. Hemor-rhagic colitis associated with a rare Escherichia coli serotype. N.Engl. J. Med.308:681-685.

28. Strockbine, N.A.,L. R.M.Marques, R.K.Holmes,and A. D. O'Brien. 1985. Characterization of monoclonal antibodies against Shiga-like toxin from Escherichia coli. Infect. Immun. 50:695-700.

29. Strockbine, N. A., L. R. M. Marques, J. W. Newland, H. W. Smith, R. K. Holmes, and A. D. O'Brien. 1986. Two toxin-converting phages from Escherichia coli 0157:H7 strain 933 encode antigenically distinct toxins with similar biologic activ-ities. Infect. Immun. 53:135-140.

30. Svennerholm, A.-M., and G. Wiklund. 1983. Rapid GM1-en-zyme-linked immunosorbentassaywithvisualreading for iden-tification of Escherichia coli heat-labile enterotoxin. J. Clin. Microbiol. 17:596-600.

31. Waddell, T., S. Head, M.Petric, A. Cohen, and C. Lingwood. 1988. Globotriosyl ceramide is specifically recognized by the Escherichia coli Vero cytotoxin 2. Biochem. Biophys. Res. Commun. 152:674-679.

on April 11, 2020 by guest



Related documents

Although the entire male age groups exhibit the same major cause patterns of farming injury as female for overexertion and strenuous movements, cutting or piercing instrument or

It is observed that the bio-inspired algorithm has outperformed AODV on different evaluation criteria such as average end to end delay, throughput, average jitter,

KEYWORDS: Electric vehicles (EVs), Bipolar DC bus architecture; Neutral point clamped (NPC) converter, three level DC-DC converter, Comprehensive DC power

The proposed spiral slot is compared with conventional rectangular slots and exhibits comparable performance in terms of total efficiency (representing coupling from waveguide mode

In the globalization era, international trade and commerce is the main pillar for rapidly growth and development for an economy...

Field crop production is the most prevalent source of livelihood in the study area, indicated by 35 (or 97 per- cent) of the 36 focus groups (Column 10), followed by livestock

Considering the above mentioned, the objective of this study was the morphometric characterization of two sparrow species of the genus Ammodramus during the wintering season

The annual natural unemployment rate’s growth and annual potential GDP’ growth rate for Romania during 1992-2004 can not be estimated because the economic evolution of