Broadsheet 135 J 1993 Isolation identification methods ... · VTEC. Some patients have features resem-bling thrombotic thrombocytopenic purpura (TTP), such as neurological disorders

J Clin Pathol 1993;46:10-1710

ACP Broadsheet 135 January 1993

Isolation and identification methods forEscherichia coli 0157 and otherVero cytotoxinproducing strains

H R Smith, S M Scotland

IntroductionStrains of Escherichia coli that produce a

cytotoxin active on Vero cells, termed Verocytotoxin (VT) or Shiga-like toxin, are now

recognised as important aetiological agents ofdiarrhoeal diseases in man and animals. Twomain types of VT, VT1 and VT2, have beendefined. The properties of Vero cytotoxinproducing E coli (VTEC) and infection bythese organisms have been reviewed recently. 2

Clinical featuresIn man VTEC are associated with diarrhoea,haemorrhagic colitis, and haemolytic uraemicsyndrome (HUS). Haemorrhagic colitis ischaracterised by grossly bloody diarrhoea,usually without pyrexia. It is frequentlypreceded by abdominal cramps and waterydiarrhoea. HUS is defined by three clinicalfeatures acute renal failure, microangiopathichaemolytic anaemia, and thrombocytopenia.HUS can occur in all age groups but is morecommon in infants and young children and is a

major cause of renal failure in childhood.There are two subgroups of HUS-typicalHUS associated with a prodromal bloodydiarrhoea and an atypical form without a

diarrhoeal phase. It is only the typical form ofHUS that is associated with infection byVTEC. Some patients have features resem-

bling thrombotic thrombocytopenic purpura

(TTP), such as neurological disorders andfever.

This Broadsheet has beeniprepared by the authors at theinvitation of the Association ofClinical Pathologists whoreserve the copyright. Furthercopies of this Broadsheet niaybe obtained froni thePublishing Manager, Journalof Clinical Pathology, BAJAHouse, Tavistock Square,London WClH 9JR

Laboratory of EntericPathogens, CentralPublic HealthLaboratory, ColindaleAvenue, London NW95HTH R SmithS M ScotlandAccepted for publication24 February 1992

EpidemiologyDisease associated withVTEC has been docu-mented in many countries, with most reportscoming from North America and Britain. Mosthuman infections are associated with strainsbelonging to serogroup 0157, usually possess-

ing the flagellar antigen H7, although some

non-motile 0157 strains have been isolated.These cases are found over a wide geographicalarea with the infections peaking in summer andautumn. Outbreaks of 0157 VTEC haveoccurred in the community, in nursing homesfor the elderly, hospitals and in daycare centresfor young children.' The most severe clinicaldisease is usually seen in children and theelderly and a significant mortality has beenobserved in several outbreaks. Outbreaks have

been associated with beefburger meat andunpasteurised milk. 0157 VTEC have beenisolated from cattle in several countries includ-ing Britain,2 suggesting that cattle are a reser-voir of infection in man. Examination of meatsin North America showed that 0157 VTECwere present in samples of beef, pork, poultryand lamb but, so far,VTEC of serogroup 0157have not been isolated from food in Brit-ain.

Laboratory diagnosisThe methods used to provide evidence ofVTEC infection can be divided into thefollowing categories: isolation of VTECincluding E coli 0157; demonstration of spe-cificVero cytotoxin; and presence of antibodieswith VT neutralising ability or antibodies to Ecoli 0157 lipopolysaccharide (LPS). Techni-ques for the isolation of VTEC and partic-ularly strains of serogroup 0157 and for thedetection ofVT are described in detail with a

brief description of methods primarily used in

the reference laboratory.

SPECIMENS

The appearance of the faeces is variable, fromsoft specimens to frank blood. The absence ofred blood cells or the presence of leucocytes infaeces does not exclude VTEC infection. Thecollection and testing of early faecal specimensis most important. Previous studies haveshown that 0157 VTEC are rapidly clearedfrom the gut with a low recovery of theorganism from specimens tested seven or moredays after the onset of symptoms.3

Tests for 0157 VTECMost VTEC belonging to serogroup 0157 donot ferment sorbitol within 24 hours of incuba-tion whereas about 95% of E coli from faecalsamples are prompt sorbitol fermenters underthese conditions.4 However, in a recent reportfrom Germany 0157 VTEC strains fermentedsorbitol within 24 hours.5 The prevalence ofsuch strains is unknown but they would bediscarded using the standard screening methoddescribed here.MacConkey agar plates with 1% D-sorbitol

instead of lactose are used to screen faecalspecimens for non-sorbitol fermenting (NSF)

on May 15, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.46.1.10 on 1 January 1993. Dow

nloaded from

http://jcp.bmj.com/

Isolation and identification of vero cytotoxin producing strains

colonies of E coli. These colonies are thentested for agglutination with an 0157 anti-serum or with an 0157 latex agglutinationkit. Any colonies that give agglutination mustbe confirmed as E coli. NSF strains of serotype0157:H16 that do not produce VT have beenreported.6 In view of these findings it isrecommended that presumptive 0157 isolatesshould be sent to a reference laboratory andtested for VT production and for flagellarantigens.

PROCEDURE

Faecal specimens are resuspended in an equalvolume of phosphate buffered saline (PBS).Suspension of the sample may not be requiredfor liquid specimens, but in some cases faecalsamples require dilution up to 10 fold. Wherepossible up to 1 g of faeces should be exam-ined. Samples are streaked out on to sorbitolMacConkey agar (SMAC) and the plates areincubated at 37°C for 18 hours. VTEC ofserogroup 0157 produce NSF colonies that aresmall, round, smooth and may look greyish (fig1). Each colonial type of NSF colony isselected for testing as possible E coli 0157; atotal of five to 10 colonies is usually exam-ined.

Colonies are emulsified in a drop of saline ona glass slide and then mixed with one drop of0157 antiserum or 0157 specific latex (seeAppendix). The slide is rocked and observedfor agglutination; a positive test is indicated byrapid clumping and clearing of the solution.False positive results should be excluded bytesting colonies with a non-0157 antiserum.Some commercial kits have a negative controlreagent.

CONFIRMATION

NSF colonies that agglutinate with the 0157antiserum must be confirmed as E coli using aset of biochemical tests. These tests shouldexclude any non-E coli that give false positiveagglutination tests with the 0157 antiserum.Escherichia hermanii is biochemically and sero-

logically similar to E coli 01577 and is positivein most screening tests for E coli 0157,although at least one latex reagent (Pro-LabDiagnostics) does not cross-react with E her-manii. However, E coli, unlike E hermanii, doesnot ferment cellobiose and does not grow inthe presence ofpotassium cyanide. Strains ofEhermanii producing Vero cytotoxin have notbeen identified. Another characteristic of 0157VTEC is that, with very few exceptions, theydo not produce fi glucuronidase, whereas mostother E coli are positive in this test.3 5 8 Testingfor ,B glucuronidase activity has therefore beenproposed to aid the identification of 0157VTEC.8 The hydrolysis of 4-methylumbelli-feryl-fl-D-glucuronide (MUG) to produce afluorescent compound is a commonly usedmethod and can be performed by differenttechniques. The preferred method in our labo-ratory9 is as follows:The MUG reagent is prepared by diluting100 mg MUG (Sigma) in 100 ml distilledwater containing 2 drops of Triton X-100(Sigma). The solution is filter sterilised(0 45,m Minisart MNL; Sartorius) andstored at 4°C. Drops of the MUG reagent areapplied to circles of filter paper (Whatman No1) in Petri dishes to obtain even distribution ofthe reagent, and air dried. These dishes arewrapped in aluminium foil and refrigerateduntil used. Cultures for testing are grown onblood agar plates (Oxoid No 2 nutrient brothcontaining 5% vol/vol unwashed horse bloodsolidified with 2% agar). Growth from a singlecolony is applied to the paper to form a smallpatch. At least 16 strains can be tested on each8-5 cm paper. The growth is moistened with adrop of saline near each patch and the plate isincubated at 37°C for 20 minutes. The papersare examined for fluorescence using ultravioletlight in a darkened room (fig 2).

IMPROVED MEDIUM FOR THE ISOLATION OF 0157VTEC

A recent paper has reported the inclusion ofcefixime and rhamnose into sorbitol MacCon-

t"s ~J

0_

Figure 1 Sorbitol MacConkey agar plate showing colourless colonies ofE coli 0157.

Figure 2 Paper method to test glucuronidase activity of strains of Ecoli showingfluorescence of 4-methylunmbelliferyl-1/ D-glucuronide(MUG).

..A..W

I11

I* 0



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

Smith, Scotland

key agar (CR-SMAC) to improve the detectionofVT producing E coli 0157.10 Using tests onagar plates, 0157 VTEC do not fermentrhamnose, whereas 60% of non-sorbitol fer-menting E coli belonging to other serogroupsferment rhamnose. In contrast to these resultson agar, 0157 VTEC do ferment rhamnosewithin one day in standard tube sugar fermen-tation tests.3'" Cefixime is included as it ismore active against Proteus spp than against Ecoli and Proteus strains account for about 15%of NSF. The use of CR-SMAC rather thanSMAC showed a significant improvement inselectivity for the isolation of 0157 VTEC.'° Ina study of 1763 samples 397 required testingfor E coli 0157 using SMAC but only 176needed to be investigated using CR-SMAC. Inthe same study strains ofE hermanii fermentedrhamnose and were therefore not confusedwith 0157 VTEC.

Cefixime (Cyanamid, Gosport, Hants) isdissolved in ethanol (25 mg in 10 ml) anddiluted further in broth. The CR-SMACmedium is prepared by adding cefixime (0 05mg/i) and 0 5% rhamnose to sorbitol Mac-Conkey agar. Faecal samples are streaked outon to the medium and the plates are incubatedat 37°C overnight. Non-fermenting coloniesare selected and tested as described above withan 0157 antiserum or 0157 latex reagent.A special agar medium containing sorbitol

and MUG has been developed (see Appendix).A disadvantage of using a medium containingMUG is that it can be difficult to identify aglucuronidase negative colony when it is sur-rounded by a number of fluorescing colonies.

COLONY IMMUNOBLO-llNGAnother method used for testing for presenceof E coli 0157 is that of colony immunoblot-ting. Bacterial growth to be tested can be in theform of spotted cultures in a grid pattern on anagar plate, streaked out growth on a plate or aplate with several hundred separated coloniesfrom a single sample (GA Willshaw, personalcommunication).An 82 mm diameter membrane (Hybond C

extra; Amersham) is marked with an orienta-tion line and placed on the plate with bacterialgrowth for 5 minutes. The membrane isrubbed gently with a glass spreader to ensuregood contact with the surface and then peeledoff and placed in a plastic box containing PBSwith 5% dried milk. This is shaken for 30minutes at room temperature. This blockingbuffer is poured off and discarded by autoclav-ing. The alkaline phosphatase conjugated 0157antiserum (Kirkegaard and Perry Laborator-ies, Gaithersburg, Maryland, USA) is diluted 1in 1000 with blocking buffer. For eight mem-branes, 20 ml of diluted antibody reagent isprepared. The membranes are transferred to aplastic bag so that the sides with bound antigenface outwards. The diluted antibody reagent isadded and the bag is sealed and incubated withshaking for two hours at room temperature.The membranes are removed and washedthree times for 10 minutes in PBS at roomtemperature. After rinsing the membranes in

alkaline phosphatase buffer (0 1MTRIS-HC 1,01M NaCl, 0-05M MgCl2) the substrate isadded: 69,ul nitroblue tetrazolium chloride(Sigma Chemical) and 54 ,ul of 5-bromo-4-chloro-3-indolyl phosphate (NorthumbriaBiologicals, Cramlington, Northumberland)per 15 ml of alkaline phosphatase buffer foreight membranes. A positive control gives adark blue coloration in 5 to 10 minutes andsamples or spots that react similarly mayindicate the presence of E coli 0157. Thereaction is stopped after 20 minutes by wash-ing the membranes in distilled water. Themembranes are blotted dry and stored in thedark. Any colonies that appear to be E coli 0157must be confirmed biochemically and sero-logically (see above) and tested for VT pro-duction.A recent paper has described a rapid sand-

wich enzyme linked immunosorbent assay(ELISA) for the detection of 0157 VTEC infood.'2 In this test a polyclonal 0157 antibodyis used as the capture antibody and a mono-clonal antibody, specific for VTEC belongingto serogroups 0157 and 026, as the detectionantibody. The reagents will be available in kitform in the near future.

Tests for presence of Vero cytotoxin (VT)Vero cytotoxin present in faeces can be detec-ted directly by its cytotoxic effect on Vero cells.Alternatively, VT production by E coli strainsisolated from the faeces can be examined,either by testing culture filtrates or, in a rapidscreening method, by testing live cultures. Insome early studies of VT production bacteriawere grown in iron restricted media and suchgrowth conditions may increase production ofVT1, but not VT2. For routine testing, con-centrations of VT are adequate in ordinarybroth media as described below. A consider-able amount of VT is not liberated into themedium but remains cell bound. It can bereleased by sonication, with the use of a Frenchpress or by polymyxin treatment, and thesetechniques have been used for preparing largequantities of VT. To detect VT production byclinical strains these techniques are not neces-sary. However, a method using polymyxinextraction for the screening of sweeps ofcolonies from faeces for VT production isdescribed. Some workers have used HeLa cellsfor the detection of VT but this cell linecannot be recommended as variants of VT2are known that do not have a cytotoxic effecton HeLa cells.

Several ELISAs have been described for thedetection of VT but as the reagents are notcommercially available the methods are notdescribed in detail here. Some ELISAs bindVT to glycolipids containing a terminala-D-Gal-(l -+4)-D-Gal; purified globotriosylceramide (Gb3), lyso-Gb3, and hydatid cystfluid have been used.'3 14 Other ELISAs haveused monoclonal antibodies againstVT to bindthe toxin initially. 5 In both assay systemsbound toxin is then detected using monoclonalor polyclonal antiserum against VT, followedby an appropriate alkaline phosphatase-

12



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/


labelled goat antiserum and then the enzymesubstrate. In general, these tests have notproved to be as sensitive as the Vero cell test. Inaddition, as the toxins show considerablevariation in their antigenicity and bindingproperties (even within theVT2 class of toxins)care must be taken in the choice of reagents ifthe aim is to detect all VT producing strainsfrom a clinical specimen.

TESTS WITH CULTURE SUPERNATANTS

The method used is a modification'6 of theoriginal method of Konowalchuk et al.'7

Preparation of culture supernatantTrypticase soy broth (BBL, Becton Dickinson,Cowley, Oxfordshire), 10 ml, in a 250 ml flaskis inoculated with the bacterial strain to betested. The flask is incubated at 37°C for18-24 hours with shaking (120 rpm). Theculture is then centrifuged (17 000 x g for 30minutes), with cooling if possible, and thesupernatant fluid sterilised by filtration using apore size of 0-45,um. This filtrate is useddirectly for-VT tests. A sample of the culturesupernatant is also heated at 100°C for 15minutes.

Vero cell testMonolayers of Vero cells are prepared in96-well tissue culture plates. A 25 ,ul portion ofthe test filtrate is added to duplicate wellswithout changing the medium. Control fil-trates should also be included. A sterility checkof the filtered supernatant is useful as growthof any remaining bacteria will result in death ofthe monolayer. After the addition of the filtratethe plates are covered and incubated at 37°C.

Cells round up and become detached in the.presence of VT. The monolayers can beexamined after one day for early cytotoxiceffects using an inverted microscope (figs 3Aand C). However, final readings should bemade after the plates have been incubated forthree to four days (compare figs 3C and 3E).For ease of viewing and to obtain a permanentrecord the monolayer can be stained (comparefigs 3A and 3B, 3C and 3D, and 3E and 3F).The tissue culture medium is removed and thecells fixed with methanol for 5 minutes. Themethanol is removed and Giemsa stain (5%w/v in phosphate buffer) added. After 45minutes the monolayer is washed three timeswith distilled water and air. dried. The cells areexamined microscopically.To obtain toxin titres two-fold or five-fold

dilutions in tissue culture medium can betested. Toxin titres are expressed either as thehighest dilution that causes any cytotoxic effectin the monolayer or more usually as thedilution at which 50% detachment of cells inthe monolayer occurs. A unit ofVT is definedas the amount present in this dilution.

TEST USING FAECAL PREPARATION

Faecal specimens can be examined directly forthe presence of VT.'8 The specimen is cen-trifuged; a microcentrifuge is suitable using ascrew capped tube for the sample. The super-natant fluid is filtered and tested for VT as

described above for culture supernatants. Theaddition of PBS to the specimen is oftennecessary, followed by thorough mixing, toobtain a sufficiently liquid sample. This dilu-tion is considered in the final estimation of anytoxin titre.

TEST OF COLONY SWEEPS AFTER

POLYMYXIN EXTRACTION

As an altemative to testing individual colonies,broth inoculated with sweeps-that is, loopfulsof confluent bacterial growth from agarmedia-can be examined. To increase thesensitivity of the test it is recommended thatVT obtained after polymyxin release is deter-mined."' For this, sweeps from MacConkeyagar plates are inoculated into 20 ml volumesof Penassay broth (antibiotic medium No 3:Difco Laboratories) and incubated for 5 hoursat 37°C with shaking. This broth culture iscentrifuged at 10 000 x g for 10 minutes andthe supernatant fluid discarded. The cell pelletis resuspended in 1 ml PBS (Dulbecco A;Oxoid) containing polymyxin B (0-1 mg/mland incubated at 37°C for 30 minutes. Thesuspension is centrifuged at 10 000 x g for 10minutes and the supernatant fluid is filteredand tested for VT as described above. Whenartificial mixtures of known VT positive andVT negative colonies were tested in this wayVT could be detected when the proportion ofVT positive organisms was at least 1 25%. Thespecificity of the toxin can be determined asdescribed below.

If faeces are screened by this method, onlythose that are positive need to be examinedfurther, if necessary, for VT producing colo-nies.

SPECIFICITY OF THE TOXIN

To confirm that cytotoxic effects on Vero cellsare indeed due to the presence of VT (orVTs) additional tests are needed. These areparticularly important in the examination ofVT in faeces. Preferably, neutralisation testsusing antisera against VT1 or VT2 should beused, but these antisera are not commerciallyavailable. VT tests on a portion of the sampleheated at 100°C for 15 minutes should benegative, confirming the heat lability of thetoxin. The specimens can also be tested on acell line which is not sensitive toVT, such asYlmouse adrenal tumour cells.'8 A faecal prepa-ration that killsYl cells to the same titre asVerocells has to be considered negative for VT,which if also present at a lower titre would notbe detected in the presence of the otheruncharacterised toxin. Vero cells are also sensi-tive to the heat labile enterotoxin (LT) pro-duced by some strains of E coli. LT is notcytotoxic but causes rounding of the cells after24 hours. ' This is unlikely to be mistaken forVT but as LT also causes rounding of Y1cells this allows the toxins to be clearlydifferentiated.To perform neutralisation tests'8 25 ,ul volu-

mes of appropriate dilutions of the prepara-tions to be tested are added to 0-2 ml volumesof complete Vero cell tissue culture medium ina 96-well plate. At least two series are prepared

13 on M

ay 15, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.46.1.10 on 1 January 1993. D

ownloaded from

http://jcp.bmj.com/

Smith, Scotland

,. i,.:: ' ..

:iw.;.; 2.:

.: X Y'* . :, . ;, t,. . , . ''° . .: ........ S'.... ':.e e '.; ,'8iF ...... ;A >; ;'_ ^ > % j ,Xlggi

*e .sU;.. ....,. Fi* . ... .. :. .. : ;.'.e. 'S i ......... '.,.X.,' . '. ';;* .. .... . 59. ; :-e i :: : .i .... .n:. 2' ':.' .: !e w. :: .............. ,!: .: .: :.*. ..... .. } ......... .... ' , ' i. . ; . i: ! s. ..

.. .: :e* . . .:: ": .. . : . . ...... .. ,j: ,,< . ' Zi.

:} .: ::

.: '. :, .:

* S.:. :. :. . .. :

.. .... SS ., . . .. . .. s ..

.... '_: .i . '.':

o ...... ...... .i.;:. S :':

' .. ,.9:.t . Z X .. ... ,: ::';: :'

.: .. {

,.* : . x . : . .i.

v '::

... ..

*~ ~ ~ ~~~~ ~ ~ ~ ~ ~~~~....

i. 4

* *, ew t$ .!.W .

A}its Ai E+t+*i'.P .,

, ;.&< % #

W. Z

*:o

4:

Figure 3 Tests of culture filtrates ofVT+ or VT- strains of E coli on Vero cell monolayers. These were examined byphase contrast microscopy or after staining at the times indicated. 3A VT- 1 day unstained; 3B VT- 1 day stained;3C VT+ 1 day unstained; 3D VT+ 1 day stained; 3E VT+ 4 day unstained; 3F VT+ 4 day stained.

so that one can be kept as a control withoutantiserum. Anti-VT serum (25 ,l) is added toeach well of a series. This can be anti-VT 1,anti-VT2, or both together. The plates areincubated at 37°C for one to three hours andthen placed at 4°C overnight. The medium iscarefully removed from a monolayer of Verocells that has been prepared earlier and thecontents of the neutralisation wells transferredinto them. The plates are incubated and read asabove. The dilution of antiserum used isdetermined in preliminary experiments so as tocontain at least 20 units. A unit is defined asthe amount present in the highest dilution ofantiserum which neutralised 1-5 units ofVT.

20

SIMPLIFIED TEST USING LIVE BACTERIAL CULTURES

A monolayer of Vero cells growing in a96-well tissue culture plate is prepared. Thecolonies to be tested are grown in 05 mltrypticase soy broth at 37°C for 18 hourswithout shaking. The tissue culture medium is

carefully removed from the cells. Live bacterialculture (50 pl) is added to a test well. After5-10 minutes the medium and bacteria areremoved and the cells washed once with PBS.Tissue culture medium (0-2 ml) with pen-icillin, streptomycin, and gentamicin (finalconcentration of 40 ,ug/ml) is then added. Theplate is covered again, incubated at 37°C, andexamined for cytotoxic effects as describedabove. Filters, a centrifuge and facilities toshake cultures are not needed for this test.Results are very clear but it is not possible totest the specificity of the toxic effect.

MAINTENANCE OF VERO CELLS

To maintain Vero cells monolayers are washedtwice with 10 ml Dulbecco's PBS. Trypsin inVersene buffer (5 ml) is added and poured offafter 1 minute. The monolayers are thenincubated at 370C until cells begin to detach.Medium 199 (5 ml) is then added and the cellsare resuspended. Volumes (2 ml) of resus-

14 on M

ay 15, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.46.1.10 on 1 January 1993. D

ownloaded from

http://jcp.bmj.com/


pended cells are added to 12 ml of growthmedium in a tissue culture flask and incubatedat 370C. A CO2 incubator is not necessary.This procedure is repeated weekly. Vero cellsmay become less sensitive to VT afterprolonged subculture and this should bemonitored by including titrations of super-natants of control VT producing strains.

PREPARATION OF MONOLAYERSFor the VT test monolayers of Vero cells areprepared in 96-well plastic plates of tissueculture grade. A portion of the resuspendedcells after digestion with trypsin is diluted inthe complete tissue culture medium to obtain afinal concentration of about 5 x 104 cells/ml.Counting in a haemocytometer is recom-mended but usually a 1 in 20 dilution issatisfactory. The diluted suspension (0-2 ml) isdistributed in each well of a 96-well tissueculture plate. The plate is covered and incu-bated for two to three days at 37°C in anatmosphere of 5% CO2 - 95% air. However ifplates are sealed by pressure sensitive film thisspecial atmosphere is not needed.

MAINTENANCE OF Y1 CELLSMaintenance of Y1 cells and preparation ofmonolayers are as described for Vero cellsexcept for the use of Ham's FI0 mediuminstead of Medium 199.

REAGENTS AND MATERIALS

Requirements for tissue cultureThe cell lines and all materials for tissueculture can be obtained from ICN Flow, HighWycombe, Buckinghamshire, unless otherwiseindicated. For convenience the tissue culturemedia and many other reagents can be pur-chased in solid or liquid form. Careful refer-ence should be made to the formulationsgiven. Glutamine is usually added to the tissueculture media immediately before their use,but the formulations should be consulted tosee whether NaHCO3 is already included orhas to be added.1 Cell line Vero, African green monkey kidney,ATCC No. CCL81.Cell lineY1, mouse adrenal cortex tumour,ATCC No. CCL79.

2 Dulbecco's phosphate buffered saline, with-out calcium and magnesium.Trypsin (2.5% w/v) is diluted 10 fold inVersene buffer for use.EDTA (Versene) buffer (0-02% w/v) in0-85% saline.L-glutamine solution (200 mM).Amphotericin B (Fungizone) (stock solution250 ,ug/ml).Penicillin and streptomycin (stock solution5000 IU/ml and 5000 ,ug/ml, respectively).Gentamicin (stock solution 10 or 50 mg/ml).Fetal bovine serum.Donor horse serum.Medium 199 (Modified with Earle's salts).Ham's F10 medium.

3 Plastic materials from numerous stockistsincluding ICN Flow, Nunc Life Technolo-gies, Glasgow and Falcon, Marathon Labo-

ratory Supplies, London NW10 7JP.96-well tissue culture grade plastic plateswith lids.75 cm2 plastic tissue culture flasks.Vero cells are grown in Medium 199 (Mod-

ified) with Earle's salts. To 100 ml are added10 ml fetal bovine serum and 0 5 ml glutaminesolution. Ifnot already incorporated, NaHCO3(2-2 g/l) must also be added. Penicillin-streptomycin (2 ml) solution to reduce bacte-rial contamination and 0 1 ml Amphotericin Bto reduce fungal contamination may be addedif required.Y1 cells are grown in Ham's F-10 medium.

To 100 ml are added 12-5 ml donor horseserum, 2-5 ml fetal bovine serum and 0 5 mlglutamine. If not already in the formulationNaHCO3 (1-2 g/l) must be added. Penicillin-streptomycin (2 ml) solution to reduce bacte-rial contamination and 0 1 ml Amphotericin Bto reduce fungal contamination may be addedif required.

DNA probe tests forVTECThe genes encoding production of VT1 andVT2 have been cloned and probes for thedetection of these genes were developed.20 21The use of such probes will detect all VTECand not only strains of serogroup 0157. Inaddition to the polynucleotide probes from thecloned genes, synthetic oligonucleotide probesfor the detection of different VT genes havealso been developed. Amplification of part ofthe VT gene, using the polymerase chainreaction, has also been used to test for thepresence of VTEC. The present description ofDNA probe tests will be restricted to the use ofnon-radioactively labelled polynucleotideprobes as these methods are the most applica-ble to a wide range of laboratories. Full detailsof the methods have been published else-where.22 Probe fragments are prepared fromrecombinant plasmids and labelled by therandom primer method with digoxigenin-1 I-dUTP (Boehringer Corporation Limited,Mannheim, Germany). Unincorporatednucleotides are removed using QlAGEN-tip5(DIAGEN, Dusseldorf) and labelled probe isstored at - 20°C. The target DNA is preparedin one of the following ways.

1 Faecal suspensions are spotted directly onto nylon membranes (82 mm in diameter,Hybond-N, Amersham) supported on Mac-Conkey agar and the plates are incubatedovernight at 37°C.

2 In order to test a large number of colonies,specimens are resuspended in an equal volumeof PBS, and 0 1 ml samples of 10 fold dilutionsare spread on MacConkey agar plates. Severalhundred well separated colonies on a singleplate are replicated using velvet on to a nylonmembrane placed on agar and the replicatedcolonies are grown for about five hours. Alter-natively, growth from streaking a faecal speci-men on a MacConkey plate can be replicatedon to a nylon membrane.

3 For testing of purified E coli isolatesnutrient broth cultures after overnight incuba-tion at 37°C are spotted on a nylon membranein a grid pattern and grown for about five

15



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

Smith, Scotland

hours. Up to 50 test strains and a positive andnegative control are tested on a single mem-brane.Membranes are prepared for hybridisation

by placing on a series of Whatman 3 MMpapers saturated with the following solutions:10% sodium dodecyl sulphate (5 minutes),lysis solution containing 0 5M NaOH and1 5M NaCl (5-10 minutes), neutralising solu-tion containing 1 5M NaCl and 0-5M TRIS-HC1, pH8-0 (5 minutes), and finally with 2 xSSPE (5 minutes). 2 x SSPE contains 0-3MNaCl, 20 mM NaH2PO4, and 2mM EDTA.Membranes are dried and the DNA is boundby baking for two hours at 80°C or placed onan ultraviolet transilluminator (wavelength302 nm) for four minutes. The procedures forhybridisation and detection of homologybetween probe and target have been describedin detail previously.22 Results of DNA probetests with polynucleotides show whetherstrains carry the genes for production ofVT 1,VT2, or both toxins.

Tests in the reference laboratoryStrains presumed to be VTEC belonging toserogroup 0157 or other serogroups should besent to a reference laboratory for confirmationand further characterisation.

SEROTYPING

Full serotyping ofVTEC requires the facilitiesof a reference centre as investigation withantisera against 173 0 antigens and 56 Hantigens is needed. VT production has beenreported in strains belonging to many differentserotypes. '2 Strains of 0157 are most com-mon, usually with the flagellar antigen H7;non-motile strains have also been reported andthis may be a useful distinguishing feature. "1 23Some VTEC belong to classic enteropatho-genic serogroups such as 026, 055, 0111 and0128. Therefore, strains of these serogroupsisolated from cases of haemorrhagic colitis or

HUS or outbreaks of diarrhoea should be sentto a reference laboratory for further tests.

VT TYPE

Two major types of Vero cytotoxin VT1 andVT2 have been defined. VT1 is virtuallyidentical to Shiga toxin produced by Shigelladysenteriae type 1. VT2 is antigenically distinctfrom VT1 and Shiga toxin. A number ofvariants of VT2 have also been defined andthey are found in strains of both human andanimal origin.' 22 Neutralisation experimentswith specific VT antisera can be used todetermine the type ofVT. However, such testscannot detect strains that produce both VT2and a variant ofVT2. Specific oligonucleotideprobes have been developed to identify thegenes of the different VT2 variants. VT genes

can be differentiated using different primers inPCR amplifications. These variations in VTgenes should provide a useful basis for thesubdivision ofVTEC strains in future studies.

BIOTYPING

The biochemical properties of VTEC are ingeneral those characteristic of E coli. Occa-

sional strains unable to ferment lactose havebeen reported. As described above, most 0157VTEC are unusual in their inability to fermentsorbitol or to produce fi glucuronidase. Thesetests have been suggested as aids in theconfirmation of 0157 VTEC. A few 0157VTEC with other unusual properties havebeen reported including indole negative orurease positive strains and strains able to usecitrate.23 These characteristics could proveuseful markers and care must be taken not toeliminate such atypical strains of E coli in theearly investigation of faecal samples. Althoughdifferent biotypes of 0157 VTEC with respectto fermentation patterns have been reported,most workers have considered the tests tooirreproducible to be of use in the differ-entiation of this group.

MULTILOCUS ENZYME ELECTROPHORESIS

Strains of E coli 0157 have been included inmultilocus enzyme electrophoresis studies ofthe genetic relatedness of E coli.24 Manyenzymes were examined for differences in theirrate of migration and this was used to assessthe clonal relations of the strains. Using 19enzyme systems the 01 57:H7 strains wereshown to be unrelated to 0157 strains belong-ing to 11 other H types. 0157 VTEC (eitherH7 or non-motile) isolated in the USA fromoutbreaks or sporadic cases of haemorrhagiccolitis or from healthy cattle formed a group ofvery closely related clones. It was noted thatthis group was characterised by the presence ofa distinct fast migrating electromorph of aspar-tate aminotransferase and this property wassuggested as a useful marker for the recogni-tion of the group.

PHAGE TYPING OF E COLI 0157A phage typing scheme for VT producingstrains ofE coli 0157 was developed in Canadafor epidemiological investigations.25 Thescheme which uses 16 phages now recognises66 types (H Lior, personal communication).Twenty four phage types have been identifiedin Britain although most strains belong tophage types 1, 2, 4 and 49.2

PLASMID ANALYSIS

Plasmid analysis of 0157 VTEC can be used toidentify strains in outbreaks and sporadic casesof infection.2 311 Virtually all 0157 VTECisolates carry a plasmid with a molecularweight of about 60 x 106 but certain strainscarry additional plasmids. Identification ofphage type, VT type, and plasmids provides avery useful combination for the character-isation of 0157 VTEC.

SERODIAGNOSTIC TESTSStudies in Canada showed that patients withVTEC infection developed rising titres ofVTneutralising antibodies and this was used todiagnoseVTEC infection in additional patientswhen other evidence was lacking.' Patientswith known 0157 VTEC infection developrises in antibody titre to 0157 LPS. High titreserum antibodies to 0157 LPS can also bedetected in patients when it is not possible to

16



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

17Isolation and identification of vero cytotoxin producing strains

isolate 0157 VTEC.Y Subjects in the controlgroup were negative for such antibodies and itwas concluded that testing sera from patientswith HUS and haemorrhagic colitis is a veryuseful method to provide evidence of infectionby E coli 0157. An ELISA is used to screen serawith confirmatory tests, where necessary, usingimmunoblotting.26

ConclusionTests for 0157 VTEC based on the use ofsorbitol MacConkey agar, or improved media,combined with an 0157 antiserum remain themost practical for clinical laboratories. Furtherstudies by the reference laboratory are requiredto confirm the 0157 strains and providecharacterisation of the isolates for epidemio-logical investigations. Where tests for 0157VTEC are negative, specimens should betested for VT or VTEC using toxin tests,immunological methods, or DNA probes.Such tests are hampered at present by the lackof widely available commercial reagents.Examination of sera for antibodies to 0157LPS has proved very useful and should beinvestigated in appropriate cases when tests for0157 VTEC are negative.

Appendix1 Suppliers of E coli 0157 Latex tests or

specific 0157 antiserum.(a) Laboratory of Microbiological

Reagents, Central Public Health Labo-ratory, 61 Colindale Avenue, LondonNW9 5HT.

(b) Oxoid Ltd, Wade Road, Basingstoke,Hampshire.

(c) Pro-Lab Diagnostics Ltd, Wirral,Merseyside.

(d) Mercia Diagnostics Ltd, MerciaHouse, Broadford Park, Shalford,Guildford, Surrey.

(e) Difco Laboratories Limited, PO Box148, Central Avenue, East Molesey,Surrey KT8 OJE.

(f) Kirkegaard & Perry Laboratories Inc.,2 Cessna Court, Gaithersburg, Mary-land, 20879, USA.UK Distributor: Dynatech Labora-tories Ltd., Daux Road, Billingshurst,Sussex RH14 9SJ.

2 Suppliers of MacConkey Sorbitol agar.(a) Oxoid Ltd.(b) Difco Laboratories.

3 Suppliers of MacConkey Sorbitol MUGagar.Biolife, Viale Manza 272, 20128, Milan,Italy.

1 Karmali M. Infection by Verocytotoxin-producing Escher-ichia coli. Clin Microbiol Rev 1989;2:15-38.

2 Vero cytotoxin-producing Escherichia coli 0157. Proceed-

ings of a seminar. I-IHLS Microbiology Digest 1990;7(sup-plement).

3 Wells JG, Davis BR, Wachsmuth IK, et al. Laboratoryinvestigation of hemorrhagic colitis outbreaks associatedwith a rare Escherichia coli serotype. 7 Clin Microbiol1983;18:512- 20.

4 Farmer JJ, Davis BR. H7 antiserum-sorbitol fermentationmedium for detecting Escherichia coli 0157.H7 asso-ciated with hemorrhagic colitis. 7 Gln Microbiol 1985;22:620-5.

5 Gunzer F, Bohn H, Russman H, et al. Molecular detectionof sorbitol fermenting Escherichia coli 0157 in patientswith hemolytic-uremic syndrome. _7 Clini Microbiol1992;30:1807- 10.

6 Borczyk AA, Lior H, Thompson S. Sorbitol-negativeEscherichia coli 0157 other than H7. .7 Infect 1989;18:198-9.

7 Lior H, Borczyk AA. False positive identification of Escher-ichia coli 0157. Lancet 1987;i:333.

8 Thompson JS, Hodge DS, Borczyk AA. Rapid biochemicaltest to identify Verocytotoxin-positive strains of Escher-ichia coli serotype 0157. _7 Clin Microbiol 1990;28:2165-8.

9 Scotland SM, Cheasty T, Thomas A, Rowe B. Beta-glucuronidase activity ofVero cytotoxin-producing strainsof Escherichia coli, including serogroup 0157, isolated inthe United Kingdom. Lett Appl Microbiol 1991 ;13:42-4.

10 Chapman PA, Siddons CA, Zadik PM, Jewes L. Animproved selective medium for the isolation of Escher-ichia coli 0157. . Med Microbiol 1991;35:107-10.

11 Scotland SM, Willshaw GA, Smith HR, Rowe B. Propertiesof strains of Escherichia coli belonging to serogroup 0157with special reference to production of Vero cytotoxinsVT1 and VT2. Epidemiol Infect 1987;99:613-24.

12 Padhye NV, Doyle MP. Rapid procedure for detectingenterohemorrhagic Escherichia coli 01 57:H7 in food. AppEnviron Microbiol 1991;57:2693-8.

13 Basta M, Karmali M, Lingwood C. Sensitive receptor-specified enzyme linked immunosorbent assay for Escher-ichia coli verocytotoxin. .7 Clin Microbiol 1989;27:1617-22.

14 Acheson DWK, Keusch GT, Lightowlers M, Donohue-Rolfe A. Enzyme-linked immunosorbent assay for Shigatoxin and Shiga-like toxin II using P1 glycoprotein fromhydatid cysts. .7 Infect Dis 1990;161:134-7.

15 Downes FP, Green JH, Greene K, Strockbine N, Wells JG,Wachsmuth IK. Development and evaluation of enzyme-linked immunosorbent assays for detection of Shiga-liketoxin I and Shiga-like toxin II. . Clin Microbiol 1989;27:1292-7.

16 Scotland SM, Day NP, Rowe B. Production of cytotoxinaffectingVero cells by strains of Escherichia coli belongingto traditional enteropathogenic serogroups. FEMS Micro-biol Lett 1980;7:15-7.

17 Konowalchuk J, Speirs JI, Stavric S. Vero response to a

cytotoxin of Escherichia coli. Infect Iminiun 1977;18:775-9.

18 Scotland SM, Rowe B, Smith HR, Willshaw GA, Gross RJ.Vero cytotoxin-producing strains of Escherichia coli fromchildren with haemolytic uraemic syndrome and theirdetection by specific DNA probes. .7 Med Microbiol1988;25:237-43.

19 Karmali MA, Petric M, Lim C, Cheung R, Arbus GS.Sensitive method for detecting low numbers ofVerotoxin-producing Escherichia coli in mixed cultures by use ofcolony sweeps and polymyxin extraction of Verotoxin. 7Clin Microbiol 1985;22:614-9.

20 Willshaw GA, Smith HR, Scotland SM, Rowe B. Cloning ofgenes determining the production of Vero cytotoxin byEscherichia coli. Gen Microbiol 1985,131:3047-53.

21 Willshaw GA, Smith HR, Scotland SM, Field AM, Rowe B.Heterogeneity of Escherichia coli phages encoding Verocytotoxins: comparison of cloned sequences determiningVT1 and VT2 and development of specific gene probes..7Gen Microbiol 1987;133:1309-17.

22 Thomas A, Smith HR, Willshaw GA, Rowe B. Non-radioactively labelled polynucleotide and oligonucleotideDNA probes for selectively detecting Escherichia colistrains producing Vero cytotoxin VT1, VT2 and VT2variant. Mol Cell Probes 1991;5:129-35.

23 Bopp CA, Greene KD, Downes FP, et al. UnusualVerotoxin-producing Escherichia coli associated withhemorrhagic colitis. . Clin Microbiol 1987;25:489-97.

24 Whittam TS, Wachsmuth IK, Wilson RA. Genetic evidenceof clonal descent of Escherichia coli 0157:H7 associatedwith hemorrhagic colitis and hemolytic uremic syndrome.

Infect Dis 1988;187:1 124-33.25 Ahmed R, Bopp C, Borczyk A, Kasatiya S. Phage-typing

scheme for Escherichia coli 0157:H7. . Iifect Dis1987;155:806-9.

26 Chart H, Scotland SM, Rowe B. Serum antibodies toEscherichia coli 0157:H7 in patients with hemolyticuremic syndrome. . Clin Microbiol 1989;27:285-90.



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

Documents

Broadsheet 135 J 1993 Isolation identification methods ... · VTEC. Some patients have features resem-bling thrombotic thrombocytopenic purpura (TTP), such as neurological disorders