Time and media saving testing and identification of microorganisms by multipoint inoculation on undivided agar plates

(1)

JOURNAL OF CLINICAL MICROBIOLOGY, Aug. 1978,p.219-227 0095-1137/78/0008-0219$02.00/0

Vol. 8, No. 2

Printedin U.S.A.

Time- and Media-Saving Testing and Identification of

Microorganisms

by

Multipoint Inoculation on

Undivided Agar Plates

LARS G. BURMAN* AND RUNE OSTENSSON

Departmentof Clinical Bacteriology, University ofUmeà,S-90185Umeà, Sweden Received for publication 23 November 1977

Motility and various biochemical activities of isolates of bacteria and yeasts were tested on undivided agar plates by using a simple, manually operated

multipoint inoculation apparatus that allowed the analysis of 25 isolates per 9-cm-diameter petri plate. Fermentation ofall17carbohydrates tested aswellas 13

other biochemical activities commonly used for identification of bacteria were

readilydemonstrated by the multipoint inoculation plate method, and the

results

agreed very well with those of conventional tube tests. In addition to speedy

inoculation and low costof materials, the multipoint inoculation plate method

offersseveral other advantages when compared withconventional tube tests or

with some of the manufactured test kits currently available for recognizing membersof the family Enterobacteriaceae.

Inrecent years,there has been a rapidly

grow-ing interest among clinical

microbiologists

in

meeting thesteady increases inspecimenstobe analyzed andin costsoflaborandmaterials by introducing various time- andmedia-saving

pro-cedures, suchascombining severaltests in one

tube, microscale methods, and attemptsto au-tomate laboratory testing ofclinical isolates of

microorganisms.

Furthermore, a number of

manufactured multitest kits designed for iden-tification of bacteria belonging to the family Enterobacteriaceae on the basis of their bio-chemical activities have become commercially available.Whereasseveral suchsystemsstillare

beingevaluated,theimprovedAPIand

Entero-tube identification systems appear to be suffi-ciently reliable in comparisons with

conven-tional tubetestsandare

presently

accepted and widely used (e.g., in Sweden). However, these systemsalsohavedisadvantages,such aslackof

flexibility, contributing to a

fairly high

cost of materialsperisolateidentified. Inaddition, the

API system requiresaconsiderable amountof

work formanual inoculation of its 20 separate

wells, andintheflexible Miniteksystemmanual addition of bacteria and also

media,

reagents,

and sometimesoiltothe chosennumber ofwells

isnecessary.

Since thefirstmultipointinoculation appara-tus was introduced

by

S. D. Garrett in 1946, several devices for

microbiological

work based

onthe

multipoint

principle

have beendescribed

(see review inreference 7). Their main clinical

useiscurrentlyinphageandbacteriocin

typing

ofbacterial strains,and inseveral clinical labo-ratoriesmultipoint inoculators alsoenable

large-scale routine antibiotic

susceptibility

testing of

microorganisms

by using the agar dilution

methodinstead of the morevariable disk

diffu-sion method(14, 15). Suchinoculators have also beenused foralong timeinmicrobial genetics forwork withphagesorantibioticsaswellasfor

testingofauxotrophy, carbohydrate

utilization,

or other genetic markers of bacterial mutants,

recombinants,transductants,etc.

(i.e.,

in

simple

situations wheredefinedmedia canbe used and either growthor no growth is tobe recorded). Multipoint inoculators have also been

sug-gested fordemonstrationofmetabolic activities ofclinical isolates of bacteria. The

organisms

to

be analyzed are inoculated into the compart-ments of divided

petri

dishes

(13)

or Teflon

plates

(7).Fromthesemaster

plates

thebacterial

strains are transferred to

multicompartment

plates containing test media

by

using

a multi-point inoculator. Thesemethods had not been

developed or much used until

recently,

when

multipoint testing with microtiter trays was

again

adopted

forevaluation

(6,

12).

We wish to report our

positive

experience

from large-scale testing of clinical isolates of bacteria and yeasts on undivided agar

plates,

usinga

simple

multipoint

inoculationapparatus.

On each

regular

9-cm-diameter

plate,

oneortwo

activities of 25 isolates have been tested, thus

making

expensive

multicompartment trays

and elaborate

media-dispensing

routines unneces-sary. The system is

entirely

basedon

generally

219

on February 7, 2020 by guest

http://jcm.asm.org/

(2)

220

BURMAN AND OSTENSSON

accepted tests. When necessary, however, the

conventional

test substrate or the inoculation

procedure

hasbeenmodifiedtofit theagarplate

condition.Incomparison with classical tube test-ing, this systemallows considerable reductions

incosts of media andreagents,aswellasinlabor withglassware service, media preparation, and inoculation, with little loss ofreliability. Since the system offers flexibility and several other advantages in addition, it constitutes, in our opinion, an attractive alternativetothe manu-factured

identification

kits currently available. This testmethod, described and evaluated be-low, is hereafter referred to as the multipoint inoculation plate

(MIP>

method.

MATERIALS AND METHODS

Microorganisms. The strains of bacteria and

yeasts tested were aerobic or facultative

organisms

isolated fromclinicalspecimensreceivedatthe Bac-teriological Laboratory at the Regional Hospital in

Umeâ,

Sweden.

Replication equipment. The replication equip-mentwascommunicatedby G. Bertani. The pin rep-licator (replicator A, Fig.

lA)

was made by alocal workshop and consists ofasquareacrylicplastic (Plex-iglas) handle (70 by70by15mm)towhich25stainless

steel pins (40 by 3mm) have been attached 13 mm apart. Thetips of thepinsmustbelevel.

Fortesting of

HeS

production and motility,wehave developedasecondreplicator (replicatorB,Fig. 1B), equippedwithinjectionneedles(50 by0.8mm)instead of steel pins and by which themicroorganisms tobe testedareinoculated deeply into the agar.

Thecommerciallyavailableautoclavable nylon mi-croculture container(ELESA, CasellaPascoli,1-20100,

Milan, Italy) measures 60 by 60by 12 mm and is dividedinto 25cubical compartments (volume, about

i

1. I

I

r l

t aa _ _

I

FIG. 1. Multipointinoculation equipment. (A)Steel pinreplicator; (B)

needle

replicator;(C) autoclavable microculture container;(D)disposable container.

9

JK

.... .. - jur,

on February 7, 2020 by guest

http://jcm.asm.org/

(3)

MULTIPOINT INOCULATION PLATE IDENTIFICATION 1 cm3 each), one for each replicator pin (Fig. 1C).

During autoclaving and subsequent handling,the

con-tainer wasstored inaglass petri dish whose lid did

not touch the top of the culture container. Water condensing under the lid might otherwise have caused cross-contamination between thecompartmentsof the container when in use. While this study was being

finished, a thin-walled and inexpensive disposable

plastic microculture container of similar size (Kemi-intressen, Sundbyberg, Sweden) became available in Sweden (Fig. 1D). This containercanalso beused for MIPtesting.

The distance between the plated culture samples obtained with this equipment isanappropriate

com-promise between mediaeconomyand speedy inocula-tionon onehand and the risk of interferencebetween too-large neighborzonesof reactionontheother.

Use of equipment. Before use, about 0.5 ml of

sterilesalinepercompartmentwasdispensed into the

culture container. When identifying gram-negative bacteria, tryptophan broth (see below) wasused

in-stead of saline. Parts ofacolony of each straintobe tested were then transferred to individual compart-mentsby using autoclaved toothpicksoraplatinum

loop, andarecord of strainnumbers andcorresponding

container andcompartment (numbered 1to25)

num-berswaskept. (It is advisabletoleaveacompartment,

e.g.,number25[the lower-right corner], uninoculated toavoid laterconfusionconcerningtheidentityof the spots ofgrowth onthetestplates.) The pins of the replicator were dipped into ethanol, flamed, and

al-lowed to coolfor afewseconds before the toolwas

ready for rapid inoculation of thetest plates desired with strains presentinthecontainer. To avoid later flowing and mixing of the culture drops applied, we

foundit advisable todrytheplates (1h at37°C with their lidsoff)beforeuse.

The strainnumber,results of the testschosen,and fmal straindiagnosiswererecordedon anappropriate

datasheet (25strains persheet).Each batch of test mediumwascontrolled by including standard strains

in thesettobe testedto ensurethat thenegativeas

well as the expected weakly and strongly positive

reactionsweredemonstrated. If testsrequiring

anaer-obicincubationwereused(aminoaciddecarboxylases

oroxidation/fermentationofglucose),astrain of Pseu-domonasaeruginosawasincluded to check that

sat-isfactory anaerobicconditionswereachieved.

General cultivation conditions. Unless other-wise stated, thetestswere incubated in air at37°C

and read18 to 24 h after inoculation. Whenneeded, anaerobic conditionswereobtainedby usingGasPak jarsor agloveboxcontainingH2 (10%), C02 (5%),and N2 (85%).

By using dried testplates (see above) with ahigh

concentrationofagar(Oxoidagarno. 1,3% [wt/vol]) problemsduetoswarmingbyProteus wereavoided. Thus,except for themotilitytestmedium(see below), allcommercially available testagars used hadto be

further solidified.

The mediaweredispensedinto 9-cm-diameter plas-ticpetridishes. The resultsofalltestswereeasier to

readifplateswith extra-thickagarwereused(e.g.,40 to50mlperplate),since therisk of too-wide reaction

zones was then minimized and the optical contrast

between positive reactions and the negative back-groundwas enhanced.

Tests. Forconvenience, the review by Cowan and Steel (2) is usedas ageneral reference fortestmedia instead oftheoriginal references. When Cowan and Steel include alternative methods, the particular method here used is stated.

(i)Esculin hydrolysis. Blackeningin and around thegrowthspots onregularesculin agar(2) indicated hydrolysis.

(ài) Carbohydrate utilization. Proteose peptone no. 3(Difco)with carbohydrate (D-arabinose, D-cello-biose, dulcitol, D-galactose, D-glucose, inositol, D-lac-tose, D-maltose, D-mannitol, D-mannose,D-melibiose, D-raffinose, D-ribose, D-sorbitol, D-sucrose, D-treha-lose, orD-xylose [10 g/liter]) and bromothymol blue (50mg/liter)added wasused for staphylococci as well as for gram-negative bacteria to test carbohydrate utilization. Although not necessary, a doubled concen-trationofbromothymol blue facilitated reading of the tests.Ifanaerobic incubationwasperformedto facii-tatedifferentiation between oxidationand very weak fermentation, NaNO3 (425 mg/liter) had to be in-cludedin the test medium (see Results). For yeasts, the basalmedium described by Martin and Schneidau (11) butwith bromocresol purpleas apH indicator (33 mg/liter)wasused.Utilization of a carbohydrate was demonstratedby heavy growth andachange in color of the pH indicator in and around the growthspot. Fermentation usually producedasizable haloof indi-catorcolor change around eachspot(Fig.2).

(iii) Citrate utilization.Utilization of citrate as a carbon and energy source was indicated by visible growthand a shift inpHtowardsalkalinity (dark-blue halo)onSimmond citrateagar(Oxoid).

(iv) Decarboxylases (arginine, lysine,and or-nithine). Moeller decarboxylase broth base (Balti-moreBiologicalLaboratory [BBL]) wassupplemented with the appropriate amino acid (10 g/liter) plus NaNO3 (425 mg/liter), adjustedtopH5.8 to 6.0, and solidified asdescribed above. After inoculation, the plateswereincubated anaerobically. The results were interpretedasdescribedin Results.

(v)Deoxyribonuclease. Deoxyribonuclease agar (Oxoid)wasused forgram-negative bacteriaaswellas forstaphylococci,and after incubation theplateswere flooded with toluidineblue (1 g/liter) orwith hydro-chloric acid (1 M) to obtain purple halos against a dark-blue background or zones of clearing, respec-tively, around deoxyribonuclease-positive spots of growth.

(vi) Gelatin liquefaction. Gelatin agar (2) was inoculated anddeveloped byfloodingit withthe acid mercuricchloride solutionofFrazier(2) after incuba-tion. Distincthalosofclearingindicatedhydrolysisof gelatin.

(vii) Hippuratehydrolysis.Hippuratehydrolysis bygram-negative bacteriawas clearly demonstrated bygrowthandpinkcoloronhippurateagar(2).

(viii)Hydrogen sulfideproduction. Totestfor hydrogen sulfideproduction, regulartriplesugar iron (TSI) agar (2) mustbeused,e.g., that of Oxoid

em-ployedhere(seeResults).Extra-thickagar gavemore

distinctpositive reactions.Afterinoculation with rep-licatorA,replicatorBwaspunchedthroughthe

inoc-VOL. 8, 1978 221

on February 7, 2020 by guest

http://jcm.asm.org/

(4)

2a

t

2b

3

222

on February 7, 2020 by guest

http://jcm.asm.org/

(5)

MULTIPOINT INOCULATION PLATE IDENTIFICATION 223

ula and halfwaythrough theagarlayer.H2S

produc-tionwasindicatedbyblackening inthedepth of the

agar,bestobserved from thebottomsideof the plate

(Fig.3).

(ix)Indole production.Totestfor indole produc-tion, the broth and plates usedwerebasedontryptone

(Difco, 10g/liter),andindolewasdetected byusing

Kovacsreagent(2).

(x) Motility. The motility testmedium ofDifco containing0.5%agarcouldbe used, but

supplementa-tion with beef extract (Difco, 3 g/liter) promoted growth and simplifiedreading of the results. Alterna-tively, 2,3,5-triphenyltetrazolium chloride could be added (0.05 g/liter)toimprove visualization of bacte-rialgrowth. If strains ofProteuswerepresent among

the bacteriato betested, their swarming had to be preventedbyincludingchloral hydrate(1g/liter[3]) in the test medium (see Results). The plates were

inoculated using replicator B, which was punched

halfwaythroughtheagarlayer.Incubation for40hat

22°Cwas necessaryfordemonstration ofmotility in somestrainsin MIPaswellasin tube testing.

(xi) ONPGtest. Forthe o-nitrophenyl-fl-D-galac-topyranoside (ONPG)test,solidifiedONPGbroth (2)

wasused, andayellow haloaroundabacterial

inocu-lum, due to the o-nitrophenol released, indicated a

positivereaction.

(xii)Phenylalanine deaminase.Totestfor phen-ylalanine deaminase, afterincubation phenylalanine

agar(Oxoid) plateswerefloodedwith ferric chloride

(100mg/ml), and thepresenceofphenylpyruvicacid

resultedin rapid development ofgreencolor inand around thegrowthspots.

(xiii) Urease. Bright-red haloswereproduced by

urease-positivestrainsgrowingon ureaagarbase (Ox-oid) plusthe amount ofurearecommended. This MIP

testhadtobeincubatedat22°C (seeResults). Chemicals.Chloralhydratewaskindlysuppliedby

Ferrosan AB, Malmo, Sweden. ONPG wasfrom E.

Merck A.G., Darmstadt, West Germany, and 2,3,5-triphenyltetrazoliumchloridewasfrom BDH Chemi-calsLtd., Poole,England.

RESULTS

MIP tests not requiring evaluation.

Sev-eral conventional testsare normally performed on agar plates. Therefore, the corresponding

MIPtestswerecarriedoutmainlytomakesure

that the growth spots created clearly visible

zonesof reactionand,aboveall, that the latter

did not become too wide and interfere with reading of theadjoiningtests.With thefollowing

tests, theresultswereso clear-cut thata closer

evaluationwasregardedassuperfluous: esculin

hydrolysis, deoxyribonuclease, gelatin liquefac-tion, and hippuratehydrolysis.

Evaluation ofMIP tests. In the following,

TABLE 1. Comparison between conventionaltesting and MIPtesting of biochemicalandphysiological

activities ofgram-negative bacteria

No. of Conven-

Agree-Activity _com-tests tional ment pared test (

Esculinhydrolysis AP EOb

Carbohydrate&utilization:

Glucose 350 B 100

Lactose 350 B 100

Mannitol 350 B 100

Citrateutilization 350 AT 100 Decarbosylases:

Arginine 100 B 100

Lysine . 100 B 100

Ornithine 100 B 100

Deoxyribonuclease AP EO

Gelatinliquefaction AP EO

Hippuratehydrolysis AP EO

Hydrogen sulfideproduction 350 AT 98

Indoleproduction 350 B 95-100d Motility in:

Regularmedium 350 AT 100d

Regular medium+chloral 200 AT 100d hydrate

ONPG(8-galactosidase) 350 B 98

Phenylalanine deaminase 350 B 100

Urease 350 B 100

'Media: B,broth; AT, agar intesttube;AP,agar inpetri dish(seealso text).

bEO,Evaluationomitted (see text).

'Forcarbohydrates other than those listed (see text), eval-uation wasregardedassuperfluous.

dSeetext.

the application of a number of tests to MIP performance is described. The comparisons of

MIP results with those of the corresponding conventional tests are summarized in Table 1

and refertoanalysesof strains ofgram-negative bacteria isolatedfrom urine, wound, and chest infections and most oftenbelongingtothe

family

Enterobacteriaceae.

Carbohydrate

utilization. The results of

MIPfermentationsof

mannitol,

lactose, and

glu-cose by enteric bacteria were compared with

those of tube tests, and 100% agreement was

obtained

(Table

1). A mannitol and a lactose

plateareshown in

Fig.

2.Sincefermentation of all 17 carbohydrates tested (see Materials and Methods)

appeared

to bedemonstrated

equally

well with the MIPmethod,furthercareful eval-uationswere omitted. Itthus appears that

fer-mentation testing of almost any

carbohydrate

can beappliedtothismethod.

Toavoidconfusionbetweenvery slow fermen-tation and the weak acid reaction

resulting

from

FIG. 2. Carbohydrate fermentation plates inoculated with strains of enteric bacteria and incubated overnight. (Top) Lactose;(bottom)mannitol.

FIG. 3. TSIplate inoculated with strains ofenteric bacteria and incubated overnight, viewed from the bottomside.

VOL. 8, 1978

on February 7, 2020 by guest

http://jcm.asm.org/

(6)

oxidationof acarbohydrate bycertainobligate

aerobic bacteria (e.g., pseudomonads), the

fer-mentation plates could be incubated

anaerobi-cally. By usingaerobically as wellas anaerobi-cally incubatedplates,the MIPglucosetestthus

became equivalent to the classical

oxida-tion/fermentation(OF)testofHughand Leifson

(2).

Mannitol fermentation byStaphylococcus au-reus and the fermentations included in the

clas-sification system for other

staphylococci

re-cently describedbyKloos and Schleifer(glucose, galactose,lactose,maltose, mannitol, rhamnose,

ribose, and trehalose

[8]),

as well as a large number of carbohydrate utilization tests used

for identification ofyeasts

(1,

9,10),all appeared

to be equally suitable for MIP performance. Occasional yeast strains, however, grew very

slowlyand needed more than 48 h of incubation. In such situations, the halosofacidity created

by the strongly positive strains could become confluent and obscure the reaction of theslow strain. The latterthen had to be retestedwithout

anyorganismsinoculated into theneighbor com-partments.

Citrate utilization. Inthecitrateutilization

test,bacterialgrowthwasslowandfalse

negative

resultscould occur. However, this riskwas not greater in MIP tests than inregularagarslope

tests, since a comparison showed 100% agree-mentbetween the two methods (Table 1).

Decarboxylases (arginine,lysine, and

or-nithine).Inthedecarboxylasetests,differences

inappearancebetweenMIPand tube tests were

found. Some of the acid metabolites from the

smallamountsofglucose fermented apparently evaporated from the plates (e.g.,

C02).

Thus, growth of enteric bacteria resultedin a higher

pH value on plates than in the conventional petrolatum-covered broth.

Decarboxylase-nega-tive strainstherefore didnotcreateyellow halos

of acidity in MIP tests. On the other hand,

positive strains created violet halos against an

indifferently colored background. The results

agreedvery

weil

withthose of Moellertube tests

(Table 1), provided that all degrees of violet were scored as positive. Furthermore, after in-cubationthe MIPplates should not be left in air formore than 30 min before reading, since

ail

negative tests eventually became weakly

pseudo-positiveinair, possibly due to an accel-erated evaporation of

C02

andvolatileorganic

acids.

Another important difference between tube and MIP

decarboxylase

testing was that NaNO3 had to beadded to the MIP medium (see Ma-terials and Methods). Otherwise, growth was very poor on MIP platesincubated during

par-ticularly successful anaerobiosis. Presumably,

NaNO3 served as an additional electron acceptor necessary in the Moeller medium during very strictanaerobicagarsurfaceconditions.

Hydrogen sulfide production. Initially,we

hopedtobeable to combine the H2S test with any fermentation test desired. However, this turned out to be very difficult. For anunknown reason, all positive reactions were completely

suppressed by bromothymol blue, and the

re-sults withvarious sugar-phenol red agars

incu-bated aerobically or anaerobically were often different from those with TSI agar. Thus, bac-terial formation of H2S appeared to be variable and easily influenced by the nutritional and atmospheric environment of thecells. Since

re-sults with TSI agar presently are the basis of

taxonomyin thismatter, theevaluationof MIP H2S tests wasfinallyperformed with TSI agar (Fig. 3). The agreement with tube tests was then good(Table 1). The halos of acidity(if formed) weredue to fermentation of lactose or sucrose, whereas utilization of the low concentration of glucose present did not result in any indicator change. Thus, the MIP TSI test differed from the tube test in that fermentation of glucose could be demonstrated in the bottom of a tube testbut notin MIP plates.

Indole production. Since indole is volatile and also diffusesrapidly in agar, it could not be demonstrated in or around the MIP growth spots on atryptoneplate. This plate,however,

contained bacteria whichwereinduced for

tryp-tophanase and suitable for further analysis of this enzyme activity. By using replicator A,

heavy inocula ofsuchbacteriaweretransferred

from the overnightplate to a second tryptone

platewithonly 1%agar. Thereplicator pinswere punched through the agar layerso thatcrypts containing bacteria were formed. The second

platewasincubated for0.5 to 1 hbeforeflooding

with reagent. This plate then had to be read

immediately, and any pink color developed in thecrypts was scored aspositive. The agreement

withtube tests wasacceptable (95%)among 350

consecutive gram-negative isolates tested. Due tothemanyindole-positive strains (e.g.,

Esche-richia coli) analyzed, the tendency towards false-negativeMIPresultswas hereemphasized,

resulting in 3.5% doubtful and 1.5%

false-nega-tive MIP tests (Table 1).

Ourattempts tomakethis MIPtest simpler

andmore sensitive have so far been unsuccess-fui. Atpresent, we therefore routinely use the broth testperformed in the culture container. The bacteria are suspended in tryptone broth instead of insaline, and afterinoculationof the testplatesthe containerisalso placed at 37°C. After incubationovernight, a drop of reagent is added to each of itsmicroculture compartments

J. CLIN. MICROBIOL.

on February 7, 2020 by guest

http://jcm.asm.org/

(7)

byusing apipetteor apipettor (i.e.,

semiauto-matic

pipette).

This performance of course

yields results which are identical withthose of

tubetests. Thecarry-overoftryptone broth to the test plates (about 1.5 ulper spot) does not influence the outcome of the other MIP tests.

Since the indole reagent is strongly acidic and may beharmfultothe polymer of the container,

the latter shouldbe rinsed soon after use.

Alter-natively,adisposablecontainer can be used (see

MaterialsandMethods.)

Motility.

The MIP

motility

testis

essentially

the same as the conventional test, and with nonswarming bacteria thetwoalso yielded

iden-ticalresults(Table 1). However, because ofthe

low concentration of agar in the motility

me-dium, strains ofProteus present amongthe

bac-teria to be tested could swarm and invade all

testsrun onthe same

plate.

The

problem

was

circumvented by including 1 mg ofchloral hy-drateperml in themedium. Thisconcentration of chloral hydrate was optimal and effectively prevented swarming without inhibiting motility

inall 34strainsofProteus tested.Similarly, no

negative effectof 1 mgof chloralhydrateperml

onmotilitywasnotedamong 94strainsof other

motile gram-negative bacteria studied. Thus, 100%agreementbetween tubeand MIP testing ofmotilitywasobtained,

despite

thepresenceof chloralhydrateinthelattercase(Table 1).

ONPG

test. Since the ONPGtestis

capable

ofvisualizingverylow

/3-galactosidase

activities, bacteriathatutilized lactose

rapidly

gavea vis-iblereaction afteronlyafew hours ofincubation

of MIP

plates,

whereas slow strains

required

overnight

growth. By

that

time,

however, the

reaction zones of the most active strains had diffusedoverthewholeplate.The MIPONPG

testwastherefore

practical

only

for isolatesthat

appeared negative

on lactose fermentation

plates. However, sinoe the ONPG test is

rou-tinely used

mostly

for further

analysis

of such

gram-negative

bacteria (e.g., from fecal

speci-mens),itis fortunate that theMIP ONPGtest

turnedout tobeuseful inthis

particular

situa-tion.

We have been satisfiedwith theMIP ONPG

test in the feces

analysis laboratory

ever since

the

following

routine was

adopted.

Clones that

appearlactose

negative

onbrilliantgreen agar

(Oxoid) or on

deoxycholate

agar

(BBL)

are

transferredtothe container for further

testing.

Thelactose-negativecharacter of theisolatesis

firstrecheckedona

regular

lactosefermentation plate (see

above)

incubated

overnight. Thus,

sizableamountsof bacteriainduced for

,B-galac-tosidase

(if

present)

are obtained.

Using

repli-cator A, heavy inocula of such

cells

are then

transferredtothe ONPG

plate,

and the testis

read after 2 to 4 h at 37°C. The results agree well withthose of overnight ONPG broth tests

(Table 1).

Phenylalanine deaminase activity and pigment production. After incubation, phen-ylalanine plateswere first inspected for halos of

colordue to bacterial pigments, as this whitish

platenicely visualized diffusible pigments

pro-duced byPseudomonasaeruginosa (blue-green or yellow), P. fluorescens (yellow), P.

malto-phila

(brown), etc. This plate was also suitable

for detection offluorescent pigments by using

UV light. The plates were then flooded with reagent,and phenylpyruvicacidformed was in-dicated by additional distinctly green

reactions.

Thehalos never became too large, and the agree-mentwith the Henriksen tubetest (2) was 100%

(Table 1).

Urease. After overnightincubation at 37°C,

strains ofProteus usually showed strong

reac-tions on MIP urea plates, and the bright-red halos produced were large and tended to obscure

the results of the neighbor strains. The urea

plateswere

therefore

incubated at 22°C. At this

temperature,

however, bacteria with weaker

urease activities (strains ofKlebsiella

pneumo-niae, Enterobacter, etc.) gave rather small and

lessintensely coloredhalos. Therefore,allsuch

discrete

reactions

had to be scored as positive. If

this iskept in mind, theagreement with the urea

brothtestofStuartetal. (2)is complete (Table 1). Theurease testwastheonly MIP test which

couldnotbe incubated for more than 24 h.This was due to the confluence of strong

reactions

mentioned.

MIP

method in routineuse.After theabove

evaluation of MIP tests had been completed, additional studies were performed to ascertain

whether the MIP method was practical in the routine laboratory. Even if various MIP tests

appeared reliable when evaluated individually,

the method might have been invalidated by

cross-contamination, loss of order of

strains,

or othertechnical

problems

when in

practical

use. The limited setof tubetests

previously

used

inthis laboratory for first-stage identificationof

gram-negative bacteriainurine

specimens (TSI,

mannitol-Durham

tube,

indole,

ONPG,

and Voges-Proskauer) wasused as areference in a

preliminary

practicality

trial of a set of MIP tests (citrate, glucose, lactose,

mannitol,

H2S,

indole, and

phenylalanine deaminase).

The

fol-lowing 222 strains of

gram-negative

bacteria were tested: 102E. coli, 49 Klebsiella-Entero-bacter,5

Citrobacter,

32Proteus

mirabilis,

4P.

vulgaris,

15Pseudomonas, and 15Salmonella.

The twomethods

disagreed

on 4

strains,

which were subject to an extended series oftests. It

turnedoutthat 2

H2S-negative,

indole-positive

VOL. 8, 1978

on February 7, 2020 by guest

http://jcm.asm.org/

(8)

strains ofCitrobacter and 2 Voges-Proskauer-negative, indole-positive strains of Klebsiella

wereidentifiedasE.coli

by

the tubesetbecause

it lacked acitrate utilizationtest. Most

impor-tant, however, was the fact that no technical problems were encountered when several MIP

tests were inoculated from the same culture container. The latter has also been further

con-trolled withregardtocontamination

during

han-dling. In our

experience,

occasional

contamina-tion occurs as aresult of failurein

isolating

the original clone inapurestate.

Therefore,

the risk ofcontamination is no greaterin MIP than in tubetesting. Such

problems

are,

however,

often visiblein MIPgrowthspots, in contrasttotube andkit tests.

Recently, theMIP methodwasintroducedas

routine in this laboratory, and

microbiologists

and technicians have

rapidly

become familiar with themethodand

accepted

it.

Furthermore,

we have found that MIP test

plates

yield

un-changedresultsafter1monthofstorage in

plas-ticbagsat4°C. This circumstance will

simplify

media preparation routines when the MIP

methodisadopted for

regular

use.

DISCUSSION

Anysystemintroducedto

simplify laboratory

testing and identification of

microorganisms

must offer a sufficient degree of

reliabiity

in

comparison withaccepted reference methods.A number of

steadily

improving manufactured

multitestkitsare nowavailable for

recognition

ofmembers of the family Enterobacteriaceae

onthe basisof their biochemicalactivities.Many evaluationsofsuch kits havebeen

performed

in recentyears, and the API,Enterotube, and also

Minitek andR/Btestshavebecome

increasingly

accepted. Althoughthey require less laborthan the classical tube tests, each of these systems

also has disadvantages (see introduction). In very smalllaboratories, kitsareparticularly

use-ful,sincetheyarealmost independentofmedia

preparation

facilities and still enable a rather

high degreeofdiagnosticsophistication and ac-curacy. On the other hand, the

microtiter

mul-tipointmethod(seeintroduction) might become useful in very large laboratories, although the elaborate routines for dispensing media are a

problemandrequireautomation.

In our opinion, the MIP method here pre-sented and evaluatedcan also offer a sufficient degree of accuracy at each level of diagnostic

sophistication, although independent

evalua-tionsin otherlaboratoriesare of course desira-ble. The method is practical in smallas well as in large larboratories and is also economical. Table 2 shows an approximate comparison of costs of prepared media inconventional tubes,

TABLE 2.Approximate materialcostcomparison betweendifferentEnterobacteriaceaeidentification

methods

Relative material

Method cost/testa

Conventional tube testing . 100

Enterotube ... 26

API ... 14

MIP ... 3

a For the Enterotube and API tests, the current pricesinSweden were used. The calculations for the conventional and MIP methods were based on the pricescharged by the NationalBacteriological Labo-ratory,Stockholm, Sweden (andbymostother Swed-ish laboratories as well) forprepared media of the typesused(containedin tubes andplastic petridishes, respectively).Theconventional tube and MIP meth-odsallow various combinations ofmorethanone test perunit. Tosimplify comparison, the degree of com-bination used in theEnterotube method (11testsin 8 units) has herebeenapplied alsotothe conventional tube and MIP methods.For each of themethods,the material cost per test is given relative to that per conventional tube test (defined as 100). The costs during later stages of testing (inoculation and reading) are notincluded in the table.

MIPplates,and twomanufactured kitsavailable in Sweden (Minitek is temporarily withdrawn from theSwedish market dueto"technical prob-lems," andnopricecanbeobtainedatpresent).

Accordingto ourcalculations, the materialcost per MIP test isabout 3% ofthat per tube test

and apparently also less thanthe costper En-terotube and API test. Inaddition, thecostper

isolate identified canbefurthercutwhen using MIP,since thesetof tests used canbe reduced andadjusted accordingtothetypeofspecimen and theparticular level of diagnostic refinement desired. With a rigid system, such as API, 20 tests arespentoneachisolate.Furthermore, the MIP method requires little time for handling since 25inoculationsareperformedin 1

manip-ulation.

Inadditiontoversatility,speedy inoculation, and low cost, MIPalsooffersotheradvantages herebrieflypointed out.

(i) MIP enables inspection of colony-like growthspots, so that "colony" appearance,

pig-mentation, poor growth (risk of false-negative

results), and sometimes contamination are ob-served.

(ii) ReadingMIPiseasy.No color comparison cardsarenecessary, as theresults generallylook like thoseof the conventionaltests. Reading is

further simplified because many positive and

negative reactions on the same plate are

com-pared againstabackground of uninoculated test medium. Thus, controls are automatically

per-formed and any significant decline in medium

on February 7, 2020 by guest

http://jcm.asm.org/

(9)

quality will be detectedas adeviationfrom the

ordinary visual impression.

(iii) Theage,quality, and composition of

me-dia andreagents areknownandcanbechanged.

(iv) No precaution against carry-over of test media has to be taken during inoculation (cf.

microtiter).

(v) Antibioticsusceptibility testingbythe agar dilution method can easily be added to MIP identification.

(vi) The MIP method can be used also for

other tasks,such aslarge-scale testing of

micro-bial growthrequirements,response to tempera-tures, pH,atmosphericconditions, phages,

bac-teriocins,and toxicagents(tellurite, KCN,azide,

bile salts,disinfectants,etc.and theagentsused

for resistogramtyping [4, 5]).

(vii) TheMIPmethodcanbeappliedtomany

species of bacteria in addition to members of Enterobacteriaceae, aswellastoyeasts.

(viii) If desired, the strains tested by the MIP method can be saved without spending much extraeffort or space.Bysealingoneof the MIP

plates, the strainscanbe storedat4°C for

sev-eralweeks to be furtheranalyzedinnosocomial orotherfollow-ups. The MIPdata sheetsthen serve as atemporary strainrecord.

(ix) In parallel with identification and anti-biotic susceptibility testing of gram-negative bacteria, efficient screening for the presence of conjugative R

plasmids

in the strains can be

performed

by

using the MIP

equipment (L.

G. Burman and R. Ostensson,

Plasmid,

in press). The drawbacks with theMIP system should

also be pointed out. One is that certain tests have so far resisted application to MIP (e.g., malonate, methyl red, Voges-Proskauer, andgas

formation).

Secondly,

MIPis

impractical

if

sev-eral days of incubation of tests are

required.

Chloral

hydrate

mustthen be addedto

all

plates,

since 3% agar is not an absolute guarantee

against lateswarming ofProteus.

Furthermore,

strongpositivereactionsontheurea

plate might

becomeconfluent.

However, we believe that the MIP method

could becomeavaluable toolin

microbiological

workand thatfurther improvementsand

appli-cationsofthemethodarefeasible.

ACKNOWLEDGMENT

This work wassupported by theSwedish MedicalResearch

Council (grants4508 and5217).

LITERATURE CITED

1. Ahearn, D. G. 1970. Systematics ofyeastsofmedical interest. Proc. Internat. Symposium on Mycoses 205:64-70.

2. Cowan, S. T., and K.J. Steel. 1976. Manual for the

identification of medical bacteria. Cambridge Univer-sity Press, London.

3. Davis, B. D., R. Dulbecco, H. N. Eisen, H. S. Ginsberg, andW. B.Wood, Jr. 1973. Microbiology, 2nd ed., p.

772.Harper &Row,Publishers, New York.

4. Elek,S. D., and L. Higney. 1970. Resistogram typing-a new epidemiological tool: application to Escherichia coli. J. Med. Microbiol. 3:103-110.

5. Elek, S. D., and C. Moryson. 1973. Resistotyping of

Staphylococcusaureus.J. Med.Microbiol. 7:237-249.

6. Fung,D. Y.C.,andP. A.Hartman. 1975. Miniaturized microbiological techniques for rapid characterization of bacteria, p.347-370.InC.-G. Hedén and T. Illéni (ed.), New approaches to the identification of microorga-nisms. John Wiley & Sons, Inc., New York.

7.Hartman,P. A. 1968.Miniaturizedmicrobiological meth-ods. AcademicPress, New York.

8.Kloos, W. E., and K. H.Schleifer. 1975. Isolation and characterization of staphylococci from human skin.II. Descriptions offournewspecies. Int. J. Syst. Bacteriol. 25:62-70.

9.Lennette, E. H., E. H.,Spaulding, and J. P. Truant (ed.). 1974.Manual ofclinical microbiology, 2nd ed. AmericanSociety for Microbiology, Washington, D.C. 10.Lodder, J. (ed.). 1970. The yeasts. A taxonomic study,

2nd ed.North-Holland Publishing Co., Amsterdam.

11. Martin, M. V., and J. D. Schneidau. 1970. A simple and reliable assimilation test for the identification of Can-dida species. Am. J. Clin.Pathol. 53:875-879.

12.Newsom, S. W.B. 1975.Easy, economic typing of

enter-obacteria,p.435-443.InC.-G. Hedén and T. Illéni(ed.), New approaches to the identification of microorga-nisms. JohnWiley & Sons, Inc., New York.

13. Sneath, P. H. A., and M. Stevens.1967.Adivided petri dish forusewithmultipoint inoculators. J. Appl. Bac-teriol.30:495-497.

14. Steers, E., E.L. Foltz,and B. S. Graves. 1959. An

inoculareplicating apparatus for routine testing of bac-terialsusceptibilitytoantibiotics. Antibiot. Chemother.

(Washington,D.C.)9:307-311.

15. Washington,J.A.,II.1969.Antimicrobialsusceptibility of Enterobacteriaceae andnonfermenting

gram-nega-tivebacilli.Mayo Clin.Proc. 44:811-824. VOL. 8, 1978