Further modifications of the auxanographic method for identification of yeasts

(1)

Further

Modifications of the Auxanographic

Method for

Identification of

Yeasts

PATRICIA A. MICKELSEN,* LAURENCE R. McCARTHY, AND MYRA A. PROPST

ClinicalMicrobiologyLaboratories, North Carolina MemorialHospital, Chapel Hill, North Carolina27514 Received for publication 7June 1976

Amodifiedauxanographic carbohydrate assimilation procedure for the

identi-fication of medically important yeasts is described. This method employs a

heavy inoculum ofunstarved yeasts, autoclaved yeast assimilation medium,

pourplates of shallow depth, andcommercially available

carbohydrate-impreg-nated disks. The accuracy of this procedure was established in a comparison

withthe Wickerham broth method.

The role ofyeasts as agents ofopportunistic infection has steadily gained in prominence

during the last decades (7, 10, 11). Because

yeastsoften colonizeman,determining the sig-nificance ofisolates from clinical specimens is

usuallydifficult. Properidentification ofyeasts

allowsanappreciation of thepathogenic

poten-tial and epidemiological character of the

iso-late. This informationmayassistthe physician

in evaluating the medical significance of the

yeast, as well as in determining the proper

courseoftherapyand the patient's prognosis.

Carbohydrate assimilation reactions are

amongthe primary tests used todifferentiate

generaandspecies ofyeasts (9). Inmany

labo-ratories,theidentificationtothe specieslevel of yeastsother thanCandida albicans is avoided because methods avaliable for assimilation testingarecumbersome, time consuming, and often difficulttointerpret.Recently,new

proce-duresfordeterminingthecarbohydrate assimi-lation reactions ofyeastshave been described

(1, 5,8).Thesemethodsrepresentmodifications

of the auxanographic technique ofBeijerinck

(3)and/or the standard broth method of Wick-erham (14). Although these modifications

ap-pear to be satisfactory, they possess several

disadvantages.TheyusepH indicatorstodetect assimilation of carbohydrates. This requires buffering, manipulating, and monitoring the

pH of the medium. Misinterpretation of test

results may occur with reversion of the pH

indicator or from the diffusion of acid in the medium (5). Inaddition, somemethodsrequire preparation of sterilecarbohydrateoryeast

ni-trogen base (YNB)

carbohydrate-impregnated

disks (5,8), which is timeconsuming.

Wehavedevelopedamodification ofthe

aux-anographic method forcarbohydrate assimila-tiontesting thatrequiresminimal preparation

timeandiseasily interpreted. The accuracy of our modified method anditsvalue in

identify-ingyeasts isolated from clinical materials was

evaluated against the standard Wickerham

brothtechnique.

MATERIALS AND METHODS

Development of the modified auxanographic method. Several representative yeast species were used to examine selected variables in the auxano-graphiccarbohydrateassimilationtechnique. Vari-ations inthe yeast inoculum and the basal media were tested in the preliminary studies. These in-cluded: inoculum consisting of nonstarved yeasts and of yeast starved for 24 and 48 h in distilled water, lx YNB, 0.1 and 0.5%dextrose indistilled water or lx yeast nitrogenbase; an inoculum size rangingfrom0.1 mlofaMacFarlandno. 1

nephlo-metrystandard to1.0mlof a MacFarlandno.5; the useof pour platesandplatesinoculatedbyswabbing

the surface with the yeastsuspension; the alteration of agardepth,i.e., 25, 35,50,or70mlofmedium per 150-mmsterile petridish;variationinthe medium's agar concentration between1.5and2.0%; and prep-arationof the medium withfilter-sterilizedor auto-claved YNBinconcentrations oflx (6.7gofYNB/

1,000 mlofmedium)or0.1x.

Speciesexaminedinthese trialswerelaboratory

stock cultures thatincluded one isolate each of Can-didaalbicans, C.guilliermondii, C. tropicalis, Cryp-tococcus albidus var. albidus, C. albidus var.

dif-fluens, C. neoformans, C. uniguttulatus, Rhodoto-rula rubra, Saccharomyces cerevisiae, Torulopsis

g&abrata,andTrichosporoncutaneum.Themodified

auxanographicmethoddescribed belowwasderived

from theinformation obtained from these studies. Modifiedauxanographic method. Mediumfor the modifiedauxanographicprocedurewasprepared by

dissolving6.7g of YNB(Difco)in 100mlofdistilled water and then adding this to 1 liter of a 2.0% aqueoussolution of Noble agar(Difco)thathad been boiled for1to 2min.Aliquots (25ml)of this medium were dispensed into tubes (25 by 150 mm), which 297

on February 7, 2020 by guest

http://jcm.asm.org/

(2)

298 McCARTHY, AND PROPST

were thenautoclavedat 15lb/in2at1210Cfor15min. Medium either wasusedafter coolingto 45 to50°Cor storedinarefrigeratorat5°C foramaximumof 6 weeks. Before use, two tubes of medium for each isolate to be identified were melted in a boiling

waterbathand thencooledto45 to500C.

Growth from a48-h culture of the yeaston Sa-bouraud dextrose agarwassuspended in 5.0ml of sterile distilled waterandhomogenizedwitha Vor-texmixer. The turbiditywas adjustedtothat ofa MacFarland no. 5 nephlometry standard. Each tube ofcooledmediumwasinoculated with1.0mlof thestandardized yeast suspension,mixed,and then poured into a sterile petri plate (150 by 15 mm). After the mediumwas allowedto harden at room temperature, six Minitek (BBL)

carbohydrate-im-pregnated disks wereplacedoneach of theplates. Carbohydrates tested include cellobiose, dextrose,

dulcitol (galactitol), galactose, inositol, lactose, maltose, melibiose, raffinose, sucrose, trehalose, andxylose. Inoculatedplateswereincubatedat25°C andexamined by indirectlight every otherdayfor 14days. Assimilation ofa carbohydrate was indi-catedbyazoneofgrowtharound thecarbohydrate

disk.

Evaluation of the modified auxanographic method. Yeastisolates used inthisstudywere ob-tained from clinical material submittedtothe bacte-riology andmycologylaboratories of the North Car-olina MemorialHospital, Chapel Hill. Two hundred and thirteen yeastswereexamined(Table 1), which represented 7 genera and 15 species of medically importantyeasts. Afterconfirmation of its purity, each yeast was tested by the modified auxano-graphicmethoddescribed aboveand by the Wicker-ham broth method without agitation (13). To pre-parethe Wickerham assimilationmedium,4.0mlof distilled water was dispensed into test tubes and sterilized. To thiswasadded 0.5 ml ofasolution (6.7

g/100ml) of YNB (Difco) which had been filter

steri-lized. A 6% solution of each carbohydrate to be tested (raffinose and dulcitol, 12%)wasfilter steri-lized, and0.5mlwasaddedtoeach tube. A control tube containing YNB butnocarbohydrate was used foreach isolate. Each tubewasinoculated with0.1 ml of a MacFarland no. 1 suspension of yeast that had been starved in distilled water at250Cfor48h. Tests were incubated at250Cand were observed for 21days. Assimilation testswereconsidered positive whenmoregrowthwasobserved inthe tubes con-tainingcarbohydrate thaninthe control tube.

Isolates that gave conflicting reactions for a par-ticularcarbohydrate wereretested by both assimila-tionmethods todeterminewhich procedure resulted inthe initial incorrect reaction. Other parameters usedto identify yeasts included fermentation of dex-trose, galactose, lactose, maltose, sucrose, and tre-halose (13); nitrate assimilation; urea hydrolysis; pellicle formationonSabouraud dextrosebroth; pro-duction of germ tubes inhuman serum; chlamydo-sporeproductiononcorn meal agar; and characteris-tic microscopic morphology (12). Identification was basedonthereactions foryeast species described in reference 9.

RESULTS

Method development. During preliminary studies, variations in theauxanographic

proce-dure that affected the accuracy or

interpreta-tionof resultswerenoted. Mediumcontaining autoclaved 1 x YNB gave readings comparable tothoseof mediacontaining filter-sterilized 1 x

and O.lx YNB. Although medium containing

autoclaved 0.1x YNBwassatisfactoryfor

test-ing yeasts, itfailedtosupport abundantgrowth

ofCryptococcus uniguttulatus.

Variationininoculumdensity didnot

signifi-cantly reducethe completiontime of

assimila-tion tests. However, theuse of 2% agar anda

heavyinoculum, 1.0ml ofaMacFarlandno. 5

suspension, resultedinthemostcompact,

well-definedzonesaround carbohydrate disks. Nonstarvedyeastsandyeastsstarvedin var-iousmediafor24and48h gavethesame

read-ingsforassimilationtests.Backgroundgrowth, which made the interpretation of some tests

difficult, occurred with certain genera of

yeasts, notably Cryptococcus and Trichospo-ron. This could not be eliminated by altering inoculum size or by prolonged starvation of

yeasts. CommerciallypreparedYNBcontains a

small amount ofinositol,whichcan be

assimi-lated bymembers of these genera. The use of

YNB prepared without inositol in our

labora-tory failed to reduce background growth.

De-creasingthe volume ofagar in pour plates to 25

ml minimizedbackground growth, thus permit-ting unambiguous and accurate readings.

Be-cause assimilations were generally complete

withinafewdays, desiccation of thepourplates

was not a problem. Inoculation of plates by

swabbing the surface, despite the aforemen-tioned variations, invariably resulted in

trou-blesome background growth, particularly with Cryptococcus and Trichosporon.

Evaluation of the modified auxanographic method.Twohundred andthirteen yeasts were

tested for their abilitytoassimilate12

carbohy-drates by the modified auxanographic and

Wickerham broth methods. A total of 2,556

paired assimilationswereperformed (Table1). Agreement between the two methods on initial

testing was 98.2%. Ofthe 45 (1.8%) reactions

that gave conflicting results initially, 44

showedagreementby both methods when the

discrepantcarbohydrate assimilations were

re-peated.In oneinstance, thedifferencein assim-ilation reactions could not be resolved by

retest-ing. Final agreementbetween the two methods

was99.9%. The Wickerham procedure was

re-sponsiblefor 37 (82.2%) of the 45 initial

incor-rect results, whereas the auxanographic

http://jcm.asm.org/

(3)

methodwasinerrorin 7(15.6%)instances. The

auxanographic method averaged 3.3 days for

completion of assimilation reactions, whereas

the Wickerham broth procedure averaged 5.0

days. No changes in the assimilation patterns

of the isolates tested were observed after 10

dayswith theauxanographicmethodorafter14

days with the Wickerham method.

The 45 assimilation reactions yielding initial

disagreement between the two methods are

shown in Table 2. Incorrectassimilationresults

for certain carbohydrates were frequently

re-lated to an assimilation method. In addition,

similar errors occurred with several speciesof

yeasts. Cellobiose most frequently resulted in

initial errors with both methods (40.0%).

Six-teenof the 18 initial incorrect cellobiose

assimi-lation reactions were negative, with 14

occur-TABLE 1. Assimilation results obtained by the Wickerham brothandmodifiedauxanographic methods

No. ofincorrect results with: No. of No. of No. of tests in No. of con- i

Organism strains paired Mdfe

Organism strains sma- agreementon flicting Wicker-

No.

unre-tested tioests_{tion tests}a- initial test tests _mthodham _graphica solved by_solved method method retest

Candida albicans 45 540 538 2 0 2 0

C. guilliermondii 6 72 71 1 1 0 0

C. krusei 11 132 132 0 0 0 0

C. parapsilosis 18 216 214 2 1 1 0

C. pseudotropicalis 7 84 81 3 1 2 0

C. tropicalis 54 648 634 14 14 0 0

Cryptococcus neoformans 8 96 93 3 3 0 0

C.albidus var. albidus 2 24 23 1 1 0 0

C.laurentii 2 24 24 0 0 0 0

C.terreus 1 12 12 0 0 0 0

Geotrichumcandidum 1 12 12 0 0 0 0

Rhodotorula rubra 5 60 55 5 2 2 1

Saccharomycescerevisiae 11 132 131 1 1 0 0

Torulopsisglabrata 35 420 420 0 0 0 0

Trichosporon cutaneum 7 84 71 13 13 0 0

Total 213 2,556 2,511 (98.2)a 45(1.8)a 37(82.2)b 7(15.6)" 1(2.2)b

aPercentage of total number ofassimilation testsperformed.

Percentage of 45conflictingtests.

TABLE 2. Assimilation reactionsforyeastspeciesthatwereincorrectoninitialtesting

Assimilation of:a No.of

unre-Organism Dulci- Inosi- Lac- Melibi- Raffi- Su- Treha- solved Total

Cellobiose tol tol tose ose nose crose lose reac-tions

Candida albicans 1*/0 1*/0 2

C.guilliermondii 0/1 1

C.parapsilosis 1*/1* 2

C.pseudotropicalis 1/1 0/1* 3

C. tropicalis 0/12 0/1* 0/1 14

Cryptococcus neofor- 0/1* 0/1* 0/1 3

mans

C. albidus var. albidus 0/1 1

Rhodotorula rubra 1/1+ 1** 0/1* 1/0 1 5

Saccharomyces cerevi- 0/1 1

siae

Trichosporon cutaneum 0/4* 0/1 0/1 0/3* 0/4* 13

Subtotal

False-positive reac- 1*/0* 0*/4* 0*/0* 1*/1* 0*/5* 1*/6* 0*/0* 0*/1* tions

False-negative reac- 2/14 0/0 0/1 0/2 0/1 1/0 0/1 0/2 tions

Unresolved 1 0 0 0 0 0 0 0

Total 18 4 1 4 6 8 1 3 45

aDataexpressedasincorrectmodified auxanographic result/incorrectWickerham result.Symbols: *, False-positive

reactions;**,unresolved.

http://jcm.asm.org/

(4)

300 MICKELSEN, McCARTHY, AND PROPST

ring inthe Wickerham method.Candida

tropi-calisaccountedfor 12 of these 14negative

reac-tions. Additional discrepancies withcellobiose

occurred with C.albicans, C.pseudotropicalis

and Rhodotorula rubra. The unresolved

dis-crepancyoccurringbetween both methodswas

observed with cellobiose.

A total of 18 incorrect assimilationtestswere

attributedtodulcitol, melibiose, and raffinose.

Sixteen of the 18 reactions werefalsepositive,

occurring almost exclusively bythe Wickerham

method. Eleven of the 16 incorrect positive

as-similation reactions for these carbohydrates

were observed with Trichosporon cutaneum.

The remaining nine initial discrepancies

be-tweenthe twomethodswerenoted with

inosi-tol, lactose, sucrose, and trehalose. Other

dis-crepancies occurredrandomlywith various

or-ganisms.

Inthisstudy, assimilationreactions for each

isolate were considered correct (i) only after

they wereobtainedby both methods after

ini-tial testing or retesting and (ii) when these

results were consistent with published

reac-tions (9). Correct assimilation patterns were

obtainedoninitialtestingfor 209(98.1%)of the

214 isolates by the auxanographic procedures

and for 185 (86.4%) with the Wickerham

method (Table3). Initialincorrectresultswere

observed most frequently with carbohydrates

that may or may not be assimilated by the

particular speciesof yeast tested. Assimilation

of cellobiosebyC.tropicalis andR. rubra and

of melibiose, raffinose,and dulcitolbyT. cuta-neum is variable within these species.

There-fore, in most instances an incorrect

assimila-tion reacassimila-tion obtainedby either methodwould

nothave resulted in the incorrect identification

ofanisolate.

DISCUSSION

Our method for carbohydrate assimilation

testingusesmodificationsof the basic

auxano-graphic technique described by Silva-Hutner

and Cooper (12), which are similar to those

used by otherinvestigators.Adams and Cooper

(1)obtainedsatisfactoryresultswhenYNBwas

sterilized by autoclaving. Although Land and

co-workers (8) have recommended the use of

assimilation agar prepared with autoclaved

0.1x YNB, weobservedreduced growth of

Cry-tococcus uniguttulatus on this medium and

would recommend the use of 1x YNB. Our

method uses a heavy inoculum of unstarved yeasts in pour plates of shallow agar depth.

Huppert et al. (5) also recommend the use of

plates with shallow agardepthinorderto

ob-tain more clearly defined endpoints. Huppert

et al. (5) and Land et al. (8) have alsoused a

heavy inoculum of unstarved yeasts without

encountering difficulties due to carry-over of

stored carbon sources. Thus, ourmodifications

andthose of otherinvestigators represent

pro-cedural alterations that do not comprise, and

apparently enhance, the accuracy of

assimila-tion determinaassimila-tions while allowing these test

TABLE 3. Comparison ofcorrectassimilationpatternsfor yeast obtainedoninitialtestwiththemodified

auxanographic and Wickerham broth methods

No. of correct assimilationpatternsoninitial test

No. of with: No. unresolved

Organism strains byrest

tested Both methods Modifiedauxan- Wickerham by retest

ographic method method

Candida albicans 45 44 44 45

C.guilliermondii 6 5 6 5

C. krusei 11 11 11 11

C. parapsilosis 18 16 17 17

C.pseudotropicalis 7 4 6 5

C. tropicalis 54 40 54 40

Cryptococcus neoformans 8 6 8 6

C.albidus var. albidus 2 1 2 1

C.laurentii 2 2 2 2

C. terreus 1 1 1 1

Geotrichum candidium 1 1 1 1

Rhodotorula rubra 5 2 3 3 1

Saccharomyces cerevisiae 11 10 11 10

Torulopsis glabrata 35 35 35 35

Trichosporon cutaneum 7 0 7 0

Total 213 178 (83.5)a 209 (98.1) 185 (86.9) 1 (0.5)

a Numbersinparenthesesarepercentages.

http://jcm.asm.org/

(5)

procedurestobe moreconvenientandpractical.

In our evaluation of the Wickerham broth

and modified auxanographic methods, agree-ment between the two methods was greater than 98% on initial testing. However, when inconsistent results occurred, they were more

commonly observedwith certain organisms and

substrates. These results were obtained most

frequently with the Wickerham assimilation

method. The majority of false-positive errors

obtainedoninitial testing ofT. cutaneum

prob-ablyresulted from the poorly defined end points observed with nonassimilated carbohydrates whenthe broth methodwas used. Despite the

use of a carbohydrate-free control tube,

dis-criminationbetween weak-positive results and background growth ofT. cutaneum wasoften difficult. The useof plates witha shallow agar

depth with the auxanographic method

mini-mized the problem of background growth so

thatpositiveand negative results with thisand other organismswereeasily differentiated.

With both methods, initial negative cello-biose results were observed more frequently thananyother incorrect reaction. Initial

nega-tive Wickerham assimilation tests for cello-biose were mostoften obtained with Candida tropicalis. We did not observe this problem with C. tropicalis when we used our modified auxanographic procedure. The dependence of cellobiose assimilationuponthepresence ofan

inducibleenzyme system (4, 6) may accountfor the frequency of false-negative reactions. The

discrepancy with cellobiose between the two

assimilation methods also may reside in the

degree of aeration of cultures. Agitation of Wickerham broth tubes, which was not done

inthis study, has been shown toenhance the assimilation ofcarbohydrates byyeasts(2).The

more successful detection of positive cellobiose assimilation results for C. tropicalis by the

auxanographicmethodmaybe duetogrowthof

theyeast onthe pourplate surface. These

in-consistencies, whichwereencounteredwith the

Wickerhamassimilation method,maybeof

in-terest to those using the procedure for yeast

identification.

Theagreementbetweenthemodified

auxan-ographic and Wickerham broth methods was

foundtobe 98.2% oninitial testing and 99.9%

whendiscrepancieswerereevaluated. The

aux-anographicprocedure, inourhands, wasmore

reproducible than the reference Wickerham

method, yielding correct assimilation results

andaccurateidentificationof isolatesmore

fre-quentlyon initialtesting.

Thehigh degree ofreproducibilityevidenced by themodified auxanographic

procedure

may

be attributed to its simplicity and the ease of interpreting results. Other advantages of the

auxanographic method include: theelimination

ofinoculum starvation; use of easily prepared reagents that can be sterilized by autoclaving

and stored before use; and minimal time and

skill requirements for the performance ofthe

test. The shortertime required forcompletion

of assimilation reactions represents an

addi-tional advantageof our method over the refer-encemethod. The modified auxanographic

pro-cedure is an accurate and practical means of assimilation testing thatmaybe useful for the identification of medically important yeasts in

diagnostic laboratories.

LITERATURE CITED

1. Adams, E. D., and B. H. Cooper. 1974. Evaluation of a modified Wickerham medium for identifying medi-cally important yeasts. Am. J. Med. Technol. 40:377-388.

2. Ahearn, D. G., F. J. Roth, Jr., J. W. Fell, and S. P. Meyers. 1960. Use of shaken cultures in the assimila-tion testfor yeast identification. J. Bacteriol. 79:369-371.

3. Beijerinck, M. W. 1889. L'auxanographic ou la methode

del'hydrodiffusion dansla gelatine appliqueeaux re-cherches microbiologiques. Arch. Neerl.Sci.Exactes Nat. 23:367-373.

4. Fiol, J. B. 1975. A critical study of the taxonomic value

ofsome testsofassimilationusedfor the

classifica-tionofthe sporogenous yeasts. Mycopathology 57:79-88.

5. Huppert, M., G. Harper, S. H. Sun, and V. Delanerolle. 1975. Rapid methods for identification of yeasts. J. Clin. Microbiol. 2:21-34.

6. Kaplan, J. G. 1965. An inducible system for the hydroly-sisand transport of8-glucosidesinyeasts. I. Charac-teristics of the ,3-glucosidase activityof intact and

lysedcells. J. Gen. Physiol.48:873-886.

7. Klainer, A. S., and W. R. Beisel. 1969. Opportunistic infection: a review. Am. J. Med. Sci. 258:431-456. 8. Land,G. A., E. C. Vinton, G. B. Adcock, and J. M.

Hopkins.1975. Improved auxanographic method for yeast assimilations: comparison with other

ap-proaches.J. Clin.Microbiol.2:206-217.

9. Lodder, J. (ed.). 1970. The yeasts. North-Holland

Pub-lishing Co., Amsterdam.

10. McGowan, J. E., Jr., M. W.Barnes,andM. Finland. 1975. Bacteremia at Boston City Hospital. Occur-renceandmortality during12selected years (1935-1972), with special reference to hospital acquired cases.J.Infect. Dis. 132:316-335.

11. Rifkind, D., T.L.Marchioro,S.A.Schneck,and R. B. Hill, Jr. 1967. Systemicfungalinfections

complicat-ing renal transplantation and immunosuppressive

therapy.Am. J.Med. 43:28-38.

12. Silva-Hutner, M., and B. H.Cooper. 1974.Medically

important yeasts, p. 491-507. In E. H. Lennette,E.

Spaulding,andJ. P.Truant(ed.),Manual ofclinical

microbiology,2nd ed. AmericanSocietyof Microbiol-ogy,Washington,D.C.

13. Webb, C. D., C. Papageorge, and C. T. Hall. 1971. Identification of yeasts. Center for DiseaseControl,

Atlanta, Ga.

14. Wickerham,L.J.1951.Taxonomyofyeasts. U.S.Dept. ofAgriculture Tech. Bull. 1029. U.S. Dept.of Agri-culture, Washington,D.C.

http://jcm.asm.org/