Evaluation of the API ATB 32C system for the rapid identification of foodborne yeasts

International Journal of Food Microbiology, 11 (1990) 215-224 Elsevier

215

FOOD 00326

Evaluation of the API ATB 32C system for the rapid identification of foodborne yeasts H. Rohm, Frieda Lechner* and Marietta Lehner Department of Dairy Science and Bacteriology, University of Agriculture, Vienna, Austria

(Received 20 November 1989; accepted 18 April 1990)

The commercial API ATB 32C identification kit was compared with a standard method for identifying 11 reference strains and 53 yeast strains isolated from fermented milk products. Approx. 50% of the species considered in the API ATB 32C database were identified on a level of good, very good, and excellent identification. The numerical profile of 25 strains was not found in the API ATB 32C index. Low discrimination or misidentification was observed in seven strains. The low reliability of the API ATB 32C system may be ascribed to the incomplete nature of the profile index. A majority (91%) of the strains, however, were identified correctly by the API ATB 32C strip test results in combination with the commercial computer program of Barnett et al. (1985). This combined procedure offers the possibility to identify any out of 497 species considered by Barnett et al. (1985). Key words: Yeast identification; Yeast, foodborne; Yeast identification kits

Introduction D u e to the increasing i m p o r t a n c e of yeasts in the p r o d u c t i o n a n d spoilage of foods, commercially available systems p r i m a r i l y developed for the i d e n t i f i c a t i o n of medically i m p o r t a n t yeasts b e c a m e an object of interest for n o n - c l i n i c a l purposes. The majority of these systems are based m a i n l y o n c a r b o n c o m p o u n d assimilation tests a n d a few other biochemical assays. Some of the systems have to be used in c o n j u n c t i o n with morphological e x a m i n a t i o n s for definitive identification, T h e databases of the systems include always a limited n u m b e r of certain yeast species c o m m o n l y associated with the clinical e n v i r o n m e n t . These m e t h o d s offer the adv a n t a g e of easy storage, rapid p r e p a r a t i o n a n d simple m a n i p u l a t i o n . I d e n t i f i c a t i o n is possible within 4 8 - 7 2 h, whereas traditional procedures as described b y V a n der W a l t a n d Y a r r o w (1984) require approx. 2 - 3 weeks. I n general, the reliability of the c o m m e r c i a l testing systems is satisfactory a n d matches the results of c o n v e n t i o n a l m e t h o d s w h e n clinical isolates are identified. By

Correspondence address: H. Rohm, Department of Dairy Science and Bacteriology, University of Agriculture, Gregor Mendel Strasse 33, A-1180 Vienna, Austria. * Present address: Government Research Institute for Dairy Science, A-6200 Rotholz, Tyrol, Austria.

0168-1605/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

216

using the Uni-Yeast-Tek system, Bowman and Ahearn (1975, 1976) and Cooper et al. (1978) identified 89-99% of clinical isolates correctly. In similar investigations, percentages of 80-92% were reported for the Quantum II yeast identification system (Kiehn et al., 1985; Salkin et al., 1985), 96% for the Abbott yeast identification system (Cooper et al., 1984), 85-96% for the AutoMicrobic system yeast biochemical card (Hasyn and Buckley, 1982; Land et al., 1984; Oblack et al., 1981), 94-100% for the AP! 20C method (Bowman and Ahearn, 1976; Roberts et at., 1976; Kiehn et al., 1985; Salkin et al., 1985), and 97% for the modified API 20C system (Buesching et al., 1979; Land et al., 1979). Several testkits have been evaluated with respect to their reliability in the identification of foodborne yeasts (Busse, 1987; Deak and Beuchat, 1988). Correlation with conventional methods was not as good as it was when using clinical isolates. In many cases, reaction profiles failed to appear in the databases, and sometimes, erroneous reactions caused misidentification. The present study was undertaken to evaluate the identification reliability of the newly developed API ATB 32C rapid identification system and to compare the results with the standard identification method as described by Barnett et al. (1983, 1985).

Experiments Yeast strains

Strains of established identity used in this study were obtained from the Deutsche Sammlung fiir Mikroorganismen, GSttingen, F.R.G. (Table I). Yeasts isolated from sour milk products (yoghurt, sour milk and kefyr) by using yeast extract-glucose-chloramphenicol agar (International Dairy Federation, 1980) were purified and maintained on 2% malt extract (ME)-agar (OXOID no. CM 59) and transferred to slants of ME-agar prior to identification (van der Walt and Yarrow, 1984). Some strains isolated from an analogue environment were kindly provided by Dr. H. Seiler (Dept. of Microbiology, Munich-Weihenstephan, F.R.G.).

TABLE 1 Reference strains used in the study Yeast species

DSM a no.

Yeast species

DSM no.

Candida guilliermondii Candida intermedia Candida kefyr Kluyveromyces marxianus Pichia farinosa Pichia membranaefaciens

70057 70753 70073 70804 70362 70373

Rhodotorula glutinis Saccharomyces cerevisiae Torulaspora delbrueckii Williopsis saturnus Zygosaccharomyces rouxii

70 398 70 449 70 483 70 278 2 531

Deutsche Sammlung fiir Mikroorganismen, G~Sttingen, F.R.G.

217

Standard identification routine All references and isolates were characterized according to the standard procedures (Barnett et al., 1983; van der Walt and Yarrow, 1984). The routine testing scheme included fermentation of glucose, assimilation of carbon sources (Larabinitol, D-arabinose, z-arabinose, arbutin, cellobiose, citrate, erythritol, D-galactose, D-glucosamine, glycerol, myo-inositol, 2-keto-D-gluconate, DL-lactate, lactose, maltose, D-mannitol, melezitose, melibiose, methyl-a-D-glucoside, raffinose, Lrhamnose, ribitol, D-ribose, salicin, L-sorbose, surcrose, a-trehalose, xylitol and D-xylose), assimilation of nitrogen sources (cadaverine, creatine, creatinine, ethylamine, L-lysine, nitrate and nitrite), vitamin requirements (growth without vitamins, and without biotin, thiamin, and pyridoxine, respectively), growth at 25, 30, 35, 37 and 42 ° C, 0.01 and 0.1% cycloheximide resistance, growth at 50 and 60% glucose, splitting of urea, pink colony appearance, and the presence of pseudohyphae, true mycelium, and arhtrospores. The test results were entered into the Barnett et al. (1985) computer program (BCP). In the identification routine, this software corn-

TABLE

II

Y e a s t s p e c i e s c o n s i d e r e d i n the d a t a b a s e o f t h e A P I A T B 3 2 C s y s t e m a

Candida albicans Candida boidinii Candida catenulata Candida glabrata Candida inconspicua Candida intermedia Candida magnoliae Candida melibiosica Candida membranaefaciens Candida norvegica Candida parapsilosis Candida rugosa Candida sake Candida tropicalis Candida steatolytica Candida zeylanoides Citeromyces matritensis al Claoispora lusitaniae a2 Cryptococcus albidus Cryptococcus curoatus Cryptococcus humicolus

Cryptococcus laurentii Cryptococcus terreus Debaryomyces hansenii a3 Debaryomyces marama Debaryomyces polymorphus Filobasidiella neoformans a4 Filobasidium uniguttulatum a5 Geotrichum s p p . Geotrichum capitatum Hanseniapsora guilliermondii a6 Hanseniaspora uvarum a7 Hanseniaspora valbyensis a8 Hansenula anomala a9 Hansenula holstii alO Hansenula jadinii all lssatchenkia orientalis a12 Kluyveromyces marxianus a13 gluyveromyces marx. vat. lactis a14 gluyveromyces thermotolerans a~5 Metschnikowia pulcherrima a16 Pichia carsonii

Pichia etchellsii Pichia farinosa Pichia fermentans a17 Pichia guilliermondii al 8 Pichia membranaefaciens a19 Pichia norvegensis a2o Pichia ohmeri Pichia spartinae Rhodotorula glutinis Rhodotorula minuta Rhodotorula mucilaginosa Saccharomyces cerevisiae Saccharomyces exiguus a21 Saccharomyces kluyveri Sporidiobolus s p p . a22 Stephanoascus ciferrii a23 Torulaspora delbrueckii a24 Trichosporon beigelii Williopsis saturnus b Yarrowia lipolytica a25 Zygosaccharomyces s p p .

( B a r n e t t e t al., 1 9 8 3 ) o f : al Candida globosa; aZCandida lusitaniae; a3Candida famata; a4 Cryptococcus neoformans; aSCryptococcus uniguttulatus; a6Kloeckera apis; a7 KIoeckera apiculata; a8 Kloeckera japonica; agcandida pelliculosa; alo Candida silvicola; a11 Candida utilis; a12 Candida CFusei; a13 Candida kefyr; a14 Candida sphaerica; a15 Candida dattila; al6Candida pulcherrima; a17 Candida lambica; alSCandida guilliermondii; a19 Candida valida; a20 Candida norvegensis; 21 Candida holmii; 2 l Sporobolomyces s p p . ; a23 Candida ciferrii; a24 Candida colliculosa; a25 Candida lipolytica. b Hansenula saturnus, r e n a m e d b y B a r n e t t et al. ( 1 9 8 5 ) a Teleomorph

218 pares the entered results with the descriptions of the 497 species in its database and provides a list of all yeasts which may give the specified results. When necessary, i.e. indicated by the BCP, additional reactions to distinguish between species were performed.

API A TB 32C method The database of the ATB 32C test kit (API system, 1987), which has been developed for rapid yeast identification in the clinical environment, includes 63 species. To allow further comparisons, Table II shows the species included in this database in a modified form according to the nomenclature used by Barnett et al. (1983). The ATB 32C system consists of 30 dehydrated carbon compound assimilation media in plastic cupules. Control cupules contain a glucose basal medium and a zero-growth control medium. All tests were performed in line with the manufacturer's instructions (API system, 1987). Inocula were prepared by removing several identical colonies from ME-agar plates and transferring them into the suspension medium, which was adjusted to a visible density of approx. 2 McFarland units. 250/xl of the suspension were diluted in an ampoule of C Medium supplied with the system. By using sterile Pasteur pipettes, each of the 32 plastic cupules was inoculated with approx. 135/~1 C Medium. Strips were incubated at 30 ° C, and growth was assessed after 48 h. Based on the test results, an eight-digit profile number was constructed as outlined in the instructions (AP1 system, 1987). Identification was based on the Analytical Profile Index in conjunction with morphological characteristics. When necessary, results of additional reactions required by the Analytical Profile Index were taken from the standard identification routine as described above.

Accuracy of the Analytical Profile Index Accuracy of the Analytical Profile Index was evaluated by means of the BCP. ATB 32C test results considered in the BCP (Table III) as well as morphological reactions used for API identification were entered into the computer. Additional reactions required for the API identification method were included as well.

TABLE II1 Test reactions (assimilations) of the API ATB 32C system in alphabetical order N-acetyl-glucosamine ~ g-arabinose Blank b Cellobiose Cycloheximide c

Erythritol Esculine~,b Galactose Gluconate Glueosamine a.b Glucoseb

G l u c u r o n a t e Levulinate Me-a-D-glucoside Sorbitol Glycerol Maltose Palatinosea Sorbose b lnositol b Mannitol b Raffinose Sucrose 2-KetogluconateMelezitose R h a m n o s e a-Trehalose Lactate Melibiose Ribose D-xylose Lactose b a

Reactions not included in the Barnett et al. (1983) computer program. b Reactions not used for constructing the API ATB 32C numerical profile. c Cycloheximideresistance.

219

Results and Discussion

Table IV gives the 64 yeast strains identified according to Barnett et al. (1983, 1985) in alphabetical order and numbered consecutively as well as the corresponding API ATB 32C identification results. Isolates with identical description are

T A B L E IV Identification of reference strains and foodborne yeasts with the API ATB 32C system Strain no(s),

Species (Barnett et al., 1983, 1985)

Result a ATB 32C

Strain no(s),

Species (Barnett et al., 1983, 1985)

Result a ATB 32C

Candida guilliermondii b Candida intermedia Candida intermedia

0 + ex.i. ( + ) 1.d.

34 35,36 37-39

Kluyveromyces marxianus b Kluyveromyces marxianus c Metschnikowia reukauffii d

4,5 6 7 8 9 10 11 12 13 14

Candida intermedia Candida intermedia b Candida kefyr b Candida norvegiea Candida norvegica Candida parapsilosis c Candida parapsilosis Candida rugosa c Candida tropicalis Candida tropicalis c

40 41--44 45 46 47 48 49 50 51 52

Pichiafarinosa b Pichia guilliermondii Pichia membranaefaciens b Piehia membranaefaeiens Rhodotorula glutinis b Rhodotorula glutinis Rhodotorula minuta Rhodotorula mucilaginosa Rhodotorula mucilaginosa Saecharomyces eerevisiae b

15,16 17 18-21 22 23 24,25 26 27 28 29 30

Candida zeylanoides Clavispora lusitaniae Debaryomyees hansenii Debaryomyces hansenii Debaryomyces hansenii ~ Debaryomyees hansenii Debaryomyces hansenii Debaryomyees hansenii Endomycesfibuliger d,e Geotrichum candidum c Hanseniaspora guilliermondii Hanseniaspora uvarum Hansenula anomala Hansenula anomala

+ g.i. 0 q- ex.i. + g.i. 0 + ex.i. + v.g.i. + g.i. 0 - Candida sake + ex.i. + g.i. r 0 0 + g.i. f + g.i. f 0 0 0 + ex.i. ~

0 + ex. i. g -- Candida pulcherrima + ex.i. ( + ) I.d. 0 0 0 0 0 + g.i, + ex.i. 0

53,54 55 56 57 58 59 60 61 62 63 64

Saecharomyees cerevisiae Torulaspora delbrueckii b Torulaspora delbrueekii Torulaspora delbrueckii Triehosporon beigelii Trichosporon beigelii Triehosporon beigelii Williopsis saturnus b Yarrowia lipolytica Zygosaeeharomyees rouxii b Zygosaccharomyces rou,xii

+ q+ 0 + (+ 0 0 + 0 0

1 2 3

31 32 33

ex.i. ex.i. ex.i. ex.i. ) l.d.

ex.i.

+ v.g.i, f + g.i. f'g + ex.i. f 0

a A P I ATB 32 C results as given by the Analytical Profile Index: + , correct identification; - , incorrect identification; 0, profile not found; I.d., low discrimination; g.i., good identification; v.g.i., very good identification; ex.i., excellent identification. b Reference strain. c Identification confirmed by the Centraalbureau voor Schimmelcultures, Baarn, Netherlands. d Species not considered within the API ATB 32C database. e Isolated from leaven. A n a m o r p h s of the species given by the Analytical Profile Index (see Table II). g Identification on genus niveau by the Analytical Profile Index.

220

grouped. All reference strains and 49 out of 52 yeasts isolated from sour milk products were found to be part of the API ATB 32C database. Three isolates of Metschnikowia reukauffii as well as Endomyces fibuliger isolated elsewhere are also included in Table IV.

TABLE V I d e n t i f i c a t i o n of yeasts not d i s c r i m i n a t e d by the A P I ATB 32C s y s t e m w h e n using AP1 ATB 32C test results and the i d e n t i f i c a t i o n p r o g r a m of Barnett et al. (1985) Strain no(s). ~ 1

Identification b

False A P I A T B 32C reactions c

Debaryomyces hansenii Leucosporidium scottii d Leucosporidium scottii d

+

3 6 9 13 14 18 21 22 26 27 28 33 34 37-39

+ + +

41-44 45 46

+ + +

47 48

+ +

49 52 57

+ +

59 60 61 63 64

+ + + -

A d d i t i o n a l species characterized by the same test reactions (Barnett et al., 1985)

+ + CEL +

None INO None

Leueosporidium scottii d Candida haemulonii d, Leucosporidium scottii d XYL GAL ERY GRT None

+ + +

Brettanomyces clausenii d Candida sake, Metschnikowia bicuspidata d Metschnikowia pulcherrima, Pichia carsonii Debaryomyces hanseniL Debaryomyces vanrijae d Arxiozyma telluris d, Saccharomyces cerevisiae Arxiozyma telluris d, Candida ethanolica, Leucosporidum antarcticum d Saccharomyces cerevisiae, Zygosaccharomyces bailii Rhodotorula mucilaginosa Rhodotorula minuta, Sporidiobolus pararoseus, Sporobolomyces roseus d Rhodotorula aurantiaca d Sporobolomyces foliicola d

+

None

Candida halonitratophila d Zygosaccharomyces bailii, Zygosaccharomyces bisporus, Zygosaccharomyces rouxii LCT None None None SORMAN

a Strain identity, see T a b l e IV. b + , correct identification; - , incorrect identification. A b b r e v i a t i o n s of reactions: C E L , cellobiose a s s i m i l a t i o n (ass.); E R Y , erythritol ass.; G A L , D-galactose ass.; G R T , D-glucuronate ass.; INO, myo-inositol ass.; LCT, DL-lactate ass.; M A N , m a n n i t o l ass.: SOR, sorbitol ass.; X Y L , D-xylose ass. a Species not i n c l u d e d in the A P I ATB 32C database.

221

Out of the reference strains, Candida kefyr, Pichia farinosa and Torulaspora delbrueckii were identified excellently by the ATB 32C system. The numerical profiles of the other references failed to appear in the Index. Approx. 53% of the species isolated and considered in the API ATB 32C database were classified on a different level of certainty (10 were good, 2 were very good, and 16 were excellent identification). Low discrimination was observed when using results of Candida intermedia, Pichia guillierrnondii and Trichosporon beigelii. One isolate of Candida tropicalis with confirmed identity (Centraalbureau voor Schimmelcultures, The Netherlands) was found to be Candida sake when analysed by the API ATB 32C method. In case of 16 isolates, the numerical profiles failed to appear in the index. Out of the strains not included in the API ATB 32C database, Metschnikowia reukauffii was misidentified as Metschnikowia (Candida)pulcherrima which is related closely. The profile of Endomyces fibuliger was not found. To obtain information on the reasons for poor selectivity, the API ATB 32C test results (except those not included in the Barnett et al. (1983) computer program as outlined in Table III) as well the additional reactions used for API identification were entered into the BCP. Table V shows the results for the 35 references and isolates included in Table IV which could not be identified by the ATB 32C Index (i.e., incorrect identification, low discrimination, and profile not found). When using the ATB 32C test results and the BCP, however, 29 of the 35 yeast strains could be identified correctly. Depending on the particular combinations of the test results, additional species characterized by the same reaction patterns were sometimes given by the BCP. When entering the ATB 32C results of Candida guilliermondii (strain

TABLE VI Identification of yeasts discriminated by the API ATB 32C system when using API ATB 32C test results and the identification program of Barnett et al. (1985) Strain no(s), a

2

Additional species characterized by the same test reactions (Barnett et al., 1985)

Strain no(s), a

Additional species characterized by the same test reactions (Barnett et al., 1985)

Leucosporidium scottii b Leucosporidium scottii b

30 32 35,36 40 50 51

Hanseniaspora uoarurn

4,5 7

None

8

Pichia optuniae b

10 11 12 15,16 17 23 24,25 29

None

Pichia carsonii Candida pararugosa b None None None

Cryptoeoecus albidus Geotriehum armillariae, Geotrichum klebahnii

53,54 55 56 58 62

a Strain identity, see Table IV. b Species not included in the API ATB 32C database.

None None None

Rhodotorula glutinis, Rhodotorula minuta Rhodosporidium sphaerocarpurn b, Rhodotorula glutinis, Rhodotorula minuta None None None None None

222

No. 1), e.g., the BCP output included Candida quilliermondii and Debaryomyces hansenii. In six species the BCP identification using ATB 32C test results failed due to erroneous cupule reactions. The same procedure was applied to the strains identified by the API ATB 32C system as shown in Table VI. 16 out of 28 strains were identified definitely when combining the API ATB 32C test reactions and the BCP. In the other cases, the BCP output contained 1-3 additional species characterized by the same reaction patterns. Generally, there was a very good agreement for assimilation results between the API ATB 32C method and the standard procedure (detailed results not included). Due to the longer incubation period, especially when species are described to utilize carbon compounds slowly or weakly (Barnett et al., 1983), more tests tend to become positive with the conventional method. The results indicate that the poor performance of the API ATB 32C system is mainly caused by the incomplete nature of the Index. The Index seems to consider mainly the variability of strains found in the clinical environment. Therefore it is not surprising that identification of strains isolated from other biotopes fails. We conclude that the API ATB 32C system is very useful for identification of foodborne yeasts if the test results are computed with the BCP instead of the API ATB 32C Index. This procedure allows test results to be compared with the descriptions of 497 yeast species instead of being restricted to the limited number of species of the API ATB 32C database.

Acknowledgements Parts of the study were supported by Mr. E. Liftnegger, Bio-Merieux Austria.

References API System S.A. (1987). ATB 32C Analytical Profile Index. Montalieu Vercieu, France. Barnett, J.A., Payne, R.W. and Yarrow, D. (1983) Yeasts: Characteristics and identification. Cambridge University Press, Cambridge. Barnett, J.A., Payne, R.W. and Yarrow, D. (1985) Yeast identification program. Cambridge University Press, Cambridge. Bowman, P.I. and Ahearn, D.G. (1975) Evaluation of the Uni-Yeast-Tek kit for identification of medically important yeasts. J. Clin. Microbiol. 2, 354-358. Bowman, P.I. and Ahearn, D.G. (1976) Evaluation of commercial systems for the identification of clinical yeast isolates. J. Clin. Microbiol. 4, 49-53. Buesching, W.J., Kurek, K. and Roberts, G.D. (1979) Evaluation of the modified API 20C system for identification of clinically important yeasts. J. Clin. Microbiol. 9, 565-569. Busse, M. (1987) Testkits zur Hefenidentifizierung. Jahresbericht Siiddt. Versuchs- und Forschungsanstalt, Miinchen, Weihenstephan. Cooper, B.H., Johnson, J.B. and Thaxton E.S. (1978) Clinical evaluation of the Uni-Yeast-Tek system for rapid presumptive identification of medically important yeasts. J. Clin. Microbiol. 7, 349-355. Cooper, B.H., Prowant, S., Alexander, B. and Brunson D.H. (1984) Collaborative evaluation of the Abbott yeast identification system. J. Clin. Microbiol. 19, 853-856.

223 Deak, T. and Beuchat, L.R. (1988) Evaluation of simplified and commercial systems for identification of foodborne yeasts. Int. J. Food Microbiol. 7, 135-145. Hasyn, J.J. and Buckley, H.R. (1982) Evaluation of the Auto Microbic system for identification of yeasts. J. Clin. Microbiol. 16, 901-904. International Dairy Federation (1980) IDF-Standard 94. Kiehn, T.E., Edwards, F.F., Tom, D., Lieberman, G., Bernard, E.M. and Armstrong, D. (1985) Evaluation of the Quantum II yeast identification system. J. Clin. Microbiol. 22, 216-219. Land, G.A., Harrison, B.A., Hulme, K.L., Cooper, B.H. and Byrd, J.C. (1979) Evaluation of the new API 20C strip for yeast identification against a conventional method. J. Clin. Microbiol. 10, 357-364. Land, G., Stotler, R., Land, K . and Staneck, J. (1984) Update and evaluation of the AutoMicrobic yeast identification system. J. Clin. Microbiol. 20, 649-652. Oblack, D.L., Rhodes, J.C. and Martin, W.J. (1981) Clinical evaluation of the AutoMicrobic System Yeast Biochemical Card for rapid identification of medically important yeasts. J. Clin. Microbiol. 13, 351-355. Roberts, G.R., Wang, H.S. and Hollick, G.E. (1976) Evaluation of the API 20 C Microtube system for the identification of clinically important yeasts. J. Clin. Microbiol. 3, 302-306. Salkin, I.F., Schadow, K.H. Bankaitis, L.A., McGinnis, M.R. and Kemna, M.E. (1985) Evaluation of Abbott Quantum II yeast identification system. J. Clin. Microbiol. 22, 442-444. Van der Walt, J.P. and Yarrow, D. (1984) Methods for the isolation, maintenance, classification and identification of yeasts. In (N.J.W. Kreger-Van Rij, Ed.), The yeasts - a taxonomic study, 3rd edn., Elsevier, Amsterdam, pp. 45-104.