Yeast identification: reassessment of assimilation tests as sole universal identifiers

Transcription

1 Letters in Applied Microbiology ISSN ORIGINAL ARTICLE Yeast identification: reassessment of assimilation tests as sole universal identifiers J. Spencer 1, S. Rawling 1, M. Stratford 2,3, H. Steels 3, M. Novodvorska 2, D.B. Archer 2 and S. Chandra 1 1 GSK GlaxoSmithKline, Nutritional Healthcare R&D, Royal Forest Factory, Coleford, Gloucestershire, UK 2 School of Biology, University Park, University of Nottingham, Nottingham, UK 3 Mologic Ltd, Colworth Science Park, Sharnbrook, Bedford, UK Keywords assimilation test, D1 D2 sequencing, identification, species, yeast. Correspondence Sachin Chandra, GSK GlaxoSmithKline, Nutritional Healthcare Future Group, 980 Great West Road, Brentford, Middlesex, TW8 9GS UK. sachin.x.chandra@gsk.com : received 1 December 2010, revised 20 July 2011 and accepted 21 July 2011 doi: /j x x Abstract Aims: To assess whether assimilation tests in isolation remain a valid method of identification of yeasts, when applied to a wide range of environmental and spoilage isolates. Methods and Results: Seventy-one yeast strains were isolated from a soft drinks factory. These were identified using assimilation tests and by D1 D2 rdna sequencing. When compared to sequencing, assimilation test identifications (MicroLogÔ) were 18Æ3% correct, a further 14Æ1% correct within the genus and 67Æ6% were incorrectly identified. The majority of the latter could be attributed to the rise in newly reported yeast species. Conclusions: Assimilation tests alone are unreliable as a universal means of yeast identification, because of numerous new species, variability of strains and increasing coincidence of assimilation profiles. Assimilation tests still have a useful role in the identification of common species, such as the majority of clinical isolates. Significance and Impact of the Study: It is probable, based on these results, that many yeast identifications reported in older literature are incorrect. This emphasizes the crucial need for accurate identification in present and future publications. Introduction The identification and naming of a species are dependent upon its properties. Conversely, knowledge of a species name provides access to its phylogeny and taxonomy and enables prediction of its physiological characteristics. It is essential that new species are correctly identified, especially as that prevents erroneous information appearing in databases and perpetuating confusion. Since the discovery, more than 100 years ago, that yeast exist in a variety of species, the correct identification of yeasts has been a prime requisite for researchers in a number of fields. Medical practitioners, for example, require correct identification to allocate appropriate treatments to pathogenic species (Pfaller et al. 2010), and patent legislation requires accurate determination of the relevant species. In the food context, it is essential to differentiate the small number of spoilage yeasts from the numerous environmental isolates (Pitt and Hocking 1997). Historically, yeasts were first identified by their macroscopic and microscopic morphologies and assigned to a mould-based taxonomy. Early research showed that different yeast species varied in the sugars fermented. This was greatly expanded into a battery of assimilation, fermentation and physiological tests that could be used to identify individual yeast species (Barnett et al. 2000). A number of inexpensive commercial assimilation test kits are now available for yeast identification. More recently, yeast identification and taxonomy have been transformed by DNA-based methods. It was found that rdna sequences varied consistently amongst yeast species, and the divergence was sufficient to resolve species (Kurtzman and Robnett 1998, 2003). As a general rule, conspecific yeast species varied by <1% in the Letters in Applied Microbiology 53, ª 2011 The Society for Applied Microbiology 503

2 Yeast identification methods J. Spencer et al. D1 D2 region of the 26S ribosomal subunit (Kurtzman and Robnett 1998), although ITS spacer sequences were more variable, and useful for strain identification within species. At present, it is possible to identify yeasts accurately by DNA sequencing; however, this is time-consuming and expensive. Faster and cheaper pyrosequencing may be applied to limited sequences of defined species (Borman et al. 2010). Commercial assimilation test methods are cheaper, especially for large numbers of identifications. However, in nonexpert hands, assimilation tests can be difficult to interpret, particularly as certain strains of single species are known to give variable results to some tests. It has been estimated that up to 50% of spoilage yeasts in the literature are incorrectly identified (Stratford 2006). Furthermore, the number of recognized yeast species is rising steeply. In the year 1979, 437 yeast species were recognized; in 2000, this had risen to c. 800 (Barnett et al. 1979, 2000). Ten years later, the figure is c yeast species (Kurtzman et al. 2011), while the numbers of assimilation tests has remained largely static. Therefore, the statistical probability of multiple yeast species having the same assimilation pattern is steadily increasing. In this article, a large number of yeast strains were collected from the environment of a soft drinks factory and subjected to identification by assimilation tests and by D1 D2 sequencing. The results were used to compare methods and indicate areas of accord or discrepancy. Materials and methods Yeast isolation An ecological survey carried out in a UK soft drinks factory and its environs, yielded >3200 yeast isolates. Some were from air-settle plates, whilst others were isolated from surface swabs, from wooden pallets, from drains and from spilled products, ingredients and sugar syrup samples. Isolates were cultured on OMEA agar (0Æ01% w v oxytetracycline malt extract agar). Yeast colonies were counted, picked off and individually resuspended in sterile water in 96-well microtitre dishes. Resistance to preservatives was assessed by replica-plating from the 96-well dishes onto YEPD agar ph 4Æ0 containing progressively higher concentrations of sorbic acid, 0 4 mmol l )1 in 0Æ5 mmol l )1 increments. Yeasts were assessed by macroscopic and microscopic morphology, for colony colour shape; cell size, shape; and bud morphology, and by location of isolation. From these tests, yeast isolates were grouped into 71 strains. Multiple isolates from the same location, identical in all respects were assumed to be clones, grown from the same strain. Eight groups from sugar-rich locations contained isolates. Each strain was re-streaked on OMEA agar to ensure purity. All multiple isolate groups were tested using API ID 32C kits (Biomerieux, Marcy l Etoile, France) on 10 50% of isolates showing that isolates within groups had identical carbon-assimilation patterns. The yeast flora used in this study is typical of that found in soft drinks factories (unpublished results). Yeast assimilation test method Cultures were grown on malt extract agar for 48 h at 25 C. Suspensions of yeast were prepared in 15 ml sterile water at inoculum density transmittance level 47 ± 2%. For all strains, a BiologYT MicroPlateÔ (Biolog, Hayward, CA, USA) was inoculated with 100 ll of cell suspension. The microplates contain 94 biochemical tests, listed in Table S1. Microplates were incubated for 24, 48 and 72 h at 25 C. Wells were colourless when inoculated. During incubation, yeast respiration in wells containing compounds that can be utilized will either reduce the tetrazolium dye forming a purple colour or initiate growth leading to an increase in turbidity. Each metabolic pattern was read by a MicroStation (Bio Tek, Bedfordshire, UK) and interpreted by Biolog Micro- LogÔ 3 software ver (Biolog, Hayward, CA). Colorimetric or turbidity change in each well was referenced against negative control wells. Microplate wells are scored as negative ()), as positive (+) or as borderline (\). The pattern is cross-referenced to a library of species. If an adequate match is found, an identification of the isolate is made (standard Biolog methodology). 26S D1 D2 rdna sequencing Yeasts were cultured on YEPD agar for 2 days. Approximately 200 ll of yeast was removed, added to 0Æ5 g sterile glass beads (1 mm) in 2 ml Eppendorf tubes and frozen at )80 C. Eight hundred microliters of extraction buffer (50 mmol l )1 Tris ph 7Æ5, 10 mmol l )1 EDTA, 50 mmol l )1 NaCl, 1% w v SDS) was added to thawing yeast and vortexed for 2 min. Samples were heated to 65 C for 30 min and cooled to 22 C, before performing two phenol extraction steps and ethanol precipitation of the DNA. Following DNA quantification, the 26S rdna D1 D2 sequences were amplified by PCR, as described by Kurtzman and Robnett (1998). The primers used were NL1: 5 -GCATATCAATAAGCGGAGGAAAA-3 and NL4: 5 -GGTCCGTGTTTCAAGACGG-3. Amplified D1 D2 DNA was sequenced as described by Kurtzman and Robnett (1998) using the aforesaid primers. The sequences were compared with the EMBL-EBI database using blast (standard fungi) ( 504 Letters in Applied Microbiology 53, ª 2011 The Society for Applied Microbiology

3 J. Spencer et al. Yeast identification methods Because of the many misidentified fungi on the database, species type strains listed by Kurtzman et al. (2011) were used as reference points. Results A total of 3202 yeast isolates were obtained from a soft drinks factory. Most were in low numbers and diverse in appearance, indicating nongrowing isolates. Samples from the few sugary locations, where yeasts proliferate, yielded hundreds of identical isolates. Yeast isolates were grouped into 71 strains, from their location, morphology and preservative resistance, before identification by assimilation test (BiologÔ) and D1 D2 sequencing. The majority of groups were also tested using API 32C test kits. Raw identification results are shown in Table S2. In summary (Fig. 1), the two methods were in agreement for 13 strains (18Æ3%), and to within the genus for an additional 10 strains (14Æ1%) allowance being made for perfect (sexual) and imperfect (asexual) genera such as Cryptococcus and Filobasidium. Contradictory results were obtained for 48 strains (67Æ6%). In deciding which of the two identification methods yielded the correct result, note was taken of yeast morphology and behaviour, such as the red coloration of Rhodotorula spp. and the preservative resistance of Zygosaccharomyces spp. These further tests indicated that the sequence results appeared correct in every case. Assimilation test results were compared at 24, 48 and 72 h. In general, the identification did not change over this time. But percentage probability of identification increased. Further tests at 1 week showed little difference to the 72 h results. It appears that problems with identification by assimilation test could not be overcome by Percentage of strains Correct species Correct genus Wrong ID - moulds Wrong ID - not on database Wrong ID - on database Figure 1 Assimilation test identifications compared with D1 D2 sequencing results as a proportion of the 71 yeast strains tested. Yeasts were correctly identified to the species level, to the genus level or misidentified. Many misidentifications were because of newly reported species or moulds, not on the assimilation database. prolonged incubation. Assimilation tests using the API 32C method gave similarly unsatisfactory results. Overall, the sequences gave a yeast flora composed primarily of Cryptococcus, Rhodotorula, Candida and Zygosaccharomyces spp. but when the isolate numbers were examined the majority of the flora was from within the Candida genus. The yeast strains correctly identified by assimilation tests are shown in Table 1. These are again predominantly from the Candida genus, Candida ernobii being the imperfect form of Pichia holstii, together with two strains of Zygosaccharomyces rouxii. The Candida genus contains most of the common pathogenic yeasts, and it is possible that assimilation identification systems may be focused primarily on this important area. Misidentification appeared to be because of several causes. First, new or newly reported yeast species may not be currently recognized on the database. Second, there may be variability of individual strains to certain assimilation tests and, third, several yeast species have near-identical assimilation profiles. Examination of the yeasts listed on the Biolog database showed that, of the 48 wrongly identified strains, 28 were not on the yeast database (Fig. 1). Three strains of Aureobasidium were listed on the database as filamentous fungi, not yeast. Aureobasidium is a mould genus but is always isolated amongst yeasts as initial growth on agar yields yeast-like colonies, typically white or pinkish, which only become filamentous (sometimes black) after several days have passed. A further four strains, not listed on the database, were putative, new yeast species. The sequences of two, a Starmerella sp. and a Rhodotorula sp., differed by >1% from any sequence present on the EMBL-EBI database, generally indicating a new species. The third new species (two strains) had previously been entered on the EMBL- EBI database, and named as Candida flavicola, but this specific name had not been formally described in the literature. A new species is only formally recognized the following publication in hard copy in the scientific literature, of its name, properties and Latin description. One strain of a new species of Aureobasidium was found, and one strain, Cryptococcus sp. nov., closely related to Cryptococcus albidus, was identified as the latter by assimilation tests. These six strains, >8% of the 71 isolated, serve to demonstrate the potential extent of un-discovered species in the fungal kingdom. For such species, no assimilation test results have been published, rendering identification by assimilation tests impossible. Of the other wrongly identified yeast strains not on the assimilation database, the majority have been relatively recently reported, examples being Zygosaccharomyces lentus and Starmerella meliponinorum. Clearly, updating the assimilation databases is not keeping pace with reporting of new species. Letters in Applied Microbiology 53, ª 2011 The Society for Applied Microbiology 505

4 Yeast identification methods J. Spencer et al. Biolog identification D1 D2 sequence identification Correct identity 5 Strains Candida boidinii 100% Candida boidinii (99Æ6% NRRL Y-2332 U70242) 2 Strains Candida parapsilosis 95% Candida parapsilosis (99Æ4% NRRL Y U45754) Cryptococcus albidus 90% Cryptococcus sp. nov. (98Æ6% C. albidus CBS 142 AF075474) 2 Strains Pichia holstii 95% Candida ernobii (99Æ2% NRRL Y U70241) Williopsis californica 100% Williopsis californica (99Æ8% NRRL Y U75957) 2 Strains Zygosaccharomyces rouxii 87% Zygosaccharomyces rouxii (99Æ8% NRRL Y-229 U72163) Misidentified Rhodotorula acheniorum 80% Candida vartiovaariai (99Æ6% NRRL Y-6701 U69875) 9 Strains Mixed ID s (see Table 2) Cryptococcus magnus (100% CBS 140 AF181851) Candida multis-gemmis 99% Debaryomyces hansenii (99Æ8% NRRL Y-7426 U45808) Pichia ohmeri 73% Lachancea fermentati (99Æ8% NRRL Y-1559 U84239) Endomycopsella vini 84% Pichia guilliermondii (99Æ8% NRRL Y-2075 U45709) 2 Strains Mixed ID s Rhodotorula graminis (100% CBS 2826 AF070431) 5 Strains Mixed ID s Rhodotorula mucilaginosa (100% CBS 316 AF070432) Table 1 Yeast strains correctly identified by assimilations tests, and misidentified strains which are present on the assimilation test database. Biolog identifications include a figure for percentage probability of identity. D1 D2 identifications include percentage DNA identity with a designated type strain and large subunit EMBL-EBI accession number Twenty other yeast strains were wrongly identified, whose profiles were present on the assimilation database (Table 1). Sixteen of these 20 were environmental Basidiomycete yeasts. The primary cause of the misidentification of these strains may be due to strain variability within the species. D1 D2 sequencing revealed a total of 12 strains of Cryptococcus magnus, with three different sequence patterns. However, these 12 strains possessed a wide variety of assimilation profiles (Table 2), leading to a corresponding variety of misidentifications. Seven different assimilation profiles were found, only one of which was close to that listed for C. magnus in the assimilation database. Others were similar to species of Bulleromyces, Candida, Pichia or Zygosaccharomyces. The third possibility for misidentification by assimilation tests concerns the possibility that strains of widely differing species may still give the same assimilation profile. Here, it appears probable that such a case exists. Four strains of S. meliponinorum were isolated and not correctly identified because this species is not on the assimilation database. However, the four S. meliponinorum strains all gave assimilation profiles indistinguishable from Cryptococcus albidus var. aerius (99% probability or greater). Discussion Clearly, there is a need for a universal system for the correct identification of all yeast species, from all environments. In the past, when the number of recognized yeast species was limited to a few hundred, assimilation tests provided such a system. In pioneering work, Barnett et al. Biolog identification D1 D2 sequence identification 2 Strains Bulleromyces albus 79% C. magnus (100% CBS 140 AF181851) Bulleromyces albus 94% C. magnus (99Æ6% CBS 140 AF181851) Candida glabrata 73% C. magnus (99Æ6% CBS 140 AF181851) Candida vartiovaariai 86% C. magnus (99Æ8% CBS 140 AF181851) Cryptococcus albidus var. aerius 96% C. magnus (100% CBS 140 AF181851) 2 Strains C. albidus var. aerius 99% C. magnus (99Æ6% CBS 140 AF181851) Pichia mexicana 96% C. magnus (99Æ6% CBS 140 AF181851) Zygosaccharomyces bailii 77% C. magnus (99Æ6% CBS 140 AF181851) 2 Strains No identification possible C. magnus (99Æ6% CBS 140 AF181851) Table 2 Assimilation test results of 12 yeast strains identified by D1 D2 sequencing as Cryptococcus magnus. Biolog identifications include a figure for percentage probability of identity. D1 D2 identifications include percentage DNA identity with a designated type strain and large subunit EMBL-EBI accession number 506 Letters in Applied Microbiology 53, ª 2011 The Society for Applied Microbiology

5 J. Spencer et al. Yeast identification methods (1979) used assimilation tests and physiological parameters to assemble a key for identification of 437 yeast species, a system readily lending itself to computer-aided probabilistic identification. Thirty years on with c recognized yeast species and no significant changes in the numbers of assimilation tests, it appears that assimilation tests alone cannot be relied upon as an identification system for all yeasts from any environment. The results shown here suggest that 67% of such species are misidentified by assimilation tests. Arias et al. (2002) came to a similar conclusion when examining yeasts isolated from orange juice, with only 13 35% of yeasts correctly identified by assimilation methods, while only 25% of rarer clinical isolates were correctly identified (Cendejas-Bueno et al. 2010). The majority of the 67% misidentifications by assimilation testing appear to be caused by new and newly reported yeast species. Is it possible that this problem could be resolved simply by regular updating of new species into assimilation databases? The data shown here suggest that such a course of action would decrease the misidentification rate considerably, from 67 to 32%, but still leaving a high level of misidentification. The discovery of new yeast species is increasing rapidly, and even in the limited survey of a normal environment shown here, 8Æ4% of strains were putative new species. New species are likely to cause progressively more misidentification with time. Furthermore, on discovery of a new species, many researchers are increasingly not carrying out the time-consuming tests required for formal announcement of a new species and are merely registering the relevant DNA sequences. By statistical probability, the great majority of yeasts discovered in the last century are the most common species. Newly discovered species are increasing likely to represent rarer yeasts, from increasingly obscure habitats. It appears that assimilation tests are most effective in the identification of the most common species, i.e. the tests were designed to identify the yeasts commonly isolated at the time. Similar conclusions have been noted in yeast identification of medical isolates. The most common medical yeast isolates are Candida albicans, Candida glabrata, Candida tropicalis, Candida parapsilosis and Candida krusei = Issatchenkia orientalis (Putignani et al. 2008; Pfaller et al. 2010). These common isolates can be readily identified by commercial assimilation tests such as API Candida, Auxacolor or Vitek (Verweij et al. 1999; Linton et al. 2007), but many rare and unusual medical species cannot be identified by assimilation tests (Linton et al. 2007; Putignani et al. 2008; Cendejas-Bueno et al. 2010). Future developments are likely to yield new highly specific, molecular methods targeted at pathogenic yeasts, such as pyrosequencing methodology (Montero et al. 2008; Borman et al. 2010) or species-specific PCR primers. Other microbes, such as clinical bacteria, have been similarly successfully identified using assimilation methods (Holmes et al. 1994; Morgan et al. 2009), although atypical bacteria were problematic. In filamentous fungi, such as Zygomycetes, most carbon-assimilation profiles correlated with species (Schwarz et al. 2007), but certain species, e.g. Rhizopus oryzae, gave highly variable profiles. In the current publication, C. magnus showed similar strain variability. Such yeast strain variability has long been recognized for certain tests (Barnett et al. 1979, 2000) with results being marked positive, negative, variable and slow or delayed. One difficulty found using growth or assimilation of different nutrients concerns the viability and vitality of the yeast concerned, which affects duration of incubation time (Preston-Mafham et al. 2002). Having a fixed inoculum level will certainly help, but with unknown species, the time and temperature of incubation must be, to some extent, guesswork. Slow-growing species may only visibly grow on the most favoured nutrients and give false-negative results for other nutrients. In such a case, identification will default to the species growing on fewest types of carbon sources, most likely Zygosaccharomyces bisporus and Zygosaccharomyces bailii. It is clearly a failing of the assimilation method in the food context, that the default identification for a slow-growing species, is the most significant of all spoilage yeast species (Zygosaccharomyces bailii). To conclude, assimilation tests alone are unreliable as a sole, universal means of yeast identification but used beside methods, such as sequencing, still have a useful role in primary screening and identification of the most common clinical isolates. Acknowledgements We gratefully acknowledge the help of Sue Redwood, GSK Royal Forest Factory for assistance in the gathering of yeast samples and a BBSRC Defra Link project (FQI28) for part funding this work. References Arias, C.R., Burns, J.K., Friedrich, L.M., Goodrich, R.M. and Parish, M.E. (2002) Yeast species associated with orange juice: evaluation of different identification methods. Appl Environ Microbiol 68, Barnett, J.A., Payne, R.W. and Yarrow, D. (1979) A Guide to Identifying and Classifying Yeasts. Cambridge, UK: Cambridge University Press. Letters in Applied Microbiology 53, ª 2011 The Society for Applied Microbiology 507

6 Yeast identification methods J. Spencer et al. Barnett, J.A., Payne, R.W. and Yarrow, D. (2000) Yeasts: Characteristics and Identification. Cambridge, UK: Cambridge University Press. Borman, A.M., Linton, C.J., Oliver, D., Palmer, M.D., Szekely, A. and Johnson, E.M. (2010) Rapid molecular identification of pathogenic yeasts by pyrosequencing analysis of 35 nucleotides of internal transcribed spacer 2. J Clin Microbiol 48, Cendejas-Bueno, E., Gomez-Lopez, A., Mellado, E., Rodriguez- Tudela, J.L. and Cuenca-Estrella, M. (2010) Identification of pathogenic rare yeast species in clinical samples: comparison between phenotypical and molecular methods. J Clin Microbiol 48, Holmes, B., Costas, M., Ganner, M., On, S.L.W. and Stevens, M. (1994) Evaluation of Biolog system for identification of some gram-negative bacteria of clinical importance. J Clin Microbiol 32, Kurtzman, C.P. and Robnett, C.J. (1998) Identification and phylogeny of ascomycetous yeasts from analysis of nuclear large subunit (26S) ribosomal DNA partial sequences. Antonie Van Leeuwenhoek 73, Kurtzman, C.P. and Robnett, C.J. (2003) Phylogenetic relationships among yeasts of the Saccharomyces complex determined from multigene sequence analysis. FEMS Yeast Res 3, Kurtzman, C.P. and Fell, J.W. and Boekhout, T. (2011) The Yeasts, a Taxonomic Study, 5th edn. Amsterdam: Elsevier. Linton, C.J., Borman, A.M., Cheung, G., Holmes, A.D., Szekely, A., Palmer, M.D., Bridge, P.D., Campbell, C.K. et al. (2007) Molecular identification of unusual pathogenic yeast isolates by large ribosomal subunit gene sequencing: 2 years of experience at the United Kingdom Mycology Reference Laboratory. J Clin Microbiol 45, Montero, C.I., Shea, Y.R., Jones, P.A., Harrington, S.M., Tooke, N.E., Witebsky, F.G. and Murray, P.R. (2008) Evaluation of pyrosequencing technology for the identification of clinically relevant non-dematiaceous yeasts and related species. Eur J Clin Microbiol Infect Dis 27, Morgan, M.C., Boyette, M., Goforth, C., Volpe Sperry, K. and Greene, S.R. (2009) Comparison of the Biolog Omnilog identification system and 16S ribosomal gene sequencing for accuracy in identification of atypical bacteria of clinical origin. J Microbiol Methods 79, Pfaller, M.A., Diekema, D.J., Gibbs, D.L., Newell, V.A., Ellis, D., Tullio, V., Rodloff, A., Fu, W. et al. (2010) Results from the ARTEMIS DISK global antifungal surveillance study, 1997 to 2007: a 10.5-year analysis of susceptibilities of Candida species to fluconazole and voriconazole as determined by CLSI standardized disk diffusion. J Clin Microbiol 48, Pitt, J.I. and Hocking, A.D. (1997) Fungi and Food Spoilage, 2nd edn. London, Weinheim, New York, Tokyo, Melbourne, Madras: Blackie Academic and Professional. Preston-Mafham, J., Boddy, L. and Randerson, P.F. (2002) Analysis of microbial community functional diversity using sole-carbon-source utilization profiles a critique. FEMS Microbiol Ecol 42, Putignani, L., Paglia, M.G., Bordi, E., Nebuloso, E., Pucillo, L.P. and Visca, P. (2008) Identification of clinically relevant yeast species by DNA sequence analysis of the D2 variable region of the 25 28S rrna gene. Mycoses 51, Schwarz, P., Lortholary, O., Dromer, F. and Dannaoui, E. (2007) Carbon assimilation profiles as a tool for identification of Zygomycetes. J Clin Microbiol 45, Stratford, M. (2006) Food and beverage spoilage yeasts. In Yeasts in Food and Beverages ed. Querol, A. and Fleet, G.H. pp Berlin, Heidelberg: Springer Verlag. Verweij, P.E., Breuker, I.M., Rijs, A.J.M.M. and Meis, J.F.G.M. (1999) Comparative study of seven commercial yeast identification systems. J Clin Pathol 52, Supporting Information Additional Supporting Information may be found in the online version of this article: Table S1 List of the 94 assimilation biochemical tests used in the Biolog YT MicroPlate designed to identify characterise a broad range of yeasts Table S2 Yeast strain identification made using assimilation tests (Biolog) and by D1 D2 rdna sequencing Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. 508 Letters in Applied Microbiology 53, ª 2011 The Society for Applied Microbiology