Comparison of killer toxin-producing and non-producing strains of Filobasidium capsuligenum: Proposal for two varieties

ARTICLE IN PRESS

Microbiological Research 163 (2008) 267—276

0944-5013/$ - sdoi:10.1016/j.

�CorrespondE-mail addr

(A. Keszthelyi)

www.elsevier.de/micres

Comparison of killer toxin-producing andnon-producing strains of Filobasidiumcapsuligenum: Proposal for two varieties

Andrea Keszthelyi�, Zsuzsanna Hamari, Ilona Pfeiffer,Csaba Vagvolgyi, Judit Kucsera

Department of Microbiology University of Szeged, H-6726 Szeged, Hungary

Received 27 October 2007; received in revised form 17 December 2007; accepted 5 January 2008

KEYWORDSFilobasidium;Killer toxin;ITS;D1/D2;RAPD

ee front matter & 2008micres.2008.01.002

ing author.ess: andrea.keszthelyi@.

SummaryThe basidiomycetous yeast, Filobasidium capsuligenum, produces killer toxin againstthe opportunistic pathogen Cryptococcus neoformans. Not every strain isolated sofar is able to produce the anti cryptococcal toxin. The aim of the present work wasto study the relationship between the toxins and the toxin-producing and non-producing isolates. The toxin was coded on chromosomal DNA in each producingstrain as molecular analysis revealed. In addition, both the killing spectra andbiochemical properties of the toxins proved to be identical, thus intraspecificvariation in the toxin was not found. For molecular typing of the isolates, the D1/D2region of 26S rDNA, partial sequences of internal transcribed spacer (ITS) regions,PCR fingerprinting RAPD and mtDNA-RFLP patterns were examinated. Phylogeneticanalyses based on the different approaches showed that strains with the ability ofkiller-toxin production and those without it differ significantly and cluster into twodistinct groups. The differences between the two groups and the similarity withinthem suggest the authority to separate the species into varieties.& 2008 Elsevier GmbH. All rights reserved.

Introduction

Filobasidium capsuligenum was isolated first in aSouth-African winery and described as Torulopsis

Elsevier GmbH. All rights rese

gmail.com

capsuligena by van der Walt and van Kerken in1961. Fell et al. (1969) classified this species asLeucosporidium capsuligenum, but later Rodriguesde Miranda (1972) reclassified it as F. capsuligenum.Additional strains had been isolated from differentlocation of the world (Japan, Great Britain,Germany, Italy and Turkey) and from varioussources (brewery, soil and fruits). Up to now only

rved.

ARTICLE IN PRESS

A. Keszthelyi et al.268

a small number of isolates are available fromculture collections. Some strains were publishedto secrete extracellular enzymes, such as amylase(De Mot and Verachtert, 1985, 1987) and protease(Yamamoto, 1991), and another one (IFM 40078/CBS 4736) to secrete a killer toxin (FC-1) (Golubevand Kuznetsova, 1991; Keszthelyi et al., 2006).FC-1 was strongly active against all varieties ofCryptococcus neoformans derived from eitherclinical sample or from the environment.F. capsuligenum mycocin is active not only onCryptococcus species but also against the species ofthe genera Atractogloea, Fellomyces, Filobasidiel-la, Filobasidium, Sebacina and Tremella (Golubevand Kuznetsova, 1991).

The ability of yeasts to produce killer toxin(mycocin), by which other yeasts could be de-stroyed, was first found in Saccharomyces cerevi-siae (Markower and Bevan, 1963). Since then manykiller strains from both ascomycetous and basidio-mycetous yeast genera have been described. In thecase of S. cerevisiae the killer phenotype could berelated to either chromosomal DNA or extrachro-mosomal elements; in the case of other yeastspecies killer properties could be coded exclusivelyby extrachromosomal elements (dsRNA viruses orDNA plasmids) or chromosomal DNA. It has longbeen known that mycocin production is strain butnot species dependent. Moreover, different strainsof one species can produce different mycocins(Cryptococcus laurentii, Pichia membranifaciens,Rhodotorula mucilaginosa, S. cerevisiae).

As only limited number of comparisons on theproperties of killer toxins and their geneticdetermination within the same species could befound in the literature, the aim of the present workwas to compare 10 different F. capsuligenum strains

Table 1. Origin of Filobasidium capsuligenum strains

Strains Location of isolation

IFM 40078 (CBS 4736) South AfricaNCAIM Y-01177 TurkeyNCAIM Y-01226 TurkeyVKM Y-1513 (CBS 1906) JapanVKM Y-2623 (CBS 4381) Great BritainNCAIM Y-00469 UnknownNCAIM Y-00472 UnknownCBS 6122 GermanyCBS 4173 GermanyCBS 6219 Italy

IFM: Culture Collection of the Research Centre for Pathogenic FungiVKM: All-Russian Collection of Microorganisms, Institute of BiocheSciences, Puschino, Russia).NCAIM: National Collection of Agricultural and Industrial MicroorganiCBS: Centraalbureau voor Schimmelcultures, Utrecht, Netherlands.

originating from yeast culture collections. Fouramong them were found to produce the anticryptococcal toxin.

Materials and methods

Yeast strains, media and toxin production

F. capsuligenum strains used in this study are listedin Table 1. C. neoformans strains tested for toxinsensitivity are shown in Table 2. Other yeast speciestested for sensitivity were Cryptococcus laurentii(VKM Y-1665), Cryptococcus albidus (IFO 378),Cystofilobasidium bisporidii (VKM Y-2700), Rho-dothorula mucilaginosa (SZMC-0832), Candida krusei(IFM 46834), Candida albicans (ATCC 10231), Debar-yomyces hansenii (CBS 767), Kluyveromyces lactis(CBS 683), Pichia membranifaciens (CBS 5557),Saccharomyces cerevisiae (CBS 432). Methods fortoxin production, determination of the killer-sensi-tive relationships and characterization of the toxinwere described earlier (Keszthelyi et al., 2006).

Determination of mating types

Cells of different F. capsuligenum strains weremixed in pairs on cornmeal agar, and incubated at20 1C for 1–2 weeks. Hyphal growth was visuallydetected, and basidiospore formation was micro-scopically detected. Strain IFM 40078 was used asreference for a, VKM Y-1513 for a mating type.

Competition assay

Commercially available yeast cell wall polysac-charides, at a concentration of 5mgml�1 or

Origin Isolated by

Wine cellar van der WaltDried fig G. PeterDried fig G. PeterSake moto Dikkens and LodderCider F.W. BeechUnknown T. DeakUnknown T. DeakSoil G. KraepelinGrape juice E. FriedrichUnknown A. Fernandez

and Microbial Toxicoses, Chiba University, Chiba, Japan.mistry and Physiology of Microorganisms, Russian Academy of

sms, Budapest, Hungary.

ARTICLE IN PRESS

Table 2. Varieties, serotype, molecular subtype and original source of Cryptococcus neoformans strains tested onsensitivity to toxins produced by Filobasidium capsuligenum

Varieties Serotype Molecular subtype Original source

IFM 5844 neoformans D NIH B-3501IFM 5845 neoformans D NIH B-3502WM 148 grubii A VN I Sydney, clinicalWM 626 grubii A VN II Sydney, clinicalWM 628 Hybrid A/D VN III Melbourne, clinicalWM 629 neoformans D VN IV Melbourne, clinicalWM 179 gattii B VG I Australia, clinicalWM 178 gattii B VG II Australia, clinicalWM 161 gattii B VG III California, eucalyptus sp.WM 779 gattii C VG IV Africa, veterinaryH-23 (IFM 51702) neoformans D Hungary, pigeon droppingsH-28 (IFM 51706) grubii A Hungary, pigeon droppingsH-30 C Hungary, pigeon droppingsH-62 (IFM 51710) grubii A Hungary, clinicalH-63 (IFM 51711) grubii A Hungary, clinical353/2004 Hungary, clinical2679/2003 Hungary, clinical2710/2003 Hungary, clinical60203A Hungary, clinical60209C Hungary, clinical60467B Hungary, clinical62680A Hungary, clinical62680C Hungary, clinical

Comparison of killer toxin-producing and non-producing strains of F. capsuligenum 269

10mgml�1, were mixed with 1ml crude toxin,supplemented with glucose. Samples were shakengently at 24 1C for 24 h. The residual activity wasdetermined in bioassay experiments. The testedcell wall polysaccharides were chitin oligosacchar-ides (Sigma-Aldrich, C-7170), mannan (Sigma-Aldrich, M-3640), laminarin (Sigma-Aldrich, L-9634)and pustulan (Calbiochem, 540501). The residualtoxin activities were tested in bioassays.

Determination of sensitivity to proteinasetreatment

Crude toxin preparations were incubated over-night at 20 1C in the presence of different protei-nases: proteinase K (Sigma-Aldrich, P-2308), papain(Reanal, 25781-1-99-33), pepsin (Boehringer Mann-heim GmbH, 108057) and pronase E (Merck KGaA,1.07433) of 10mgml�1. The residual toxin activ-ities were tested in bioassays.

Analysis of nucleic acids

DNA was isolated according to Leach et al.(1986). DNA fragments were separated in 0.8%or 1.0% agarose gels using standard methods

(Sambrook et al., 1989), stained with ethidiumbromide and examined under UV light.

Analysis of ITS and D1/D2 regions

Fragments of the internal transcribed spacer(ITS) region were amplified using primers ITS1(TCCGTAGGTGAACCTGCGG) and ITS4 (TCCTCCGC-TTATTGATATGC) (White et al., 1990). Fragmentscontaining the D1/D2 region of about 600–650 bp atthe 50 end of the LrDNA were amplified usingforward primer F63 (GCATATCAATAAGCGGAG-GAAAAG) and reverse primer LR3 (GGTCCGTG-TTTCAAGACG) (Fell et al., 2000). The 50 ml reactionmixtures contained 200 mM of each dNTP, 1 mM ofeach primer, 2.5 U Taq DNA polymerase and 200 nggenomic DNA in an amplification buffer suppliedby the manufacturer (Zenon-Bio Ltd., Szeged,Hungary). The thermal cycler was programmed foran initial denaturation step (94 1C, 5min), 35 stepsconsisting of denaturation at 94 1C for 1min,annealing at 55 1C for 1min and extension at 72 1Cfor 2min, followed by a final extension step at 72 1Cfor 5min. The PCR products were separated byelectrophoresis and the DNA fragments purified.Direct sequencing was performed on an ABI 373ADNA sequencer (Applied Biosystems Inc., Foster

ARTICLE IN PRESS

Table 3. Accession numbers of sequences of Filobasi-dium capsuligenum strains

Strains GenBank accession nos.

ITS1-5.8S rDNA-ITS2 D1/D2 region

IFM 40078(CBS 4736)

AF444369 AF075501

NCAIM Y-01177 EF532829 EF532833NCAIM Y-01226 EF532830 EF532834CBS 6219 AF444334 AF181506VKM Y-1513(CBS 1906)

AF444381 AF363642

VKM Y-2623(CBS 4381)

AF444382 AF444695

NCAIM Y-00469 EF532831 EF532835NCAIM Y-00472 EF532832 EF532836CBS 4173 EF611232 EF611233CBS 6122 EF611231 EF611234

A. Keszthelyi et al.270

City, CA, USA). Accession numbers of ITS and D1/D2sequences are listed in Table 3.

RAPD analysis

The following decamers of Operon RandomPrimer Kits (Operon Technologies, Alameda, CA,USA) were used for RAPD analysis: OPO-12(CAGTGCTGTG), OPG-19 (GTCAGGGCAA), OPR-15(GGACAACGAG). PCR fingerprinting using the mi-crosatellite-specific primer M-13 (50-GAGGGTGG-CGGTTCT-30) and minisatellite-specific primer(GACA)4 were also performed. Amplifications werecarried out in 25-ml volumes containing the appro-priate buffer supplied by the manufacturer, 100 mMeach of the nucleotides, 0.2 mM primer, 50 nggenomic DNA and 0.5 U of Taq DNA polymerase(Zenon-Bio Ltd., Szeged, Hungary). The program-mable thermal controller (Model PTC-100-60, MJResearch Inc., Waltham, MA, USA) was programmedfor 45 cycles (denaturation at 92 1C for 1min, low-stringency annealing of the primer at 35 1C for1min, extension at 72 1C for 2min). The amplifica-tion products were analyzed by gel electrophoresis.

Figure 1. Killer-toxin production of Filobasidium capsu-ligenum strains. The strains tested are indicated bynumbers: 1, NCAIM Y-01177; 2, IFM 40078; 3, NCAIMY-01226; 4, CBS 6219; 5, VKM Y-2623; 6, NCAIM Y-00469;7, CBS 4173; 8, NCAIM Y-00472; 9, CBS 6122; 10, VKMY-1513. (A) Activity against Cryptococcus neoformans IFM5844 strain. (B) Cross reactivity of F. capsuligenumstrains. (C) Activity against Candida albicans and Sac-charomyces cerevisiae.

Phylogenetic analyses

The D1/D2 region of the 26S rDNA and the DNAsequences of part of the ITS region of the eight

ARTICLE IN PRESS

Comparison of killer toxin-producing and non-producing strains of F. capsuligenum 271

F. capsuligenum isolates were sequenced andcompared to sequence data from GenBank. Theaccession numbers of sequences of F. capsuligenumstrains and related species are listed in Table 3.Phylogenetic analyses were conducted by programsof the PHYLIP software package version 3.65(Felsenstein, 2004). Sequences were aligned usingCLUSTALW (Thompson et al., 1997). Indices ofsupport (bootstrap values) for internal brancheswere generated by 1000 replications of the boot-strap procedure using Phylogenetic Computer ToolsVersion 1.32 for windows program (Buntjer, 2001).Distance matrices were also generated by thisprogram with Jaccard coefficient. Neighbor-joininganalyses of the data sets were carried out with theprogram NEIGHBOR of the PHYLIP package.

Figure 2. The residual activity of the crude toxinpreparation after co-incubation with different cell wallpolymers, for 24 h. 1, Pustulan; 2, mannan; 3, control; 4,laminarin; 5, chitin. C. neoformans IFM 5844 was used assensitive lawn.

Results and discussion

Killer-sensitive relationships

First we clarified the killer-sensitive relationshipsamong 10 F. capsuligenum and different C. neofor-mans strains. These tester strains represented thefive serotypes (A, B, C, D and AD) and the eightmolecular subtypes (VN I–IV, VG I–IV) of C. neofor-mans (Meyer et al., 1999) (Table 2). Four out of the10 F. capsuligenum strains (IFM 40078, NCAIMY-01177, NCAIM Y-01226 and CBS 6219) turned outto be able to produce killer toxin with antic-ryptococcal activity (Figure 1A). The others did notproduce toxins against C. neoformans but were allsensitive to the toxins of the producing strains(Figure 1B). The specific activity of FC-1 againstthe opportunistic human pathogen C. neoformanscan raise the possibility to use the toxin asan antifungal agent. However, as it is shown(this paper), the toxin is unstable in the physio-logical milieu of the human body (37 1C, pH 7).Besides, its antigenic and cytotoxic effects are notclarified yet. However, exploiting the anticrypto-coccal activity of FC-1 without its undesired effectsvia raising antiidiotipic antibodies representing theinternal image of the active site of the toxin canstill lead us to evolve a useful antibiotic agent.Similar method has been described for the anti-candidal toxin (KT) of Pichia anomala (Polonelliet al., 2003).

Among 27 different C. neoformans strains all butone (WM 161) were sensitive to all the four toxins.These results showed that the specificity of thetoxins of the four strains was identical. Resistanceof the WM 161 strain is because of differences incell wall or membrane receptors. However, as the

exact mechanism of toxin activity is not yetdiscovered, alternations in other factors may alsoresult in loss of sensitivity.

The mycocin produced by F. capsuligenum strainsproved to be active against the tremellaceous yeastC. laurentii. The filobasidiaceous C. albidus, thecystofilobasidiaceous C. bisporidii, the sporidiobo-laceous R. mucilaginosa and ascomycetous yeastsfrom genera Candida, Debaryomyces, Kluyvero-myces, Pichia and Saccharomyces were not sensi-tive to the toxin (Figure 1C).

Comparison of biochemical characteristicsand genetic background of the toxins

To compare the biochemical characteristics ofthe toxins sensitivity for proteinase enzymes, heatstability and pH optimum of the crude toxins wereexamined. All toxins were rather resistant toproteinase treatment (proteinase K, papain, pepsinof 10mgml�1); only pronase E reduced the activityof all the four crude toxins. Optimal temperatureand pH for the activity of NCAIM Y-01177, NCAIMY-01226 and CBS 6219 strains corresponded withthat of IFM 40078 strain described earlier (Golubevand Kuznetsova, 1991; Keszthelyi et al., 2006). Alltoxins lost their activity during overnight incuba-tion above 30 1C, or boiling for 1min. The optimalpH for the activity of the toxins was in the rangeof 4–6.

ARTICLE IN PRESS

A. Keszthelyi et al.272

Pustulan (b-1,6-glucan) in the cell wall mightprovide a binding site for FC-1 toxin (Keszthelyi etal., 2006). To investigate its possible receptor role

Figure 4. Mating of Filobasidium capsuligenum (A) macromgrowth and clamp connections. (C–E) Basidia and basidiospor

Figure 3. Gel electrophoresis pattern of total nucleicacid extracts of the toxin-producing strains. NuclearDNA, small and large rRNAs (18S and 26S) are indicated byarrows. M.: l-Puc mix marker; 1, IFM 40078; 2, NCAIMY-01177; 3, NCAIM Y-01226; 4, CBS 6219.

in the case of other toxins as well crude toxinpreparations were incubated in the presence ofdifferent commercially available cell wall polymers(chitin oligosaccharides, mannan, laminarin andpustulan). The residual toxin activities were testedin bioassays. The results showed that in all casesonly pustulan could compete out the effectivenessof the toxins, while the others did not reduce thetoxin activity (Figure 2).

In order to reveal the genetic background of thetoxin production, total nucleic acid extracts wereprepared and separated by gel electrophoresis fromthe strains of killer phenotype. Neither RNA norDNA plasmids were detected by agarose gelelectrophoresis, thus the gene encoding the toxinis most likely located in the chromosomal DNA(Figure 3).

Mating type of the strains

To explore whether there is any relationshipbetween toxin production and mating types, weassayed strains for mating. Cells belonging todifferent mating types start to conjugate on cornmeal agar, and dikaryotic hyphae grow out from theconjugated pairs (Kwon-Chung, 1998). This hyphalgrowth can be visually detected within a few days(Figure 4). IFM 40078 is known to have a matingtype, while VKM Y-1513 strain belongs to type a.

orphology of the mating of a and a strains. (B) Hyphales. Bar: 10 mm.

ARTICLE IN PRESS

Comparison of killer toxin-producing and non-producing strains of F. capsuligenum 273

All toxin-producing strains were able to formhyphae with non-producing strains VKM Y-1513,VKM Y-2623 and NCAIM Y-00469, but were not ableto mate with NCAIM Y-00472 (Table 4). The fact thatexclusively a mating-type strains had also killerphenotypes allows the conclusion that the ability ofF. capsuligenum strains to produce the mycocinmay be associated with a mating type.

Analysis of ITS and D1/D2 regions

As the toxins are chromosomally coded, wewanted to clarify if there are other differences inthe genomes of the toxin-producing and non-

Figure 5. Alignments of D1/D2 and ITS sequences of F. caDifferences in the sequences are highlighted with red rectansequences of toxin-producing and non-producing strains are13 bp long deletion of strain CBS 6219. Green rectangles indi

Table 4. Distribution of mating types and killerphenotypes of Filobasidium capsuligenum strains

Killer phenotype Strains Mating type

k+ IFM 40078 aNCAIM Y-01177 aNCAIM Y-01226 aCBS 6219 a

k� VKM Y-1513 aVKM Y-2623 aNCAIM Y-00469 aNCAIM Y-00472 aCBS 4173 aCBS 6122 a/a

producing strains. Sequences of the D1/D2regions of the 26S rRNA genes and ITS regionsobtained by the standard ITS1-ITS4 primers wereexamined.

Altogether 498 nucleotides from the D1/D2sequences and 490 nucleotides from the ITS regionswere involved in the analysis (Figures 5A, B).Nucleotide sequences of D1/D2 regions of produ-cing strains were identical and these of non-producing strains too. However, the two groupsdiffered from each other significantly. In all, 8 outof 498 nucleotides have changed, which means1.61% divergence. Conspecific strains ordinarilyshow 0–1% substitutions in the D1/D2 region ofrRNA (Peterson and Kurtzman, 1991).

In the case of ITS regions, CBS 6219 strain of killerphenotype differed from the other producingstrains in 13 nucleotides, but this refers to acontiguous deletion rather than substitutions. Be-sides NCAIM Y-01177 (k+), VKM Y-1513 (k�) andNCAIM Y-01226 (k+) differed in one nucleotide fromother strains of their group, the latter two havingthe same substitution on the same position, thoughnot belonging to the same group. VKM Y-2623 (k�)differs from other non-killers in two nucleotides.One of them is the same substitution as in VKMY-1513 (k�) and NCAIM Y-01226 (k+), and the otheris as in NCAIM Y-01177 (k+). Still, except for the 13nucleotides-long deletion in the sequence of CBS6219 (k+), no members of each group differed from

psuligenum strains. (A) Alignment of D1/D2 sequences.gles. (B) Alignment of ITS sequences. Differences in thehighlighted with red rectangles. Blue rectangle is for thecate individual substitutions.

ARTICLE IN PRESS

A. Keszthelyi et al.274

the others in more than two nucleotides (max. 0.4%divergence). Comparison of the ITS region oftoxin-producing and non-producing groupsshowed significant differences between the twogroups. Excluding the deletion in CBS 6219,

Figure 6. RAPD analysis of F. capsuligenum strains. (A) RAPD pOP primers M13 and (GACA)4 primers. Lanes: M1:1000 bp(IFM 40078; NCAIM Y-01177; NCAIM Y-01226; CBS 6219); 5�Y-00469; NCAIM Y-00472; CBS 4713; CBS 6122); 11: Cryptococcby the neighbor-joining analysis based on RAPD patterns of thvalues, only values 450% are shown.

members of the two groups differ at 9–11 positions(1.84–2.24%).

This relatively high level of separation of the twogroups might even raise the possibility of theproposal of two different varieties.

atterns of total DNA from the strains using three differentDNA ladder, M2: 100 bp DNA ladder 1–4. Killer strains10. Non-killer strains (VKM Y-1513; VKM Y-2623; NCAIMus neoformans IFM 5844. (B) Phylogenetic tree generatede strains. The numbers above the branches are bootstrap

ARTICLE IN PRESS

Comparison of killer toxin-producing and non-producing strains of F. capsuligenum 275

RAPD analysis

To further clarify the relationships, RAPD analysisof the total DNAs was also performed. M13, (GACA)4and OP primers (OPO12, OPG19, OPR15) were used.Patterns generated by amplification with all pri-mers clearly showed that toxin-producing and non-producing strains are entirely different. The killerstrains produced many common bands, whileamong the non-producing strains the patterns wereless similar (Figure 6A). Based on these patterns, aphylogenetic tree was generated by the neighbor-joining method using C. neoformans IFM 5844 asoutgroup (Figure 6B). The tree suggests that thetwo groups were distinct, which corresponds withthe results from the ITS and D1/D2 sequenceanalysis. The close relationship of the producingand non-producing strains is supported by highbootstrap values, while the relationships amongstrains belonging to one group are not confirmeddue to the low values.

RFLP patterns of the mitochondrial DNA

To see if the two groups differ at the extrachromosomal level as well mtDNA-RFLP was per-formed using HhaI and HaeIII enzymes. The resultedpatterns reinforced the results of RAPD analysissince the producing strains also showed isomorphicpatterns clearly distinct from non-producingstrains, and members of the latter group sharedless similarity (Figure 7). The phylogenetic tree

Figure 7. mtDNA RFLP patterns of the strains generatedby using restriction enzymes HaeIII and HhaI. Lanes: M:1000 bp DNA ladder, 1–4. Killer strains (IFM 40078 NCAIMY-01177; NCAIM Y-01226; CBS 6219); 5–10. Non-killerstrains (VKM Y-1513; VKM Y-2623; NCAIM Y-00469; NCAIMY-00472; CBS 4713; CBS 6122; 11. Cryptococcus neofor-mans, IFM 5844.

based on RFLP shows similar topology than the onebased on RAPD data, thus it also confirms thedifferentiation of the two groups (data not shown).

Conclusion

A comparison among 10 different F. capsuligenumstrains available from yeast culture collectionsrevealed that the strains differed in the ability toproduce killer toxin against C. neoformans: 40%were killer positive, 60% were not able to producethe toxin. Thus the killer phenotype could beconsidered relatively common in this taxon. Theyshared common specificity on other yeast species.The toxins were biochemically similar, probablyidentical. b-1,6-glucan may serve as receptor onthe wall of the sensitive cells. The genes coding forthe toxin were located on the chromosomal DNA ineach strains. As all the toxin-producing strains hadmating type a, the toxin-coding ability could beconnected to mating-type genes. The strains belong-ing to the producing groups were identical in theirnucleotide sequences in the D1/D2 region, andslightly differed (0.4%) in their ITS regions. Thesewere also true for the members of the non-producinggroup. But high rate of divergence (1.6–2.24%) wasobserved between the two groups in both rDNAregions. Phylogenetic trees generated on the basis ofRAPD and mtRFLP patterns pointed out that in spiteof the observed homogeneity within the member ofthe groups, the non-producing strains were a littlemore heterogeneous. The high rate of differencebetween the producing and non-producing strainsindicates existence of two varieties, althoughfurther investigations (DNA–DNA reassociation ki-netics or studying the fertility of the progeny ofcrossings) would be necessary for their descriptions.

Acknowledgments

The authors thank Prof. W. I. Golubev for helpfuldiscussion and comments on the manuscript. Thework was financially supported by the HungarianScientific Research Fund (OTKA) T035194.

References

Buntjer JB. Phylogenetic computer tools v. 1.3. TheNetherlands: Wageningen University; 2001.

De Mot R, Verachtert H. Purification and characterizationof extracellular amylolytic enzymes from the yeastFilobasidium capsuligenum. Appl Environ Microbiol1985;50:1474–82.

ARTICLE IN PRESS

A. Keszthelyi et al.276

De Mot R, Verachtert H. Regulation of the amylasesecretion by the yeast Filobasidium capsuligenum anda 2-deoxy-D-glucose resistant mutant. Appl MicrobiolBiotechnol 1987;26:258–62.

Fell JW, Boekhout T, Fonseca A, Scorzetti G, Statzell-Tallman A. Biodiversity and systematics of basidiomy-cetous yeasts as determined by large-subunit rDNAD1/D2 domain sequence analysis. Int J Syst EvolMicrobiol 2000;50(Pt 3):1351–71.

Fell JW, Statzell AC, Hunter IL, Phaff HJ. Leucosporidiumgen. n., the heterobasidiomycetous stage of severalyeasts of the genus Candida. Antonie Van Leeuwen-hoek 1969;35:433–62.

Felsenstein J. PHYLIP version 3.65. Distributed by theauthor. Seattle, WA 98195-7730, USA: Department ofGenetics, University of Washington; 2004.

Golubev WI, Kuznetsova L. The taxonomic specificity ofthe action spectrum of Filobasidium capsuligenummycocin. Microbiology (in Russian) 1991;60:530–6.

Keszthelyi A, Ohkusu M, Takeo K, Pfeiffer I, Litter J,Kucsera J. Characterisation of the anticryptococcaleffect of the FC-1 toxin produced by Filobasidiumcapsuligenum. Mycoses 2006;49:176–83.

Kwon-Chung KJ. Filobasidium Olive. In: Cletus PK, JackWF, editors. The yeasts. fourth ed. Amsterdam:Elsevier; 1998. p. 663–9.

Leach J, Finkelstein DB, Rambosek JA. Rapid miniprep ofDNA from filamentous fungi. Fungal Gen Newslett1986;33:32–3.

Markower M, Bevan EA. The physiological basis of thekiller character in yeast. In: Proceedings of the XIthInt. Cong. Genet., 1963. p. 202–203.

Meyer W, Marszewska K, Amirmostofian M, Igreja RP,Hardtke C, Methling K, et al. Molecular typing ofglobal isolates of Cryptococcus neoformans var.

neoformans by polymerase chain reaction fingerprint-ing and randomly amplified polymorphic DNA-a pilotstudy to standardize techniques on which to base adetailed epidemiological survey. Electrophoresis 1999;20:1790–9.

Peterson SW, Kurtzman CP. Ribosomal-RNA sequencedivergence among sibling species of yeasts. Syst ApplMicrobiol 1991;14:124–9.

Polonelli L, Magliani W, Conti S, Bracci L, Lozzi L.Therapeutic activity of an engineered synthetic killerantiidiotypic antibody fragment against experimentalmucosal and systemic candidiasis. Infect Immun2003;71:6205–12.

Rodrigues de Miranda L. Filobasidum capsuligenum nov.comb. Antonie Van Leeuwenhoek 1972;38:91–9.

Sambrook J, Fritsch EF, Maniatis T. Molecular cloning.Cold Spring Harbor, New York: Cold Spring HarborLaboratory Press; 1989.

Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F,Higgins DG. The CLUSTAL_X windows interface: flex-ible strategies for multiple sequence alignment aidedby quality analysis tools. Nucleic Acids Res 1997;25:4876–82.

van der Walt J, van Kerken A. Three new anascospor-ogenous yeasts. Antonie Van Leeuwenhoek 1961;27:206–12.

White TJ, Burns T, Lee S, Taylor JW. Amplification anddirect sequencing of fungal ribosomal RNA genes forphylogenetics. In: Gelfand MADH, Sninsky JJ, WhiteTJ, editors. PCR protocols: a guide to methods andapplications. New York: Academic Press, Inc.; 1990.p. 315–22.

Yamamoto H. (Hyogo, JP), Protease, United Statespatent, 50 53333, 1991.