Phylloplane yeasts from Portugal: Seven novel anamorphic species in the Tremellales lineage of the Hymenomycetes (Basidiomycota) producing orange-coloured colonies

Authors

  • João Inácio,

    1. Centro de Recursos Microbiológicos (CREM), Biotechnology Unit, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
    Search for more papers by this author
  • Laura Portugal,

    1. Centro de Recursos Microbiológicos (CREM), Biotechnology Unit, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
    Search for more papers by this author
  • Isabel Spencer-Martins,

    1. Centro de Recursos Microbiológicos (CREM), Biotechnology Unit, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
    Search for more papers by this author
  • Álvaro Fonseca

    Corresponding author
    1. Centro de Recursos Microbiológicos (CREM), Biotechnology Unit, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
    Search for more papers by this author

*Corresponding author. Present address: SABT, FCT/UNL, Quinta da Torre, 2829-516 Caparica, Portugal. Tel.: +351 212948500; fax: +351 212948530., E-mail address: amrf@fct.unl.pt

Abstract

A survey of epiphytic yeasts on leaves of selected Mediterranean plant species collected at the ‘Arrábida Natural Park’ (Portugal) yielded about 850 isolates, mostly of basidiomycetous affinity. Amongst the basidiomycetes, 35 strains showed the following characteristics: production of orange-coloured colonies, ability to produce starch-like compounds, assimilation of d-glucuronic acid and/or inositol, inability to utilize nitrate, and formation of ballistoconidia by many of the isolates. This group of yeasts was assigned to the Tremellales lineage of the Hymenomycetes and was further characterised using a combination of conventional phenotypic identification tests with molecular methods, namely PCR fingerprinting and rDNA sequencing. Eight additional strains presumptively identified as Bullera armeniaca, B. crocea or Cryptococcus hungaricus were also studied. Twenty-eight strains could be assigned to or were phylogenetically related to recognised species of Dioszegia in the ‘Luteolus clade’, but the 15 remaining strains belonged to other clades within the Tremellales. Ten phylloplane isolates were identified as Dioszegia hungarica, one as D. aurantiaca, another as D. crocea and three others were ascribed to the recently described species D. zsoltii. Seven novel species, viz. Cryptococcus amylolyticus, C. armeniacus, C. cistialbidi, Dioszegia buhagiarii, D. catarinonii, D. fristingensis and D. takashimae, are proposed for the remaining strains that did not correspond to any of the hitherto recognised species.

1Introduction

The surface of leaves, normally referred to as the phylloplane, is known to harbour a variety of yeast species but has not been thoroughly investigated as a natural habitat for these microorganisms [1,2]. Yeasts have been reported as major components of the phylloplane microbiota on plants from temperate regions (e.g. [3–5]), as well as on tropical plants (e.g. [6,7]). However, little is known about the epiphytic mycobiota of Mediterranean plants [8,9]. There appears to be a marked dominance of basidiomycetous yeasts on the phylloplane, namely of species belonging to the genera Sporobolomyces, Rhodotorula (collectively referred to as the ‘pink yeasts’) and Cryptococcus (‘white yeasts’) [2,3,10]. Moreover, members of the genera Bullera, Sporobolomyces and Tilletiopsis are commonly isolated from leaves and are thought to be especially adapted to this environment due to the production of forcibly ejected ballistoconidia (e.g. [11]).

Here, we report on the phenotypic and molecular characterisation of Portuguese isolates of hymenomycetous phylloplane yeasts that produced orange-coloured colonies. Additional isolates from different sources that had been previously assigned to the species Bullera armeniaca, B. crocea and Cryptococcus hungaricus (sensu Barnett et al. [12]) were also studied. These species are known for their conspicuous orange-coloured colonies, an uncommon characteristic among the hymenomycetous yeasts, and were recently transferred to the genus Dioszegia reinstated by Takashima et al. [13]. Seven novel species in the Tremellales lineage of the Hymenomycetes are proposed to accommodate the isolates that could not be assigned to any of the currently recognised taxa [12,14,15].

2Methods

2.1Cultures

A list of the strains used in this study is presented in Table 1. Reference strains were obtained from the ‘Centraalbureau voor Schimmelcultures’, Utrecht, The Netherlands (CBS). Phylloplane isolates originated mainly from a survey of mycobiota on selected plant species of the ‘Arrábida Natural Park’, a Mediterranean-type ecosystem in the Setúbal region, south of Lisbon, Portugal [16] (Table 1). Leaves from five plant species, namely the deciduous trees Acer monspessulanum and Quercus faginea, and the evergreen shrubs Cistus albidus, Pistacia lentiscus and Osyris quadripartita, were collected at two distinct locations (northern and southern slopes of “Serra da Arrábida”) and on six consecutive dates (three in early spring, one in summer and two in autumn) spanning a two-year period [9,17]. Yeasts were isolated by two methods: the conventional method (CM), in which leaf washings are plated onto solid medium (MYP agar: malt extract 0.7%, yeast extract 0.05%, Soytone 0.25%, agar 1.5%, supplemented with chloramphenicol 0.05% and Rose Bengal 0.004%), and the spore-fall method (SFM) using MYP agar without supplements. The latter is used for the isolation of yeasts that produce forcibly discharged spores such as ballistoconidia. Additional details of this study can be found in reference [9]. Strains were maintained on MYP agar slants at 4 °C.

Table 1.  List of cultures used in this study
SpeciesIsolateaStrainbOrigincGenBank Accession No.d
    D1/D2ITS
  1. aOriginal designations of the isolates.

  2. bStrains deposited in or obtained from culture collections: PYCC, Portuguese Yeast Culture Collection (Portugal); CBS, Centraalbureau voor Schimmelcultures (The Netherlands); T, type strain.

  3. cIsolation source and geographic location; N, northern slope of ‘Serra da Arrábida’; S, southern slope of ‘Serra da Arrábida’; 1, Spring; 2, Summer; 3, Autumn; * isolate obtained using SFM.

  4. dAccession numbers of sequences determined in the present study; (GB), sequences determined by other authors and available from GenBank (see Figs. 3 and 4).

Cryptococcus amylolyticus sp. nov.3CVF16PYCC 5850TLeaves of Cistus albidus (Arrábida, Portugal; N1)AY562134
  (CBS 10048)   
C. armeniacus sp. nov.Cont.NALPYCC 5855TAir, culture contaminant (Caparica, Portugal)AY562140
  (CBS 10050)   
C. cistialbidi sp. nov.1CSF5PYCC 5851TLeaves of Ci. albidus (Arrábida, S1)AY562135
  (CBS 10049)   
 2CSFm8PYCC 5852Leaves of Ci. albidus (Arrábida, S3)AY562136
 2CSF7Leaves of Ci. albidus (Arrábida, S3)
 2CVFm15Leaves of Ci. albidus (Arrábida, N3)
 3CSF1Leaves of Ci. albidus (Arrábida, S1)
 5CSF3Leaves of Ci. albidus (Arrábida, S3)
 5CSFe14PYCC 5853Leaves of Ci. albidus (Arrábida, S3)AY562137
 5CVFe1PYCC 5854Leaves of Ci. albidus (Arrábida, N3)AY562138
 A2CSV7Leaf of Ci. albidus (Arrábida, S3)*
 A2CVV5Leaf of Ci. albidus (Arrábida, N3)*
 A3CSVI3bLeaf of Ci. albidus (Arrábida, S1)*
 PYCC 4561Flower of Acacia sp. (Estoril, Portugal)
 PYCC 4702Stagnant pool water (Sintra, Portugal)AY562139
Dioszegia aurantiacaCBS 6980TStem of Urtica sp. (Canada); type strain of Bullera aurantiaca(GB)(GB)
  (PYCC 5455)   
 ZP 362PYCC 5856Leaf of Andromeda polifolia infected with Exobasidium karstenii (Tübingen, Germany)*AY562141AY562153
D. buhagiarii sp. nov.4AVF6PYCC 5866TLeaves of Acer monspessulanum (Arrábida, N2)AY562151AY885687
  (CBS 10054)   
D. catarinonii sp. nov.A2AVV2PYCC 5857TLeaf of A. monspessulanum (Arrábida, N3)*AY562142AY562154
  (CBS 10051)   
 A2AVS8Leaf of A. monspessulanum (Arrábida, N3)*AY885688
 2AVF10Leaves of A. monspessulanum (Arrábida, N3)
 5AVFs3PYCC 5858Leaves of A. monspessulanum (Arrábida, N3)AY562143AY562155
 5CVFe14PYCC 5859Leaves of Ci. albidus (Arrábida, N3)AY562144AY562156
 5CSFe15Leaves of Ci. albidus (Arrábida, S3)
 6PVF13Leaves of Pistacia lentiscus (Arrábida, N1)
D. crocea CBS 6714TFruit of Fragaria abanassa (UK); type strain of Bullera crocea(GB)(GB)
  (PYCC 5433)   
 ZP 373PYCC 5860Leaf of fern (Tübingen, Germany)*AY562145AY562157
D. fristingensis sp. nov.ZP 359PYCC 5861TLeaf of Arum maculatum infected with Melanotaenium ari (Fristingen, Germany)*AY562146AY562158
  (CBS 10052)   
D. hungarica CBS 4214TSoil (Külsó-tó, Hungary); type strain of Cryptococcus hungaricus(GB)(GB)
  (PYCC 3954)   
  CBS 7091Brassica oleacea var. capitata (UK); type strain of Bullera armeniaca(GB)(GB)
  (PYCC 5432)   
 2AVF18PYCC 5862Leaves of A. monspessulanum (Arrábida, N3)AY562147
 5QSF24Leaves of Quercus faginea (Arrábida, S3)
 6CVFm9Leaves of Ci. albidus (Arrábida, N1)
 A1CSV2Leaf of Ci. albidus (Arrábida, S1)*
 A1PSV1Leaf of P. lentiscus (Arrábida, S1)*
 A2AVV1Leaf of A. monspessulanum (Arrábida., N3)*
 A2AVS7Leaf of A. monspessulanum (Arrábida, N3)*
 A2QVV1Leaf of Q. faginea (Arrábida, N3)*
 A2QVS1Leaf of Q. faginea (Arrábida, N3)*
 3F1Senescent leaf of Betula sp. (Caramulo, Portugal)*
 ZP 49Condensation water in refrigerator (Oeiras, Portugal)AY562148
D. takashimae sp. nov.A2QSS4PYCC 5864TLeaf of Q. faginea (Arrábida, S3)*AY562149AY562160
  (CBS 10053)   
 A2QSS2PYCC 5863Leaf of Q. faginea (Arrábida, S3)*AY562159
 A2AVV3PYCC 5865Leaf of A. monspessulanum (Arrábida, N3)*AY562150AY562161
D. cf. zsoltii2PSF1PYCC 5867Leaves of P. lentiscus (Arrábida, S3)AY562152AY562162
 A2CSV3PYCC 5868Leaf of Ci. albidus (Arrábida, S3)*AY562163
 5CSFE13PYCC 5869Leaves of Ci. albidus (Arrábida, S3)AY562164

2.2Conventional characterisation

Micromorphology was investigated by phase contrast light microscopy. Production of ballistoconidia was checked on MYP agar, corn meal Agar (CMA) and Yeast Carbon Base (YCB) agar. The strains were subjected to a preliminary physiological characterisation on plates with solidified Wickerham media (selected carbon or nitrogen sources with Yeast Nitrogen Base or YCB, respectively), which included selected tests that could discriminate the orange-coloured species Bullera armeniaca, B. crocea and Cryptococcus hungaricus (see [12] and chapters on Bullera and Cryptococcus in [14]). Starch hydrolysis was tested on solid YNB medium with soluble starch as sole carbon source; after one week the plates were sprayed with Lugol's iodine solution to reveal possible halos of hydrolysed starch around yeast growth. Other tests followed the methods described by Yarrow [18]. Results from a total of 25 phenotypic tests (see Fig. 1) were integrated by means of numerical taxonomy. In this analysis the Simple Matching coefficient and the UPGMA clustering method were used and the resulting phenogram was computed using the NTSYS software package (version 2.0, Applied Biostatistics, Setauket, NY, USA). Complete physiological characterisations [18] were performed for selected strains of the newly proposed species. In addition to assessing the utilisation of standard carbon compounds, six aldaric acids [19] and 11 aromatic compounds [20] were also tested, the latter at a concentration of 0.1% (w/v) since some of those compounds can be toxic for yeasts.

Figure 1.

Phenotypic characterisation of phylloplane isolates and Dioszegia spp. reference strains (Table 1): phenogram resulting from the analysis of 25 test responses (Simple Matching coefficient, UPGMA clustering method). Responses to potentially discriminating tests are listed on the Table. The following tests yielded uniform results: assimilation of glucose, DBB and urease tests – positive; assimilation of protocatechuic acid and nitrate – negative. In a few strains production of ballistoconidia was only detected in the first isolation plates but appears to have been lost upon subculturing (see text) – denoted by (+).

2.3Molecular methods

Genomic DNA was isolated from 1-wk-old cultures on MYP agar plates by a simplified method using glass beads for cell disruption following the protocol used by Sampaio et al. [21]. PCR-amplification of polymorphic regions of genomic DNA using minisatellite primer M13 (GAGGGTGGCGGTTCT, [22]; MSP-PCR fingerprinting) followed the protocol described in [21] with 0.25 mM of each of the four dNTPs and used a Uno II Thermal Cycler (Biometra, Göttingen, Germany). Gel electrophoresis images were acquired with the Kodak Digital Science 1D Image Analysis Software (Rochester, NY, USA). Amplified DNA fragment profiles were analysed with GelCompar (version 4.1, Applied Maths, Sint-Martens-Latem, Belgium) using the Pearson's correlation coefficient and dendrograms were computed using the UPGMA clustering method. PCR amplification prior to sequencing employed forward primer ITS5 (5′ GGA AGT AAA AGT CGT AAC AAG G) and reverse primer LR6 (5′ CGC CAG TTC TGC TTA CC) and the resulting amplicon was purified with the GFX Band Purification Kit (Amersham Biosciences, Piscataway, NJ, USA). Cycle sequencing employed standard protocols with the following primers: forward primer NL1 (5′ GCA TAT CAA TAA GCG GAG GAA AAG) and reverse primer NL4 (5′ GGT CCG TGT TTC AAG ACG G) for the D1/D2 variable domains of the 26S rRNA gene (D1/D2 region); forward primer ITS1 (5′ TCC GTA GGT GAA CCT GCG G) and reverse primer ITS4 (5′ TCC TCC GCT TAT TGA TAT GC) for the ITS1 and ITS2 spacers, including the 5.8S rRNA gene (ITS region). Primers used for amplification and sequencing of rDNA were developed by other authors (see, e.g. http://www.biology.duke.edu/fungi/mycolab/primers.htm). Sequences were obtained with an ALFexpress II DNA Analyser (Amersham Pharmacia Biotech, Uppsala, Sweden), aligned with MegAlign (DNASTAR Inc., Madison, WI, USA) and visually corrected. Phylogenetic trees were computed with PAUP version 4.0b8 (Sinauer Associates Inc., Sunderland, MA, USA) using the neighbour-joining method and the Kimura two-parameter model for calculating distances, or the maximum parsimony analysis (full heuristic search with the following options: random stepwise addition with 10 replications, branch swapping using tree bisection–reconnection and maximum of 100 trees). Gaps were treated as missing data. Nucleotide sequences were deposited in GenBank under the accession numbers listed in Table 1. Additional sequences were retrieved from GenBank (accession numbers are indicated on the phylogenetic trees), the majority of which were determined by Fell et al. [23] and Scorzetti et al. [15]. Sequence alignments used to produce the phylogenetic trees are available upon request to the authors.

3Results and discussion

3.1Conventional characterisation

The majority of the strains investigated in the present study were isolated from leaves of Mediterranean plants in the Arrábida Natural Park [9,17]: eighteen from Cistus albidus, nine from Acer monspessulanum, five from Quercus faginea and three from Pistacia lentiscus (Table 1). Positive results for the assimilation of inositol and/or d-glucuronate and for the production of extracellular starch-like compounds (besides positive DBB and urease tests) suggested a hymenomycetous affinity [23] for those isolates, all of which produced orange colonies. Moreover, their failure to utilize nitrate as N-source pointed to an affiliation with the Tremellales but not with the Filobasidiales or Cystofilobasidiales (e.g. [24]), the latter containing the only other hymenomycetous taxa with deeply pigmented colonies. The dendrogram presented in Fig. 1, which resulted from the preliminary phenotypic characterisation (see 2.1), shows the full set of strains split into two major groups. One group consisted of exclusively non-ballistoconidiogenic strains that did not assimilate any of the aldaric acids tested. Members of this group were provisionally designated as Cryptococcus spp. and formed two sub-groups (named A and B). The other group comprised mostly ballistoconidiogenic strains that assimilated at least one of the aldaric acids tested and which were thought to be related to either Bullera armeniaca or B. crocea. Members of this second group were provisionally designated as Dioszegia spp. and formed five sub-groups (named C–G). Eleven strains in sub-group C were phenotypically similar to the type strains of B. armeniaca (CBS 7091) and C. hungaricus (CBS 4214) (Fig. 1). These two species have been recently found to be conspecific and were re-classified in the species Dioszegia hungarica[13]. All the other isolates shown in Fig. 1 could not be positively assigned to any of the orange-coloured tremellaceous species included in the most recent treatises on yeast taxonomy [12,14], except strains PYCC 5856 and 5860, which were similar to the type strain of B. crocea (transferred to the genus Dioszegia as D. crocea[13]). It is worth mentioning that some of the strains isolated by the SFM (Table 1) were apparently non-ballistoconidiogenic (Fig. 1) and we have repeatedly made this observation in our laboratory (J. Inácio and Á. Fonseca, unpublished data). One possible explanation for the SFM isolates in sub-group B is that yeast cells reached the agar surface attached to small loose particles that fell from leaf surfaces by gravity. An alternative hypothesis for the observed behaviour of strains in sub-groups C–G is that the capacity to produce ballistoconidia appears to be an unstable feature in members of the genus Dioszegia[13,25,26] (see below).

3.2Molecular characterisation

In a further attempt to ascertain the taxonomic affiliation of the phylloplane isolates we employed MSP-PCR fingerprinting. Previous studies (e.g. [21,22]) have shown that this technique provides a fast, reliable and reproducible measure of genetic relatedness among closely related taxa and is especially useful as a first approach to species delineation in studies that involve large numbers of strains. Usually, conspecific strains display DNA banding patterns (fingerprints) with high overall similarity and give rise to well-defined clusters in dendrograms based on numerical analysis (e.g. [21]). Further decisions on species boundaries are normally based on rDNA sequencing (D1/D2 or ITS regions) of representative strains from each cluster of the MSP-PCR dendrogram (e.g. [21]). Analysis of the DNA banding patterns resulting from MSP-PCR with primer M13 yielded the dendrogram presented in Fig. 2. Some of the strains formed clusters that agreed with those apparent in Fig. 1 and were given the same designation. Specifically, the strains in clusters B, C and G of Fig. 2 coincided with sub-groups B, C and G of Fig. 1, each of which most likely represents a distinct species. On the other hand, all strains in sub-groups A, D and E of Fig. 1 had different PCR-fingerprints and no conclusion on species delimitation was reached. The fingerprints of the type strains of B. armeniaca and of C. hungaricus were again similar to those of the isolates in cluster C, thus providing additional evidence for their inclusion in the species D. hungarica. However, the remaining strains in Fig. 2 could not be assigned to any of the named species for which reference strains were included in the MSP-PCR analysis. Moreover, strains of sub-group F of Fig. 1 formed three different clusters according to their PCR-fingerprints (named F1, F2 and F3; Fig. 2). Clarification of their taxonomic status was attempted by selecting representative strains of the different clusters apparent in Fig. 2 for sequencing of the D1/D2 and ITS regions of the rDNA. These genomic regions have been used successfully for the identification of basidiomycetous yeast species (e.g. [15]). Analysis of the D1/D2 sequence data by neighbour-joining yielded the phylogenetic tree depicted in Fig. 3. Additional sequences of phylogenetically related taxa and of representatives of the major clades in the Tremellales [15] were retrieved from GenBank and included in the analysis. Tree topology from maximum parsimony analysis of the same dataset did not differ significantly (data not shown). The strains of sub-groups C–G in Fig. 1 all fell in a strongly supported subclade within the larger ‘Luteolus clade’ of the Tremelalles lineage [15] containing all recognized Dioszegia spp., whereas the strains of sub-groups A and B appeared in separate and distantly related clades of the Tremelalles (Fig. 3). A second phylogenetic tree based on neighbour-joining analysis of ITS sequences is depicted in Fig. 4 and was intended to improve the resolution of some of the species belonging to the ‘Dioszegia subclade’ of Fig. 3. Decisions on species delimitations were based mainly on congruence between clustering of strains in the MSP-PCR analysis (Fig. 2), their phylogenetic position on the D1/D2 and ITS trees (Figs. 3 and 4) and their morphological and physiological characteristics (Fig. 1). In cases of conflict between the different analyses, high levels of sequence identity (>99%) in D1/D2 and ITS were given priority in deciding for conspecificity.

Figure 2.

MSP-PCR fingerprinting of phylloplane isolates and Dioszegia spp. reference strains (Table 1): DNA banding patterns obtained with primer M13 and resulting dendrogram using Pearson's coefficient and the UPGMA clustering method (co-phenetic correlation coefficient, r= 0.74). Clusters that coincided with the sub-groups of Fig. 1 were given the same labels (except clusters F1–F3; see text).

Figure 3.

Phylogenetic tree of representative phylloplane isolates and of selected species in the Tremellales lineage (Hymenomycetes, Basidiomycota) obtained by neighbour-joining analysis of 26S rRNA gene (D1/D2 domains) sequences using PAUP 4.0b8. The numbers given on the branches are the frequencies (>50%) with which a given branch appeared in 1000 bootstrap replications. Filobasidium spp. were used as outgroup. Sequences determined by the authors of the present study are typed in boldface. Additional sequences were retrieved from GenBank (species names followed by the corresponding accession number).

Figure 4.

Phylogenetic tree of representative phylloplane isolates and of selected species in the ‘Dioszegia subclade’ of the Tremellales lineage obtained by neighbour-joining analysis of ITS region (ITS1 + 5.8S rRNA gene + ITS2) sequences using PAUP 4.0b8. Bullera oryzae, B. sinensis and Cryptococcus luteolus were used as outgroup. Other details as for Fig. 3.

3.3Discussion of species assigned to Cryptococcus

Five strains from sub-group B of Fig. 1, which belonged to an equally homogenous cluster in Fig. 2, had identical D1/D2 sequences that differed from those of any known species (Fig. 3). All these results taken together suggest that strains in cluster B (Fig. 2) represent an undescribed species, for which the name Cryptococcus cistialbidi sp. nov. is proposed. One remarkable aspect concerning this putative novel species is that all eleven phylloplane isolates assignable to C. cistialbidi originated from the same plant species, viz. the evergreen shrub Cistus albidus (Table 1). In fact C. cistialbidi was found consistently only on the leaves of this shrub, sampled on both slopes of ‘Serra da Arrábida’ and throughout the year (Table 1), and reached densities as high as 5–6 × 102 CFU cm−2[17]. In contrast, other equally abundant species found in the same study (viz. Cryptococcus laurentii, Erythrobasidium hasegawianum, Lalaria inositophila, Rhodotorula bacarum, Rh. slooffiae, Sporobolomyces roseus) were always isolated from at least four of the plant species sampled, at both sampling sites and on different dates [9,17]. The possible association with a particular plant, which appears to be rare among epiphytic yeasts (e.g. [9]), was confirmed by direct detection using fluorescent in situ hybridisation (FISH, data not shown; [17]). This issue deserves further study since two additional strains (PYCC 4561 and 4702) with apparently unrelated origins (Table 1) appear to belong to the same species (Figs. 1 and 2).

The two strains in sub-group A of Fig. 1 had unique PCR-fingerprints (Fig. 2) as well as unique D1/D2 sequences (Fig. 3) suggesting a separate taxonomic position for each strain, for which the names Cryptococcus amylolyticus sp. nov. (PYCC 5850) and C. armeniacus sp. nov. (PYCC 5855) are proposed. According to the D1/D2 data (Fig. 3) the three novel Cryptococcus spp. that produce orange colonies are not closely related to Dioszegia spp.: C. cistialbidi appears to be loosely connected to the ‘Aurantia clade’, whereas C. amylolyticus and C. armeniacus seem to be sister species (they differed at eighteen nucleotide positions) but are not clearly linked to any particular clade of the Tremellales (Fig. 3). In spite of the recognized polyphyletic nature of the genus Cryptococcus (e.g. [23,24]) we chose to include the novel species under this genus since they fit well within its current circumscription. A discussion of genus concepts is outside the scope of this paper. However, in our view, the lack of support of many clades within the Tremellales lineage (e.g. [15]) and the unresolved issue of the relationship of many anamorphic species to teleomorphic Tremellales, for which molecular data are not (yet) available, makes the proposal of novel genera to accommodate the newly described anamorphic species premature.

3.4Discussion of species assigned to Dioszegia

As mentioned above, isolates in cluster C (Fig. 2) probably belong to D. hungarica. Nucleotide sequences of the D1/D2 region from strains PYCC 5862 and ZP 49 offered further support for this hypothesis (Fig. 3). The limited information on the ecological distribution of this species derives from the origins of the type strains of B. armeniaca (cabbage) and C. hungaricus (soil), since additional strains assigned to the latter species may not belong to D. hungarica (see below). The isolates obtained in this study suggest that D. hungarica may be more common as a leaf epiphyte than previously thought since it apparently occurred on four of the plants sampled at both locations in ‘Serra da Arrábida’; an additional isolate originated from a birch leaf in northern Portugal (Table 1). Incidentally, all the isolates examined produced ballistoconidia (Fig. 1), albeit in variable amounts, which means that the type strain of D. hungarica (CBS 4214) remains the sole strain assigned to this species that is unable to produce this type of cell (e.g. [13]). Other strains originally identified as C. hungaricus (CBS 5124, 6324, 6576, 6953) were studied by Gáczser et al. [27,28] who demonstrated that they do not belong to D. hungarica but may represent distinct and as yet unnamed species according to their D1/D2 sequences (Fig. 3). However, the quality of the sequence data produced by Gáczser et al. for those strains is doubtful since some of them included ambiguous bases (N) and alignment with the remaining taxa in Fig. 3 generated many insertions/deletions (data not shown). Moreover, the D1/D2 sequence of CBS 6324 (GenBank Accession No. AF314233) was so divergent that we decided not to include it in the phylogenetic analysis (Fig. 3). This apparent divergence is in complete disagreement with all other data presented by Gáczser et al. [27,28] for that strain, including the corresponding ITS sequence (Fig. 4), which point to its conspecificity with strain CBS 6576 (both strains were isolated from seawater). A final decision about the taxonomic position of those strains will probably require them to be re-sequenced. Another D1/D2 sequence deposited in GenBank (AJ511333) and corresponding to an unidentified soil isolate (HB 1047, labeled as ‘basidiomycete yeast sp.’) suggests that the latter belongs to D. hungarica (Fig. 3).

All four phylloplane isolates in sub-groups D and E of Fig. 1 had distinct PCR-fingerprints (Fig. 2) and D1/D2 sequences (Fig. 3). Strains PYCC 5861 (isolated in Germany) and PYCC 5866 (Arrábida) could not be assigned to any of the named species included in the phylogenetic tree depicted in Fig. 3. Two novel species, D. fristingensis sp. nov. and D. buhagiarii sp. nov., are thus proposed to accommodate those isolates. This proposal is fully supported by the ITS sequences obtained for both novel species (Fig. 4). D. fristingensis may represent a sister-species to C. statzelliae, from which it differed at seven nucleotide positions in D1/D2; and D. buhagiarii a sister-species to D. hungarica (six nucleotide substitutions in D1/D2) (Fig. 3). The phylogenetic position of C. statzelliae (Figs. 3 and 4) justifies transfer to Dioszegia as a new combination (see below) as has been suggested by other authors [26]. On the other hand, strains PYCC 5856 and PYCC 5860 (both isolated in Germany) appear to belong to D. aurantiaca and D. crocea, respectively, according to their nucleotide sequences from the two rDNA regions analysed (Figs. 3 and 4). It is interesting to note that we could not find any representatives of these two species among the phylloplane isolates from Arrábida.

In a recent publication, Renker et al. [29] have reported on the amplification of ITS sequences from mycorrhizal roots in Germany that included 37 clones attributed to putative Dioszegia spp. or Cryptococcus spp. phylogenetically related to the ‘Dioszegia subclade’. We added their ITS data to our own database and found that 30 sequences (GenBank Accessions Nos. AJ581051–AJ581080) clustered with D. crocea, six (AJ581039–AJ581044) with D. fristingensis (incidentally, our single isolate of this species was isolated in Germany) and AJ581081 had no immediate neighbour (data not shown). Surprisingly, there were no two sequences alike within the two large groups of clones but it is not clear to us whether the observed differences are genuine or may have resulted from PCR, cloning and/or sequencing errors or artefacts (e.g. presence of scattered nucleotide substitutions at invariable positions in all other members of the subclade). Nevertheless, the clones assignable to D. crocea and D. fristingensis suggest that these species occur in soils and/or in association with root systems and warrant further studies. In another study based on direct amplification of fungal DNA from the wheat rhizosphere in The Netherlands [30], D. crocea was also identified among clones of amplified 18S rDNA fragments.

Three strains (PYCC 5857, 5858 and 5859) representing clusters F1 and F3 in Fig. 2 (strains that belonged to group F of Fig. 1) had almost identical D1/D2 sequences, which differed in four nucleotide substitutions from those of cluster F2 strains (PYCC 5864 and 5865). In turn, the latter differed in a single nucleotide position from the D1/D2 sequences of the type strains of both varieties of D. zsoltii[25] (Fig. 3). Moreover, three other D1/D2 sequences retrieved from GenBank and corresponding to unnamed species (Cryptococcus sp. KCTC 17077, Dioszegia spp. TY-211 and TY-217) were identical or differed in less than three nucleotides from the sequences of cluster F2 strains or of D. zsoltii (Fig. 3). Clarification of the taxonomic status of these strains and those from cluster G (Fig. 2) was only possible with the analysis of the respective ITS sequences (Fig. 4). Strains PYCC 5863, 5864 and 5865 (cluster F2) had identical ITS sequences and were clearly separated from the strains of clusters F1, F3 or G and from D. zsoltii by differences in six to ten nucleotide positions (Fig. 4). A novel species, D. takashimae sp. nov., is proposed to accommodate the former three strains (also based on the results depicted in Figs. 1 and 2). Four representatives of clusters F1 (PYCC 5857, PYCC 5859) and F3 (A2AVS8, PYCC 5858) had identical ITS sequences that differed from the cluster G and D. zsoltii strains in five to six nucleotide positions (Fig. 4). In view of all the data obtained for the seven strains that make up clusters F1 and F3, we propose to accommodate them in a novel species, D. catarinonii sp. nov. Formation of the two clusters in the MSP-PCR analysis (Fig. 2) suggests the existence of two different genotypes for this species. However, they cannot be distinguished by the respective ITS sequences (Fig. 4). Regarding the three strains in cluster G (PYCC 5867, 5868 and 5869), their ITS sequences differed from those of the two varieties of D. zsoltii in up to three nucleotide positions (Fig. 4). Since we did not obtain PCR fingerprinting data for the reference strains of the latter species, they were provisionally ascribed to D. zsoltii. However, it was not possible to define the variety with certainty based on the sequence data (Figs. 3 and 4). Bai et al. [25] stated that the only physiological difference between the two varieties is the assimilation of d-mannitol, which was found to be positive for our three isolates (the result expected for the variety zsoltii). With respect to the unnamed species in the D1/D2 tree (Fig. 3) mentioned above, an ITS sequence is available only for Dioszegia sp. strain TY-217 (Fig. 4), which appears to be unique and is thus a good candidate for a novel species. Elucidation of the status of the other two strains (TY-211 and KCTC 17077) should await further studies. Finally, none of the strains for which sequence data are available appear to represent the most recently described species of Dioszegia, D. changbaiensis[26], which has an isolated position on both phylogenetic trees (Figs. 3 and 4). Formal descriptions of the novel taxa are presented in Section 3.5.

3.5Taxonomy: Latin diagnoses and standard descriptions

3.5.1Cryptococcus amylolyticusÁ. Fonseca, J. Inácio et I. Spencer-Martins, sp. nov.

Latin diagnosis. Status teleomorphosis incognitus. In agaro “MYP” post dies 7 ad 20 °C cellulae ovoideae ad elipsoidae, (2.6–3.7) × (4.0–4.8) μm. Cultura in agaro “MYP” post dies 7 ad 20 °C aurantiaca, glabra, nitida, mucosa, margine integro. Ballistoconidia nullae. Fermentatio nulla. Characteres biochemici et physiologicique in tabula adjuncta (Table 2) describuntur. Characteres moleculares (culturae typi): sequentia acidi nucleici rDNA 26S (D1/D2), AY562134, in collectione sequentiarum acidi nucleici NCBI (GenBank) deposita est. Typus: PYCC 5850 (=CBS 10048), ex foliis Cisti albidi in Lusitania isolatus, in collectione zymotica lusitanica praeservatus.

Table 2.  Physiological/biochemical test responses of the newly proposed species
Test responsesaCryptococcus amylolyticusCryptococcus armeniacusCryptococcus cistialbidibDioszegia buhagiariiDioszegia catarinoniicDioszegia fristingensisDioszegia takashimaed
  1. aTest results: +, positive; D, delayed positive; W, weak; −, negative; V, variable; ND, not determined.

  2. bResponses for the following strains: PYCC 5851, PYCC 5853, PYCC 5854, 2CVFm15, 3CSF1, A2CSV7, A3CSVI3b.

  3. cResponses for the following strains: PYCC 5857, PYCC 5859, A2AVS8.

  4. dResponses for the following strains: PYCC 5864, PYCC 5865.

C-sources
d-Glucose++++++V
d-Galactose+D+++++
l-SorboseDDDD
d-Glucosamine+++DVDW
d-Ribose+D+++D+,D
d-Xylose+++++++
l-Arabinose+++++++
d-Arabinose+D++,DDD
l-Rhamnose+D+++D+
Sucrose+++++++
Maltose+++++++
α,α-Trehalose+++++++
Methyl-α-d-glucoside+D+,DD−,DD−,D
Cellobiose+++++++
Salicin+++++DD
Melibiose++++++
Lactose+++W,D+,D
Raffinose+++++++
Melezitose+++++++
Inulin
Soluble starch+W
GlycerolDWDD
Erythritol+(-)
Ribitol+D+DD
Xylitol+++++,W++
d-Glucitol+D+D+
d-Mannitol+D++−,D++,D
GalactitolWD++−,D++,D
Inositol+D+−,DD
Glucono δ-lactone+W,D++W+
d-Gluconic acidDD−,D++++
d-Glucuronic acid++++++V
DL-Lactic acid+,DD+,D
Succinic acid+++++++
Citric acid+D+++++
Methanol
Ethanol
l-Malic acid+++(-)++++
l-Tartaric acidW++
d-Tartaric acid−,D+
m-Tartaric acidD+,DD+
Saccharic acidD+V
Mucic acidD+,DV
Protocatechuic acidND
Vanillic acidND
Ferulic acidND
Veractric acidND
p-Hydroxybenzoic acidND
m-Hydroxybenzoic acidND
Gallic acidND
Salicylic acidND
Gentisic acidND
CatecholND
PhenolND
N-sources
Nitrate
Nitrite++V++
Ethylamine++
l-Lysine+D+++,D++,D
Cadaverine
Creatine
Creatinine
Other tests
Growth in vitamin-free medium+
Growth with 0.01% cycloheximideDD+,DV
Growth with 0.1% cycloheximide
Splitting of arbutin+++++++
Formation of starch-like compoundsWWW++++
Hydrolysis of urea+++++++
Colour reaction with Diazonium Blue B+++++++
Growth at 25 °C++++++
Growth at 30 °C

Description. Teleomorph: unknown.

Growth on MYP agar. After 7 d at 20 °C, cells oval to ellipsoidal (2.6–3.7) × (4.0–4.8) μm, single or with buds. After 1 wk colonies light orange coloured, surface glossy and smooth, texture mucous and margins entire. Ballistoconidia not produced.

Fermentation of carbohydrates: None.

Other physiological characteristics: see Table 2.

Molecular characteristics: nucleotide sequence of the D1/D2 domains of 26S (LSU) rDNA, AY562134, deposited in GenBank (type strain).

Etymology: Named for its strong starch-hydrolysing ability, a rare trait among the orange-coloured yeasts studied.

Deposits: Strain PYCC 5850 (=3CVF16), isolated in March 1998 by J. Inácio from leaves of Cistus albidus (CM) at ‘Fonte do Veado’ (northern slope of ‘Serra da Arrábida’) in the Arrábida Natural Park (Portugal) and designated as type culture, was deposited in the Portuguese Yeast Culture Collection (Caparica, Portugal) and in the Centraalbureau voor Schimmelcultures, CBS 10048 (Utrecht, The Netherlands).

3.5.2Cryptococcus armeniacusÁ. Fonseca et J. Inácio, sp. nov.

Latin diagnosis. Status teleomorphosis incognitus. In agaro “MYP” post dies 7 ad 20 °C cellulae elipsoidae ad cylindraceae, (2.9–3.7) × (4.6–6.7) μm. Cultura in agaro “MYP” post dies 7 ad 20 °C aurantiaca, glabra, nitida, mucosa, margine integro. Ballistoconidia nullae. Fermentatio nulla. Characteres biochemici et physiologicique in tabula adjuncta (Table 2) describuntur. Characteres moleculares (culturae typi): sequentia acidi nucleici rDNA 26S (D1/D2), AY562140, in collectione sequentiarum acidi nucleici NCBI (GenBank) deposita est. Typus: PYCC 5855 (=CBS 10050), ex aere in Lusitania isolatus, in collectione zymotica lusitanica praeservatus.

Description. Teleomorph: unknown.

Growth on MYP agar. After 7 d at 20 °C, cells ellipsoidal to cylindrical (2.9–3.7) × (4.6–6.7) μm, single or with buds. After 1 wk colonies light orange coloured, surface glossy and smooth, texture mucous and margins entire. Ballistoconidia not produced.

Fermentation of carbohydrates: None.

Other physiological characteristics: see Table 2.

Molecular characteristics: nucleotide sequence of the D1/D2 domains of 26S (LSU) rDNA, AY562140, deposited in GenBank (type strain).

Etymology: Named for the light orange colour of its colonies on agar media.

Deposits: Strain PYCC 5855 (=Cont.NA L), isolated in April 1997 by Á. Fonseca as a single colony that appeared as contaminant on a nutrient agar plate (Caparica, Portugal) and designated as type culture, was deposited in the Portuguese Yeast Culture Collection (Caparica, Portugal) and in the Centraalbureau voor Schimmelcultures, CBS 10050 (Utrecht, The Netherlands).

3.5.3Cryptococcus cistialbidiÁ. Fonseca, J. Inácio et I. Spencer-Martins, sp. nov.

Latin diagnosis. Status teleomorphosis incognitus. In agaro “MYP” post dies 7 ad 20 °C cellulae elipsoidae ad cylindraceae, (3.3–4.2) × (5.6–7.0) μm. Cultura in agaro “MYP” post dies 7 ad 20 °C aurantiaca, glabra, nitida, mucosa, margine integro. Ballistoconidia nullae. Fermentatio nulla. Characteres biochemici et physiologicique in tabula adjuncta (Table 2) describuntur. Characteres moleculares (culturae typi): sequentia acidi nucleici rDNA 26S (D1/D2), AY562135, in collectione sequentiarum acidi nucleici NCBI (GenBank) deposita est. Typus: PYCC 5851 (=CBS 10049), ex foliis Cisti albidi in Lusitania isolatus, in collectione zymotica lusitanica praeservatus.

Description. Teleomorph: unknown.

Growth on MYP agar. After 7 d at 20 °C, cells ellipsoidal to cylindrical (3.3–4.2) × (5.6–7.0) μm, single or with buds. After 1 wk colonies light orange coloured, surface glossy and smooth, texture mucous and margins entire. Ballistoconidia not produced.

Fermentation of carbohydrates: None.

Other physiological characteristics: see Table 2.

Molecular characteristics: nucleotide sequence of the D1/D2 domains of 26S (LSU) rDNA, AY562135, deposited in GenBank (type strain); other strains studied had identical D1/D2 sequences (GenBank accession numbers in Table 1).

Etymology: From Cistus albidus, the species name of the evergreen shrub from which many of the studied strains were isolated.

Deposits: Strain PYCC 5851 (=1CSF5), isolated in March 1997 by J. Inácio from leaves of Cistus albidus (CM) at ‘Mata do Solitário’ in the Arrábida Natural Park (Portugal) and designated as type culture, was deposited in the Portuguese Yeast Culture Collection (Caparica, Portugal) and in the Centraalbureau voor Schimmelcultures, CBS 10049 (Utrecht, The Netherlands). Additional strains were deposited in the Portuguese Yeast Culture Collection or are maintained at CREM (see Table 1).

3.5.4Dioszegia buhagiariiÁ. Fonseca, J. Inácio et I. Spencer-Martins, sp. nov.

Latin diagnosis. Status teleomorphosis incognitus. In agaro “MYP” post dies 7 ad 20 °C cellulae ovoideae, (3.3–4.0) × (3.8–4.3) μm. Cultura in agaro “MYP” post dies 7 ad 20 °C aurantiaca, glabra, nitida, butyracea, margine integro. Ballistoconidia nullae. Fermentatio nulla. Characteres biochemici et physiologicique in tabula adjuncta (Table 2) describuntur. Vitamina externa crescentiae necessaria sunt. Characteres moleculares (culturae typi): sequentia acidi nucleici rDNA 26S (D1/D2), AY562151, et rDNA ITS, AY885687, in collectione sequentiarum acidi nucleici NCBI (GenBank) deposita est. Typus: PYCC 5866 (=CBS 10054), ex foliis Acer monspessulani in Lusitania isolatus, in collectione zymotica lusitanica praeservatus.

Description. Teleomorph: unknown.

Growth on MYP agar. After 7 d at 20 °C, cells mostly oval (3.3–4.0) × (3.8–4.3) μm, single or with buds. After 1 wk colonies light orange coloured, surface glossy and smooth, texture butyrous and margins entire. Ballistoconidia not produced.

Fermentation of carbohydrates: None.

Other physiological characteristics: see Table 2.

Molecular characteristics: nucleotide sequences of the D1/D2 domains of 26S (LSU) rDNA, AY562151, and of the ITS region, AY885687, deposited in GenBank (type strain).

Etymology: Named in honour of R.W.M. Buhagiar, the author who isolated and described Bullera armeniaca and B. crocea.

Deposits: Strain PYCC 5866 (=4AVF6), isolated in June 1998 by J. Inácio from leaves of Acer monspessulanum (CM) at ‘Fonte do Veado’ (northern slope of ‘Serra da Arrábida’) in the Arrábida Natural Park (Portugal) and designated as type culture, was deposited in the Portuguese Yeast Culture Collection (Caparica, Portugal) and in the Centraalbureau voor Schimmelcultures, CBS 10054 (Utrecht, The Netherlands).

3.5.5Dioszegia catarinoniiÁ. Fonseca, J. Inácio et I. Spencer-Martins, sp. nov.

Latin diagnosis. Status teleomorphosis incognitus. In agaro “MYP” post dies 7 ad 20 °C cellulae globosae ad elipsoidae (2.6–3.7) × (4.0–4.8) μm. Cultura in agaro “MYP” post dies 7 ad 20 °C aurantiaca, glabra, nitida, butyracea, margine integro. Ballistoconidia exiguae vel nullae. Fermentatio nulla. Characteres biochemici et physiologicique in tabula adjuncta (Table 2) describuntur. Characteres moleculares (culturae typi): sequentiae acidi nucleici rDNA 26S (D1/D2), AY562142, et rDNA ITS, AY562154, in collectione sequentiarum acidi nucleici NCBI (GenBank) deposita est. Typus: PYCC 5857 (=CBS 10051), ex folio Acer monspessulani in Lusitania isolatus, in collectione zymotica lusitanica praeservatus.

Description. Teleomorph: unknown.

Growth on MYP agar. After 7 d at 20 °C, cells globose to ellipsoidal (2.6–3.7) × (4.0–4.8) μm, single or with buds. After 1 wk colonies light orange coloured, surface glossy and smooth, texture butyrous and margins entire. Rotationally symmetrical ballistoconidia scantily produced or not produced at all.

Fermentation of carbohydrates: None.

Other physiological characteristics: see Table 2. Strain 5AVFs3 (PYCC 5858) grew poorly on liquid media and test responses deviated significantly from those listed in Table 2.

Molecular characteristics: nucleotide sequence of the D1/D2 domains of 26S (LSU) rDNA, AY562142, and of the ITS region, AY562154, deposited in GenBank (type strain); other strains studied had identical ITS sequences and D1/D2 sequences that differed from those of the type strain in one nucleotide position (GenBank accession numbers in Table 1).

Etymology: Named in honour of Prof. Fernando Catarino, former director of the Botanical Garden of the University of Lisbon and a profound connoisseur of the flora of the Arrábida Natural Park.

Deposits: Strain PYCC 5857 (=A2AVV2), isolated in November 1997 by J. Inácio from a leaf of Acer monspessulanum (SFM) at ‘Fonte do Veado’ (northern slope of ‘Serra da Arrábida’) in the Arrábida Natural Park (Portugal) and designated as type culture, was deposited in the Portuguese Yeast Culture Collection (Caparica, Portugal) and in the Centraalbureau voor Schimmelcultures, CBS 10051 (Utrecht, The Netherlands). Additional strains were deposited in the Portuguese Yeast Culture Collection or are maintained at CREM (see Table 1).

3.5.6Dioszegia fristingensisÁ. Fonseca, J. Inácio et J.P. Sampaio, sp. nov.

Latin diagnosis. Status teleomorphosis incognitus. In agaro “MYP” post dies 7 ad 20 °C cellulae elipsoidae ad cylindraceae, (2.4–3.4) × (5.2–6.3) μm. Cultura in agaro “MYP” post dies 7 ad 20 °C aurantiaca, glabra, nitida, butyracea, margine integro. Ballistoconidia flabelliformes aut castaniformes (5.0–7.0) × (4.4–5.2) μm. Fermentatio nulla. Characteres biochemici et physiologicique in tabula adjuncta (Table 2) describuntur. Characteres moleculares (culturae typi): sequentiae acidi nucleici rDNA 26S (D1/D2), AY562146, et rDNA ITS, AY562158, in collectione sequentiarum acidi nucleici NCBI (GenBank) deposita est. Typus: PYCC 5861 (=CBS 10052), ex folio Ari maculati in Germania isolatus, in collectione zymotica lusitanica praeservatus.

Description. Teleomorph: unknown.

Growth on MYP agar. After 7 d at 20 °C, cells ellipsoidal to cylindrical (2.4–3.4) × (5.2–6.3) μm, single or with buds. After 1 wk colonies dark orange coloured, surface glossy and smooth, texture butyrous and margins entire. Napiform to flabelliform ballistoconidia (5.0–7.0) × (4.4–5.2) μm are abundantly produced (MYP, CMA).

Fermentation of carbohydrates: None.

Other physiological characteristics: see Table 2.

Molecular characteristics: nucleotide sequence of the D1/D2 domains of 26S (LSU) rDNA, AY562146, and of the ITS region, AY562158, deposited in GenBank (type strain).

Etymology: From Fristingen (Germany), the Bavarian locality where the type strain was isolated.

Deposits: Strain PYCC 5861 (=ZP 359), isolated in May 1997 by J.P. Sampaio from a leaf of Arum maculatum (SFM) infected with Melanotaenium ari at Fristingen (Germany) and designated as type culture, was deposited in the Portuguese Yeast Culture Collection (Caparica, Portugal) and in the Centraalbureau voor Schimmelcultures, CBS 10052 (Utrecht, The Netherlands). A spontaneous, colourless mutant of strain PYCC 5861 is also maintained at CREM (its genetic identity to the original strain was confirmed by MSP-PCR, data not shown).

3.5.7Dioszegia takashimaeÁ. Fonseca, J. Inácio et I. Spencer-Martins, sp. nov.

Latin diagnosis. Status teleomorphosis incognitus. In agaro “MYP” post dies 7 ad 20 °C cellulae elipsoidae (2.8–3.6) × (4.6–5.2) μm. Cultura in agaro “MYP” post dies 7 ad 20 °C aurantiaca, glabra, nitida, butyracea, margine integro. Ballistoconidia subglobosae aut napiformes (4.6–5.8) × (5.2–6.7) μm. Fermentatio nulla. Characteres biochemici et physiologicique in tabula (Table 2) describuntur. Characteres moleculares (culturae typi): sequentiae acidi nucleici rDNA 26S (D1/D2), AY562149, et rDNA ITS, AY562160, in collectione sequentiarum acidi nucleici NCBI (GenBank) deposita est. Typus: PYCC 5864 (=CBS 10053), ex folio Querci faginea in Lusitania isolatus, in collectione zymotica lusitanica praeservatus.

Description. Teleomorph: unknown.

Growth on MYP agar. After 7 d at 20 °C, cells mostly ellipsoidal (2.8–3.6) × (4.6–5.2) μm, single or with buds. After 1 wk colonies light orange coloured, surface glossy and smooth, texture butyrous and margins entire. Subglobose ballistoconidia (4.6–5.8) × (5.2–6.7) μm, are abundantly produced (MYP, CMA, YCB).

Fermentation of carbohydrates: None.

Other physiological characteristics: see Table 2.

Molecular characteristics: nucleotide sequence of the D1/D2 domains of 26S (LSU) rDNA, AY562149, and of the ITS region, AY562160, deposited in GenBank (type strain); other strains studied had identical D1/D2 and ITS sequences (GenBank accession numbers in Table 1).

Etymology: Named in honour of Dr. Masako Takashima, in recognition of her many studies on the systematics of ballistoconidiogenic yeasts, including leading authorship of the paper in which the genus Dioszegia was re-instated [13].

Deposits: Strain PYCC 5864 (=A2QSS4), isolated in November 1997 by J. Inácio from a leaf of Quercus faginea (SFM) at ‘Mata do Solitário’ (southern slope of ‘Serra da Arrábida’) in the Arrábida Natural Park (Portugal) and designated as type culture, was deposited in the Portuguese Yeast Culture Collection (Caparica, Portugal) and in the Centraalbureau voor Schimmelcultures, CBS 10053 (Utrecht, The Netherlands). Additional strains were deposited in the Portuguese Yeast Culture Collection or are maintained at CREM (see Table 1).

3.5.8Dioszegia statzelliae (Thomas-Hall, Watson & Scorzetti) Á. Fonseca, comb. nov.

Basionym: Cryptococcus statzelliae Thomas-Hall, Watson & Scorzetti; in Int. J. Syst. Evol. Microbiol. 52 (2002) 2306; type strain: CBS 8925.

3.6Concluding remarks

Surveys of phylloplane yeasts from different plants and climatic regions have been reported by several authors but the full extent of species diversity in this particular habitat has not yet been unveiled (e.g. [9]). About 850 yeast isolates from our two-year survey of the epiphytic fungi on leaves of Mediterranean plants at the ‘Arrábida Natural Park’[9] may represent at least 70 distinct species, half of which seem to correspond to undescribed species [17]. Here, we focused on the hymenomycetous yeasts that produced orange colonies (results for some of the other taxonomic groups studied were presented elsewhere [17,34,35]). The 35 phylloplane isolates from Arrábida discussed herein were found to represent seven distinct species, five of which are newly described (Table 1). This finding illustrates the importance of this type of study in which culture-dependent isolation methods were followed by molecular-taxonomy techniques that enabled an adequate assessment of species diversity among the isolates.

Another interesting finding of this study was that production of deeply pigmented colonies (orange, pink or red) by members of the Tremellales is apparently not restricted to members of the ‘Luteolus clade’, namely the species transferred to the genus Dioszegia[13]. In fact, we have now found three novel Cryptococcus spp. that produce orange colonies but are not closely related to Dioszegia spp. (Fig. 3). Concerning the production of ballistoconidia we consistently found this type of cell in the isolates belonging to D. aurantiaca, D. crocea, D. fristingensis, D. hungarica and D. takashimae, whereas they were neither observed in D. buhagiarii nor in the three isolates ascribed to D. zsoltii, a ballistoconidiogenic species according to the original description [25]. As for D. catarinonii, only some of the strains produced ballistoconidia (Fig. 1) but sparsely and only immediately after their isolation, that capacity having apparently been lost upon sub-culturing. As mentioned before, this inconsistency in ballistoconidiogenesis appears to be common in the genus Dioszegia (e.g. [13]). In effect Nakase and Suzuki [31] had reported that only five out of thirteen strains freshly isolated from dead rice leaves and identified as B. crocea produced ballistoconidia vigorously while the remaining eight produced them only poorly or not at all. The ecological and taxonomical implications of ballistoconidiogenesis in basidiomycetous yeasts are outside the scope of the present paper but have been discussed by other authors (e.g. [11,32,33]).

We have added four new members to Dioszegia, thus almost doubling the number of recognized species in this recently reinstated genus. Altogether, Dioszegia species appear to be present either on plant surfaces (mainly later in the season or on senescing leaves, e.g. D. catarinonii, D. takashimae and D. zsoltii) or soil and some species may be restricted to areas with colder climates, namely D. aurantiaca, D. crocea, D. fristingensis and D. statzelliae. However, two isolates from seawater (CBS 6324 and 6576) likely represent a yet unnamed Dioszegia species (Figs. 3 and 4) and suggest that members of this genus could have a more widespread ecological distribution. Differentiation of the newly described taxa from named species in the Tremellales that also produce orange colonies is in some cases possible based on a combination of physiological characteristics (Table 3). However, responses to some of those tests may be doubtful or variable (e.g. discrepant responses in liquid or solid media, see Fig. 1 and Table 2) and distinction of many species pairs relies on just one or two tests (e.g. D. aurantiaca/D. crocea) or is not at all possible (e.g. D. catarinonii/D. takashimae). Reliable identifications can thus only be achieved by sequencing of the D1/D2 domains of the 26S rRNA gene and/or of the ITS region.

Table 3.  Discriminating physiological/biochemical characteristics of orange-coloured tremellaceous speciesa
SpeciesMethyl-α-glucosideMelibioseLactoseRibitolGlucitolInositolErythritoll-Tartaric acidSaccharic acidMucic acidNitrite (N)Ethylamine (N)Growth at 25 °C
  1. aData retrieved from the CBS Yeast Database (http://www.cbs.knaw.nl/databases/index.htm) or obtained in the present study for the newly proposed species (in bold) and for the aldaric acids; data for D. changbaiensis taken from [26]; test results as in Table 2.

D. aurantiaca−/D++D+++
D. changbaiensis+DNDNDND++
D. crocea−/D++++++++
D. hungarica++/D−/D+/D+/D++
D. statzelliaeW++/WW+WWWW
D. zsoltii−/D++−/D++++
D. buhagiarii/DDD/W+/D+/D++
D. catarinonii/D+/D/D+++/DV+
D. fristingensisD+D+V++
D. takashimae/D++/D/D+++++
C. amylolyticus+++++++++
C. armeniacusD++DDD+
C. cistialbidi+/D+++++++

Acknowledgements

This work was partly funded by ‘Fundação para a Ciência e Tecnologia’ and FEDER (Praxis XXI/2/2.1/BIA/413/94. J.I. received a Ph.D. grant (Praxis XXI/BD/19833/99). We are indebted to Dr. Alan J.L. Phillips (CREM, Portugal) for checking the English and for critical reading of the manuscript.

Ancillary