Introduction

Methanol-assimilating yeasts have been observed from the various plant-related substrates, like slime fluxes (Péter et al. 2006), tree bark, rotten wood material (Péter et al. 2003), and leaves (Péter et al. 2007). Recently they have been detected on about 45% of leaves collected in Hungary (Péter et al. 2007). Methanol produced inside the leaves and emitted primarily through stomata (Fall and Benson 1996) can obviously support development of these ascomycetous yeasts although the most common phylloplane colonisers are of basidiomycetous affinity (Péter et al. 2007). Consequently, inner plant tissues could be even more promising habitats for detection of the yeasts with rare nutritional abilities, including the assimilation of methanol. Indeed, endophytous yeast communities have been reported recently to be significantly distinct from those formed on plant surfaces (Isaeva et al. 2009). Abnormal growths formed from plant tissues and due to parasitic activity of another organism, known as galls (Redfern and Shirley 2002) are annually observed on leaves of deciduous plants. They can be induced by fungal or bacterial infections as well as the activity of different invertebrates. Sawflies are well known gall-inducing insects colonising willows in Eurasia with an extensive and opportunistic host shifts (Price 2005). The wide ecological distribution of willows and the proximity of plant species in the field enable frequent host shifts and radiation of sawflies, but also make galls, as a substrate, more abundant in temperate zone. While plant surfaces have been recognized as an important habitat for yeasts and filamentous fungi for a long time, galls, as a substrate received very little attention so far.

During the survey aimed to analyse yeasts found as endophytes in insect galls of willow leaves in comparison with regular epiphytic yeast community, a novel methanol-assimilating yeast was isolated repeatedly. This species is described here as Ogataea cecidiorum.

Materials and methods

The survey was performed at Losiny Ostrov (Elk park) National Park, Moscow, Russia (approximate coordinates: 55.85N; 37.75E). Particular attention was given to the analysis of leaves that formed galls on their surface, which were induced by sawflies (Pontania spp. and Euura spp., Order Hymenoptera, Familia Tenthredinidae). Leaves of willows (Salix myrsinifolia Salisb.) were collected between May and September 2008, as long as galls were not destroyed due to elaboration of adult insects. Seventy intact leaves and 70 leaves containing a gall were collected and analysed.

Phylloplane yeast community was analysed by plating of leaf washings. Leaves were placed in 50 ml plastic tube, suspended 1:50 (w/v) in sterile demineralised water and vortexed. An aliquot of 0.1–0.2 ml was distributed on the surface of acidified with 4 ml/l lactate (final pH 4–4.5) glucose-yeast extract-peptone (GPY) agar, containing 20 g/l glucose, 5 g/l yeast extract, 10 g/l peptone, 20 g/l agar. Each sample was plated in triplicates. Pre-cultivation sample treatment was performed to isolate yeasts from the internal gall layers. For this, leaf surfaces were sterilised with ethanol and the epidermis was removed. The internal part of each gall was placed in a 50 ml plastic tube, diluted 1:50 (w/v) in sterile demineralised water, vortexed and plated in triplicates on acidified GPY agar as described above. Plates were incubated at room temperature and checked after one and two weeks. Grown yeast colonies were differentiated into macromorphological types using a dissection microscope, counted, and 1–2 representatives of every colony type per plate were purified. Phenotypic characterization of purified isolates was carried out according to Yarrow (1998).

DNA was extracted using the protocol of Sampaio et al. (2001) with the modification of cell lysate centrifugation at room temperature for 15 min, with 14,500 rpm. Mini- and microsatellite-specific PCR fingerprints and Restriction Fragments Length Polymorphism analysis were performed as described before (Yurkov and Chernov 2005; Kachalkin et al. 2008). For sequencing of the D1/D2 region of the 26S ribosomal gene and the internal transcribed spacers (ITS), DNA fragments were amplified with the ITS1f (5′-CTT GGT CAT TTA GAG GAA GTA) and NL4 (5′-GGT CCG TGT TTC AAG ACG G) primer pair using a program consisting of an initial denaturation step of 2 min at 96°C, followed by 35 cycles of 20 s at 96°C, 50 s 52°C and 1.5 min at 72°C and a final extension step of 7 min at 72°C. The 18S ribosomal RNA gene region was amplified according to White et al. (1990). PCR products were purified with my-Budget Double Pure kit (Bio-Budget Tech., Germany) and sequenced using an ABI3130xl sequencer using the amplification primers. The assembly and editing of sequence data was performed using Sequencher 4.7 (Gene Codes Corp., USA). Alignments were made using the MAFFT algorithm (Katoh et al. 2002). The model of DNA substitution (GTR + G+I) and parameters for maximum likelihood analysis were derived by using Modeltest, version 3.7 (Posada and Crandall 1998). Maximum likelihood analysis was performed using RAxML, version 7.0.3 with 1,000 rounds of bootstrap replicates (Stamatakis et al. 2008; Felsenstein 1985).

Constrained trees based on several hypotheses for the phylogenetic relationships of the novel species and the other members of the Ogataea genus were constructed with PAUP 4.0b10 (Swofford 2002) and the Kishino–Hasegawa test (Kishino and Hasegawa 1989) was performed to evaluate the significance of those hypotheses. Therefore congruence between best maximum likelihood and constrained tree topologies also was examined by the Kishino–Hasegawa test.

Results and discussion

Ecology

The total abundance of yeasts inhabiting the leaf surface estimated by the plating technique was 2.5 × 105 CFU/g, in average and maximal 5.0 × 105 CFU/g in June. Galls contained less yeasts, 5.0 × 104 CFU/g in average, but the total yeast quantity gradually increased through the sampling period reaching a maximum of 1.5 × 105 CFU/g in August.

The internal gall yeast community observed at the beginning of gall formation resembled very closely the epiphytic community with a dominance of filobasidious cryptococci. One month later, galls contained mainly ascomycetous yeasts, in contrast to the phylloplane, which was inhabited mainly by the basidiomycetous yeasts Cryptococcus spp. (Magnus clade, Filobasidiales), Rhodotorula glutinis sensu lato and Cystofilobasidium capitatum. Ascomycetes were very rare on the leaf surface. Besides Candida oleophila, which had been reported to be a regular phylloplane yeast (Glushakova et al. 2007), Wickerhamomyces anomalus (Pichia anomala) was observed on the leaf surface from June to September.

The internal gall yeast community was dominated by the novel methanol-assimilating yeast with an average relative abundance of 94–100% in June, July and August. However, the novel species was rarely observed at the initial stage of the gall formation in May (<1%) and after elaboration of adult insects, when galls dried out (9%). These observations imply that the environmental conditions formed in the galls during the development of sawfly’s larvae are favourable for methanol-assimilating yeast. This could be, partly, explained by the fact that methanol produced in the plant tissues (Fall and Benson 1996) can support the development of the yeast population in a confined environment like a gall. Péter et al. (2007) recently reported the regular occurrence of methanol-assimilating yeasts on leave surfaces in Hungary, however, often in a low abundance. Regular isolation of O. cecidiorum and its high relative abundance in galls allow us to treat this habitat as a probable ecological niche for this species. A possible function of the gall-inducing insects as vectors for the yeast is considered and will lead to additional studies.

Identification, characterisation and distinction from related species

A total of about 100 cultures were isolated and studied in the survey. The physiological properties of the novel species resembled those of methanol-assimilating yeasts, including members of genera Ogataea, Pichia and related Candida species. The most similar species was Pichia methanolica, from which it differed in the assimilation of galactose, dulcitol and salicin (Kurtzman and Fell 1998).

We reduced the number of strains involved in the molecular studies by taking two to three isolates when they were obtained from the same sampling time, but also displayed identical growth responses. Thereby, 10 isolates obtained from different galls collected at different times were chosen to represent the novel species and were used in further analyses. They were grouped by PCR-fingerprinting with minisatellite (MS)-specific oligonucleotide derived from the core sequence of bacteriophage M13 with the sequence given by Sampaio et al. (2001) and the microsatellite (MS)-specific oligonucleotides (GTG)5, (ATG)5 and (GAC)5 as single PCR primer. This technique has been repeatedly demonstrated to distinguish closely related species and populations of ascomycetous yeasts (Inácio et al. 2004; Naumova et al. 2004; Yurkov and Chernov 2005; Fidalgo-Jiménez et al. 2008). All of the studies strains showed high similarity of MS-fingerprints suggesting conspecificity (Fig. S1). Their conspecificity was additionally tested by Restriction Fragment Length Polymorphisms of the 5.8S-ITS rDNA region (Esteve-Zarzoso et al. 1999). All strains showed identical restriction profiles, namely, fragments of 540 + 250 (HaeIII), 480 + 320 (HinfI), and 400 + 330 (CfoI) nucleotides. However, no match was found between observed and previously recorded restriction profiles (Esteve-Zarzoso et al. 1999; Villa-Carvajal et al. 2006).

The nearest sequence match among currently recognized species was obtained with Pichia trehalophila showing four nucleotide substitutions and one gap in the D1/D2 domains of LSU rRNA, one nucleotide substitution and one gap in SSU rRNA gene, and 16 nucleotide substitutions and three gaps in the 5.8S-ITS fragment. Phylogenetic analysis of pairwise sequence similarity for the small subunit rRNA gene, ITS region (ITS1-5.8S rRNA gene-ITS2) and the large subunit rRNA gene (D1/D2 domains) suggested relatedness of the newly described species with Pichia methanolica, Pichia trehalophila, and Williopsis salicorniae (Fig. 1, Fig. S2), also referred to the Pichia methanolica group (Nagatsuka et al. 2008). This group has repeatedly been mentioned to belong to the phylogenetic clade of methanol assimilating yeasts, comprising, amongst others, also Pichia angusta (Ogataea polymorpha) (Kurtzman and Robnett 1998; Morais et al. 2004). Recently, Kurtzman and Robnett (2010) proposed the transfer of the Pichia methanolica group and the sister species, Pichia pilisensis, P. naganishii, P. methylivora and P. ramenticola to the genus.

Fig. 1
figure 1

Unrooted tree illustrating phylogenetic placement of O. cecidiorum obtained by maximum likelihood analysis of 18S rRNA, 26S rRNA (D1/D2 domains) genes and the ITS region. The numbers given on branches are frequencies (>50%) with which a given branch appeared in 1,000 bootstrap replications. The scale indicates the number of expected substitutions accumulated per site. Sequence accession numbers of type strains are listed; sequences determined in this study are given in bold

Full size image

Analysis of the large subunit (D1/D2 domains) and the small subunit rRNA genes showed high similarities among the type strains of the “Pichia methanolica” group (Nagatsuka et al. 2008), including the proposed species O. cecidiorum, while ITS sequences clearly distinguished them (Table 1). High sequence similarities of ribosomal gene regions, not exceeding five nucleotide positions in the D1/D2 domains, were also reported for the species pair Pichia pilisensis–Ogataea nitratoaversa (Péter et al. 2008) and species, forming another subgroup in the Ogataea clade, related to Ogataea glucozyma (Nagatsuka et al. 2008).

Table 1 Pairwise comparison of 26S (D1/D2 domains, 570 bp), 18S (1670 bp) rRNA genes and the ITS1-5.8S-ITS2 region (730 bp) sequence data derived from alignment (MAFFT algorithm): substitutions and gaps, illustrating relationship within P. methanolica cluster
Full size table

The following physiological tests, namely positive assimilation of methyl-α-d-glucoside, inability to ferment trehalose and the maximal growth temperature, provide phenotypic differentiation of the novel species from Pichia trehalophila. All members of the Pichia methanolica group were considered as independent species on the basis of physiological tests. The first described species, Pichia trehalophla was isolated from tree fluxes (Phaff et al. 1964). Later, Pichia methanolica and Pichia cellobiosa were isolated from the soil in Japan (Kato et al. 1974; Lee and Komagata 1980). Conspecificity of these species was demonstrated in DNA–DNA reassociation experiments by Kurtzman (1992) and P. methanolica has priority over the name P. cellobiosa. Interestingly, P. methanolica and P. cellobiosa differed in the original descriptions only in assimilation of l-sorbose and dl-lactate and fermentation of maltose. However, Lee and Komagata (1980) designated Pichia castillae as the closest species on the basis of physiological tests. In the original description Pichia methanolica significantly differs from Pichia trehalophila in assimilation of seven out of 30 carbon sources. The former species also grew above 37°C. Therefore, the recognition of Pichia methanolica was well-supported on the basis of a phenotypic approach. The last species, Williopsis salicorniae, was compared in physiological properties with the other members of Williopsis genus. This species did not assimilate methanol, and its relationship with methanol-assimilating yeasts was not recognised in the original description (Hinzelin et al. 1991). Close relationship of these taxa and other methanol-assimilating yeasts was not revealed before the analyses of partial 26S rRNA (Kurtzman and Robnett 1998) and 18S rRNA (James et al. 1998; Suzuki and Nakase 1999) genes sequence data.

Latin diagnosis of O. cecidiorum Glushakova, Maximova, Kachalkin et Yurkov sp. nov.

In medio liquido cum extracto malti (YM) post dies tres ad 25°C cellulae singulae, binae, in catenis brevis, aut congregationibus; cellulae ovoidae aut elongatae (1.8–4.8 × 2.5–5 μm). Post mensem unum sedimentum et pellicula (exigue) formantur. Cultura in striis in agaro cum dextroso et peptono et extracto levidinis (GPY) post unum mensem ad 25°C candida ad canita, obscura, laevis, margine integri. In agaro cum extracto malti post 3–5 dies ad 25°C cellulae sphaeroideae ad ovoideae, singulae, binae vel aggregatae, multilateraliter gemmantes. Post hebdomades tres in agaro farinae Zea mays vel agaro YM ad 20–25°C pseudohyphae vel hyphae verae desunt. Status teleomorphicus post cultivationem 3–7 dies ad 25°C in agaro cum extracto Solani tuberose et dextroso (PD) vel in agaro cum extractis levidinis et malti observatus, homothallicus. Asci conjugatione cellularum vel conjugatione cellularum gemmarumque oriuntur. Asci continentens 2–4 galeiformes aut pileiformes sporas. Glucosum fermentatur (procrastinater, post 3–4 dies) at non d-galactosum, sucrosum, d-maltosum, trehalosum, d-xylosum, nec raffinosum. L-sorbosum, trehalosum, d-xylosum (variabiliter), L-arabinosum (variabiliter), d-arabinosum (variabiliter), d-ribosum, L-rhamnosum (interdum exigue), methanolum, ethanolum, glycerolum, erythritolum, ribitolum, sorbitolum, d-mannitolum, methyl-α-d-glucosidum, arbutinum (variabiliter) et acidum succinicum (variabiliter) assimilantur at non d-galactosum, sucrosum, d-maltosum, cellobiosum, lactosum, melibiosum, raffinosum, melezitosum, inulinum, amylum solubile, d-glucosaminum, acidum 2-ketogluconicum, acidum 5-ketogluconicum, acidum DL-lacticum, acidium citricum, acidium acidum d-glucuronicum, nec inositolum. Assimilatio nitro-compositorum: kalium nitricum (aliquando exigue) et lisinum. Materia amyloidea iodophila non formantur. Ureum non finditur. Diazonium caeruleum B est negativum. Temperatura maxima crescentiae: 32°C. Cultura typica KBP Y-3846 isolata ex cecidium S. myrsinifolia, viva et exsiccata numero CBS 11522T ( = VKM Y-2982T = VKPM Y-3482T = MUCL 52544T = NCAIM Y.01965T) in collectione zymotica Centraalbureau voor Schimmelcultures, Trajectum ad Rhenum, Hollandia, sustentat.

Description of Ogataea secidiorum Glushakova, Maximova, Kachalkin et Yurkov sp. nov.

In liquid malt extract (YM) medium after three days at 25°C cells are ovoid to elongated (1.8–4.8 × 2.5–5 μm), occurring singly, in pairs or in small clusters. After 1 month sediment and pellicle were formed. On Glucose Peptone Yeast extract Agar (GPYA), after 1 month at 25°C, the streak culture is whitish to pale-grey, dull and smooth. The margin is entire. After growth on YM agar for 3–5 days at 25°C, cells are spheroidal to short ovoidal (1.8–4.8 × 2.5–5.0 μm), occur singly, in pairs or in small clusters and proliferate by multilateral budding (Fig. 2). Pseudohyphae and true hyphae are not observed after three weeks in plate culture on YM agar or on Corn Meal (CM) agar at 20–25°C. On modified Gorodkowa agar cells are angular. The teleomorphic stage was obtained independently for all 10 strains on different media recommended for ascospore formation (van der Walt and Yarrow 1984; Yarrow 1998). Ascospores were observed on McClary’s acetate, potato-dextrose (PD, Difco), Gorodkowa, and YM agars after 3–7 days at 25°C. Strain KBP Y-3846 produced ascospores on GPY agar after replating from PD agar. Copulation tubes were frequently observed on McClary’s acetate and potato-dextrose agars. Ascus formation may be preceded by either conjugation between a parent cell and a bud or by conjugation between independent cells. Asci contain two or four hat-shaped ascospores, and after maturation, ascospores are liberated from the ascus and tend to agglutinate (Fig. 2).

Fig. 2
figure 2

Phase contrast micrographs (a–c) and scanning electron microscopic image (d) of O. cecidiorum KBP Y-3846. Vegetative cells reproducing by multilateral budding (a) after 7 days on YM agar at room temperature. Conjugated ascus with ascospores and copulation tube (b); hat-shaped ascospores are liberating from the ascus (c); McClary acetate agar, 14 days at room temperature, bar = 10 μm. Bar on scanning electron microscopic image (d) = 2 μm

Full size image

Glucose fermentation is delayed and starts after 3–4 days. No fermentation of galactose, sucrose, maltose, trehalose, xylose and raffinose was detected. Assimilation of carbon compounds: l-sorbose, trehalose, d-xylose (variable), l-arabinose (variable), d-arabinose (variable), d-ribose, l-rhamnose (occasionally weak), methanol, ethanol, glycerol, erythritol, ribitol, sorbitol, d-mannitol, methyl-α-d-glucoside, arbutin (variable) and succinic acid (variable) are assimilated. No growth occurs on galactose, sucrose, maltose, cellobiose, lactose, melibiose, raffinose, melezitose, inulin, soluble starch, d-glucosamine, 2-ketogluconic acid, 5-ketogluconic acid, dl-lactic acid, citric acid, d-glucuronic acid and inositol. Assimilation of nitrogen compounds: potassium nitrate (occasionally weak) and l-lysine. Starch-like compounds are not produced. Growth at 20, 25, 28, 30 and 32°C is positive. Growth at 34, 37 and 40°C is negative. Urease activity is negative. The Diazonium Blue B reaction is negative. Growth on vitamin-free medium is positive. Growth on YM agar with 10% sodium chloride is negative. Growth in 50% glucose/yeast extract (0.5%) is weak. Growth on 1% acid acetic medium is negative. Growth in the presence of 0.1% cycloheximide is positive.

The habitat is galls induced by sawflies on leaves of willows. Type strain KBP Y-3846 was isolated from leaf sample collected in the Losiny Ostrov National Park, Moscow, Russia. It has been deposited in the collection of the Yeast Division of the Centraalbureau voor Schimmelcultures, Utrecht, the Netherlands, as strain CBS 11522T (= VKM Y-2982T = VKPM Y-3482T = MUCL 52544T = NCAIM Y.01965T). The specific epithet cecidiorum refers to the substrate, galls (cecidium) from which the species has been isolated.

The sequence of the ITS region (ITS1-5.8S-ITS2) and 26S rRNA (D1/D2 domain) is deposited in GenBank under the accession number FJ897742. The sequence of the 18S rRNA gene is deposited in GenBank under the accession number FN429983.

Phylogenetic placement

Despite the observation that most methanol-assimilating yeasts appear closely related (Kurtzman and Robnett 1998), the genus Ogataea has not been well recognised until recently (Suh et al. 2006). Kurtzman et al. (2008) demonstrated the polyphyletic nature of the genus Pichia from multigene sequence analysis and proposed to use the genus name Pichia only for species closely related to Pichia membranifaciens, the type species. Several studies have proposed the transfer of certain Pichia species to Ogataea (Limtong et al. 2008; Nagatsuka et al. 2008) and the genus diagnosis has been modified several times to accommodate phenotypic differences among phylogenetically related Pichia species (Péter et al. 2007, 2008; Nagatsuka et al. 2008).

Recently, a study of the Ogataea–Ambrosiozyma–Kuraishia clade resulted in the designation of three sub-clusters, even though the tree had weak branch support (Nagatsuka et al. 2008). From these results, Ogataea nitratoaversa was found to be related to a sub-cluster that included Pichia methanolica, P. trehalophila, Williopsis salicorniae, P. pilisensis, Candida piceae and C. sonorensis. The support for relatedness of that cluster to the other Ogataea species was weak. Nevertheless, due to the placement of Ogataea nitratoaversa apart from other recognised Ogataea spp., but in the “P. methanolica” clade, the authors suggested exclusion of the latter from the genus Ogataea.

Phylogenetic analysis of ribosomal small and large (D1/D2 domains) subunit genes and the ITS region performed in the current study using the maximum likelihood algorithm clearly suggests relatedness of the newly described species with a species group that includes Pichia methanolica, P. trehalophila and Williopsis salicorniae, and additionally reveals relationship between this group and C. arabinofermentans and C. ovalis (Fig. 1; Fig. S2). The clade comprised of Pichia pilisensis, Candida piceae, C. sonorensis and the recently described Ogataea nitratoaversa appeared to be the sister group to the “P. methanolica” clade.

The Kishino–Hasegawa test of constrained maximum likelihood trees was used to find out whether the novel species O. cecidiorum and the members of the “P. methanolica” clade form a monophyletic group with the other known Ogataea species (except for O. nitratoaversa) and thus could be classified in the latter genus. Therefore, the statistical estimation of phylogenetic tree topologies corresponding to the three different hypotheses of the “P. methanolica” clade placement were submitted to the Kishino–Hasegawa test: (i) when that clade was constrained with the Candida cidri, C. hungarica, Kuraishia molischiana; (ii) when it was constrained with the Ambrosiozyma clade (including Pichia angophorae); (iii) when it was constrained with the other described Ogataea species. After the statistical comparison of the tree topologies the first two constrained trees were rejected in favour the tree when “P. methanolica” clade forms a monophyletic group with the other Ogataea species (diff −ln L: 124.69 and 216.41 respectively, P < 0.001). Our results suggest that members of “P. methanolica” clade should be rather classified in the genus Ogataea, than as in any other genera. Additionally, the properties of these species do not contradict the diagnosis of the genus Ogataea emended by Péter et al. (2008) and Nagatsuka et al. (2008).

Our observations are supported by the four-gene phylogenetic study of Kurtzman & Robnett, simultaneously performed with our investigation, which revealed strong support for the separate position of Ambrosiozyma clade and brought all Ogataea species and related Pichia, Williopsis and Candida species in into a single clade (Kurtzman and Robnett 2010). Based on these results, authors suggested the transfer of Pichia methanolica, P. trehalophila, P. pilisensis and Williopsis salicorniae to the genus Ogataea (Kurtzman and Robnett 2010). Phylogenetic analysis performed in our study supports relatedness of these species (Fig. 1). Therefore, we support the novel combinations Ogataea trehalophila, O. methanolica, O. pilisensis and O. salicorniae proposed by Kurtzman and Robnett (2010) and report a novel member of this clade for which the name O. cecidiorum is proposed.