Adaptation of yeast Saccharomyces cerevisiae to grape-skin environment

Saccharomyces cerevisiae, an essential player in alcoholic fermentation during winemaking, is rarely found in intact grapes. Although grape-skin environment is unsuitable for S. cerevisiae’s stable residence, Saccharomycetaceae-family fermentative yeasts can increase population on grape berries after colonization during raisin production. Here, we addressed adaptation of S. cerevisiae to grape-skin ecosystem. The yeast-like fungus Aureobasidium pullulans, a major grape-skin resident, exhibited broad spectrum assimilation of plant-derived carbon sources, including ω-hydroxy fatty acid, arising from degradation of plant cuticles. In fact, A. pullulans encoded and secreted possible cutinase-like esterase for cuticle degradation. When intact grape berries were used as a sole carbon source, such grape-skin associated fungi increased the accessibility to fermentable sugars by degrading and assimilating the plant cell wall and cuticle compounds. Their ability seems also helpful for S. cerevisiae to obtain energy through alcoholic fermentation. Thus, degradation and utilization of grape-skin materials by resident microbiota may account for their residence on grape-skin and S. cerevisiae’s possible commensal behaviors. Conclusively, this study focused on the symbiosis between grape-skin microbiota and S. cerevisiae from the perspective of winemaking origin. Such plant–microbe symbiotic interaction may be a prerequisite for triggering spontaneous food fermentation.


Results
Interaction between grape-skin microbiota and S. cerevisiae on raisins. To investigate S. cerevisiae's behavior on grape-skin environments, we focused on raisin microbiota. Although S. cerevisiae is rarely found from sound grape berries [2][3][4] , raisins are traditionally and commonly used as a source of dough-fermenting S. cerevisiae in homemade bread making 32 . Commercially available raisins from different manufacturers were covered with water and statically incubated (Fig. 1a). In all experiments but sample #1, raisins floated, the liquid turned cloudy, and numerous small bubbles were observed within a few days. This is what is called "raisin water" or "raisin yeast". As shown in Fig. 1b, fungal internal transcribed spacer (ITS) amplicon sequence analysis indicated that common predominant fungal families in all raisin samples were (i) Aspergillaceae, composed of Aspergillus mold species, (ii) Saccharomycetaceae, including S. cerevisiae and its relative species (e.g. Zygosaccharomyces rouxii, Lachancea dasiensis), and (iii) Saccharomycodaceae, mainly Hanseniaspora uvarum, which is one of the most abundant yeast species at the beginning of natural fermentations. After two-week incubation, raisin yeast from samples #2-#5 was exclusively occupied by Saccharomycetaceae and Schizosaccharomycetaceae family fermentative yeast species. Notably, initial yeast species on raisins do not always propagate in raisin yeast, suggesting complicated fungal interaction and succession. Although sample #1 contained plenty of S. cerevisiae, raisin yeast making failed. Thus, the raisin production process is the primary step for S. cerevisiae and other fermentative yeasts to trigger colonization into grape environments.
We aseptically dried sound grape berries in an experimental oven for metagenomic analysis during the raisin production process (Fig. 2a). The berries discolored, shriveled, and eventually developed raisin-like wrinkles within 3-week fan drying process. In spite of repeated trials, we failed to detect S. cerevisiae from resulted raisins (Fig. 2b). In most cases, Sclerotiniaceae family molds (e.g., Botrytis a.k.a. "noble rot") dominated fresh berries and drastically decreased during the drying process. In the representative experiment shown in Fig. 2b, after fungal families such as Aspergillaceae, Capnodiaceae, and Pleosporaceae transiently increased the frequency, Didymellaceae, Nectriaceae, and Sporidiobolaceae dominated. Rarely detected yeast species from Saccharomycetaceae family were almost non-Saccharomyces (e.g., Candida, Pichia) and did not increase the population. In contrast, inoculation of S. cerevisiae X2180 cells (approximately 10 cells/berry) on the surface of grape-skin before drying  www.nature.com/scientificreports/ resulted in the increased abundance of Saccharomycetaceae and Saccharomycodaceae yeast families, reproducing the industrial raisin microbiota (Fig. 1b). Based on the data described above, fresh grape-skin is likely unsuitable for S. cerevisiae's stable residence, but once S. cerevisiae cells are exogenously added to the ecosystem, they can effectively adapt to the grape-skin environment. To address the possible commensal behavior of S. cerevisiae, interaction with grape-skin resident microorganisms were further examined.
Isolation of yeasts and related microorganisms inhabiting grape-skin. Among chloramphenicoltolerant, yeast-like colonies isolated from grape juice, surface-washed suspensions, or enrichment cultures, 76 clones were identified to species level by sequencing the ITS region of the nuclear rRNA gene (Supplementary  Table S1). Isolated microorganisms were classified as yeast-like fungi, basidiomycetous yeasts, or ascomycetous yeasts (Fig. 3a). The yeast-like fungi included a single ascomycetous species A. pullulans [33][34][35] , also known as "black yeast", frequently isolated from grape and wine environments in previous reports 3,4,6,8  www.nature.com/scientificreports/ and S. cerevisiae cells were cultivated in a yeast nitrogen base (YNB) minimum medium with various possible carbon sources in grape berries (Fig. 4). Both A. pullulans and S. cerevisiae vigorously grew in the presence of fermentable sugars, such as glucose and sucrose ( Fig. 4a,b). A major part of grape-skin consists of plant cell wall polymers, cellulose and pectin 17 . Although the use of carboxymethyl cellulose (CMC), a water-soluble cellulose derivative, was undetected in A. pullulans and S. cerevisiae, the main cellulose degradation product cellobiose was assimilated by A. pullulans (Fig. 4c,d). As previously reported 38 , the activity of carboxymethyl cellulase (i.e., CMCase) is deficient among some A. pullulans strains. The A. pullulans strain identified in this study may be unable to degrade cellulose by themselves but can grow by using cellobiose produced through cellulose degradation by other grape-skin fungi. The major pectic polysaccharide polygalacturonic acid and its building block galacturonic acid were assimilated by A. pullulans, but not by S. cerevisiae (Fig. 4e,f). Thus, A. pullulans likely degrades and assimilates a broad spectrum of plant cell wall-relevant materials, unmetabolized by S. cerevisiae, consistent with previous reports [19][20][21][22]34,35 . The plant cuticle, the outermost hydrophobic layer, is another major component of grape-skin 17 . Cutin in the plant cuticle is a polyester of ω-hydroxy C 16 and C 18 fatty acids and their derivatives. We discovered that A. pullulans exhibited weak but reproducible growth using ω-hydroxypalmitic acid as a sole carbon source (Fig. 4g). Since neither palmitic acid or ω-carboxypalmitic acid was assimilated (Fig. 4h,i), A. pullulans may possess a utilization system specific for ω-hydroxy fatty acids. Besides A. pullulans, three basidiomycetous yeasts, P. laurentii, S. pararoseus, and R. mucilaginosa, and two ascomycetous yeasts, M. caribbica and D. hansenii, showed significant growth in the presence of ω-hydroxypalmitic acid (Fig. 4k).
Cutinase-like esterase activity in A. pullulans. The secretion of cutin-degrading enzymes in A. pullulans, basidiomycetous yeasts, and S. cerevisiae was tested using a model polyester polycaprolactone (PCL)-plate (Fig. 5a). Known cutinases from the other species represent PCL degradation activity in previous studies [39][40][41] . The supernatant of the fully grown A. pullulans culture in YNB medium plus 2% glucose formed a clear halo, although no halo was observed using the other supernatant samples. Additionally, the supernatant of the A. pullulans culture exhibited higher esterase activity toward p-nitrophenyl butyrate (pNPB) and p-nitrophenyl palmitate (pNPP) used as substrates than the supernatant of the S. cerevisiae culture (Fig. 5b). These data suggest that the isolated A. pullulans strain from grapes secretes cutinase-like esterase to assist penetration into plant cuticles.
Previous whole-genome analysis of the A. pullulans EXF-150 strain 42 revealed nine candidate genes encoding cutinase-like enzymes, designated as ApCut1 to ApCut9 ( Supplementary Fig. S1). These gene products contained a classical α/β-hydrolase catalytic triad Ser-His-Asp and a Gly-Tyr-Ser-Gln-Gly (GYSQG) motif conserved among cutinase catalytic sites 29,35 . Additionally, two pairs of cysteine residues forming disulfide bonds, important for spatial conformation 29,43 , were also conserved in all cutinase candidates except for ApCut4, in which the amino terminus is truncated. Phylogenetic analysis indicated that ApCut1 to ApCut3 form a subgroup with yeast cutinases from Arxula adeninivorans and Cryptococcus sp. S-2 39,41 , while ApCut5 to ApCut9 form a subgroup with mold cutinases from Fusarium solani, Aspergillus oryzae, and Botrytis cinerea (Fig. 6a) [44][45][46] . Since the yeast-type cutinases were less characterized than mold cutinases so far, we decided to overexpress ApCut1 to ApCut3 to test their enzymatic activity. Whole-cell lysate of E. coli expressing recombinant ApCut1 indicated cutinase-like activity, based on the PCL-plate clearing assay and the pNPB hydrolysis assay (Fig. 6b-d). Also, the extracts of ApCut2-or ApCut3-expressing E. coli cells weakly degraded PCL ( Supplementary Fig. S2). Since the expression of ApCut2 and ApCut3 was almost undetectable in the coomassie brilliant blue (CBB)-stained gel, more attention should be paid to the protease sensitivity, expression conditions, and synonymous codon usage bias. These results revealed the activity of the yeast-type cutinase isoenzymes in A. pullulans. Which and how A. pullulans genes are expressed under plant cuticles environments are major issues to be solved. www.nature.com/scientificreports/ Alcoholic fermentation of glucose by grape-skin fungi and S. cerevisiae. To assess alcoholic fermentation performance of the grape-skin fungi and S. cerevisiae, each microorganism was statically incubated in a 5-mL liquid YNB medium, synthetic minimum medium for laboratory yeast strains, supplemented with 10% (w/v) glucose as a sole carbon source. As all species grew well in this medium (Supplementary Fig. S3; optical density of 600 nm (OD 600 ) > 1), carbon dioxide emission rate, glucose consumption, and ethanol production were quantified during the 6-d fermentation test (Fig. 7a,b). S. cerevisiae exhibited a robust peak of fermentation rate and full glucose consumption to yield approximately 5% (v/v) ethanol. In contrast, yeast-like fungus A. pul-     We mainly focused on the coculture effect of non-fermenting A. pullulans on the alcoholic fermentation ability of S. cerevisiae. Under the coculture condition of S. cerevisiae and non-fermenting species, alcoholic fermentation is solely attributed to S. cerevisiae. In the coculture experiment in 50-mL YNB medium containing 10% (w/v) glucose as a sole carbon source, the non-fermenting species A. pullulans displayed at least no significant interaction with S. cerevisiae that could be detected by carbon dioxide emission in alcoholic fermentation (Fig. 7c). The coculture of A. pullulans and S. cerevisiae may have no synergistic effect on the growth of both species (Supplementary Fig. S4). We note that the coculture exhibited a pink color, suggesting growth of A. pullulans (red) and S. cerevisiae (white) strains.
Alcoholic fermentation of intact grapes by nonfermentative, grape-skin fungi and S. cerevisiae. In 50-mL YNB minimum medium plus 50-g intact grape berries as a sole carbon source (Fig. 7d), A.
pullulans generated almost no carbon dioxide, whereas S. cerevisiae slowly progressed alcoholic fermentation, reaching the maximum carbon dioxide emission rate 4 d after inoculation. Since robust peaks of fermentation rates were typically observed within 1−2 d after inoculation in the presence of glucose (Fig. 7a,c), grape-skin may function as a physical barrier against S. cerevisiae cells to protect fermentable sugars, such as glucose, fructose, and sucrose, inside grape berries.
When A. pullulans and S. cerevisiae were co-cultured, the initial carbon dioxide emission rate increased, and a maximum fermentation rate was observed at 3 d (Fig. 7d). Considering the minor contribution of A. pullulans to alcoholic fermentation and S. cerevisiae growth, A. pullulans may specifically contribute to an increase in sugar availability for S. cerevisiae, leading to accelerated alcoholic fermentation. Coculture of S. cerevisiae with the nonfermentative basidiomycetous yeast, P. laurentii, S. pararoseus, or R. mucilaginosa, gave similar results ( Supplementary Fig. S5). A possible hypothesis accounting for these data is that S. cerevisiae cells might make use of the ability of grape-skin microbiota to access fermentable sugars in grape berries upon triggering spontaneous wine fermentation.

Discussion
Based on the experimental data, we propose that the grape-skin-resident microorganisms, including nonfermentative, yeast-like fungus A. pullulans, increase the accessibility to fermentable sugars in intact grape berries by degrading and assimilating the plant cell wall, cuticle compounds, or both. Grape-skin microbiota's high and versatile abilities to degrade and assimilate plant cell wall and/or cuticle is likely to be essential for their adaptation and proliferation on grape-skin. In contrast, S. cerevisiae cells without such abilities need the aid of grape-skin microbiota to survive in grape environments. Studies of wild populations of S. cerevisiae and its closest relative Saccharomyces paradoxus suggest that woodlands or primeval forests are natural habitats for these yeast species [47][48][49] . In this study, there was no S. cerevisiae clone isolated even from fermented juice or enriched cultures in 5% sucrose. Moreover, S. cerevisiae was unable to assimilate plant cell wall, cuticle, and their components. Since our original motivation was to search for wine yeasts responsible for spontaneous alcoholic fermentation, no bacterial screening was performed in this study. In fact, our finding supported that wine yeasts are unlikely derived from grape-skin, consistent with previous reports [2][3][4] . Altogether, nutrient-poor, intact grape surfaces may be inappropriate for wine yeasts as their stable and permanent habitats. Although Saccharomyces species might have been accidentally brought to the vineyard by wind or yeast-carrier animals 10,11 , the incidence of S. cerevisiae wine yeasts still needs to be experimentally explored. Our results provide an important clue to address how S. cerevisiae cells met and conquered grapes upon the origin of spontaneous wine fermentation.
What are the key grape-skin compounds that protect fermentable sugars in intact grape berries? Based on the assimilation tests of CMC and cellobiose (Fig. 4c,d), our isolated A. pullulans strain lacks cellulase. This is consistent with a previous report about intraspecific variations of cell wall-degrading enzymes in A. pullulans 38 . Thus, cellulose degradation may be nonessential for accelerated alcoholic fermentation of intact grapes. Among pectin-degrading enzymes, polygalacturonase is genetically encoded and expressed by grape-skin resident species and by S. cerevisiae 50,51 . Recently, we revealed the importance of pectin as an initial target for the saprophytic bacterium Bacillus subtilis to recognize the surface of dead soybeans 52 , whereas it is unlikely that pectin degradation by non-Saccharomyces microorganisms specifically accelerated alcoholic fermentation of intact grapes. As shown in Fig. 5a, the PCL-degrading activity was detected in A. pullulans, although not in the examined basidiomycetous yeast strains. Based on these data, the responsible enzymatic activity specific and common to all grape-skin residents is still unidentified. The degradation of the other plant cell wall or cuticle components should be focused on in future research. Furthermore, cooperative and synergistic degradation by whole grapeskin microbiota, including not only fungi but also bacteria, needs to be investigated. It is also noted that grapes and raisins used in this study are commercial and there is no way to know the influence of pre-and post-harvest treatments. We should start considering the impact of typical grapevine pathogenic microorganisms 53 , such as B. cinerea, with strong degradative capacity to invade into plant bodies, on the interaction with S. cerevisiae upon the origin of wine. Such complicated, highly ordered microbial interactions at the chemical, metabolic, genetic, and genomic levels will be the central issue in applied microbiology and microbial ecology. This is the first report on ω-hydroxypalmitic acid assimilation as a carbon source by A. pullulans and several yeast species. Although ω-hydroxylation of fatty acids also occurs in mammals and insects, ω-hydroxy fatty acids play a broad and vital biological role in higher plants as major components of cutin and suberin 25 www.nature.com/scientificreports/ Thus, the microbial ability to assimilate ω-hydroxy fatty acids may mainly contribute to symbiotic interactions with terrestrial plants. In the ω-oxidation process of animals and plants, known as a minor, fatty acid catabolic pathway, a hydroxy group is introduced onto the ω carbon of the medium to long-chain fatty acids 55,56 . The resultant ω-hydroxy fatty acids are oxidized to ω-carboxy fatty acids (i.e., dicarboxylic fatty acids), further degraded through the β-oxidation pathway. Grape-skin residents may probably metabolize ω-hydroxy fatty acids similarly. Based on our data, A. pullulans cells are suggested to have an ω-hydroxy fatty acid-specific transporter because they can grow using ω-hydroxypalmitic acid as a sole carbon source, not palmitic acid or ω-carboxypalmitic acid. Furthermore, this study revealed that A. pullulans may secrete cutinase to hydrolyze cutin into ω-hydroxy fatty acids and other minor components, such as glycerol. Both aspects of cutin degradation and ω-hydroxy fatty acid assimilation may characterize A. pullulans as the most abundant and persistent resident among grape-skin microbiota. Notably, the A. pullulans EXF-150 strain was first identified as a microorganism that acquired both yeast-and mold-type cutinase genes in the genome, which may be associated with the yeast-to-hyphal dimorphic transition of this species 33 . Enzymatic analysis of yeast-and mold-type cutinases in A. pullulans will reveal the significance of differences between both cutinase types. Our analysis of microbiota in raisins and raisin yeast provided important clues to the archetype of natural wine fermentation. Since dried raisins were found to contain a conciderable amount of S. cerevisiae or its relatives, a decrease in sugar concentration by mixing with water enables yeast cells to avoid high osmotic stress and induce vigorous alcoholic fermentation. As a result, ethanol-tolerant microorganisms, such as S. cerevisiae, govern the population in raisin water (Fig. 1). Although S. cerevisiae was not found on aseptically dried grape berries, we here presented the evidence that inoculated S. cerevisiae can grow on the surface of grape berries (Fig. 2). Thus, colonist S. cerevisiae may adapt to the grape-skin environments and conquer other microorganisms in response to the optimal sugar concentration for yeast growth and alcoholic fermentation. As grape-skin is unlikely the original habitat for S. cerevisiae, the initial adaptation step may be the key to understand the origin of wine. Now we are further addressing how S. cerevisiae copes and interacts with grape-skin habitants to survive and establish the growth on the grape environment, based on metabolite dynamics.
Conclusively, this study focused on the symbiosis between grape-skin microbiota and S. cerevisiae from the perspective of winemaking origin (Fig. 8). Oligotrophic microorganisms, such as A. pullulans, have developed versatile abilities to use plant cell wall polysaccharides and plant cuticular lipids as nutrient sources to establish their ecological niche. Especially, degradation and assimilation of the plant cuticle, the outermost layer interacting with the environment, may be a prerequisite for oligotrophic resident microorganisms to trigger colonization and adaptation. In contrast, eutrophic yeasts, including S. cerevisiae, yield energy through the alcoholic fermentation of sugars in grape berries with the aid of oligotrophic microorganisms. Such tripartite interaction between grape berries, oligotrophic residents, and eutrophic yeasts determines the grape-skin microbiome dynamics and spontaneous wine fermentation. Moreover, since grape bacterial microbiota 7-9 remain elusive in this study, the individual interactions between bacteria and grape berries, between bacteria and oligotrophic eukaryotic residents, or between bacteria and eutrophic yeasts should also be investigated in future. Thus, studying the origins and microbiota ecological succession in fermented foods will help elucidate the key principles governing plant-microbial ecosystems' emergence and development.

ITS amplicon sequence analysis of raisin microbiota.
To prepare raisin yeast, California Green Seedless raisins, non-oil coated, from different manufacturers were purchased from Tomizawa Shouten (Japan), Imagawa Seika (Japan), Nishiuchi Kagetsudo (Japan), and Fujimoto Shouten (Japan). Raisin microbiota were prepared from surface-washed suspensions. 100-g raisins and 300-mL sterilized water were mixed and incubated at 30 °C for 14 d. For the raisin making experiment, approximately 300-g (per batch) fresh grape berries of Green Seedless were dried in an forced air flow oven (WFO-520W, Eyela, Japan) at 37 °C for 21 d. S. cerevisiae X2180 cells, obtained from the American Type Culture Collection (USA), were fully grown in the rich medium, www.nature.com/scientificreports/ washed by sterilized water, and inoculated on the surface of individual grape berries at the cell density of 10 cells/berry. X2180 was also used for other experiments as a representative of diploid S. cerevisiae strains (Sc) in this study. Both DNA extraction and ITS amplicon sequence analysis were performed by Bioengineering Lab. Co., Ltd. (Sagamihara, Japan) based on a previously reported method 57 . In summary, the ITS1 region was amplified using the ITS1F_KYO1/ITS2-KYO2 primer sets. Sequencing was conducted using a paired-end, 2 × 300 bp cycle run on a MiSeq sequencing system (Illumina) and MiSeq Reagent Kit version 3 chemistry. After quality filtering and chimera check of sequencing reads, the operational taxonomic units (OTU) were predicted using QIIME 2 58 . The OTUs that accounted for more than 1% of the total sequence number were classified, whereas taxa with abundance < 1% were summarized as "others".
Isolation and identification of grape-skin microorganisms. To isolate yeasts or yeast-like fungi, (i) juice, (ii) surface-washed suspensions, or (iii) enrichment cultures were obtained from commercially available wine or table grape varieties belonging to Vitis vinifera species or Vitis interspecific hybrids (see Supplementary  Table S1). The use of plants in the present study complies with the IUCN Policy Statement on Research Involving Species at Risk of Extinction and the Convention on the Trade in Endangered Species of Wild Fauna and Flora. Grape juice was obtained from freshly pressed grape berries using a food-grade juicer (Panasonic, Japan) and was used either immediately or after being incubated at 30 °C for 3 d. Surface-washed suspensions were obtained by vigorously shaking the flask containing approximately 50-g grape berries and 25-mL sterilized water at 30 °C for 15 min 59 . Note that the washing did not destroy the sound grape berries used. Enrichment cultures were obtained by statically incubating the flask containing approximately 10-g grape berries and 40-mL sterilized water or 5% (w/v) sucrose at 30 °C for 3 d. Each sample was spread on a nutrient-rich,YPD (1% yeast extract, 2% peptone, and 2% glucose) medium plate with 0.1% chloramphenicol to inhibit bacterial growth, and was incubated at 30 °C. Single colonies representing yeast-like colony morphology were isolated by repeatedly streaking on YPD medium plates with 0.1% chloramphenicol.
The isolated clones were identified by DNA sequencing of the rRNA gene ITS region 60 , using the ITS_1F (5′-GTA ACA AGG TYT CCGT-3′) and ITS_1R (5′-CGT TCT TCA TCG ATG-3′) primer pair and genomic DNA as PCR templates. Genomic DNA was extracted and purified using the Dr. GenTLE (from Yeast) High Recovery (Takara Bio, Japan). Each 25 µL PCR reaction consisted of 1 µL of approximately 100 ng/µL template DNA, 12.5 µL of 2× PCR buffer for KOD FX Neo, 0.5 µL of KOD FX Neo (Toyobo, Japan), 10 µM forward primer, 10 µM reverse primer, and 10 mM dNTPs. The PCR conditions were: 94 °C for 2 min for initial denaturing, then 98 °C for 10 s, 55 °C for 30 s, and 68 °C for 30 s for 30 cycles, then 16 °C. The products were purified using Gel/PCR DNA isolation system (Viogene, USA). DNA sequencing was performed by Macrogen Japan (Japan), using both the ITS_1F and ITS_1R primers. Species were identified by searching databases using the BLAST sequence analysis tool (http:// www. ncbi. nlm. nih. gov/ BLAST/). Since it was difficult to get a clean reading of the total length of the ITS for all isolated strains, these analyses should be considered at the preliminary level. In the nine representative grape-skin resident species (Ap, Pl, Sp, Rm, Hu, Td, Mc, Dh, and Pt) used in the later experiments, the sequence of the D1/D2 region of 26S rDNA was determined by TechnoSuruga Laboratory (Japan).
Carbon assimilation test. Cells were aerobically precultured at 30 °C for 2 d in YNB minimum medium containing 2% (w/v) glucose as a carbon source, harvested, washed by sterilized water, and inoculated into 5 mL of YNB medium containing 0.5% (w/v) of carbon sources as below: glucose, sucrose, CMC, cellobiose, polygalacturonic acid, galacturonic acid, palmitic acid, 16-hydroxyhexadecanoic acid (i.e., ω-hydroxypalmitic acid), or heptadecanedioic acid (i.e., ω-carboxypalmitic acid). Test tubes were aerobically shaken at 130 strokes per minute at 30 °C for 3 d. Upon inoculation, initial OD 600 was adjusted to 0.1. For assimilation tests of palmitic acid, ω-hydroxypalmitic acid, or ω-carboxypalmitic acid, 0.05% (w/v) Tween 40 was added to the medium. In advance of measurement of OD 600 , insoluble fatty acids were removed by washing pellets with hexane three times, and cells were dissolved in sterilized water 61 . All growth tests were repeated at least three times. Data were statistically analyzed with the two-tailed Student's t-test.
Phylogenetic analysis of yeast and mold cutinases. Amino acid sequences of nine putative cutinase gene products in the A. pullulans EXF-150 strain and well-studied yeast and mold cutinases [Cut2 from Arxula adeninivorans (AaCut2) 41 , Cle1 from Cryptococcus sp. strain S-2 (CS2Cle1) 39 , Cut1 from Fusarium solani (FsCut1) 40 , CutL from Aspergillus oryzae (AoCutL) 45 , and CutA from Botrytis cinerea (BcCutA) 46 ] were obtained from UniPlotKB. Multiple sequence alignment was conducted using Clustal Omega program, and the phylogenetic tree was constructed using Molecular Evolutionary Genetics Analysis v. 10 Expression of recombinant A. pullulans cutinase. The chemically synthesized ApCUT1 gene (Eurofins, Luxembourg) was cloned into the BamHI-EcoRI site of the pET-21b(+) vector (Merck Millipore, USA) to express recombinant ApCut1p tagged with the T7 epitope at the amino terminus and the 6× His epitope at the carboxy terminus. The resultant pET-21b(+)-ApCUT1 plasmid was introduced into Escherichia coli BL21(DE3) pLysS cells (Novagen, Germany). A transformant was inoculated into 100-mL Luria-Bertani medium (0.5% yeast extract, 1% tryptone, and 1% sodium chloride) with 100-μg/mL ampicillin and 34-μg/mL chloramphenicol, and was cultured at 37 °C to an OD 600 of 0.6. Isopropyl thio-β-d-galactoside (IPTG) was added to the culture www.nature.com/scientificreports/ to a final concentration of 1 mM, and cells were further cultured at 37 °C for 4 h. Whole-cell extracts were prepared from cell pellets suspended in the xTractor buffer (TakaraBio, Japan), followed by the addition of DNase I. After protein separation by SDS-PAGE, the proteins were detected by CBB staining or using His-Detect In-Gel Stain (Nacalai Tesque, Japan). Full-length gels are displayed in Supplementary Fig. S6.
Enzymatic assay. To assay the potential cutinase activity from the culture supernatants of fully grown A. pullulans or yeasts, PCL was used as a model polyester substrate. Turbid agar plates containing 0.05% PCL (Fujifilm Wako Pure Chemical, Japan) were prepared as previously described 40,41 . The assay for determining esterase activity was conducted according to previous reports 41, 62 , using pNPB and pNPP (Sigma-Aldrich, USA) as substrates. All assays were repeated at least three times.
Fermentation test. Cells were aerobically precultured at 30 °C for 2 d in 0.67% YNB minimum medium containing 2% (w/v) glucose as a carbon source and harvested. For the alcoholic fermentation of glucose, cells were inoculated into a YNB medium containing 10% (w/v) glucose at a final OD 600 of 0.1 and were then further incubated at 30 °C without shaking. For the alcoholic fermentation of intact grapes, cells were inoculated into 50-mL YNB medium at a final OD 600 of 0.1, mixed with approximately 50-g commercially available intact grape berries of Green Seedless, and were then further incubated at 30 °C without shaking. Fermentation was continuously monitored by measuring the weight loss of evolved carbon dioxide for 5-mL test tube-scale tests or using a Fermograph II apparatus (Atto, Japan) for 50-mL-scale coculture tests. Glucose and ethanol concentrations were determined using the LabAssay glucose kit (Fujifilm Wako Pure Chemical, Japan) and the ethanol assay F-kit (Roche, Switzerland), respectively. All fermentation tests were repeated three times. Data were statistically analyzed with the two-tailed Student's t-test.

Data availability
All sequences determined in the ITS amplicon sequence analysis using raisin yeast and dried grape berry samples have been deposited in the DNA Data Bank of Japan (DDBJ) under the accession numbers DRA015192 and DRA015183, respectively. All sequences of the D1/D2 region of 26S rDNA in the representative grape-skin resident species have been deposited in the DDBJ under the accession numbers, LC765834-LC765842.