Fungal Diversity of Deteriorated Sparkling Wine and Cork Stoppers in Catalonia, Spain

Filamentous fungi are rarely reported as responsible for spoiling wine. Cork taint was detected in sparkling wine; therefore, we investigated fungal contamination as a possible cause of organoleptic alteration. Spoiled wine was filtered and membranes were plated onto potato dextrose agar (PDA). The cork stoppers used for sealing bottles were cut and also plated onto PDA. Fungal strains were phenotypically characterized and molecularly identified by sequencing of a fragment of the 28S nrRNA gene (LSU) and (occasionally) by other additional molecular markers. Twenty-seven strains were isolated and sixteen species were identified, all of them belonging to the phylum Ascomycota. The fungi isolated from wine were three species of Aspergillus section Nidulantes, a species of Penicillium section Exicaulis and Beauveria bassiana. Candida patagonica was isolated from both sort of samples, and the fungi isolated from cork stoppers were Altenaria alternata and Cladosporium cladosporioides. Surprisingly, most of the taxa recovered from the cork stoppers and/or wine were new to the science: a new genus (Dactylodendron) and seven new species belonging to the genera Cladophialophora, Dactylodendron, Kirschsteiniothelia, Rasamsonia, and Talaromyces. Future studies could let us know if these fungi would be able to produce compounds responsible for cork taint.


Introduction
Sparkling wine is one of the most economically important wine varieties in southern Europe. It is produced by the "champenoise" method, which consists of two steps: a primary alcoholic fermentation, in which the grape must is transformed to the wine base (cuvée); and a second alcoholic fermentation after the addition of sucrose, selected yeasts, and bentonite to the base wine, which is then bottled, closed with a metal cap or a cork stopper, and allowed to age in cellars for a longer period of time (at least 12 months for French champagne and 9 months in the case of the Spanish-mostly Catalonian-"cava") [1]. During fermentation, a certain diversity of environmental microorganisms, mainly bacteria and fungi, can produce organoleptic alterations that render the wine undrinkable. Some of these fungi can be present on the cork stoppers and/or be acquired by exposure of the must to bio-aerosols, perhaps because of poor environmental microbiological control at the cellar. Cork taint is a musty or mouldy off-odor in wine often caused by the presence of 2,4,6-trichloroanisole (2,4,6-TCA) among other chemical compounds [2], and between 0.5 and 7% of wines can be affected by cork taint. It is estimated that cost of cork-related wine spoilage can exceed several billions of dollars per year [3,4]. The metabolic effect of fungi living on cork in the production of 2,4,6-TCA has been described [5]. Among several fungi recovered from agglomerate cork stoppers, Acremonium strictum, Chrysonilia sitophila, Cladosporium oxysporum, Fusarium oxysporum, Paecilomyces viridis, Penicillium chrysogenum, Trichoderma longibrachiatum, Trichoderma viride, and Verticillium psalliotae have displayed such an effect [5].
We recently had the opportunity to study the fungal biota associated to wine deterioration when a local winery located in Sant Sadurní d'Anoia (Barcelona province, Spain) detected cork taint in some bottles of sparkling wine during an inspection of the cellars at its historic vineyards. We conducted a study to detect, isolate, and identify the fungi involved in the production of this sort of flavor alteration.

Fungal Isolation
Samples of sparkling wine and cork stoppers were obtained from a cellar in Sant Sadurní d'Anoia, Barcelona province, Spain. Approximately, 500 bottles of sparkling wine from five different batches were opened in situ to obtain a representative number of negative controls (without organoleptic alteration) and a panel of four experts detected taste defects in any samples that had a musty or mouldy off-odor and/or flavor. A total of 54 bottles of sparkling wine sealed by cork stoppers (15 negative control and 39 with deteriorated wines) were selected and processed. A sample of 100 mL of sparkling wine was filtered through a filter membrane of 0.45 µm diameter (Millipore SA, Molsheim, France). After filtering, the membrane was plated onto a 90 mm diameter Petri dish containing potato dextrose agar (PDA; Pronadisa, Madrid, Spain) plus 50 mg/L L-chloramphenicol. The Petri dishes were incubated at 25 • C for a time period ranging from 4 weeks to 2 months in darkness, and examined under a stereomicroscope to observe any production of mold colonies with reproductive structures. If bacteria and/or yeasts develop on the culture medium, these could be recognized by mucous to buttery colonies of reduced diameter, and by the absence of hyphae; also, slide mountings on water seen under bright field microscope allows detection of the bacterial/yeast cells. The cork stoppers were cut into small pieces using a sterile disposable scalpel and plated onto 90 mm diameter Petri dishes containing PDA, which were incubated in the same way as described above. For both sorts of samples, fungal structures from selected colonies (representative of all the morphological varieties) were transferred to 50-mm diameter Petri dishes containing PDA using a sterile insulin-type needle and incubated in the same conditions to obtain pure cultures.

Phenotypic Characterization of the Fungal Strains
For the isolates of Rasamsonia and Talaromyces, suspensions of conidia were prepared in a semi-solid agar (0.2% agar, 0.05% Tween 80) [30] and inoculated in three equidistant points onto 2% malt extract agar (MEA; Difco Inc., Detroit, USA), oatmeal agar (OA) [30], Czapek yeast extract agar (CYA) [31], yeast extract sucrose agar (YES) [32], creatine sucrose agar (CREA) [32], dichloran 18% glycerol agar (DG18) [33], and cork (cut in slices by a scalpel, placed into appropriate containers, and sterilized three times in alternative days at 121 • C during 15 minutes) onto tap water agar (TWA; 1.5% agar in tap water) into disposable Petri dishes of 90 mm diameter, and incubated at 25 • C in darkness after 14 days. Cultures on CYA were also incubated at different temperatures (5,15,25,30,35, and 37 • C) to determine the minimum, optimal, and maximum temperatures of growth. The rest of the isolates were cultured and studied onto MEA, OA, PDA, and TWA with pieces of sterile cork, incubated at 25 • C in darkness after 14 days. The determination of the cardinal temperatures of growth of these strains was determined on PDA. Color notations in parentheses are from Kornerup and Wanscher [34]. The characterization and measurements of fungal structures were performed in water and 60% lactic acid from slide cultures by using the culture media cited before. Photographs were taken by a Zeiss Axio Imager M1 light microscope (Zeiss, Oberkochen, Germany) with a DeltaPix Infinity X digital camera, using Nomarski differential interference contrast. The samples for scanning electron microscopy (SEM) were processed according to Figueras and Guarro [35], and SEM micrographs were taken at 15 keV with a Jeol JSM 840 microscope. The taxonomic descriptions and names of the fungal novelties were introduced into MycoBank (www.mycobank.org) [36].

Preliminary Identification and Phylogenetic Analysis
Preliminary molecular identification was carried out by comparing of the LSU sequences of our isolates with those of the type or reliable GenBank reference strains using the Basic Local Alignment Search Tool (BLAST; https://blast.ncbi.nlm.nih.gov/Blast.cgi). A maximum level of identity (MLI) of ≥98% was considered to allow for species-level identification. MLI values < 98% provided identification only at genus level. For identification of species of Aspergillus and Penicillium, sequences of a fragment of BenA gene were used. To determine the phylogenetic placement of all our isolates, an LSU tree was built. Additionally, three trees with a combined data set were built to distinguish among the species of Talaromyces section Trachyspermi (by using the combined dataset ITS-BenA-CaM-rpb2), the species of Rasamsonia (ITS-BenA-CaM), and the species of Cladophialophora (ITS-LSU-BenA). Candida bituminiphila and Candida patagonica for LSU tree, Talaromyces rademirici and Talaromyces dendriticus for Talaromyces section Trachyspermi tree, Talaromyces flavus and Trichocoma paradoxa for Rasamsonia spp. tree, and Exophiala oligosperma and Exophiala exophialae for Cladophialophora spp. tree were used as out-groups. For sequence alignment and to perform the maximum-likelihood (ML) and Bayesian-inference (BI) phylogenetic analyses, we followed the methodology described by Valenzuela-Lopez et al. [44]. The final matrices used for phylogenetic analysis were deposited in TreeBASE (www.treebase.org; accession number: S23148).

Growing at Different Ethanol Concentrations
Strains from cork stoppers and sparkling wine were grown on test tubes with 5 mL of 2% malt extract in tap water supplemented with different amounts of ethyl alcohol to reach 5, 10, 15, and 20% v/v final concentration. The tubes were closed with plastic caps, hermetically sealed by parafilm ® , and incubated at 15 • C for up to 13 months in darkness without agitation, trying to simulate the method employed for resting/aging of wine. The tubes were examined every month for fungal growth. If growth was absent, 0.1 mL of the broth was plated onto PDA, and incubated during 2 weeks in darkness at 25 • C to confirm absence of fungal growth.

Fungal Diversity of Cork Stoppers and Sparkling Wine Samples
None of the negative controls of sparkling wine developed fungal colonies. On the other hand, 24 out of 39 odor/flavor altered samples developed bacterial, yeasts, and/or mold colonies. All of the cork stopper samples developed fungal colonies. A total of 27 ascomycetes, representing all the morphological variability of the fungal colonies produced, were isolated from cork stoppers and from sparkling wine. Five of them were identified as Talaromyces spp., four as Kirschsteiniothelia spp., two as Rasamsonia spp. and one as Cladophialophora sp., three other strains belonged to an unknown arthrosporate fungus. Several Aspergillus spp. were recovered from sparkling wine: i.e., Aspergillus aureolatus, Aspergillus jensenii, and Aspergillus puulaauensis (all of them belonging to the section Nidulantes). Penicillium corylophilum (section Exilicaulis) and several other fungi were identified from both sparkling wine and cork stoppers. Alternaria alternata and Cladosporium cladosporioides were identified on cork stoppers, but Beauveria bassiana and Candida patagonica were only found in sparkling wine.

Molecular Phylogeny
The first phylogenetic study included 64 LSU sequences, with a total of 517 characters including gaps, from which 273 were parsimony informative. The ML analysis was congruent with the BI analysis, both displaying a similar topology. In the LSU tree, our fungal isolates were distributed across two main clades (Figure 1), the first (100% BS/1 PP), corresponding to the filamentous Ascomycota, included 24 of our isolates, and the second (100% BS/1 PP), corresponding to the class Saccharomycetales (true yeasts), included the other isolates (three). The first main clade divided into six subclades: A (unsupported, including 13 isolates), corresponding to the order Eurotiales; B (100% BS/1 PP, three isolates), representing the family Eremascaceae (of the order Onygenales); C (100% BS/1 PP, one isolate), grouped the family Herpotrichiellaceae (order Chaetothyriales); D (unsupported, five isolates), which included the family Kirschteiniotheliaceae (sister clade D1; 100% BS/1 PP) and the family Pleosporaceae (sister clade D2; 100% BS/1 PP) (both pertaining to the order Pleosporales); E (100% BS/1 PP, one isolate), with the family Cladosporiaceae (order Capnodiales); and F (100% BS/1 PP, one isolate), with the family Cordycipitaceae (order Hypocreales). Subclade A has three well-supported sister clades, representing the genera Rasamsonia and Talaromyces (sister clade A1; 90% BS/0.99 PP), Penicillium (sister clade A2; 100% BS/1 PP), and Aspergillus (sister clade A3; 88% BS/0.99 PP). In this context, seven of our isolate formed three well-supported branches within the sister clade A1: two within the genus Talaromyces and the third near to the species of Rasamsonia, but not closely related to any of the known species. Within the sister clade A2, two isolates grouped with Penicillium corylophilum. The sister clade A3 has two isolates placed together with Aspergillus aureolatus and another two together with Aspergillus puulaauensis and Aspergillus jensenii. Subclade B includes three isolates related to Arthrographis pinicola and Eremascus albus (Onygenales). Subclade C groups different species of Cladophialophora, the isolate FMR 16667 being phylogenetically closely related to the type strain of Cladophialophora mycetomatis. Sister clade D1 (Kirschsteiniothelia spp. and Dendryphiopsis spp.) includes four of our isolates, grouped within two fully supported branches. Sister clade D2 includes FMR 15666 and Alternaria alternata. Subclade E includes FMR 15660, Cladosporium silenes, C. cladosporioides, and C. grevilleae. Subclade F includes FMR 16669, Beauveria brongniartii, and B. bassiana. Finally, in the second main clade (Saccharomycetales), three of our isolates and Candida patagonica were placed into a well-supported sister branch (100% BS/1 PP). Three additional phylogenies allowed the taxonomy of Talaromyces, Rasamsonia, and Cladophialophora to be resolved. The first (ITS, BenA, CaM, and rpb2) clarified the relationships among the species of Talaromyces section Trachyspermi, which included five of our isolates ( Figure 2). The final concatenated dataset was obtained using both ML and Bayesian analyses. It contained 29 taxa with a total of 2270 characters including gaps (515 of them for ITS, 376 for BenA, 527 for CaM, and 852 for rpb2), 728 of which were parsimony informative (128 of them for ITS, 145 for BenA, 212 for CaM, and 243 for rpb2). The datasets did not conflict with the tree topologies for the 70% reciprocal bootstrap trees, which allowed the four genes to be combined for the multi-locus analysis. The support values were only slightly different between these two analyses. Within the main clade, Three additional phylogenies allowed the taxonomy of Talaromyces, Rasamsonia, and Cladophialophora to be resolved. The first (ITS, BenA, CaM, and rpb2) clarified the relationships among the species of Talaromyces section Trachyspermi, which included five of our isolates ( Figure 2). The final concatenated dataset was obtained using both ML and Bayesian analyses. It contained 29 taxa with a total of 2270 characters including gaps (515 of them for ITS, 376 for BenA, 527 for CaM, and 852 for rpb2), 728 of which were parsimony informative (128 of them for ITS, 145 for BenA, 212 for CaM, and 243 for rpb2). The datasets did not conflict with the tree topologies for the 70% reciprocal bootstrap trees, which allowed the four genes to be combined for the multi-locus analysis. The support values were only slightly different between these two analyses. Within the main clade, corresponding to Talaromyces section Trachyspermi (100% BS/1 PP), three of our isolates were placed in a distinct branch (100% BS/1 PP), related with T. affinitatimellis and T. basipetosporus; the other two isolates were located within another branch (100% BS/1 PP) of a well-supported terminal clade (100% BS/0.99 PP), which also included Talaromyces brasiliensis.
The second additional phylogenetic analysis was performed (ITS, BenA, and CaM) to resolve the taxonomical placement of two of our isolates between the genera Talaromyces and Rasamsonia ( Figure  3). The final concatenated dataset contained 15 sequences with a total of 1657 characters including gaps (707 of them for ITS, 392 for BenA, and 558 for CaM), 424 of which were parsimony informative (120 of them for ITS, 115 for BenA, and 189 for CaM). The ML analysis showed a similar topology and was congruent with the Bayesian analysis. The phylogenetic tree distinguished a main clade corresponding to the genus Rasamsonia (92% BS / 1 PP), which was divided in three subclades; our  Table 1.
The second additional phylogenetic analysis was performed (ITS, BenA, and CaM) to resolve the taxonomical placement of two of our isolates between the genera Talaromyces and Rasamsonia ( Figure 3). The final concatenated dataset contained 15 sequences with a total of 1657 characters including gaps (707 of them for ITS, 392 for BenA, and 558 for CaM), 424 of which were parsimony informative (120 of them for ITS, 115 for BenA, and 189 for CaM). The ML analysis showed a similar topology and was Microorganisms 2020, 8, 12 9 of 29 congruent with the Bayesian analysis. The phylogenetic tree distinguished a main clade corresponding to the genus Rasamsonia (92% BS/1 PP), which was divided in three subclades; our two isolates were located within one of them (83% BS/-PP) in a well-supported branch related to R. cylindrospora, R. brevistipitata, R. columbiensis, and R. pulvericola.
Microorganisms 2020, 8, 12 9 of 30 two isolates were located within one of them (83% BS / -PP) in a well-supported branch related to R. cylindrospora, R. brevistipitata, R. columbiensis, and R. pulvericola.  Table 1. ITS-LSU-BenA analysis included sequences from 23 taxa of Cladophialophora, with a total of 1554 characters including gaps (615 of them for ITS, 562 for LSU, and 377 for BenA), 377 of which were parsimony informative (209 of them for ITS, 55 for LSU, and 113 for BenA). The topologies of both ML and Bayesian analyses showed similar topologies and so were congruent. In the phylogenetic tree ( Figure 4), a main clade corresponding to Cladophialophora spp. (100% BS / 1 PP) was obtained. Within this clade, a terminal branch (100% BS / 1 PP) included Cladophialophora mycetomatis and FMR 16667.  Table 1. ITS-LSU-BenA analysis included sequences from 23 taxa of Cladophialophora, with a total of 1554 characters including gaps (615 of them for ITS, 562 for LSU, and 377 for BenA), 377 of which were parsimony informative (209 of them for ITS, 55 for LSU, and 113 for BenA). The topologies of both ML and Bayesian analyses showed similar topologies and so were congruent. In the phylogenetic tree ( Figure 4), a main clade corresponding to Cladophialophora spp. (100% BS/1 PP) was obtained. Within this clade, a terminal branch (100% BS/1 PP) included Cladophialophora mycetomatis and FMR 16667.  Table 1.

Alcohol Tolerance
All the isolates tested displayed good to excellent growth at 5% v/v ethanol, but failed to grow at higher concentrations of alcohol.

Subclade A: Eurotiales
Because our strains FMR 16662, FMR 16663, and FMR 16667 form a separate branch into a terminal clade including T. basipetosporus and T. affinitatimellis, and strains FMR 15656 and FMR 15664 form another independent branch within a terminal clade that also includes T. brasiliensis (Figure 2), and because all of them display enough phenotypic and phylogenetic differences with respect to the other species of Talaromyces section Trachispermi to be considered two new species, we propose the erection of Talaromyces speluncarum and Talaromyces subericola as follows: Talaromyces   Table 1.

Alcohol Tolerance
All the isolates tested displayed good to excellent growth at 5% v/v ethanol, but failed to grow at higher concentrations of alcohol.

Subclade A: Eurotiales
Because our strains FMR 16662, FMR 16663, and FMR 16667 form a separate branch into a terminal clade including T. basipetosporus and T. affinitatimellis, and strains FMR 15656 and FMR 15664 form another independent branch within a terminal clade that also includes T. brasiliensis (Figure 2), and because all of them display enough phenotypic and phylogenetic differences with respect to the other species of Talaromyces section Trachispermi to be considered two new species, we propose the erection of Talaromyces speluncarum and Talaromyces subericola as follows: Talaromyces Etymology: From Latin suber, cork, because of the origin of the fungus. Diagnosis: Talaromyces subericola differs from T. brasiliensis [45] in faster growing rates of the colonies on all culture media tested, and by the production of smooth-walled to verruculose conidia (coarsely verrucose in T. brasiliensis).
Diagnosis: Differing notably from other species of the genus [46][47][48][49][50] by the absence of growth on CYA at 30 °C (after one week incubation, greater than 5 mm diameter in the other species), and by the production of globose conidia (ellipsoidal, ovoid to cylindrical in the rest of the species), with the exception of R. pulvericola. However, R. frigidotolerans can be easily differentiated from R. pulvericola by its production of smooth-walled stipes and branches (verrucose in R. pulvericola), and because the conidia are connected by disjunctors (absent in the rest of the species of the genus).

Subclade B: Onygenales
Because Arthrographis pinicola is phylogenetically placed far from the type species of the genus Arthrographis (Arthrographis kalrae), which is located within the family Eremomycetaceae (class Dothideomycetes), and because of the morphological differences with Eremascus albus, we introduce the new genus Dactylodendron into the family Eremascaceae (order Onygenales, class Eurotiomycetes) ( Diagnosis: Recognized by its hyaline, hyphae-like, successively branched conidiophores, or short-stalked conidiophores ending in a verticilate arrangement of fertile branches. In both cases, fertile branches produce hyaline, cylindrical, or cuboid arthroconidia. Etymology: From Latin frigus-, cold, and -tolerans, tolerant, in reference to its ability to grow at relatively low temperatures. Diagnosis: Differing notably from other species of the genus [46][47][48][49][50] by the absence of growth on CYA at 30 • C (after one week incubation, greater than 5 mm diameter in the other species), and by the production of globose conidia (ellipsoidal, ovoid to cylindrical in the rest of the species), with the exception of R. pulvericola. However, R. frigidotolerans can be easily differentiated from R. pulvericola by its production of smooth-walled stipes and branches (verrucose in R. pulvericola), and because the conidia are connected by disjunctors (absent in the rest of the species of the genus).

Subclade B: Onygenales
Because Arthrographis pinicola is phylogenetically placed far from the type species of the genus Arthrographis (Arthrographis kalrae), which is located within the family Eremomycetaceae (class Dothideomycetes), and because of the morphological differences with Eremascus albus, we introduce the new genus Dactylodendron into the family Eremascaceae (order Onygenales, class Eurotiomycetes) (Figure 1), and design Dactylodendron pinicola (formerly Arthrographis pinicola) as the type species of the genus. Description: Colonies: slow-growing at room temperature, always with shades of yellow. Conidiophores semi-macronematous, hyphae-like, single or grouped in discrete dome-shaped or floccose conidiomata, erect, successively branched, or short-stalked, ending in a verticilate arrangement of fertile branches, fertile branches eventually producing arthroconidia. Arthroconidia: hyaline, smooth-walled, usually truncate at both ends, cylindrical or cuboid, produced by transverse septation in basipetal order, separated very late by schizolytic secession from the conidiogenous branches, without disjunctors or separating cells. Chlamydospores: occasionally seen. Sexual morphology: not observed. Description: Hyphae: septate and hyaline, (0.5-) 0.8-2.5 µm wide, bearing narrow conidiophores which branch repeatedly to form floccose conidiomata. The fertile branches are initially sparsely septate and of uniformly narrow diameter, but, as arthroconidial development begins, the apical region broadens and septation occurs in basipetal sequence to form many small cells. Arthroconidia secede by schizolysis, often remaining connected in chains of 3 to 4, which then undergo further schizolysis. There are no disjunctors or separating cells. Mature arthroconidia are smooth, hyaline, tan in mass, cylindrical, but often broader than long, 1.5-4.0 × 1.5-2.5 µm wide. Teleomorph was not observed. No yeast stage was observed.
Notes: The habitat reported for such fungus is wood of Pinus contorta var. latifolia, especially in galleries and adult beetles of Ips latidens, and from galleries of Dendroctonus ponderosae in Alberta, Canada.
Other Notes: Despite S. nodosa remains as incertae sedis, a Blast search using the ITS sequence available at the GenBank/EMBL databases (data not shown) has placed this fungus phylogenetically close to members of the order Helotiales (class Leotiomycetes), a taxon phylogenetically far from the order Onygenales (class Eurotiomycetes), where the genus Dactylodendron is located. sporulation sparse; reverse pale yellow (4A3), diffusible pigment absent. Colonies on tap water agar (TWA) with sterile cork: reddish yellow (4A6), exudate absent, sporulation abundant; reverse orange (6B7), diffusible pigment absent.
Notes: Despite S. nodosa remains as incertae sedis, a Blast search using the ITS sequence available at the GenBank/EMBL databases (data not shown) has placed this fungus phylogenetically close to members of the order Helotiales (class Leotiomycetes), a taxon phylogenetically far from the order Onygenales (class Eurotiomycetes), where the genus Dactylodendron is located. Etymology: From Latin pluri-, many, and -septatum, septate, due to the presence of many septa along the fertile branches.
Diagnosis: Characterized by the production of successively branched conidiophores, whose verticillate arrangement of fertile branches develop long chains of disarticulating, prismatic arthroconidia. These conidiophores are similar to those of D. pinicola, but D. pluriseptatum never produce conidiomata, which is seen in D. pinicola. Chlamydospores: hyaline, one-celled, smooth-and thick-walled, irregularly globose, 5-7 µm diameter, arising laterally on the vegetative hyphae. Sexual morphology: not observed.

Subclade C: Chaetothyriales
Because our strain FMR 16667 was placed within subclade C (Figure 1) corresponding to the species of the genus Cladophialophora, in a terminal branch (Figures 1 and 4) together with Cladophialophora mycetomatis, and because FMR 16667 displays enough phenotypic and phylogenetic differences from the latter and from other species of the genus, we propose Cladophialophora recurvata as a new species, described as follows.

Discussion
Candida patagonica was the only yeast retrieved in our study. This fungus has previously been reported from fermentation vats and oak barrels in the cellars of North Patagonia, Argentina [52]. Despite the ascomycetous yeasts such as Dekkera bruxellensis, Hanseniaspora uvarum, Issatchenkia orientalis, Metschnikowia pulcherrima, and some species of the genera Candida and Zygosaccharomyces [53,54] having been involved in the deterioration of wines, C. patagonica has not been reported as a spoilage organism for these sorts of alcoholic beverages.
The isolates belonging to the order Eurotiales showed the broadest fungal diversity, being distributed among the genera Aspergillus, Penicillium, Rasamsonia and Talaromyces. This latter genus was the most frequently recovered from both sorts of substrata, and two of them, particularly from sparkling wine, are new species, i.e., Talaromyces speluncarum, characterized by mostly biverticillate conidiophores and brown, spinose to verrucose, globose conidia, and Talaromyces subericola, which grows faster than T. speluncarum and produces smooth-walled conidia (spinose to verrucose in T. speluncarum). Aspergillus, another common genus in sparkling wine, was represented by A. aureolatus, A. jensenii, and A. puulaauensis, all of them pertaining to the section Nidulantes [55]. Aspergillus aureolatus [56] was originally isolated from air in Serbia, A. jensenii [57] from soil in the USA, and A. puulaauensis [57] from dead hardwood in the Hawaiian archipelago. There have not been any reports of these three species found in wines. Penicillium corylophilum [58] was isolated from sparkling wine samples. This taxon was reported mostly in damp buildings in North America and West Europe, but also from foods and from mosquitoes [59,60], vineyards, grape must, fermentation wine, and fruit juices [11,61,62]. Previously, we had isolated this fungus from the environment of the cellars where the bottles containing the sparkling wine were aging (data not published). Consequently, finding P. corylophilum might be due to the bottle not being sufficiently sealed by the cork stopper. A new species of Rasamsonia, Rasamsonia frigidotolerans, was found in the wine samples. The genus is characterized by the production of ornamented, paecilomyces-like conidiophores and olive-brown conidia, and in four of the species, the production of ascomata have been reported. Rasamsonia spp. have been reported in Asia, Europe, and North America, from substrata such as compost, conifer wood chips, cow dung, house dust, indoor air, piles of peat, rice straw, seed of Piper nigrum, soil, and human and animal clinical specimens [47,48,63,64]. None of the previous studies have reported this genus either in wine or on cork stoppers. Rasamsonia species are thermotolerant or thermophilic, with an optimum growth temperature above 30 • C and a maximum above 45 • C [65,66]. Rasamsonia frigidotolerans is characterized by the production of smooth-walled conidiophores (verrucose in all other species of the genus), by an absence of growth on CYA at 30 • C (all other species are thermotolerant), and by the production of globose conidia connected by disjunctors (absent in the rest of the species of the genus).
The new genus Dactylodendron, phylogenetically closely related to the order Onygenales, is characterized by its branched conidiophores and the production of chains of arthroconidia. The type species, Dactylodendron pinicola, is an asexual fungus previously classified phenotypically within the genus Arthrographis. It was originally isolated from insect galleries and from adult beetles of lps latidens, and of Dendroctonus ponderosae in Pinus contorta var. latifolia in Canada [67]. Dactylodendron pinicola produces conidiomata (absent in the other two species of the genus), whereas D. pluriseptatum produces long chains of prismatic arthroconidia, and D. ebriosum forms conidiophores that produce fertile branches at the apex.
In our study, we found a few isolates morphologically similar to the genus Dendryphiopsis, which in a polyphasic study demonstrated to be new species of Kirschsteiniothelia. This genus demonstrated to be phylogenetically related to the anamorphic genus Dendryphiopsis [68,69]. Species of Kirschsteiniothelia/Dendryphiopsis have been isolated principally from decaying wood and leaves [70][71][72], and even in freshwater habitats [73][74][75], but never from sparkling wine or cork stoppers. Kirschsteiniothelia ebriosa and K. vinigena differ from the other species of the genus by the absence of a sexual morphology, and the conidia arising in chains directly from the main axis of the conidiophore. These two species can be distinguished each from other by the number of septa and the length of the conidia (mostly two-celled and short in K. ebriosa, and multi-celled and long in K. vinigena).
We also isolated an interesting strain of Cladophialophora from sparkling wine. Cladophialophora recurvata produces aseptate, broadly ellipsoidal to subglobose, relatively large conidia, with inconspicuous flattened scars (more evident in the other species of the genus). The species of the genus Cladophialophora have been never reported from wine or cork.
Other species isolated during our study were Alternaria alternata and Cladosporium cladosporioides, both from cork stoppers. There are a few reports of A. alternata in grape must from the Priorat region in Spain and from Douro in Portugal [11,61]. The species of the genus Alternaria infects a broad variety of living plants, but can also be recovered from plant debris [76]. Cladosporium oxysporum (but not C. cladosporioides) has been isolated previously from cork stoppers [5]. Cladosporium species are found worldwide, and frequently occur as a secondary invader of necrotic parts of different sort of plants, but also are easily recovered from air, soil, textiles, and numerous other substrata [77]. Beauveria bassiana, a well-known entomopathogenic fungus [78], was isolated once from a sample of sparkling wine. This fungus is usually found in soil [19], but is also known to be endophytic in living plants including grapevine [79].
Because the new fungal species failed to proliferate at ethanol concentrations ≥ 10% v/v, we consider them to be from a different origin than the grape must, probably the cellar and/or the cork stoppers, despite some of them being recovered only from sparkling wine samples. We are hopeful that future studies will allow us to discover whether these fungal strains produce 2,4,6-TCA and other compounds responsible for cork taint.

Conclusions
The presence of yeasts and molds (and occasionally of bacteria) was detected in several (24 out of 39) samples of sparkling wine (Catalonian cava) affected by cork taint, with a musty, or mouldy, off-odor, and/or flavor alteration that makes the wine undrinkable. On the other hand, all negative controls (without appreciable organoleptic alteration) were free of fungi. All cork stoppers from negative controls and deteriorated wine developed fungal colonies. We isolated 27 different fungi from both substrata. Among them, we found a new genus (Dactylodendron) and eight new species (Cladophialophora recurvata, Dactylodendron ebriosum, Dactylodendron pluriseptatum, Kirschteiniothelia ebriosa, Kirschteiniothelia vinigena, Rasamsonia frigotolerans, Talaromyces speluncarum and Talaromyces subericola). All fungal taxa were able to grow on cork, but only at alcohol concentrations ≤ 10% v/v (which is lower than 11.5% strength of Catalan sparkling wines). We therefore conclude that the fungi present in sparkling wine were also present in turn on the cork stoppers and/or are part of the environment of the cellar. Although Penicillium corylophylum was found in wine samples, its presence does not represent per se a risk to the health of the consumer (this fungus is a mycotoxin producer), because all the fungi we found were unable to grow at the ethanol concentration of the sparkling wines. Future studies will allow us to find out whether these fungi form 2,4,6-TCA and/or other volatile organic compounds involved in the production of cork taint in wines. Funding: The authors are indebted to the Instituto de Ciencia, Tecnología e Innovación (Mexico) and the Consejo Nacional de Ciencia y Tecnología (Mexico) for the scholarship 440135 with scholar 277137. This work was supported by the Spanish Ministerio de Economía y Competitividad, grant CGL2017-88094-P.