The impact of fungicide treatments on yeast biota of Verdicchio and Montepulciano grape varieties

Yeast species that colonize the surface of grape berries at harvest time play an important role during the winemaking process. In this study, the use of culturable microbial techniques permitted a quantitative and qualitative inventory of the different yeast species present on the grape berry surfaces of Montepulciano and Verdicchio varieties when treated with conventional and organic fungicides. The results show that the most widespread yeast species at harvest time were Aureobasidium pullulans and Hanseniaspora uvarum, which are considered normal resident species and independent of the grape varieties and treatments applied. Specific differences when comparing the grape varieties were observed in species and were detected at a lower frequency; Pichia spp. were prevalent in Verdicchio, whereas Lachancea thermotolerans and Zygoascus meyerae were found in Montepulciano. In both vineyards, the farming treatments improved the competitiveness of A. pullulans, which was probably due to its reduced susceptibility to treatments that improved the competition toward other fungi. In contrast, the fermenting yeast H. uvarum was negatively affected by fungicide treatments and showed a reduced presence if compared with untreated grapes. Organic treatments directly impacted the occurrence of Issachenkia terricola in Montepulciano grapes and Debaryomyces hansenii and Pichia membranifaciens in Verdicchio. Conversely, a negative effect of organic treatments was found toward Metschnikowia pulcherrima and Starmerella bacillaris. Overall, the data suggest that the yeast community colonizing the grape berry surface was influenced by both grape variety and farming treatments, which characterized the yeast biota of spontaneous must fermentation.


Introduction
Grapes represent a complex ecological niche where filamentous fungi, yeasts and bacteria cohabitate. The microbiome includes species at a concentration that mainly depends on the grape ripening stage and the availability of nutrients. However, the microbial communities of grapes may be affected by many other variables, such as pedoclimatic factors, viticultural practices, diseases and pests that could modify grape integrity [1]. In general, the yeast populations a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 vineyards have employed three different management systems: organic, conventional and with no treatment. To exclude any cross-contamination between different treatments, the minimum distance between each block of rows was approximately one kilometer from each other for all of the grapes; and within the same vineyard, the grapes are exposed to the same slope, sun, and shade and have similar soil characteristics. The harvest was carried out, in both varieties, at full ripeness (15 September for Verdicchio; 10 October for Montepulciano).

Spontaneous fermentations
The grapes were placed into sterile bags and were hand-crushed and shaken at 120 rpm for 30 minutes on an MAXQ 4450 (Thermo Fisher Scientific, Waltham, Massachusetts, USA). One milliliter of each fresh must was collected and used for yeast isolation and enumeration. The remaining grape juice with skins was transferred into 250 mL sterile Erlenmeyer flasks, closed with Pasteur bungs (to allow CO 2 to escape from the system) and set up for spontaneous fermentation at 25˚C under static conditions. After 7 and 15 days from the start to the spontaneous fermentation, the samples were collected to evaluate the yeast population by viable cell counts.

Yeast enumeration and isolation
Samples from fresh musts and during fermentation were collected and used for monitoring the yeast populations at the beginning and after 7 and 15 days of fermentation. For total yeast enumeration, serial decimal dilutions in sterile water were prepared and then spread on Wallerstein (WL) nutrient agar (Merck KGaA, Germany) supplemented with 0.005% chloramphenicol (Thermo Fisher GmbH, Germany) and 0.02% biphenyl (Sigma-Aldrich, <Saint Louis, Missouri, USA) to suppress the bacteria and reduce the growth of molds, respectively. The plates were incubated at 25˚C for four days. After macro-and micro-morphological analysis and in proportion to their frequencies, the yeast isolation was conducted on plates that contained between 100 and 300 colonies. Approximately 10% of the colonies per plate were isolated on YPD agar (1% Yeast Extract, 2% Peptone, 2% D-glucose, and 2% Agar) from each sample and at each time of sampling [17,18]. The total isolates were 1,240. The yeast strains were preserved in 40% (v/v) glycerol at -80˚C.

DNA extraction and yeast identification
The 1,240 isolated strains, that showed identical macro-and micro-morphological characteristics, were grouped and representative isolates were used for genomic DNA analysis. DNA was extracted according to the method described by Stringini et al. [19]. Using primer set ITS1 (5'-TCCGTAGGTGAACCTCGCG-3') and ITS4 (5'-TCCTCCGCTTTATTGATATGC-3') [20], the ITS1-5.8S rRNA-ITS2 region was amplified by PCR. The PCR was performed as described by Esteve-Zarzoso and co-workers [21]. The PCR products were separated in 1.5% (w/v) agarose gel (stained with ethidium bromide) using 0.5x TBE buffer by horizontal electrophoresis (Bio-Rad, Hercules, USA). The identities of the representative yeasts were obtained by sequencing. The BLAST program [22] and the GenBank database (http://www.ncbi.nlm. nih.gov/BLAST) were used to compare the sequences provided with those already in the data library. The inclusion of obtained sequences into the NCBI GenBank data library has been completed under the accession numbers from MK351988 to MK352096.

Statistical analysis
The relative abundance of species was obtained by calculating the corresponding portion of each species with respect to the total yeast detected in the samples and based on the colony counts. The analysis of variance was conducted using the JMP 11 from SAS program. Furthermore, the results obtained from the analysis of microbial diversity on the grape surface of different vineyards employing different agronomic practices and yeast dynamics during spontaneous fermentation were examined with Unscrambler 7.5 software (CAMO ASA, Oslo, Norway) to obtain the Principal component analysis (PCA).

Ethics statement
For this research is not require an ethics statement, and the authors confirm that the field studies did not involve endangered or protected species.

Effect of grape variety on the yeast community at harvest time
Microbial community associated with the grape surface of the Verdicchio and Montepulciano varieties was evaluated. From the general framework (Fig 1), it was observed that both grape varieties presented an abundance of yeasts, such as Aureobasidium pullulans and Cryptococcus spp., with oxidative metabolism. Together, these represent 60% and 40% of the total yeasts present on grape berries of Verdicchio and Montepulciano, respectively. In particular, out of the 60% associated with Verdicchio grapes, 50% belong to A. pullulans and 10% to Cryprococcus spp. The 40% of yeasts with oxidative metabolism associated with Montepulciano were 30% A. pullulans, 7% Cryptococcus spp. and 3% minor representative species. In terms of relative abundance and among the fermenting yeasts, Hanseniaspora uvarum (22% Verdicchio grapes and 43% Montepulciano grapes) and Starmerella bacillaris (13% Verdicchio grapes and 7% Montepulciano grapes) were the most abundant and constantly present species found in both varieties. The Montepulciano variety was characterized by a consistent presence of Issatchenkia terricola (7%). Both grape varieties showed the presence of weak fermenting yeasts such as Metschnikowia pulcherrima and Debaryomyces hansenii (1.0-1.5%) and Candida californica (<1%). At low relative abundance (1.0-2.0%), some yeast species were found in one or the other grape varieties. Pichia sporocuriosa, Pichia fermentans and Pichia membranifaciens were found only in the Verdicchio grape variety, whereas species such as Lachancea thermotolerans, Zygoascus meyerae and Rhodotorula spp. were only found in grapes from the Montepulciano variety. In summary, no substantial differences were found among the two grape varieties regarding the main yeast species (oxidative and low fermenting species) that colonized the grape berry surface. A statistically significant difference was found for the relative abundance of H. uvarum more present on Montepulciano grape variety. This difference, in addition to the variety effect, could also due to other intrinsic variables of their cultivation management such as the different harvest time of grapes (Table A in S1 File). On the other hand, differences were detected in species found at a low frequency (often fermenting species) that were only isolated in one or another grape variety.

The influence of fungicide treatments on the yeast community at harvest time
The yeast community that colonizes grape surfaces was analyzed for the influence of the fungicide treatments. Both Verdicchio and Montepulciano varieties have been subjected to an organic and a conventional treatment. The results showed that in Verdicchio samples (Fig 2), the yeast-like A. pullulans was favorited by fungicide treatments in comparison to untreated samples (t = 0.05) probably due to the lower competition of the yeasts affected by treatments. Indeed, in both organic and conventional samples, this yeast was the most abundant species (44% and 60%, respectively) while in untreated grapes it was only 5% of the whole yeast population. The same behavior was observed in Montepulciano samples (22%, 45% and 10% in organic, conventional and untreated samples, respectively) (Fig 2). Conventional treatments seem to exert more selective pressure on the yeast community and favor the colonization of A. pullulans (around half of the yeast community). Similar to A. pullulans, the occurrence of Cryptococcus spp. seemed to be positively influenced by the treatments (and this was absent in the untreated grapes). Different from A. pullulans, the organic treatments positively affected the colonization of Cryptococcus spp. in comparison with conventional treatments in both Verdicchio and Montepulciano varieties even if only in Verdicchio variety a significant difference was found (Table B in S1 File) (16% and 3% in organic and conventional Verdicchio grapes, respectively, and 7% and 1% in organic and conventional Montepulciano grapes, respectively). Variety and fungicide effects on grape yeast H. uvarum was the second most abundant species in the treated grapes of both varieties. This apiculate yeast did not seem to be influenced by the type of treatment (26% and 20% in organic and conventional Verdicchio grapes and 36% and 40% in organic and conventional Montepulciano grapes, respectively). In untreated grapes, H. uvarum was the most abundant species (47% and 75% in Verdicchio and Montepulciano varieties, respectively) and showed significant differences in comparison with both treated grapes (organic and conventional) but only in Montepulciano variety (Table C in S1 File).
S. bacillaris species showed a wide variability among the treatments/varieties. In the Verdicchio variety, S. bacillaris decreased in the treated grapes (9% in organic samples and 16% in conventional samples) compared to the untreated ones (29%) Fig 2). In Montepulciano, S. bacillaris was more present in organic grapes (17%), absent in conventional samples and poorly present in untreated samples (4%).
Organic treatments favorably affected the presence of I. terricola and showed an abundance of 14% in organic grapes and only 5% in untreated grapes, and I. terricola was almost absent in conventional samples of the Montepulciano variety while it was present at a very low relative abundance (<1%) in the Verdicchio variety. A positive effect of organic treatments was also exerted on some low abundance species only present in the Verdicchio variety, such as P. membranifaciens (4%) and P. sporocuriosa (1%), that were only found in the grapes treated with copper and sulfur. Differently, a negative effect of organic treatments was shown toward M. pulcherrima. Indeed, this yeast species was significantly present in untreated (13 and 3% in Verdicchio and Montepulciano grapes, respectively) and poorly present or absent in treated grapes (Tables B and C in S1 File). D. hansenii was detected only in treated samples, while P. fermentans and C. californica were generally found in untreated grapes (2.5% and 3%, respectively). Rhodotorula spp., a ubiquitous yeast, was present only after the treatments (1.67% and 2.57% in Montepulciano organic and conventional samples, respectively). Fermenting yeasts detected only in the Montepulciano grape variety showed a different response toward fungicide treatments. L. thermotolerans was only present in conventional treatment samples (5.5%), whereas Z. meyerae and Zygosaccharomyces bailii were mainly detected in untreated samples and almost absent in conventional grapes.

Yeasts dynamics during spontaneous fermentation
Middle fermentation. The microbial community evolution during spontaneous fermentation was monitored through viable counts after 7 and 15 days from the start of fermentation. After 7 days (approximately middle fermentation) the yeast population increased by approximately two log (from 8.0×10 5 CFU/ml to 6.0×10 7 CFU/ml) in Verdicchio samples and only one log in Montepulciano samples (from 1.7×10 6 CFU/ml to 1.5×10 7 CFU/ml) (Tables D and E in S1 File). As expected, the environmental conditions determined a selection in favor of fermenting yeasts. Indeed, the oxidative yeasts A. pullulans and Cryptococcus spp. disappeared from all of the samples. H. uvarum, which was already well represented at the beginning of fermentation (22%), became the dominant species in all of the Verdicchio trials (Fig 3) and the only occasionally fermenting yeast species (I. terricola, S. bacillaris, C. californica, M. pulcherrima, P. fermentans, Torulaspora delbreuckii, and Candida diversa) were found at low relative abundance (all together at approximately 4%). This picture is nearly confirmed in Montepulciano treated samples where other yeast species participated in the fermentation process. In untreated samples, H. uvarum was present at only 4% of the total yeast population, but other fermenting species S. bacillaris (37%), Z. bailii (34%) and C. californica (26%) appeared. S. bacillaris and C. californica were present in all Montepulciano samples even if they were more abundant in untreated samples. Montepulciano samples can be differently recognized by a relevant presence of: i) I. terricola in organic samples, ii) L. thermotolerans in conventional samples, and iii) Z. bailii in untreated samples (Fig 3).

Principal component analysis (PCA) of the yeast community
The PCA of the overall yeast microbiome of the grape berry surface at the harvest period and during a spontaneous fermentation revealed a yeast population diversity among the samples of grapes subjected to organic and conventional fungicide practices and untreated samples coming from two different varieties/vineyards (Fig 5). The biplots were obtained by evaluating the relative yeast species abundance, the grape varieties and the fungicide treatments.
At the harvest time, PC1 (63%) showed a differentiation between untreated (right quadrants) and treated samples independently by the type of treatment while PC2 (20%) distinguished between Verdicchio (upper quadrants) and Montepulciano (lower quadrants) samples ( Fig 5A). These data suggest an evident impact of fungicide treatments on yeast biota associated with the grape berry surface. Moreover, the grape varieties showed a different yeasts colonization, although different harvest times (with consequent different climatic conditions) and agronomic management could contribute to this yeast differentiation. In this regard, M. pulcherrima and H. uvarum in Montepulciano and P. fermentans and C. californica in Verdicchio varieties were the main characterizing species of untreated samples, whereas oxidative yeasts species mainly differentiated the treated samples. The spatial distribution of PCA confirmed that the species that linked the two grape varieties were: Z. meyerae, I. terricola, Rhodotorula spp., and Z. bailii for Montepulciano and P. membranifaciens, P. sporocuriosa and P. fermentans for Verdicchio. At 7 days of fermentation (Fig 5B), a general reduction and simplification of the yeast community was observed. However, the untreated samples remained separated from the treated samples (down and upper quadrants, respectively). PC1 (61% of variance explained) distinguished Verdicchio samples from those of Montepulciano even if conventional samples were more closely related than organic ones.
At the end of fermentation (Fig 5C), all of the Montepulciano samples were grouped in the right/lower quadrant and mainly characterized by C. californica and Z. bailii. Furthermore, P. sporocuriosa characterized Montepulciano conventional samples and S. bacillaris characterized Montepulciano organic samples. In contrast, all of the Verdicchio samples were differently distributed in the graphic space and indicated remarkable differences for the relevant presence of P. kudriavzevii (untreated samples), H. uvarum and T. delbrueckii (conventional samples), and C. californica organic samples).

Discussion
In recent years, the investigation of the geographic distribution of the microbial community of wine grapes revealed a geographic delineation of the yeast communities conditioned by several factors such as cultivar, vintage, climate and agricultural practices [23][24][25][26].
The influence of farming practices used in the vineyard and on the yeast, biota associated with the grape berry surface was recently investigated [18,[27][28][29]. In the present study, the impact of organic and conventional treatments on the occurrence of yeast species in two Italian varieties was evaluated. The total yeasts recovered at harvest time were from 10 5 to 10 6 CFU/ml and in accordance with the yeast presence in grapes described in previous studies [7,30].
A. pullulans was the first and most abundant species found in both treated Verdicchio samples and in conventional Montepulciano grapes and in agreement with results obtained from Setati et al. [29]. In organic Montepulciano samples, A. pullulans represented the second most abundant species found, and this disagrees the results obtained by Renouf et al. [30] that did not find A. pullulans on the grape surface. These yeasts seem to be favorably affected by farming treatments, which is probably due to their improved competitiveness towards other fungi in the presence of fungicides and their capacity to detoxify CuSO 4 as reported by Schmid et al. [31]. Their results support the key role of the yeast-like fungus A. pullulans in explaining the functional differences between organic and conventional agricultural systems. Indeed, it has long been known that Aureobasidium can utilize inorganic sulfur and is able to absorb, and in this way detoxify, copper [32][33]. In untreated grapes, A. pullulans represents only a minor component of the whole yeast population.
H. uvarum was the most abundant fermenting species in both varieties, although the fungicide treatments significantly reduced its presence. These results, which confirmed previous results, are in accordance with the current literature reviewed by Pretorius [34].
A negative effect of both conventional and organic treatments was detected toward M. pulcherrima since a significant decrease in treated samples was found. The same results were described by Milanović and co-workers [18] but only in organic samples. This finding highlighted that M. pulcherrima, antagonistic and antimicrobial yeast [35][36][37][38], is negatively influenced by fungicide treatments and particularly organic ones. Considering that this yeast species revealed a positive contribution to the analytical and aromatic composition and Variety and fungicide effects on grape yeast complexity of wine [39][40][41], fungicide treatments may reduce this yeast's positive contribution to fermentation and wine composition.
The monitoring of fermentation conducted at the laboratory scale elucidated the relationship between yeast occurrence on the grape surface and the potential influence during applicative wine management.
According to Bagheri et al. [42], our results clearly showed a decline at the start of fermentation of the oxidative strains; this decline was probably due to the anaerobic conditions created by the fermentation process. As expected, in the middle of fermentation, H. uvarum became the most representative species in all of the samples [18]. At this stage, P. fermentans and C. californica in Verdicchio and Z. meyerae and Z. bailii in Montepulciano seems to characterize the yeast biota of the two varieties. Regarding the fermenting yeast L. thermotolerans, the results obtained seem to highlight the favorable effect of conventional treatments on this species that was only found in Montepulciano. Cordero-Bueso et al. [13] described L. thermotolerans as the predominant non-Saccharomyces species found in both organic and conventional samples without relevant differences between the treatments. In our study, this yeast was present on the grape berry surface at harvest time and survived until the end of spontaneous fermentation. Its initial concentration probably plays a significant role in establishing yeast-yeast interaction that allows itself to compete and survive during fermentation [42]. In contrast to L. thermotolerans, Z. bailii was found in all of the samples of the Montepulciano varieties; this outcome indicated that this species was not affected by organic or conventional treatments.
In Verdicchio samples, the fermenting yeasts, W. anomalus and T. delbrueckii, seems to characterize the conventional samples. The presence of T. delbrueckii agrees with the results of Cordero-Bueso et al. [13] that detected this species in Barbera musts coming from conventional grapes. T. delbrueckii was found at a low frequency only at the end of fermentation as reported by Pinto et al. [43], whereas Bagheri et al. [42] found T. delbrueckii only in samples coming from an integrated vineyard. However, W. anomalus represented the second most abundant species probably due to its capacity to persist at the typical end-fermentation conditions [44][45]. Indeed, it established its ability to tolerate up to a 12% ethanol concentration and to produce a killer toxin to compete against other yeasts [46][47].
The strong fermenting yeast S. cerevisiae was poorly detected and emerged at the end of fermentation and only in Verdicchio conventional samples. These data confirm that the best fermenting yeasts is very poorly present on the grape berry surface at harvest time [17,48]. On the other hand, fermenting yeasts present at low frequencies at harvest time, took over the spontaneous fermentation differently depending on the fungicide treatments. This is the case of C. californica in organic samples of Verdicchio variety and S. bacillaris in organic samples of Montepulciano variety (both species absent or present at very low frequencies in the conventional samples). Differences in the dominant yeast species at the end of fermentation in the grape variety were also found (high presence of Z. bailii in Montepulciano and H.uvarum in Verdicchio) and could be due to the overall differences between the varieties (characteristic of grapes, time of harvest and agronomic management).
Overall, the data suggest that the yeast community colonizing the grape berry surface was influenced by agricultural treatments. A. pullulans and H. uvarum were the dominant yeast species at harvest time even if their relative frequencies were strongly influenced by fungicide treatments. Fermenting yeast species differently colonized the grape surface and characterized microfermentations trials of Verdicchio and Montepulciano varieties. These fermenting yeast population changes from varieties are conditioned by pesticide treatments and could be expected to have some impact on the fermentation process and wine composition and their evaluation should receive further attention.
Supporting information S1 File. Supplemental Materials. (Table A) Analysis of variance (ANOVA) of Verdicchio and Montepulciano samples at harvest time. The significant differences were determined using t-Test, and the data were considered significant if the associated P values was <0.05. Data with different letters (A, B) within each row are significantly different. (Table B) Analysis of variance (ANOVA) of Verdicchio samples at harvest time and after 7 and 15 days of spontaneous fermentation. Letters O, C and NT indicated organic, conventional and untreated farming management, respectively. For each yeast species detected and for each sampling time, the different letters (A, B) indicated significant differences between the samples (p < 0.05) using t-Test. (Table C