Close this search box.

Copper-based grape pest management has impacted wine aroma – Scientific Reports


Fifty-one Saccharomyces cerevisiae from different geographical areas were characterized for their H2S production during alcoholic fermentation. The genetic group, identified in previous works indicated in the references of Supplementary Table 1, reflected the colonized ecological niche: 28 belong to the “wine” clade, 14 to “velum” group and 9 to the “oak” one. Strains were selected from our laboratory collection and maintained on solid medium (agar YPD: 2% glucose, 1% yeast extract, 2% bactopeptone, 2% agar) at 4 °C.

Fermentation conditions and H2S quantification

Fermentation experiments were conducted using synthetic must (SM), designed to mimic the characteristics of a natural grape must38. It contained a 200 g/L equimolar glucose and fructose content, and 200 mg/L assimilable nitrogen, 3.8 mg/L phytosterol, and 0.25 mg/L of Cu2+. The pH was adjusted to 3.3 with sodium hydroxide solution.

One colony of each strain was grown in 5 ml of liquid YPD at 28 °C for 24 h and then diluted 100 times in SM. After 24 h at 28 °C, cells were counted with an electronic particle counter (Multisizer 3 counter; Beckman Coulter) and 250 mL of SM, supplemented with 60 mg/L of SO2 when the impact of sulfite was evaluated, were inoculated to 1 × 106 cells/mL. Fermentations were carried out at 28 °C, under permanent stirring (280 rpm) and they were followed daily by weight loss, until the theoretical percentage of sugar consumed reached 95% (87.4 g CO2/L produced). Total H2S produced during alcoholic fermentation was collected with a zinc-based trap system and quantified with sulfide specific fluorescent probe, as described before32.

When the impact of the overexpression of CUP1 was in study, SM was supplemented with Geneticin (G418—Sigma A1720-5G) to maintain the plasmid allowing the overexpression itself. Suitable antibiotic concentrations were defined for each strain (100 µg/mL for wine strains, 40 µg/mL for the oak one), to simultaneously allow the maintenance of the plasmid and a good fermentation rate, but prevent the growth of the sensitive strain (i.e. the wild-type strain without the plasmid).

When assessing the impact of copper concentration on H2S production, SM was supplemented with copper sulfate to reach 1 or 2 mg/L of copper; control copper concentration was 0.25 mg/L in all the experiments. More details about the experiments are given in the “Experimental design and statistical analyses” section.

Drop test on copper and sulfite supplemented media

Copper resistance in presence or absence of SO2, was assessed by a drop-test for three wild type strains with different CUP1 copy number in their genome (Oak-Rom 3-2, LMD17 and L1374), and their counterpart engineered to over-express CUP1 (see below). Triplicates of these strains were grown overnight at 28 °C in 5 mL of YPD. Cells were then counted with an electronic particle counter (Multisizer 3 counter; Beckman Coulter), washed with PBS and resuspended in sterile PBS to obtain 107 cells/mL. Three successive 1/10 dilutions were prepared and 1.5 µL of each dilution was spotted on synthetic must having the same composition of the one used for the fermentations, gelled with 20 g/L agar. According to the tested modalities, copper (0, 0.5, 1, 6, 12 mM) and sulfite (0, 40, 60 mg/L SO2) were added to the media to evaluate their effect. Agar plates were incubated at 28 °C for 72 h and growth was assessed by visual examination.

CUP1 copy number evaluation

For most of the strains, CUP1 copy number was estimated from their genome sequence, obtained from previous works or from sequencing performed in this study. To obtain the values, the median sequencing depth measured at SNPs encountered between coordinates 212,500 and 213,000, and between 214,500 and 215,000 on Chromosome VIII was divided by the median sequencing depth over the entire genome (excluding mitochondria and 2 microns). For Italian strains, CUP1 copy number data had been already quantified by Real Time PCR18.

Genomic DNA extraction for sequencing

Genomic DNA was isolated from liquid yeast cultures in stationary phase, with a classical phenol–chloroform method, as described before39, with an additional purification step based on the use of silica-coated magnetic beads (GMG-252-A-100 mL—PerkinElmer), as follows. Cells were broken mechanically by shaking them in the presence of 600 µm diameter glass beads, lysis buffer (Tris 50 mM pH 8, EDTA 50 mM, NaCl 100 mM, Triton 2%, SDS 1.25%) and phenol chloroform isoamyl alcohol 25:24:1. DNA was precipitated with ispopropanol and ethanol, dried, resuspended in TE (Tris 10 mM, EDTA 1 mM) and treated with RNase A. Samples were mixed with the DNA absorption solution (for one sample: 50 μL 5 M NaCl, 15 μL magnetic beads (GMG-252-A-100 mL—PerkinElmer), 250 μL 7.8 M guanidium chloride, 800 μL isopropanol), after which metal beads with DNA absorbed on their silica surface were recovered using the DynaMag™-2 Magnet tube holder (12321D-DynaMag-2—Invitrogen) and washed twice with AMMLAV/E buffer (10 mM Tris pH 8, 0.1 mM EDTA, 60 mM potassium acetate, 65% ethanol) and twice with ethanol 75%. DNA was then desorbed and in aqueous solution.

DNA purity was checked from the 260 nm/280 nm and 260 nm/230 nm OD ratio measured with NanoDrop 1000 (ThermoScientific). The DNA was quantified by fluorescence using the QuantiFluor kit, dsDNA system (Promega) and then stored at − 20 °C.

Genome sequence and analysis

DNA samples were processed to generate libraries of 500 bp inserts. After passing quality control, the libraries were sequenced with DNBseq technology using BGISEQ-500 platform, generating paired-end reads of 2 × 150 bp.

For each library, low-quality reads were processed and filtered using the FASTX Toolkit v0.0.13.2 and TRIMMOMATIC v0.3640 with the following parameters (LEADING:10 HEADCROP:5 SLIDINGWINDOW:4:15 MINLEN:50).

Reads were then mapped to the S288C reference genome with BWA v0.6.2 with default parameters41 and genotyping made with samtools v1.11 to obtain a variant file including the sequencing depth of each variant position. Sequence positions were afterwards filtered for quality criteria: sufficient coverage position as well as genotyping and mapping quality (MQ >  − 20) were kept.

Plasmid construction and yeast transformation

CUP1 was inserted via Gibson assembly method42 between TEF promoter and terminator in a high copy Yeast Episomal plasmid (YEp352), modified to confer geneticin resistance (YEp352-G418) to the host cell. In detail, the backbone was amplified with primers P1 and P2, designed to replace the original URA3 copy of YEp with CUP1, since the strains used were not auxotrophic and the selection had been made by antibiotic. Therefore, the backbone contained a 2 um replication origin (multicopy), AmpR, ColE1, pPGK and G418 resistance cassette. CUP1 was amplified from OakGri7_1, a strain previously sequenced by our laboratory10, with a single metallothionein copy and the same sequence as laboratory reference strain S288C, used to design primers (P5–P6). TEF promoter and terminator were amplified from pCfB231243 with primers P3–P4 and P7–P8, respectively. Primer sequences are listed in Table 1.

Table 1 Primers used in this work.

Proper fragment insertion was verified by enzymatic digestion (NarI, ClaI, PacI—New England Biolabs). To assure that the phenotype was related to the overexpression of CUP1, a Yep352-G418 plasmid without CUP1 was used as control. PCRs were performed with Phusion™ High-Fidelity DNA Polymerase and validated by gel electrophoresis. Escherichia coli strain DH5α was used to maintain and amplify the plasmid; cells were selected on LB medium with ampicillin (100 µg/mL) and grown at 37 °C. Yeasts (Oak-Rom 3-2, LMD17 and L1374) were transformed with the lithium acetate method44 and strains containing the recombinant plasmids were selected on YPD agar with 200 μg/mL geneticin (G418—Sigma A1720-5G).

Experimental design and statistical analyses

Experiment 1: impact of the origin of the isolate and sulfites on H2S production

33 strains were selected randomly from our laboratory collection (Supplementary Table 1, Dataset 1). Alcoholic fermentations were performed in absence or presence of SO2, in duplicate for each strain and each condition.

The factors accounting for the variation of H2S were analyzed with the following analysis of variance model:

$$ {text{Yijk }} = mu + alpha {text{i }} + beta {text{j }} + gamma {text{ij}} + varepsilon {text{ijk,}} $$

where Yijk is the H2S production, µ the overall grand mean, αi is the fixed strain group effect, βj is the fixed SO2 effect, γij is their interaction effect, and εijk the residual error.

The analysis of the residuals showed that three values were distant from the global distribution. Since results of the statistical analysis did not change after removing all the observations of the three outlier strains, the complete dataset was kept as the method is sufficiently robust to mild deviations.

Experiment 2: impact of copper content of the media on H2S production

Fermentations were performed without SO2 in triplicate, for each strain (VL1 and LMD17) and each condition (0.25–1 and 2 mg/L of copper).

To evaluate the effect of copper and strains on H2S production, ANOVA was performed, after checking for the equality of variance with a Levene test. The most parsimonious model was kept after checking of the absence of interaction between strain and the copper content:

$$ {text{Yijk }} = mu + alpha {text{i }} + beta {text{j }} + varepsilon {text{ijk,}} $$

where Yijk is the H2S production, µ the overall grand mean, αi is the fixed strain effect, βj is the fixed copper effect and εijk the residual error.

Experiment 3: impact of CUP1 copy number on H2S production

To the strains evaluated in experiment 1, we added 18 wine strains (total strains analyzed = 55), some known to harbor a high number of copies of CUP1, and some commercial strains known to be high H2S producers (Supplementary Table 1, Dataset 2). Alcoholic fermentations were performed in absence of SO2, in duplicate for each strain.

Different polynomial models were used to describe the interaction between H2S production and CUP1 copy number (first-, second- and third-degree polynomial models); ANOVA was used to assess the significance of these models.

Experiment 4: impact of the overexpression of CUP1 on H2S production

Fermentations were performed without SO2 and with standard copper content (0.25 mg/L) in triplicate, for each strain (OAK_ROM 1–3, LMD17 and L1374) and each condition (wild-type strain, strain with the empty vector, strain with the CUP1 overexpressing vector).

ANOVA was performed to test the effect of the genetic modification in each strain. The model used was:

$$ {text{Yij }} = mu + alpha {text{i }} cdot varepsilon {text{ijk,}} $$

where Yij is the H2S production, µ the overall grand mean, αi is the fixed genetic modification effect, and εij the residual error.

Figure 6 summarizes the experimental design.

Figure 6
figure 6

Experimental design.

For all the experiments, when the impact of one (or more) factor was significant, differences between modalities were evaluated by post-hoc testing (Tukey’s HSD multiple-comparison test, p < 0.05).

Statistical analyses were performed in the R environment (R version 4.0.2 (2020-06-22)45).

Compliance with international and national regulation

Yeast strains were available from culture collection, or gifted by other authors, or provided by the company Lallemand. The yeast collection and use was in accordance with all the relevant guidelines.