Abstract
Background and goals Accurate, fast, and cost-effective methods to assess smoke taint severity are essential for wineries to make processing decisions. Levels of volatile phenols related to smoke taint may increase during winemaking via limited cleavage of conjugated precursors, though phenolic glycosides have been demonstrated to be stable during bottle aging. Release of nonvolatile precursors can occur during consumption via in-mouth enzymatic hydrolysis. Estimating the potential of smoke taint in wine requires measurement in grapes either by directly analyzing these precursors or by converting them into free forms through hydrolysis or other methods. Small-scale fermentations are also used to evaluate potential off-aromas postfermentation. This study assessed three sample-preparation methods for the release of volatile aroma compounds related to smoke taint.
Methods and key findings Microvinification was compared to acid hydrolysis and enzymatic hydrolysis in an evaluation of smoke taint risk by gas chromatography-tandem mass spectrometry (GC-MS/MS). Microvinification and enzymatic hydrolysis produced similar concentrations of volatile phenols. In contrast, acid hydrolysis released higher concentrations of free volatile phenols but showed greater variability and limited predictive accuracy.
Conclusions and significance Microvinification closely mimics real winemaking conditions by incorporating yeast activity, alcohol extraction, and maceration, making it a realistic predictor of smoke taint severity. It also provides wineries with an in-house option for sample preparation before sending samples to external labs for volatile analysis, which may help minimize losses related to smoke taint.
Introduction
Wildfires are increasing in severity and becoming more common in areas not previously prone to fires (Cunningham et al. 2024). Nearby fires cause commercial grapevines to absorb volatile smoke compounds, such as guaiacols and cresols, into the vines and fruit, imparting unpleasant smoke aromas to the resulting wine (Hayasaka et al. 2010, Szeto et al. 2020). Smoke taint aroma compounds occur in fruit as glycosides and cysteinyl/glutathionyl (S-) conjugates. Several studies have shown that these precursors can be cleaved under winemaking-relevant conditions to release volatile phenols and thiols that contribute to sensory taint, the magnitude of which appears to be matrix- and process-dependent (Hayasaka et al. 2010, Szeto et al. 2020, Favell et al. 2022, Tomasino et al. 2023, Rochfort et al. 2024).
By contrast, the role of free thiophenols remains under investigation, with reports both supporting and questioning a strong sensory contribution. During fermentation, yeast enzymes and juice acidity catalyze the hydrolysis of glycosidically-bound smoke compounds, releasing them into the wine (Caffrey et al. 2019). Within the class of volatile phenols, guaiacol and 4-methylguaiacol are widely recognized as key markers of smoke exposure (Krstic et al. 2015). Although 4-ethylphenol (4-EP), 4-ethylguaiacol (4-EG), vanillin, and eugenol are sometimes detected in wines made from smoke-affected fruit, current evidence does not establish them as primary smoke-derived markers, and their occurrence may reflect alternative sources or breakdown during hydrolysis.
There is some debate regarding the origins of specific smoke taint compounds. For instance, 4-EG is known to be a metabolic byproduct of Brettanomyces dekkera (Chatonnet et al. 1992). However, higher concentrations of 4-EP have been reported in wine made from grapes subjected to very high experimental smoke loads that exceed typical field exposures (Kennison et al. 2007). Likewise, eugenol, which is commonly associated with oak aging (Towey and Waterhouse 1996), was detected in that same high-intensity exposure study (Kennison et al. 2007). These observations likely reflect treatment severity and/or alternative sources rather than primary smoke-derived formation; accordingly, 4-EP, 4-EG, and eugenol are not considered reliable markers of primary smoke taint.
Analytical sample preparation methods for smoke taint characterization typically aim to liberate glycosidically-bound precursors (Kennison et al. 2008, Wilkinson et al. 2011, Hjelmeland and Ebeler 2015, Noestheden et al. 2017). Strong-acid hydrolysis (≈pH 1, heated) efficiently cleaves conjugated precursors but is intentionally harsher than winemaking and is known to introduce artifacts through unintended chemical reactions (Sefton 1998). Enzymatic hydrolysis is a gentler approach that partially reproduces fermentative cleavage of glycosides, although any single commercial enzyme preparation captures only a subset of the enzymatic activities present during yeast metabolism. Both approaches therefore provide useful but non-equivalent information, and neither fully replicates the complexity of fermentation.
In the present study, we used one glycosidase preparation (Miller and Block 2020); future work should evaluate a broader range of enzymes to better represent fermentative diversity. These differences formed the basis for comparing acid hydrolysis, enzymatic hydrolysis, and microvinification as complementary methods for precursor release. Although hydrolysis methods were not developed to mimic fermentation or predict final wine concentrations, they remain widely used to estimate the potential pool of bound smoke-related precursors, providing information that is complementary to small-scale fermentations.
More recently, methods have been developed for measuring glycosidically-bound compounds without freeing the volatiles. These involve the use of liquid chromatography-mass spectrometry (LC-MS) to identify and quantify nonvolatile smoke-related glycosides (Hayasaka et al. 2010, Noestheden et al. 2017, Rochfort et al. 2024). Although these conjugates are detected in the grapes, these methods do not necessarily reflect the concentration of free volatiles in the resulting wine, and the library of known smoke-related glycosides is limited (Favell et al. 2022). A further challenge with both direct (analysis of the glycosidically-bound aroma precursors) and indirect (analysis of released previously glycosidically-bound aroma precursors) methods is that vineyards and wineries may not have the capabilities to conduct these tests and must send samples out to labs for quantification.
Small-scale fermentations are suggested (as reported at https://www.awri.com.au/industry_support/winemaking_resources/smoke-taint/) and commonly used in industry as a practical screening tool for evaluating potential smoke taint. The microvinification method could be adapted for sensory assessment when larger sample volumes and appropriate panel resources are available. In the present study however, microvinification was used solely as an analytical approach to compare the release of smoke taint-related volatile compounds across preparation methods. Sensory evaluation was not included because it falls outside of the scope of this work, which focused on quantifying analytically liberated volatiles rather than assessing perceived smoke attributes. Because producers often rely on analytical thresholds of free and bound compounds when making harvest and processing decisions, the objective in this study was to compare analytical workflows rather than evaluate sensory outcomes.
This study evaluated acid hydrolysis, enzymatic hydrolysis, and 50-mL microvinification for their ability to release smoke-related volatile phenols from smoke-exposed red grapes, quantified by gas chromatography-tandem mass spectrometry (GC-MS/MS), with the goal of comparing the techniques’ effectiveness as predictive tools for assessing risk of smoke taint.
Materials and Methods
Chemicals and reagents
Materials used for winemaking, ICV GRE yeast, and Go-Ferm were purchased from Lallemand; potassium metabisulfite and tartaric acid were purchased from Presque Isle Wine Cellars (North East, PA). Solvents and reagents used for chromatography sample preparation included citrate phosphate buffer, β-glucosidase from almonds, sodium chloride (NaCl), sulfuric acid (H2SO4), and methanol (all from Sigma-Aldrich), as well as Everclear grain alcohol (Luxco). Standards used for chromatography quantification included guaiacol (Tokyo Chemical Industry); guaiacol-d7, o-cresol-d7, and 4-ethyl guaiacol-d5 (LGC Group); and guaiacol, 4-EG, 4-methylguaiacol, eugenol, m-cresol, o-cresol, and p-cresol (Sigma Aldrich). All standards were of at least 95% purity.
Smoke-exposed samples
Two separate samples of each fruit cultivar (Cabernet franc, Cabernet Sauvignon, and Petit Verdot) were sourced from a commercial vineyard (Napa Valley, CA) that was exposed to smoke from the 2020 Glass Fire in Napa Valley between 27 Sept and 20 Oct. Two vineyard locations served as replicates (~0.75 km apart), with three random clusters collected per site, and each sample weighing ~2 kg. No attempt was made to account for viticultural variables (e.g., canopy position or vineyard location). Fruit was simply sampled from vines known to have been exposed to wildfire smoke, with the objective of obtaining fruit expected to contain smoke taint precursors. Samples were collected on 8 Nov 2020, shipped overnight on ice, then stored at −80°C until analysis. The basic fruit chemistry is detailed in Supplemental Table 1. Because the sampling period coincided with region-wide smoke, site-matched, unexposed control fruit was not available. The study was therefore structured as a quasi-experimental natural experiment using an impact-only observational design. For each smoke-affected sample, we applied three preparation workflows (acid hydrolysis, enzyme hydrolysis, and microvinification) to enable a blocked, within-sample methods comparison (block = sample; fixed factor = preparation). Analytical responses were quantified by GC-MS/MS as described below.
Sample preparation
Entire grape samples (~2 kg) were crushed by hand in 3.8-L plastic Ziplock bags (26.8 × 27.3 cm) until the skin of every berry was broken, as in a commercial crusher. Supernatant (50 mL) from each cultivar and location was allotted into 50-mL centrifuge tubes to be used for acid hydrolysis, enzyme hydrolysis, and juice analysis. The remaining grape solids and juice were retained in Ziplock bags for microvinification processing.
Strong acid hydrolysis
Acid hydrolysis was conducted with a method similar to those previously published (Liu et al. 2020). However, solid-phase microextraction (SPME) was not conducted to separate the glycosylated precursors from the free fractions, so the concentration of volatiles evolved should be considered the “total” amount (free and bound). The supernatant (5 mL) was added to 21-mL pressure tubes (ACE Glass Inc.), and H2SO4 was pipetted to a final concentration of 1.25 N H2SO4 in the reaction mixture. The tubes were placed into a water bath at 100°C for 75 min for hydrolysis. After cooling to room temperature in an ice water bath, the mixtures were adjusted to pH 3.5 with 5 N sodium hydroxide (NaOH) and stored at −80°C until analysis. Acid hydrolysis was conducted in quadruplicate.
Enzymatic hydrolysis
Enzymatic hydrolysis was completed in a manner similar to methods reported elsewhere (Kennison et al. 2008). Grape supernatant (2.5 mL) was added to a 20-mL amber SPME vial (Restek). Freshly prepared enzyme solution (25 μL, 100 mg/mL β-glucosidase rehydrated in 0.1 M citrate-phosphate buffer, pH 5) was added. The vials were incubated at 40°C for 20 hr in a water bath. After cooling to room temperature, the mixtures were adjusted to pH 3.5 with 5 N NaOH and stored at −80°C until analysis. Enzyme hydrolysis was conducted in quadruplicate.
Microvinification
Fermentations were conducted following microvinification methods (Warren et al. 2025). Crushed solids (22.8 g) were separated into 50-mL centrifuge tubes. Microvinifications were performed with skins to emulate red-wine extraction and provide a release pathway comparable to hydrolysate preparations. Juice was then decanted into the tubes to a final volume of 40 mL, leaving space to allow the cap to rise during fermentation. The tubes were placed into water baths at 35°C. Fermentations were punched-down and degassed twice daily by inverting the tubes. The musts were fermented on the skins until residual sugar concentration was <0.1% (typically after 7 days), as measured by AimTab Reducing Substances Tablets (Germaine Labs). No malolactic fermentation was conducted. Samples were pressed by centrifuging the 50-mL tubes for 10 min at 4816 × g and 4°C in a Sorvall Legend XTR Centrifuge (Thermo Fisher Scientific) and vacuum filtered through 110-mm Ø filter paper (Whatman plc) into 15-mL centrifuge tubes to decrease the amount of headspace. Pressed wines were stored at −80°C until analysis. Microvinification was conducted in quadruplicate.
Headspace solid-phase microextraction-gas chromatography-tandem mass spectrometry (HS-SPME-GC-MS/MS)
The HS-SPME-GC-MS/MS system comprised a Combi-PAL autosampler on an Agilent 7890A gas chromatograph with an Agilent 7000 triple quad detector (Agilent Technologies). HS-SPME-GC-MS/MS methods were conducted by applying parameters used previously (Awale et al. 2021). A divinylbenzene/carbon wide range/polydimethylsiloxane fiber (Restek) was used for sampling and extraction. Fibers were conditioned before sampling according to manufacturer recommendations. For extraction, 5-mL samples were placed in 20-mL vials with 2.5 g of NaCl and the internal standard mix and incubated at 40°C for 5 min. The fiber was exposed at 40°C for 30 min in the headspace of the samples with agitation at 500 rpm. The fiber was desorbed in the inlet at 260°C for 16 min in splitless mode. A Stabilwax-MS column (30 m × 0.25 mm i.d., 0.25-μm film thickness) and helium gas (flow rate 2 mL/min) were used for analysis. The GC oven temperature followed the program: 35°C for 5 min, heat at 6°C/min to 240°C, and hold for 10 min. The source temperature, MS1 quadrupole, and MS2 quadrupole were set to 230, 150, and 150°C, respectively. The electron ionization source operated at 70 eV. The helium quench gas flow rate was 2.25 mL/min and the collision gas was nitrogen (99.999% purity, Praxair) at a flow rate of 1.5 mL/min. Quantitative analyses were completed using MassHunter Quantitative Analysis software ver. B.07.01 (Agilent Technologies). All samples were run in analytical duplicate.
Microvinification samples (2.5 mL) were added to a 20-mL SPME vial with 2.5 mL of model wine mixture (12.5% EtOH, 5 g/L tartaric acid, adjusted to pH 3.5 with 5 N NaOH) to normalize the pH and ethanol concentration of the samples prior to SPME. Both hydrolysis samples (2.5 mL) and crushed grape supernatant (2.5 mL) were added to a 20-mL SPME vial with 2.5 mL of model wine concentrate mixture (25% EtOH, 5 g/L tartaric acid, adjusted to pH 3.5 with 5 N NaOH) to adjust the final ethanol concentration to a level equivalent to the microvinification samples. Samples were analyzed in analytical duplicate. The analysis used multiple reaction monitoring (MRM) transitions according to a published method (Liu et al. 2020) using the three deuterated internal standards guaiacol-d7, o-cresol-d7, and 4-ethylguaiacol-d5 at 0.2 ng/mL. The MRM transitions for each compound and their corresponding deuterated internal standard are shown (Supplemental Table 2). Calibration curves of smoke taint compounds 4-EG (5 to 1000 ng/mL), 4-methylguaiacol (5 to 1000 ng/mL), eugenol (5 to 1000 ng/mL), guaiacol (5 to 1000 ng/mL), m-cresol (5 to 1000 ng/mL), o-cresol (1 to 1000 ng/mL), and p-cresol (5 to 1000 ng/mL) were established in the indicated linear ranges.
Because the objective was to compare release/sample-preparation methods (not to compile an exhaustive marker list), we employed a compact analyte panel quantified with isotopically labeled internal standards—guaiacol, 4-methylguaiacol, 4-EG/4-EP, eugenol, furfural, and 5-methylfurfural—similar to the panels used in early smoke taint fermentation studies (Kennison et al. 2008). Syringol and 4-methylsyringol were not targeted here, but the workflows are compatible with their analysis and can be extended in future studies.
Statistical analysis
A two-way analysis of variance (ANOVA) with interaction terms “cultivar” and “method” was conducted on smoke taint compounds (α = 0.05). Average coefficients of variation were computed by averaging the standard deviation divided by the mean of each cultivar for every volatile phenol. Statistical analysis and graphics were computed using the software RStudio (ver. 1.3.1073), including the R packages ‘agricolae’ (de Mendiburu 2020), ‘dplyr’ (Wickham et al. 2022), and ‘ggplot2’ (Wickham 2016).
Results and Discussion
This study isolates the effect of preparation workflow within the same smoke-affected material and is not a cross-site calibration. Our conclusions therefore concern which preparation method is more representative of fermentation-relevant release under matched biological conditions.
Cabernet franc, Cabernet Sauvignon, and Petit Verdot fruit that had been exposed to wildfire smoke in California was analyzed to monitor the evolution of smoke taint-related volatile compounds in juice and after precursor liberation techniques (hydrolysis and microvinification). Two-way ANOVA (α = 0.05) was performed for all measured volatile phenol aroma active compounds, with “method” and “cultivar” as the independent variables (Table 1). All compounds except for o-cresol differed by cultivar and method. Additionally, guaiacol and o-cresol did not show significant differences for the interaction term, indicating that the combined effects of indirect smoke taint aroma method and grape cultivar were independent.
Two-way analysis of variance of extraction method, cultivar, and interaction for smoke-related volatile compounds (α = 0.05). Df, degree of freedom.
As expected, juice (i.e., untreated and unfermented samples) contained the lowest concentrations of all compounds except eugenol (Table 2), indicating that most smoke-related compounds exist as glycosylated precursors in the fruit, as previously reported (Hayasaka et al. 2010, Szeto et al. 2020, Favell et al. 2022, Rochfort et al. 2024). Juice contained moderate levels of free volatile phenols (e.g., guaiacol, 1.5 to 75.3 ng/mL; m-cresol, 2.1 to 4.0 ng/mL), consistent with the previously reported range for smoke-affected fruit (≤10 ng/g in field surveys [Liu et al. 2020], up to ~100 ng/L under heavy smoke [Szeto et al. 2020]). As expected, concentrations rose after precursor liberation and during fermentation, reflecting release from conjugates (Kennison et al. 2008). Although acid hydrolysis generally released higher concentrations of smoke taint compounds, guaiacol was an exception for Cabernet Sauvignon, in which its concentration remained relatively unchanged postdigestion. This result contrasts with previously published findings for Pinot noir, in which free guaiacol increased significantly after acid hydrolysis of smoke-exposed fruit (Wilkinson et al. 2011). Such differences may reflect cultivar-specific variations in glycosylation pathways or timing and type of smoke exposure.
Mean volatile phenol compound concentration of each method and cultivar. CF, Cabernet franc; CS, Cabernet Sauvignon; PV, Petit Verdot.
Across the multiple cultivars, acid hydrolysis consistently yielded higher concentrations of volatile phenolic compounds, including eugenol, 4-EG, o-cresol, and p-cresol. Consequently, strong-acid hydrolysis (≈pH 1, heated) can alter hydrolysis product composition by changing reactivity, as the conditions may cleave the ether linkage as well as the glycosidic linkage (Sefton 1998). In another study, acid hydrolysis was systematically optimized for smoke-related volatile phenols, with reported spike-recovery accuracy of 64 to 124% and repeatability of 2 to 14% in spiked matrices (Noestheden et al. 2017). Nevertheless, interlaboratory studies have highlighted variability in acid-hydrolysis outcomes (influenced by acid type, vessels, and handling), underscoring reproducibility challenges (Favell et al. 2022). This presents a critical challenge when assessing risk: while some producers may wish to understand a worst-case scenario, such overprediction can misrepresent the true severity of smoke taint. In contrast, other situations may require a more accurate assessment of smoke taint risk to prevent unnecessary fruit rejection or waste.
Although results were consistent across cultivars and smoke taint aroma release methods, compound release varied depending on the hydrolysis method used (Figure 1). For example, the concentration of guaiacol formed by acid hydrolysis in Cabernet Sauvignon samples was similar to that observed in Cabernet Sauvignon juice, whereas in the other cultivars, it more closely resembled the levels produced by enzyme hydrolysis and microvinification. This could indicate pathway differences across cultivars (such as differences in glycosylation) or could merely result from our small sampling subset (Wilkinson et al. 2011). Although enzymatic hydrolysis and microvinification both release volatiles related to smoke taint, their mechanisms differ. Enzymatic hydrolysis targets specific glycosidic bonds, whereas microvinification involves broader enzymatic activity from yeast as well as alcohol extraction and other fermentation-related effects (Hayasaka et al. 2010). Across the six independent samples, enzyme hydrolysis produced volatile phenol concentrations closer to those resulting from microvinification than to those resulting from acid hydrolysis in paired within-sample comparisons. In contrast, acid hydrolysis systematically over-released relative to microvinification.
Mean concentration (ng/mL) of guaiacol and m-cresol in Cabernet franc, Cabernet Sauvignon, and Petit Verdot. Samples are untreated juice (Juice), juice after acid hydrolysis (Acid), juice after enzymatic hydrolysis (Enzyme), and juice after microvinification (MV) for grape samples from two different vineyards. The horizontal dotted lines at 84.38 ng/mL (A to C) and 47.76 ng/mL (D to E) indicate the concentration observed in a severely smoke-affected Cabernet Sauvignon wine (Oberholster et al. 2023). The horizontal dashed lines at 23 ng/mL (A to C) and 20 ng/mL (D to F) indicate the aroma thresholds of guaiacol and m-cresol, respectively (Parker et al. 2012). Each field sample was divided into four separate aliquots with each digested separately using each method (n = 4 [replicates per individual treatment]; 16 individual values if including all samples within vineyard and cultivar), with all samples analyzed in triplicate. Error bars indicate standard deviation of biological replicates.
Recently, novel classes of compounds not derived from glycosylation, such as thiophenols and naphthalenethiols, have been detected at elevated concentrations in smoke-afflicted wines (Tomasino et al. 2023). These compounds are unlikely to be released in representative amounts through indirect methods such as enzymatic or acid hydrolysis, which do not replicate the full conditions of commercial fermentation. In contrast, microvinification may enable the extraction and detection of these non-glycosidically-bound smoke taint-related compounds; however, this application warrants further investigation. The ongoing discovery of new smoke-related precursors suggests that additional unidentified compounds may also contribute to smoke taint expression. Given this uncertainty, actually fermenting grape samples under realistic winemaking conditions may currently represent the most comprehensive and reliable approach for assessing risk of smoke taint.
Although detailed mechanisms lie beyond this study, cultivar and other factors likely shape glycosylation patterns of smoke taint compounds and warrant further investigation. For red wines, skins must be included in any digestion because they contain a portion of the glycosylated precursor pool; extraction efficiency therefore directly influences measured volatile levels (Ristic et al. 2011). Microvinification could be particularly advantageous for this because alcohol strength, yeast enzymatic profile, and maceration time can be tuned to the specific fruit (cultivar, vineyard, and ripeness), whereas acid or enzymatic hydrolysis require externally adjusted conditions. By leveraging endogenous fruit and yeast chemistry, microvinification more accurately mimics commercial fermentations (Warren et al. 2025). Surprisingly, even with skins present, microferments sometimes showed the lowest concentrations of smoke taint volatiles, suggesting that skin material may play a complex role in the evolution and expression of volatile phenols during fermentation. Furthermore, minimizing skin contact can reduce extraction of smoke taint-related volatiles and glycosides localized in grape skins and has been reported to lessen smoke characters in finished wines (Ristic et al. 2011). Our data set does not test this intervention; the present study isolates differences attributable to preparation workflow within the same smoke-affected material.
The aroma thresholds of guaiacol and m-cresol are shown (Figure 1; Parker et al. 2012). Because traditional winemaking conditions are more closely mimicked by microvinification than by hydrolysis methods, microvinification is expected to release smoke taint-related aroma compounds from their glycosides in a manner more representative of commercial fermentation. Because of this difference and the other reactions mentioned above, the use of acid hydrolysis can produce information on the final concentrations of these compounds that might not accurately represent winemaking conditions (Sefton 1998). For example, in the first Cabernet franc sample, the concentration of m-cresol determined by using acid hydrolysis (58.25 ng/mL) was much higher than the detection threshold (20 ng/mL). Using this early detection method for smoke taint aroma compounds could prompt drastic corrective actions or even lead to discarding the product before fermentation. Using this same sample, the concentration of m-cresol resulting from microvinification was much closer to the detection threshold (20.54 ng/mL), illustrating that less drastic action would be needed, as blending the afflicted grapes with others could reduce the concentration to below the threshold of detection. The dotted horizontal lines in Figure 1 indicate the guaiacol and m-cresol concentrations in a wine made from severely smoke-afflicted Cabernet Sauvignon grapes (Oberholster et al. 2023). This demonstrates that the concentrations of guaiacol in our samples of Cabernet franc and Cabernet Sauvignon are quite similar to those in that severe example. Because all fruit was smoke-affected, absolute concentrations should be interpreted as impact-only measures; the primary inference is the relative effect of preparation within the same material. A limitation of the present work is the omission of syringol-class markers; assessing their response across the different preparation approaches will be valuable for future work.
A critical point of comparison among preparation workflows is within-workflow variability. Acid hydrolysis showed the highest coefficients of variation (CV), often by an order of magnitude relative to the other treatments (Figure 2). For example, across cultivars for o-cresol, the mean CVs were 111.3% for acid hydrolysis, 9.8% for microvinification, and 6.8% for enzyme hydrolysis. Sensitivity of acid hydrolysis to protocol details has been documented, including acid type (H2SO4 versus HCl) and reaction vessels (glass versus polytetrafluoroethylene), and this contributes to interlaboratory variability (Noestheden et al. 2017, Favell et al. 2022). In addition, the harshness of strong-acid conditions (≈pH 1, heated) can promote isomerization and unintended bond cleavage of precursors, which may alter apparent release of volatile phenols (Arnous and Meyer 2009, Dziadas and Jeleń 2016). Across the other smoke aroma compounds measured (except 4-methylguaiacol), acid hydrolysis consistently had the highest average CV. This is an issue both for research and commercial applications, as a larger number of sample analyses would need to be conducted to compensate for the greater variation.
Average coefficient of variation (CV) of each extraction method for volatile compounds in juice (Juice), acid hydrolysis (Acid), enzyme hydrolysis (Enzyme), and microvinification (MV) (n = 6 cultivars, each with four replicates for the CV). Error bars indicate standard error of the CV within the treatment.
It should be noted that most vineyards and wineries do not have on-site GC-MS/MS capacity and typically send samples to service laboratories. A key outcome of this work is that sample-preparation choices are not interchangeable: different preparations (e.g., direct/free volatiles versus acid- or enzyme-mediated hydrolysates) emphasize different portions of the bound precursor pool and can yield divergent risk classifications even when the analytical platform is held constant. Consistent with guidance from the Australian Wine Research Institute, microvinification can support on-site sensory risk screening and in-mouth β-glucosidase activity may increase perceived smoky/ashy attributes relative to analytical measurements (Parker et al. 2012). In the present work, 50-mL microvinifications were employed to compare preparation workflows only; at this scale, replicated sensory screening was not attempted. When sensory screening is intended, larger volumes (≈100 mL or greater) would allow evaluation by multiple assessors.
Conclusion
This study highlights the complexity and variability associated with indirect methods for assessing the severity of smoke taint-related compounds in smoke-afflicted grapes. Acid hydrolysis—although effective at releasing higher concentrations of previously glycosidically-bound precursors—often results in overestimation, poor reproducibility, and potential chemical alterations that may not precisely reflect fermentation chemistry. In contrast, microvinification and enzyme hydrolysis both offer more nuanced and cultivar-representative insights through the reduction of method variability. The cultivar-specific differences observed, particularly regarding compound release and glycoside behavior, underscore the need for specific assessment strategies that consider grape variety and the severity of the smoke affliction. As new smoke-related compounds are identified (e.g., thiophenols), more integrative approaches, including microvinification, may offer the most reliable means for predicting smoke taint in the resulting wine. Future studies should attempt to refine these methods and explore cultivar-specific biochemical pathways to improve early detection and prediction of what type of smoke alleviation methods are relevant.
Supplemental Data
The following supplemental materials are available for this article in the Supplemental tab above:
Supplemental Table 1 Basic fruit chemistry of grapes used in this study. Values represent a single measurement for each trait. TSS, total soluble solids; TA, titratable acidity.
Supplemental Table 2 Gas chromatography-tandem mass spectrometry (GC-MS/MS) standards and multiple reaction monitoring (MRM) transitions.
Data Availability
All data underlying this study are included in the manuscript and its supplemental information
Footnotes
Warren ER and Kwasniewski MT. 2026. Microvinification versus hydrolysis: A comparative GC-MS/MS study for predicting smoke taint severity in winegrapes. Am J Enol Vitic 77:0770011. DOI: 10.5344/ajev.2026.25028
By downloading and/or receiving this article, you agree to the Disclaimer of Warranties and Liability. If you do not agree to the Disclaimers, do not download and/or accept this article.
- Received June 2025.
- Accepted February 2026.
- Published online June 2026
This is an open access article distributed under the CC BY 4.0 license.
References
Vol 77 Issue 1
Jump to section
Related Articles
Cited By...
More from this TOC section
Similar Articles
