Wildfires and controlled burns generate smoke particulates that could be taken up by the berries and leaves of grapes in nearby vineyards. Wines made from smoke-exposed grapes can present undesirable sensory attributes known as smoky or smoke tainted with reduced palatability and market acceptance [1
]. While the chemical composition of smoke from wood pyrolysis is rather complex, volatile phenols (VPs) are believed to be the major aroma compounds that confer the smoke taint characteristics to grapes and wines [1
Guaiacol and 4-methylguaiacol were the first recognized VPs in relation to smoke taint [1
]. Subsequently, other compounds especially cresols and syringol, were also found to be important contributors. As a result, seven VPs, including guaiacol, 4-methylguaiacol, m-cresol, p-cresol, o-cresol, syringol, and 4-methylsyringol, are currently monitored in analytical laboratories as markers for smoke taint appraisal in grapes and wines [3
]. In addition, VP glycosides can be formed through conjugation with sugars once inside berry tissues and these non-volatile compounds are precursors from which free VPs could be released during vinification or wine storage [4
]. Consequently, measuring glycosidically conjugated VPs is also essential for estimating the potential risk of producing smoke-tainted wines.
Bush fires are expected to become more frequent in the coming years with global warming [1
]. In order to minimize the financial loss associated with producing smoke-tainted wines, obtaining information on the level of free and bound VPs in grapes in the pre- or post-harvest stage in a timely manner is of critical importance. Currently, gas chromatography-mass spectrometry (GC-MS) and liquid chromatography-mass spectrometry (LC-MS) are the main analytical tools used for VP analysis. For free VP quantification, although LC-MS was used in rare cases [6
], GC-MS has been the predominant analytical technique in combination with various sample preparation protocols, including liquid-liquid extraction (LLE) from berry homogenate and wine [7
], headspace solid-phase microextraction (HP-SPME) from beer [8
], polymer stir bar sorptive extraction from beer and wine [9
], and trimethylsilyl (TMS)-based derivatization following purification by solid-phase extraction (SPE) of berry homogenate and wine [10
In the case of glycosidically bound VPs, direct analysis has been reported only with LC-MS and in most studies following an SPE cleanup. In addition, due to the lack of authentic VP-glycoside standards, a single labelled internal standard was often used to estimate all VP-glycoside species [11
]. Given the large number of potential VP glycosides in grapes [5
], accurate quantification of even the major species is difficult to achieve. As a result, indirect quantification of the major aglycones after hydrolysis is a simpler option and such an approach was already attempted in several reports [13
Currently, there is neither a consensus method concerning sample preparation for GC-MS analysis of free VPs in grape/wine samples, nor a standardized protocol for hydrolysis of VP-glycosides. The aim of this study was to develop and validate a simple GC-MS/MS method for both free and bound VP determination in grapes with optimized extraction and hydrolysis procedures. Moreover, nine additional VPs, including phenol, 4-ethylguaiacol, 4-ethylphenol, 2,4-dimethylphenol, 4-n-propylphenol, eugenol, isoeugenol, vanillin, and acetovanillone, that may also contribute to the smoke-tainted flavor of wine were included in this method.
Developing a simple analytical method enabling a timely delivery of information regarding the VP content of smoke-exposed grapes is of increasing importance to grape growers and wine-making industries. GC-MS is known to be suitable for VP analysis and has been used for measuring smoke taint-related VPs in wine, beer, and grapes [2
]. Through optimization of GC-MS/MS parameters, we were able to achieve an LOD around 1 ng/mL for all VPs, which is appropriate to estimate the level of VPs in grapes, which could potentially lead to smoke taint flavor of wine, given that the sensory limit for smoke taint of wine was between 20 and 100 ng/mL for the most studied VPs guaiacol and 4-methylguaiacol [4
Regarding sample preparation for GC-MS analysis, Allen et al. [10
] reported a fully validated method for phenolic compound analysis in grapes, but the method is rather complex because of the extra SPE purification and TMS derivatization steps. A rather simple LLE approach was adopted in several studies to transfer VPs from the supernatant of neat berry homogenate or wines to the organic phase amenable for direct GC-MS analysis [14
]. This is by far the easiest sample preparation protocol for VP analysis by GC. However, while the LLE protocol was systematically evaluated previously [17
], the first step of extraction (i.e., extracting VPs from grape berry homogenate) remained to be optimized.
Given that aqueous methanol or ethanol is generally used for metabolite extraction in plant metabolic analysis [22
], we tried various extraction regimes by varying the amount of methanol added and the incubation temperatures; a combination of adding 20% methanol (2 mL into 10 g homogenate) with 60 min of incubation at 80 °C was found to give the best overall yield for both free and bound VPs. Indeed, the extraction of phenolic compounds from grapes by adding methanol to the homogenate was described by Allen et al. [10
]. We found that 50% methanol is more efficient for extracting free phenolic compounds, but the presence of a large proportion of methanol hinders the following LLE step; whereas the presence of up to 20% methanol in the extract matrix has no adverse effect on LLE, as evidenced by the near complete recovery (around 90%) of the majority of the spiked VP standards (data not shown).
When VP standards were spiked into the berry homogenate, a relatively low recovery (70–85%) was observed for most VPs. This can be explained by the high ratio of fruit tissues (10 g) to the liquid in the extraction medium. A recovery of 48-66% was also reported by Noestheden et al. [17
] for cresols and syringol. Indeed, a second extraction of the residues with 5 mL of 20% methanol could increase the recovery by 10–15%. Clearly, a single extraction is simpler but could underestimate (up to 10–15%) the level of free VPs; whereas two extractions will inevitably increase the sample processing time and reduce the throughput. Some VPs, such as 4-ethylphenol and 4-n
-propylphenol, showed a rather low recovery (<60%) even with the 20M extraction method; the underlying cause remains to be investigated. Allen et al. [10
] reported a satisfactory recovery (80–110%) of all VPs spiked into Merlot wine, but data is scarce in relation to the recovery of VPs from berry homogenate [17
Smoke taint-related VP-glycosides in grapes and wines have been analyzed by LC-MS in several reports [5
]. Due to the lack of authentic standards and the large number of VP-glycoside species in grapes [5
], accurately measuring even the major VP-glycosides is a challenging task. In addition, purification of crude extract by SPE is usually required to minimize the ion suppression encountered in MS detection [5
]. By contrast, the number of aglycones that are relevant to smoke taint is much smaller and the standards are mostly available. Consequently, we adopted the strategy of measuring the aglycones using GC-MS/MS following a hydrolysis step.
Various hydrolysis procedures have been reported for glycosylated phenols, and acidic hydrolysis with HCl or H2
is the most widely used method for this purpose. In the literature related to phenol glycoside hydrolysis for smoke taint analysis, the pH value rather than the acid concentration is frequently given [4
]. Given that the cleavage of glycosidic linkages depends on three factors, the acid concentration and incubation temperature and incubation time [24
], instead of adjusting the pH of samples, we tried H2
and HCl at two different concentrations each with a fixed incubation temperature and time; the concentrations were chosen based on published reports for similar purposes [10
]. We found that both 1 N HCl and 1.25 N H2
generate satisfactory and comparable results, whereas a higher concentration of these acids is not advantageous for most bound VPs. Under these hydrolysis conditions, the seven most important smoke taint-related VP compounds appear to be largely stable. However, such a hydrolysis protocol is not suitable for the determination of conjugated isoeugenol; the reason for the very low recovery of isoeugenol in heated acidic medium remains to be determined.
Incomplete hydrolysis of phenol glycosides after a 60-min incubation has been reported by some researchers [14
]. In this study, we found total disappearance of three model glycosylated VPs upon hydrolysis by 1 N HCl and 1.25 N H2
for 60 min. In addition, our acid hydrolysis was performed on samples post-LLE (i.e., using aqueous fraction instead of the whole extract) and therefore enabled a clearer differentiation between free and bound VPs. Hexose-guaiacol standard is commercially available, whereas the identity of the other two glycosides is confirmed by their MS2 spectra in comparison to the published data [5
]. Procuring more phenol glycoside standards is needed to enable a thorough investigation of the conversion rate under the current hydrolysis conditions.
A small-scale inter-laboratory comparison revealed that our method generates comparable results as compared to other commercial laboratories that provide a smoke taint analysis service to grape growers. The small difference in VP content observed across different laboratories may result from the different workflows being used.
When our method was applied to samples having different degrees of bush fire smoke exposure, the level of bound VPs appears to reflect the magnitude of the smoke effect better, while the level of free VPs was low regardless of the sample type. This indicates that measuring bound VPs is necessary for determining the potential risk of producing smoke-tainted wine from smoke-exposed grapes.