Under-vine Zulla Cover crop: Effect on Glycosidic Aroma Precursors of Vitis vinifera L. cv Syrah Musts

Scientia Horticulturae 323 (2024) 112535 Available online 3 October 2023 0304-4238/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Under-vine Zulla cover crop: Effect on glycosidic aroma precursors of Vitis vinifera L. cv Syrah musts María Pilar Segura-Borrego a , Susana Tejero a , Bel´ en Puertas b , Eva Valero c , Cristina Ubeda a , * , María Lourdes Morales a a ´ Area de Nutrici´ on y Bromatología, Dpto. Nutrici´ on y Bromatología, Toxicología y Medicina Legal, Fac. Farmacia, University Sevilla, Sevilla, Espa˜ na b Instituto de Investigaci´ on y Formaci´ on Agraria y Pesquera (IFAPA), Rancho de la Merced, Jerez de la Frontera 11471, Espa˜ na. c ´ Area de Nutrici´ on y Bromatología, Dpto. Biología Molecular e Ingeniería Bioquímica, University Pablo de Olavide, Sevilla, Espa˜ na ARTICLE INFO Keywords: Cover crop Zulla Aroma precursors Syrah grape Organic crop ABSTRACT The employment of cover cropping has been developed in vineyards to mitigate soil erosion and the massive use of herbicides, but the effects on crop yields and berry quality have been little explored. Glycosidic aroma precursors constitute the aroma potential of grapes and have remarkable relevance in neutral grapes, such as Syrah. These compounds tend to be affected by environmental conditions and stressors; therefore, the influence of cover crops on aroma precursors has been studied for the first time. For this purpose, bound aroma compounds from Syrah cv. must obtained from conventional and organic vineyards, subjected to soil tillage and to the influence of different amounts of Zulla cover crop, were extracted and analysed. This experiment was carried out over three consecutive harvests (2019, 2020 and 2021). 40 aglycones were detected, mainly C 13 -norisoprenoids and terpenes. The last harvest (2021) accounted for the highest amount of precursors, owing to the exceptional amount of isodurene, naphthalene and 2-methylnaphtalene. The year Zulla was planted (2019), treatments with lower Zulla “pressure” increased aroma precursors. In the second vintage (2020), the general impact of Zulla was negative. Decreasing the amount of aglycones with the largest amount of cover crop increased the presence of aglycones in the must. In the last harvest (2021), the two highest levels of Zulla resulted in an increase of the aromatic precursors determined in the must. These results show that the effects of the presence of Zulla are strongly vintage-dependent and probably on the development of the roots. Furthermore, the last vintage results indicate that the employment of Zulla cover crops could be favourable for varieties such as Syrah, a neutral aroma grape, where bound forms are largely responsible for the potential aroma in the resulting wines. 1. Introduction Climate change is challenging economic systems worldwide, forcing substantial investment in order to adapt to this worrying situation, however, no sector of the economy is more reliant on climate than agriculture, which is implementing several changes in order to maintain yields, production and food quality to ensure it meets consumer demand (Carter et al., 2018). Consumers, being aware of this, increasingly demand products that are produced using ecological or eco-sustainable methods. This is the case in the winemaking industry, which has made considerable efforts to adapt wine production to ecological requirements while maintaining product quality. In this context, from 2005-2019, the certified organic vineyard surface area increased by an average of 13% per year, while the ‘non-organic’ vineyard area decreased by an average of 0.4% per year within the same timeframe (OIV, 2021) reflecting consumer interest. According to the OIV, organic viticulture protects consumer health and the environment and is based on three principles: soil fertility, maintaining biodiversity and pest control, in accordance with ecological cycles and processes. Thus, herbicides cannot be used to prevent the growth of vegetation between rows of vines and, therefore, tillage is presented as the most frequently used alternative to eliminate it. However, this practice has some drawbacks, such as increased soil erosion and alteration of soil microbiota, among others (Abad et al., 2020). To avoid these negative effects, the employment of cover cropping has been developed in vineyards to mitigate soil erosion, water demands, the massive use of herbicides and to reduce the need for fertilisers (Celette et al., 2009; Le Bissonnais et al., 2004; Delpuech et al., 2018; Otto et al., 2020). Moreover, the soil quality is improved (Abad * Corresponding author. E-mail address: [email protected] (C. Ubeda). Contents lists available at ScienceDirect Scientia Horticulturae journal homepage: www.elsevier.com/locate/scihorti https://doi.org/10.1016/j.scienta.2023.112535 Received 17 May 2023; Received in revised form 5 September 2023; Accepted 24 September 2023 Scientia Horticulturae 323 (2024) 112535 2 et al., 2023). Hence, the ecological effects of cover crop are clear, but the effects on crop yields and berry quality are still opaque. Attending to plant development, cover crop competes with the grapevine for soil resources, decreasing any excessive vegetative vigour of the grapevine (Delas, 1996). Greater grapevine water stress may induce a significant decrease of grape growth and yield (Smart et al., 2006) and, therefore, of quality. These yields decreased and a slight increase in the berry quality was observed by Pou et al., (2011). Gatullo et al. (2020) observed that after the study of three consecutive vintages, only in the driest one was the yield reduced by the presence of the cover crop. Compatible with this are other studies reporting slight differences in the yield obtained using cover crops compared to the control (Ingels et al., 2005, Ferrara et al., 2021). Thus, the study of P´ erez-´ Alvarez et al. (2015) with the grape cv. Tempranillo showed that, in the midterm (3 years), a gramineous cover crop reduced the nitrogen grapevine status and both YAN and free amino acids in the grapes, being potentially useful when a reduction in vigour and an improvement in the must quality are required. Also, this competition leads to a reduction in leaf area and may be useful in diminishing the vine water consumption at later stages (veraison and ripening), which can be of interest to ensure grape yield during dry years or to reduce irrigation necessities (Pou et al., 2011) and can also delay veraison (Parker et al., 2014). However, it has many other implications, such as a possible reduction in grape phenolic compounds (Kliewer and Dokoozlian, 2005) or an increase in herbaceous aromas in grapes and wines. Moreover, it has also been reported in table grapes that cover crops could be useful in improving soil nitrogen availability, and, consequently, contribute to a more efficient nitrogen uptake and utilisation overall (Ferrrara et al., 2021). Hence, it has been observed that the use of cover crops reduces must acidity and increases berry sugar accumulation, tartaric acid/malic acid ratio and berry skin total phenols and anthocyanin content, enhancing wine quality (Spayd et al. 2002; Monteiro and Lopes, 2007) whilst other studies reflect only few differences in juice Brix, pH, or titratable acidity (Ingels et al., 2005). In reference to the study of secondary metabolites responsible for wine quality, such as phenolic and volatile compounds, only a slight influence of cover crops on colour has been reported (Gatullo et al., 2020) or no improvement in the anthocyanins, total polyphenols and tannins content (Pou et al., 2011) has been observed. Volatile compounds are also secondary metabolites that have a major influence on wine aroma. Volatile composition of the grape is comprised of free volatile compounds, but also by non-volatile precursors, of which there are several types, glycosidic being the most abundant (Bayonove, 2003). Few authors have explored the influence of cover crops on the free aroma of grapes. Xi et al. (2011) analysed the effects of permanent green cover crops on Cabernet Sauvignon vines from China and reported that wines produced from grapes with cover crops had higher contents of volatile compounds. Conversely to these results, Coletta et al. (2021) worked with the Negroamaro grape cv. from Italy and observed that the employment of cover crops was counterproductive in terms of the amount of volatile compounds found in the grape, with tillage being the best option. The study of Yuan et al. (2015), who also described a diminution effect on aroma precursors employing grass cover crop with Pinot noir grapes, agreed with these negative results. In Spain, to date, only two studies about this subject have been published, one evaluated wines from Mencía grape cv. with permanent cover crops, concluding that this practice slightly influenced the concentration of volatiles (Bouzas-Cid et al., 2018) and, very recently, the other, carried out by our research group, analysed the volatile composition of must subjected to different levels (inter-row series) of Zulla (Hedysarum coronarium L.) a wild legume from the Mediterranean basin (Tava et al., 2021), which grows spontaneously in the area of the experiment. Here we found that the effect of the harvest on the musts volatile composition of the Syrah grape variety is more important than agronomic practices during the three years studied (Valero et al., 2022). There is no existing information about bound volatile fraction. These compounds are predominantly located on the grape skin and are comprised of an odorous molecule (volatile compound) named aglycone associated with a disaccharide. The β-glycosidic bond linking them is split during the winemaking process, by the action of yeast during the fermentation, or by acid hydrolysis during the aging process, releasing the volatile compound into the wine matrix (Bayonove, 2003). This odourless fraction in the berries has special interest because mature berries have more bound forms than free forms of these compounds (Sanchez Paloma et al., 2007, Fenoll et al., 2009). However, as far as we know, there are no studies of the analysis of precursors, despite the fact that glycosidic aroma precursors are one of the compounds that are most influenced by viticultural practices and are also highly influenced by water, sunlight, and other types of vine stressors (Hernandez-Orte et al., 2015; Hjelmeland and Ebeler, 2015; Alem et al., 2019). The importance of these aroma precursors lies in the fact that they constitute the aroma potential of grapes, and have remarkable relevance to neutral grape varieties, such as Syrah, which have no aroma themselves, and the glycosidic fraction of the wine aroma is proportionally more important than the free fraction, revealing its aromatic potential during and after the fermentation, in part owing to these molecules (L´ opez-Tamames et al., 1997; Bureau et al., 2000; Segurel et al., 2009) being indicated as essential contributors to high quality Shiraz wine (Abbott et al., 1991). Considering all these previous data we hypothesise that the employment of Zulla cover crop would produce certain stress to the vine, increasing aroma precursors in the grape must for vinification. To validate or refute it, firstly, Syrah cv. musts from conventional and organic vineyards (subjected to soil tillage) were compared in terms of their aroma precursors composition. Secondly, the influence of different amounts of Zulla in organic vineyards was studied. This study represents the first, analysing the effects of cover crops on the glycosidic precursors profile in grapes. 2. Materials and methods 2.1. Materials and chemicals Dichloromethane, ethanol, and methanol were supplied by Merck (Darmstadt, Germany) and ethyl acetate by Fluka (Buchs, Switzerland). Sodium dihydrogen phosphate 1-hydrate and citric acid were purchased from Panreac (Barcelona, Spain). Pure water was obtained from a Milli- Q purification system (Millipore, Bedford, MA, USA). 4-Methyl-2-penta- nol, used as internal standard (IS) was purchased from Merck (Darmstadt, Germany) and sodium chloride from Panreac (Barcelona, Spain). An alkane solution (C10–C40) was used to calculate the linear retention index (LRI) (Fluka; Madrid, Spain)). Chromabond® HR-P cartridges (MACHEREY-NAGEL, Düren, Germany), with 200 mg of resin, were purchased from Merck VWR. 50/30 μ m DVD/CAR/PDMS (1 cm) SPME fibre was supplied by Supelco (Bellefonte, USA). 2.2. Samples The study was conducted over a period including three consecutive vintages, 2019, 2020, and 2021. The samples of this study were musts from Vitis vinifera L. cv Syrah manually harvested grapes from two vineyards (conventional and organic), located in the IFAPA Centre “Rancho de La Merced”, in the Jerez winegrowing region, Spain (36:45:29N, 06:00:58W, 35 m altitude) and are the same as those employed by Valero et al. (2022) for free volatile profile determination. The vineyards were planted in 2014 and the vines grafted on 140-Ru rootstocks with a planting distance of 2.5 ×1.5 m, a North-South row orientation, and the trained system was single cordon double Guyot. The soil called “Albariza”, characteristic of this zone is a Leptosol (IUSS Working Group WRB, 2022) with a clayey texture (69% clay, 20% silt and 9% sand), with pH of 8.0 and organic matter content of 1.16%. These two vineyards were separated by a distance of 2 km and are included in region V according to the climatic classification of viticultural regions by Amerine & Whinkler (1944). No irrigation was administered in the vineyards. Climatological data was taken from the M.P. Segura-Borrego et al. Scientia Horticulturae 323 (2024) 112535 3 vineyard weather station (Table S1). As a control, a vineyard with bare soil by tillage was employed (SC). The conventional vineyard was treated with agricultural chemicals used as herbicide, insecticide, nematicide and fungicide. The organic vineyard (SE) was treated only with micronised sulphur and copper oxychloride as fungicide. This same organic vineyard was used for the experiments with Zulla (Hedysarum coronarium L.) cover crops which were planted (20 Kg/Ha) in lanes as described in Fig. 1 in October 2018, and were maintained until March 2019. At this time, it was cleared and a week later the lanes were tilled to avoid the cover ground competing with the vineyard. In the second and third years (2019-20 and 2020-21), the cover crop was allowed to grow spontaneously, with no new planting of Zulla, and the same process of clearing and soil tillage was carried out. Bunches were always collected from the same plant at a similar ripening stage (approx. 24 ◦Brix). Thus, every year, 15 grape samples were collected, 12 grape samples in the organic vineyard (SE), for every cover crop density (3 samples for every different density) and 3 grape samples in the conventional vineyard (SC). The different cover crop levels studied were cover crop on one side and tillage on the other side of the row (LZ), cover crop on both sides of the row (ZZ), cover crop on two rows on both sides of the sampled row (4Z) and tillage on both sides of the vine row as a control (LL). Approximately 2 kg of grapes, stems included, were harvested in aseptic conditions from each sampling point and placed directly into sterile bags, which were transported to the laboratory in portable refrigerators with plastic ice blocks. At the laboratory, grapes were squeezed by hand in the plastic bags, opened and about 50 mL of juice was poured into a glass vial and immediately frozen at -20◦C for further analyses (Valero et al., 2022). 2.3. Extraction of glycosidically bound aroma compounds by SPE The night before the extraction, must samples were thawed in the refrigerator at 5◦C and were centrifuged at 4000 rpm for 10 minutes. For the extraction and analysis of precursors, the method employed by Ubeda et al. (2017) was followed with some modifications. Thus, 10 mL of the supernatant were passed through 3 mL Chromabond® HR-P cartridges (MACHEREY-NAGEL, Düren, Alemania) containing 200 mg of resin previously conditioned (5 mL dichloromethane +5 mL methanol +5mL of milliQ water) employing a Visipred SPE vacuum manifold (Supelco, Bellefonte, PA, USA). Once the glycosidic aroma precursors were adsorbed into the resin, 10 mL of Milli-Q water were passed through the cartridge to eliminate interferences and vacuum dried. Subsequently, 10 mL of dichloromethane were employed to eliminate free volatile compounds and finally the precursors were eluted with 10 mL of a solution with ethyl acetate:methanol 9:1 (v/v) collected in a conical flask, and subsequently removed with a rotary evaporator (IKA, Staufen, Alemania). To recover the precursors, 10 mL of citrate buffer were used (0.2 M, pH 2.5) and were transferred to a 20 mL screw cap vial. The oxygen contained in the headspace (HS) of the vial was removed with a nitrogen stream prior to acid hydrolysis which was carried out by incubating for 1 hour at 100 C◦in a stove (Selecta, Barcelona, Spain) releasing the aglycones to the matrix. After this stage, vials were tempered to subsequently proceed to the analysis. 2.4. Analysis of volatiles released from precursors by HS-SPME-GC-MS For the analysis of aglycones, 7.5 mL of hydrolysate sample were placed in a 20 mL SPME vial and 2.25 g of sodium chloride were subsequently added to the sample joined to 10 µL of 4-methyl-2-pentanol (1045 mg/L) as internal standard. The vials were kept at 20 ◦C in a thermostatised autosampler tray until analysis. For this purpose, an Agilent 8890 gas chromatograph coupled to an Agilent 5977B simple quadrupole mass spectrometer (Agilent, Santa Clara, CA, US) equipped with a multipurpose autosampler MPS Robotic Pro LS (Gerstel, Müllheim an der Ruhr, Alemania) was used. Samples were incubated at 45◦C for 10 minutes at 300 rpm and the fibre was subsequently exposed to the vial headspace for 40 minutes (Ubeda et al., 2017). Desorption of the compounds adsorbed to the fibre was performed for 3 minutes in the injection port at 250 ◦C, on splitless mode at 50 mL/min. A J&W CPWax-57CB 50 m x 0.25 mm and 0.20 μ m film thickness (Agilent, Santa Clara, CA, US) was used. Helium was employed as carrier gas at 1 mL/min. The oven temperature programme was as follows: started at 35◦C for 2 minutes, increasing by 5◦C/min to 50◦C, followed by a temperature ramp at 2◦C/min to 200◦C, maintained for 1 minute. Finally, the temperature was raised by 5◦C/min to 220◦C, for 2 min. Detection of volatile compounds was performed on full scan mode at 70 eV and a mass registration from 29 to 300 m/z. The MS quadrupole, source, and transfer line temperatures were maintained at 150 ◦C, 230 ◦C and 280 ◦C, respectively. 2.5. Identification and quantitation of volatile compounds Compound identification was performed using the NIST Mass Spectral Search program for the NIST/EPA/NIH EI and NIST Tandem Mass Spectral Library (v.2.3, 2017) (Gaithersburg, MD, US). LRI values were calculated by injecting an n-alkanes mixture (C10–C40) under identical conditions as the samples. Subsequently, identification was performed at three levels. Identification was carried out in accordance with the matching of the linear retention indices (LRIs), and the LRIs from the literature obtained with authentic standards and their mass spectra, with those from the compounds NIST library (A). The second level was for the compounds that match with the NIST library and the LRIs from the literature (B). The third level was for the compounds that match with the NIST library, however their LRI does not coincide with the data found in the literature (C). Data were expressed as relative peak area values with respect to the internal standard (4-methyl-2-pentanol). Fig. 1. Diagram of the treatments carried out in the organic vineyard. (SELZ) cover crop on one side and tillage on the other side of the row; (SEZZ) cover crop on both sides of the row; (SE4Z) cover crop on two rows on both sides of the sampled row; (SELL) tillage on both sides of the vine row as a control. M.P. Segura-Borrego et al. Scientia Horticulturae 323 (2024) 112535 4 2.6. Statistical analyses The values of the relative peak area of the diverse volatile compounds found in each agronomic treatment were subjected to analysis of variance (ANOVA) (Tukey’s test) (p>0.05)) using INFOSTAT software (FCA, Universidad Nacional de C´ ordoba, Argentina). The Principal Component Analyses (PCA) were performed using SIMCA® Multivariate Data Analysis software 14.1.0.2047 (Umetrics Umea, Sweden). Pareto scaling was the method selected to normalise the data of the technique before the PCA analysis. Main effects ANOVA was also performed using Statsoft Statistical, version 7.0 (Statsoft, Tulsa, OK). 3. Results and discussion A total of 40 aglycones were detected, 13 norisoprenoids, 11 terpenes, 7 aldehydes, 4 alcohols, 2 ketones and 3 non-identified compounds. This distribution was expected, given that C 13 -norisoprenoids and terpenes are the glycosidically-bound precursors most usually present in the grape and in higher amounts, and are mainly accumulated in the berry skin during the ripening process. The number of compounds found in each harvest was slightly different, hence, in 2019 and 2020 there were 36 precursors, and 32 in the 2021 harvest (Table 1). It is important to consider the variability between the 2019 and 2020 harvests and that of 2021. As can be observed at the end of Table 1, the 2021 vintage accounted for a much higher amount of total volatile precursors, contrary to the results obtained in the free volatile fraction of the same samples (Valero et al., 2022). This is mainly because in 2021 an outstanding amount of isodurene, naphthalene and 2-methylnaphtalene was observed. The results are organised firstly by comparing the effect of the organic and conventional crop, with both subjected to soil tillage, and, secondly, by analysing the results of the glycosidic precursors composition of the musts from the organic vineyard grapes grown with different proportions of Zulla cover crop. Regarding the compounds observed, the 2019 and 2020 harvests were similar; however, the 2021 harvest was the most different, principally owing to the presence of the above-mentioned amounts of isodurene and naphthalenes. On the contrary, in 2021 a smaller number of terpenes was found, with rose oxide, neo-allo-ocimene, terpinen-4-ol, β-cyclocitral, myrcenol, and α -terpineol present only in the 2019 and 2020 harvests. Moreover, some of the compounds found in Syrah must have been detected in grapes for the first time as far as we know. This is the case of the 2,5,8-Trimethyl-1,2-dihydronaphthalene, 2-methyl-1-decanol and undecanol. 3.1. Comparison of conventional and organic crop The volatile precursors composition of musts of the Syrah grape variety obtained from conventional and organic culture (SC and SELL) during three vintages was compared. As can be observed in Fig. 2, the total amount of terpenes present in the must from the conventional cultivation vineyard was higher than in the must from the organic crop, but only significant in the first harvest. These significant differences were due to the high concentration of isodurene, cyclocitral, myrcenol, menthol, ocimenol, humulene and L- α -terpineol. These last two compounds also presented significantly higher contents in the must from the conventional crop in the 2020 harvest (Table 1). Aldehydes presented the same tendency, with more total contents in the musts from conventional cultivation vines, but only significant in the first harvest, as well as the ketones group, being, in this case, significant in the 2019 and 2020 harvests. For the alcohols, no significant differences were found between the two vineyard systems. Only in the case of norisoprenoids did the trend differ from the rest. The first year the must from the conventional vineyard accounted for a significantly higher amount of these compounds than from the organic vineyard, in the second year a significantly higher quantity of norisoprenoids was found in the organic must, mainly due to TDN and naphtalene, and finally, in the last studied year, the presence of these isoprenoids in the must from both vineyards was quite similar, with no significant differences between them. Very little research has been carried out on the comparison of wine composition of organic vs conventional crops. Moyano et al., (2009) observed that sherry wines from Pedro Ximenez grapes from organic crops showed, in general, a lower concentration of the determined volatile compounds. With respect to the analysis of free volatile compounds in musts, specifically of these same samples, Valero et al. (2022) observed that, in general, in the 2019 and 2020 harvests the total volatile compounds of every group were higher in the organic crop compared to the conventional crop. This is the inverse result to that of the aroma precursors, and a possible explanation could be a higher level of glycosylation of volatile compounds in the conventional crop vs organic, in response to chemicals applied to the vine. In the last year, 2021, this tendency was not so clear, and no significant differences were found in the total volatile compounds of every group sum between conventional and organic crops. This divergent behaviour among vintages was reflected in the results of multivariate statistical analysis (Fig. 2). Thus, in the PCA, the first two principal components were able to explain a 96.1% of the total variance of the data with PC1 separating the samples from the 2019 and 2020 harvests from the samples of the 2021 harvest. However, no differences concerning the vineyard system was observed, although there seems to be more distance between the musts from different harvests in the case of the conventional crop (Fig. 3). Moreover, according to the main effect ANOVA, the influence of harvest was more important than the crop type for a total of 28 compounds (Tables S2A and S2B). These results are in accordance with Valero et al. (2022) who concluded that an important influence of the harvest year on agronomic practices did exist. This strong dependence of the glycosides on the vintage, especially terpenes and C 13 -nor- isoprenoids, has been previously observed by other authors (Koundouras et al., 2009). 3.2. Study of Zulla cover crop effect in organic culture As described above, the control employed did not have Zulla and used tillage to avoid the growth of other plants near the vine. The Zulla treatments are described in Fig. 1 and include 3 treatments with different quantities (pressure/influence) of Zulla. The presence of Zulla produced significant differences in several volatile compounds from the musts obtained from the grapes grown in the organic vineyards (Table 1). As can be observed in Fig. 2, the tendencies in the influence of every treatment assayed in the different chemical groups described a very similar pattern within each vintage. Therefore, the Zulla “pressure” affected most of the compounds every year in a very similar way (except ketones). However, the effect of each treatment was not the same with respect to the control treatment. At the end of Table 1, the global effects of every Zulla treatment can be observed. Thus, in the first harvest (2019), a light Zulla pressure (LZ) significantly increased the total precursors present in the must. The next level of Zulla (ZZ) applied also increased precursors production, but not significantly, however, the higher level of Zulla (4Z) implied an important decrease in the volatile precursors determined versus tillage, but without significance. It is important to consider that, as previously mentioned, during this first harvest, the Zulla was planted and maintained until March 2019. At that point, the cover crop was cleared and a week later the lanes were tilled to avoid the cover ground competing with the vineyard. Therefore, the effect of the Zulla was perhaps still low. In the second and third years (2020 and 2021), the cover crop was allowed to grow spontaneously, with no new planting of Zulla. In the second year, 2020, the effect was different with respect to the preceding year. In this case, the presence of Zulla was significantly counterproductive for the aromatic precursors in all conditions. The musts from the last harvest presented significantly M.P. Segura-Borrego et al. Scientia Horticulturae 323 (2024) 112535 5 Table 1 Glycosidic aroma precursors in Syrah grape must from the 2019, 2020 and 2021 harvests grown under different agronomic conditions. SC SELL SELZ SEZZ SE4Z Compounds LRI ID HY Mean ±SD A Mean ±SD A Mean ±SD A Mean ±SD A Mean ±SD A Terpenes (T) cis-Rose oxide (T1) 1349 A 2019 8.53 1.43 ab 10.09 2.17 b 8.72 0.84 b 11.72 0.90 b 3.30 0.80 a 2020 17.59 5.57 a 10.42 3.09 a 15.17 0.63 a 12.36 2.01 a 8.25 1.39 a 2021 nd nd nd nd 0.05 0.01 nd nd nd nd Neo allo-ocimene (T2) 1358 B 2019 21.37 2.89 d 15.57 0.54 cd 13.06 1.98 bc 9.00 0.42 ab 5.12 0.15 a 2020 10.07 3.73 ab 18.40 2.47 b 6.79 1.44 a 4.65 0.86 a 10.29 2.70 ab 2021 nd nd nd nd nd nd nd nd nd nd Nerol oxide (T3) 1465 B 2019 22.89 5.70 b 22.48 1.63 b 13.57 4.29 ab 6.98 0.49 a 4.91 0.24 a 2020 20.64 3.14 b 25.85 0.61 b 8.83 0.95 a 5.77 0.69 a 7.47 2.07 a 2021 36.27 3.69 c 8.83 3.81 a 9.77 0.57 a 12.59 0.06 ab 20.32 1.25 b Isodurene (T4) 1473 B 2019 357.36 20.76 b 199.41 34.65 a 431.42 4.11 b 596.36 13.64 c 217.80 14.47 a 2020 205.50 31.06 a 238.64 62.99 a 170.63 34.95 a 124.34 1.53 a 129.92 3.68 a 2021 8228.17 1648.42 ab 5689.62 998.67 a 4245.98 283.61 a 10025.21 1313.00 b 7241.83 135.84 ab Terpinen-4-ol (T5) 1605 A 2019 0.94 0.10 a 3.69 0.03 d 3.03 0.30 c 2.18 0.04 b 1.03 0.13 a 2020 4.92 0.22 c 3.19 0.00 b 2.05 0.33 a 1.53 0.09 a 1.72 0.49 a 2021 nd nd nd nd nd nd nd nd nd nd β-Cyclocitral (T6) 1616 B 2019 13.22 0.24 c 8.16 1.16 ab 5.52 1.40 a 9.77 0.21 bc 6.13 1.35 ab 2020 9.86 1.84 bc 11.86 1.28 c 5.97 1.07 ab 4.02 0.71 a 6.67 0.19 ab 2021 nd nd nd nd nd nd nd nd nd nd Myrcenol (T7) 1619 B 2019 54.06 3.65 d 24.20 0.08 b 35.03 3.29 c 20.49 0.77 ab 10.99 2.04 a 2020 45.25 16.94 a 37.75 1.42 a 19.82 4.34 a 10.70 3.05 a 26.03 7.38 a 2021 nd nd nd nd nd nd nd nd nd nd Menthol (T8) 1646 A 2019 12.53 0.71 b 7.10 1.10 a 8.55 0.90 a 12.00 0.23 b 6.66 0.80 a 2020 9.55 1.29 b 6.58 0.75 ab 7.00 1.18 ab 3.96 0.17 a 5.30 0.84 a 2021 92.83 10.15 ab 94.93 1.98 ab 85.50 11.24 a 142.31 3.70 c 121.82 11.10 bc α -Humulene (T9) 1663 A 2019 60.04 5.38 d 28.93 0.48 bc 39.41 3.39 c 22.86 1.14 ab 12.87 0.54 a 2020 104.43 4.07 c 41.83 0.28 b 23.49 6.06 ab 12.39 3.77 a 30.32 8.75 ab 2021 13.05 0.97 b 18.17 3.02 b 3.80 2.00 a 2.74 1.12 a 15.76 1.49 b Ocimenol (T10) 1687 B 2019 51.37 0.97 d 22.43 0.37 b 33.60 1.61 c 18.51 1.84 b 9.72 1.95 a 2020 76.96 21.43 b 37.54 3.81 ab 20.90 3.92 a 10.40 2.71 a 27.06 8.75 a 2021 24.35 3.21 b 8.15 1.14 a 21.29 1.76 b 21.87 4.34 b 25.85 4.50 b L- α -Terpineol (T11) 1708 A 2019 69.74 1.30 b 42.90 0.87 a 42.05 5.95 a 46.87 1.77 a 39.29 0.14 a 2020 74.02 16.23 b 29.72 3.36 a 21.24 3.66 a 14.21 0.13 a 15.70 5.57 a 2021 nd nd nd nd nd nd nd nd nd nd Norisoprenoids (N) Vitispirane A (N1) 1517 A 2019 988.72 42.72 b 395.91 33.76 a 959.12 289.45 b 559.58 35.04 ab 309.79 47.44 a 2020 423.60 108.40 a 507.08 91.62 ab 250.65 6.05 a 256.78 60.59 a 747.72 7.55 b 2021 413.64 16.59 b 529.12 52.52 b 250.45 23.85 a 519.72 2.47 b 460.99 39.55 b Vitispirane B (N2) 1520 A 2019 677.39 15.98 b 302.97 22.79 a 706.52 196.26 b 388.24 27.65 ab 215.04 27.80 a 2020 313.95 79.98 a 361.97 66.69 ab 186.54 11.70 a 176.39 41.66 a 523.81 5.66 b 2021 390.48 34.45 b 482.37 63.90 b 217.53 16.65 a 435.28 18.17 b 371.49 39.37 ab α -Ionene (N3) 1545 B 2019 220.63 8.62 c 125.67 17.59 ab 170.18 31.90 bc 152.86 11.17 abc 83.26 10.71 a 2020 36.41 7.14 a 210.03 17.01 c 96.52 22.92 b 30.68 8.93 a 62.16 6.84 ab 2021 116.84 6.37 ab 159.19 18.60 b 62.64 6.58 a 77.76 24.19 a 67.24 5.93 a Naphthalene (N4) 1725 A 2019 361.18 46.17 a 270.49 36.19 a 415.11 94.17 a 518.46 30.74 a 307.72 97.06 a 2020 83.51 1.22 b 433.03 17.66 c 32.63 1.04 a 30.09 2.37 a 53.43 1.52 ab 2021 24410.25 1594.00 a 19528.04 6479.96 a 19907.57 1017.05 a 39893.14 976.47 b 30488.71 272.71 ab TDN (N5) 1731 A 2019 3316.72 220.06 b 2775.62 325.86 ab 3046.05 119.40 b 2910.12 71.00 ab 1760.38 506.52 a 2020 1764.15 49.16 ab 3737.89 535.49 c 1568.75 205.90 ab 1161.96 111.21 a 2755.05 491.10 bc 2021 385.99 4.44 a 748.97 61.99 ab 400.21 67.83 a 792.83 77.79 b 758.55 176.21 ab β-Damascenone (N6) 1826 A 2019 225.39 6.86 c 64.29 0.84 a 104.04 13.78 b 85.68 0.42 ab 56.45 8.40 a 2020 179.03 41.06 b 178.33 9.84 b 127.66 39.22 ab 57.16 18.59 a 89.12 0.94 ab (continued on next page) M.P. Segura-Borrego et al. Scientia Horticulturae 323 (2024) 112535 6 Table 1 (continued) SC SELL SELZ SEZZ SE4Z Compounds LRI ID HY Mean ±SD A Mean ±SD A Mean ±SD A Mean ±SD A Mean ±SD A 2021 168.36 2.51 ab 217.29 21.13 b 108.43 12.43 a 120.61 23.76 a 135.76 3.64 a 2-Methylnaphthalene (N7) 1845 A 2019 nd nd nd nd nd nd nd nd nd nd 2020 nd nd nd nd nd nd nd nd nd nd 2021 1835.06 23.41 a 2393.40 427.39 ab 2403.32 147.21 ab 4509.13 661.68 c 3735.09 638.65 bc 1,1,6,8-Tetramethyl-1,2-dihydro-naphthalene (N8) 1896 C 2019 38.52 0.54 b 19.64 1.88 a 32.92 0.31 b 30.88 3.00 b 18.20 4.25 a 2020 33.24 6.57 bc 45.96 0.70 c 26.44 5.75 ab 14.32 2.36 a 19.27 0.02 ab 2021 10.27 0.35 a 18.17 3.02 b 12.26 0.66 ab 14.68 1.23 ab 16.01 0.27 ab β-Ionone (N9) 1947 A 2019 2.82 0.01 c 1.33 0.02 a 2.31 0.05 b 2.30 0.02 b 1.56 0.15 a 2020 2.28 0.55 bc 3.58 0.09 c 1.87 0.35 ab 0.78 0.04 a 1.35 0.33 ab 2021 58.99 1.52 a 56.78 14.91 a 52.04 1.71 a 65.42 4.81 a 73.16 6.77 a 2,5,8-Trimethyl-1,2-dihydronaphthalene (N10) 1985 C 2019 263.24 11.45 d 103.16 4.73 a 183.38 32.46 bc 234.72 4.17 cd 170.14 5.13 b 2020 25.44 5.16 a 177.57 15.30 c 86.95 16.92 c 51.65 11.88 ab 60.45 14.64 ab 2021 70.29 5.50 a 104.62 10.45 ab 77.89 7.80 ab 88.88 20.82 ab 121.81 4.12 b Dehydro-β-ionone (N11) 2009 C 2019 30.75 1.51 c 8.13 0.17 a 19.15 0.97 b 33.11 1.30 c 11.08 1.15 a 2020 28.13 9.77 b 16.56 2.03 ab 4.06 0.69 a 2.45 0.07 a 3.25 0.28 a 2021 21.02 2.41 c 0.06 0.01 a 0.06 0.02 a 0.14 0.00 a 6.96 0.88 b β-Methylionone (N12) 2143 C 2019 nd nd nd nd nd nd nd nd nd nd 2020 nd nd nd nd nd nd nd nd nd nd 2021 567.94 22.07 a 995.40 450.02 a 427.17 11.15 a 986.72 94.29 a 700.52 22.26 a 1,6-Dimethyl-4-(1-methylethyl) naphthalene (N13) 2211 B 2019 187.41 12.88 b 34.66 1.43 a 35.62 0.63 a 33.64 3.09 a 20.64 7.63 a 2020 36.80 7.39 ab 45.72 5.36 b 32.32 2.06 ab 16.06 3.42 a 32.38 10.64 ab 2021 33.18 0.13 a 62.78 28.33 a 23.17 4.64 a 67.51 15.58 a 44.64 2.25 a Aldehydes (A) Nonanal (A1) 1400 A 2019 18.30 0.44 b 10.78 1.64 a 17.48 0.75 b 11.86 2.87 ab 7.95 1.45 a 2020 33.38 9.09 b 23.64 0.33 ab 26.33 1.98 ab 11.66 0.47 ab 16.58 3.97 ab 2021 158.94 48.07 a 125.83 34.10 a 87.75 33.02 a 166.67 68.64 a 166.56 7.33 a Decanal (A2) 1492 A 2019 19.08 0.53 d 7.39 1.44 b 10.64 0.15 c 4.67 0.32 ab 4.00 0.59 a 2020 13.26 5.09 a 8.02 1.32 a 4.46 0.24 a 3.88 0.82 a 10.59 0.20 a 2021 186.69 14.05 ab 180.86 4.79 ab 131.42 17.17 a 215.85 0.74 ab 289.20 56.69 b Benzaldehyde (A3) 1507 A 2019 nd nd nd nd nd nd nd nd nd nd 2020 nd nd nd nd nd nd nd nd nd nd 2021 1074.44 327.04 a 332.77 104.70 a 506.62 86.91 a 894.01 136.46 a 674.25 248.27 a (E)-2-Nonenal (A4) 1531 A 2019 23.05 0.15 b 9.87 0.58 a 9.48 0.01 a 10.03 1.27 a 7.35 1.56 a 2020 9.33 0.85 ab 21.89 2.54 bc 13.44 3.63 abc 8.66 0.92 a 23.97 5.74 c 2021 51.80 1.33 a 96.51 34.72 a 47.31 4.71 a 74.89 20.24 a 84.02 4.66 a 5-Methylfurfural (A5) 1567 A 2019 66.39 8.65 b 27.55 0.73 a 29.60 1.57 a 22.26 1.48 a 14.28 0.93 a 2020 51.61 16.67 b 51.50 6.62 b 26.68 2.93 ab 17.41 0.14 a 30.12 2.87 ab 2021 19.99 0.92 ab 15.44 3.55 a 13.71 0.16 a 21.23 1.68 ab 26.08 1.47 b 2,4-Dimethyl-benzaldehyde (A6) 1806 C 2019 192.06 4.16 ab 132.03 0.62 a 205.01 19.90 ab 216.47 5.70 ab 156.83 38.46 a 2020 425.57 30.71 b 304.59 52.84 ab 306.10 81.32 ab 209.13 46.29 a 178.43 34.98 a 2021 663.26 100.13 a 671.81 448.98 a 631.16 26.10 a 1096.67 120.46 a 551.64 32.65 a 2,4,5-trimethyl benzaldehyde (A7) 1878 B 2019 nd nd nd nd nd nd nd nd nd nd 2020 nd nd nd nd nd nd nd nd nd nd 2021 732.38 52.17 b 539.35 47.17 a 723.84 14.95 ab 924.79 36.46 c 969.78 64.85 c Alcohols (AL) 2-Ethyl-1-hexanol (AL1) 1498 A 2019 121.01 14.00 b 81.02 1.35 a 74.88 6.15 a 74.91 2.09 a 60.78 4.23 a 2020 220.83 33.07 c 80.16 15.30 ab 130.87 17.73 b 48.26 0.63 a 120.47 6.96 ab 2021 2128.75 706.11 a 2447.21 636.25 a 2746.22 645.83 a 3069.18 147.36 a 4296.99 775.12 a 2-Methyl-1-decanol (AL2) 1808 B 2019 2.17 0.66 a 7.83 1.05 bc 2.89 0.37 a 10.72 1.13 c 5.79 1.06 ab 2020 9.46 2.06 b 21.10 1.58 c 4.12 0.01 a 16.36 0.73 c 17.81 1.04 c 2021 27.19 7.69 a 314.64 178.40 a 14.33 1.55 a 264.20 10.89 a 42.56 4.86 a Undecanol (AL3) 1883 B 2019 4.37 0.04 a 34.94 4.38 b 12.12 1.43 a 60.27 2.11 c 32.91 1.39 b 2020 18.27 0.29 a 65.36 5.66 c 12.49 0.79 a 40.84 0.37 b 66.82 10.91 c 2021 130.24 7.29 a 128.85 49.49 a 139.23 2.54 a 309.86 15.15 b 202.09 40.94 ab (continued on next page) M.P. Segura-Borrego et al. Scientia Horticulturae 323 (2024) 112535 7 Table 1 (continued) SC SELL SELZ SEZZ SE4Z Compounds LRI ID HY Mean ±SD A Mean ±SD A Mean ±SD A Mean ±SD A Mean ±SD A Dodecanol (AL4) 1980 A 2019 20.42 2.08 a 19.32 2.90 a 20.44 4.38 a 41.67 1.33 b 20.90 2.56 a 2020 41.11 3.91 b 40.27 2.05 b 33.85 7.93 ab 17.46 1.06 a 23.60 1.46 a 2021 543.22 18.99 a 580.30 256.95 a 492.37 27.80 a 913.67 91.46 a 937.20 153.19 a Ketones (K) 2,6-Dimethyl-4-heptanone (K1) 1164 B 2019 51.75 1.42 c 23.51 4.53 ab 18.38 2.58 ab 31.61 6.06 b 11.32 2.63 a 2020 61.35 4.25 c 10.50 2.28 ab 14.02 3.45 ab 29.55 1.26 b 15.64 0.90 a 2021 18.64 0.49 a 15.09 4.06 a 26.27 1.02 ab 35.05 6.42 b 25.37 0.69 ab 2-Undecanone (K2) 1594 A 2019 58.33 4.20 c 16.62 3.23 ab 16.71 2.55 ab 22.12 1.06 b 8.82 0.91 a 2020 5.17 0.90 a 6.76 1.06 a 5.82 1.55 a 2.72 0.22 a 3.50 1.00 a 2021 128.57 17.08 ab 114.09 12.36 ab 89.00 7.77 a 195.20 0.57 b 180.15 42.33 b Unknown (U) n.i. (m/z 159, 174) (U1) 1408 2019 156.89 1.25 c 77.41 20.10 ab 104.02 4.77 b 78.34 7.19 ab 40.99 6.09 a 2020 71.86 17.19 ab 110.79 24.01 b 47.65 7.84 a 19.75 3.23 a 38.84 4.09 a 2021 nd nd nd nd nd nd nd nd nd nd n.i. (m/z 159, 174) (U2) 1640 2019 75.90 7.58 a 56.98 8.12 a 72.47 1.21 a 56.07 0.62 a 61.77 11.34 a 2020 141.01 9.75 c 78.70 11.73 b 48.35 15.21 ab 20.60 4.95 a 31.28 3.86 a 2021 nd nd nd nd nd nd nd nd nd nd n.i. (m/z 159, 174) (U3) 1697 2019 108.49 1.74 c 59.62 5.12 b 57.98 2.30 b 72.82 4.77 b 39.15 5.11 a 2020 66.25 11.65 b 101.57 7.15 c 37.35 9.79 ab 19.78 5.53 a 27.16 4.23 a 2021 44.50 4.27 ab 76.85 18.47 b 34.95 1.18 a 42.87 5.64 ab 44.02 4.86 ab TOTAL PRECURSORS SUM 2019 7449.60 376.06 c 4791.65 211.29 ab 6613.80 210.68 c 6132.89 148.39 bc 3558.89 765.00 a 2020 4424.52 158.94 bc 6704.83 407.63 d 3235.32 434.59 ab 2397.69 109.48 a 5089.25 432.56 c 2021 41936.36 2774.20 ab 36056.93 7367.74 a 33206.81 2181.18 a 64998.15 812.73 c 51859.41 344.29 bc LRI: Experimental linear retention index values estimated by linear regression. ID: identification of the compound; A. mass spectrum and LRI agreed with standards; B. mass spectrum agreed with mass spectral database and LRI agreed with the literature data; C. mass spectrum agreed with mass spectral database. HY: Harvest year. Sample codes of grape musts: SC: conventional vineyard. Organic vineyard: SELL: tillage on both sides of the vines row; SELZ: cover crop on one side and tillage on the other side of the vines row; SEZZ: cover crop on both sides of the vines row. SE4Z: cover crop on two rows on both sides of the sampled row. nd: non-detected compound. Mean: mean values. SD: standard deviation. A: ANOVA test results (Different lower case letter in the same row indicates significant differences according to Tukey’s test (p>0.05) between crop treatments of the same harvest). M.P. Segura-Borrego et al. Scientia Horticulturae 323 (2024) 112535 8 Fig. 2. Total amount of volatile compounds determined in the three consecutive harvests (2019, 2020 and 2021) of the main chemical families in all conditions tested (SC, SELL, SELZ, SEZZ, SE4Z). Data are expressed as relative area. M.P. Segura-Borrego et al. Scientia Horticulturae 323 (2024) 112535 9 higher amounts of precursors when the highest pressure of Zulla was applied (SEZZ and SE4Z). In the third year this cover crop was possibly starting to settle and the vine was being stressed by the competition of nutrients and water and reacting by storing these compounds in their glycosylated form. Previous studies have pointed out that the insertion of a permanent cover crop in vineyards in a water-limited environment would generate both water and nitrogen stress in the grapevine (Celette and Gary, 2013). Water competition can lead to water deficits in the vine. The influence of water deficit has been previously observed in Cabernet Sauvignon and Merlot grapes and wines (Koundouras et al., 2009; Qian et al., 2009; Song et al., 2012; Talaverano et al., 2018). Furthermore, water deficit reduces canopy size, which subsequently increases cluster exposure to sunlight (Intrigliolo and Castel, 2011) and this is a key external factor that has been demonstrated to increase glycosidically bound aroma compounds in grapes (Marais et al., 1992; Meyers et al., 2013). This effect is mainly due to the increase of C 13 -norisoprenoids in grapes, which mostly come from the β-carotene oxidative degradation. With more sunlight exposure, more carotenoids are present in the berry and consequently more C 13 -norisoprenoids (Baumes et al., 2002). Considering that, despite competition for nutrients, the employment of cover crops has been shown to enrich the soil in nitrogen and carbon, with potential effects on the berry quality of Vitis vinifera grapes and table grapes (Gatullo et al., 2020; Ferrara et al., 2021). Analysing the results in detail by chemical group to see the effects of the cover crop by year, it can be clearly observed that the initial planting of Zulla in the first vintage, 2019, had a positive effect on the production of terpene precursors with light cover crop influence (SEZZ) (Fig. 2). Glycosidic terpenes are of great importance to non-aromatic varieties, such as Syrah, as they confer a characteristic varietal aroma on wines (Maicas and Mateo, 2005). Hence, from the 11 terpenes, myrcenol (fresh floral/lavender/citrus), ocimenol (citrus/lemon) and isodurene (camphor-like odour) were significantly higher in SELZ musts, with isodurene being the most impacted (Table 1). When more Zulla was planted close to the vine, SEZZ, the significantly higher terpenes with respect to the control were menthol (mint) and, again, isodurene, which tripled its Fig. 3. Principal Component Analysis (PCA) showing the data scores (a) and loadings (b) biplot on the plane of the first two principal components (PC1 against PC2) comparing the samples of conventional vs. organic crop of the three vintages (2019, 2020 and 2021) including all the precursors determined as variables. M.P. Segura-Borrego et al.