Materials and methods

An experimental vineyard, consisting of ‘Chambourcin’ winegrapes, was established in 2008 at the University of Missouri’s Southwest Research, Extension, and Education Center at Mt. Vernon, MO, USA. The site is in southwest Missouri, situated at lat. 37.07409°N, long. 93.87962°W, elevation 1240 ft, and averages 1175 mm precipitation annually. The soil is mapped as a Hoberg silt loam (fine-loamy, siliceous, mesic Mollic Fragiudalfs) that is upland, deep, gently sloping, and moderately well-drained to a fragipan at 40 to 90 cm (Hughes 1982). To improve soil drainage and root-penetrable soil depth over the fragipan, soil was mechanically pushed from the alleys to form raised berms ∼10 inches high. After soil analyses in 2008, amendments of lime, phosphorus (P), potassium (K), and zinc (Zn) were applied and incorporated into the soil according to recommendation for winegrape production (Dami et al. 2005a). Soil pH was evaluated midstudy (June 2013), and indicated an excellent pH of 6.3. Significant additional details on climate, vineyard establishment, pest management, soil fertility management, and irrigation are outlined in Thomas et al. (2017, 2020).

The ‘Chambourcin’ vines were produced and donated by Wonderful Nurseries (Wasco, CA, USA), grown either as own-rooted vines or bench-grafted to nine rootstocks: ‘Couderc 3309’ (3309C), ‘Couderc 1616’ (1616C), ‘Paulsen 1103’ (1103P), ‘Sélection Oppenheim 4’ (SO4), ‘Millardet et de Grasset 420A’ (420A), ‘Millardet et de Grasset 101-14’ (101-14), ‘Kingfisher’, ‘Matador’ (all Vitis sp.), and ‘Gloire de Montpellier’ (Riparia Gloire) riverbank grape (Vitis riparia). The rootstocks ‘Matador’ and ‘Kingfisher’ are recent nematode-resistant releases from the US Department of Agriculture (USDA) (Cousins 2011), whereas the other seven are in common use worldwide (Pongrácz 1983). Vines were planted 25 Jun 2008 with graft unions positioned ∼5 inches above the berms to prevent scion rooting. Vines were trained with double trunks to a single-curtain, high-wire, bilateral cordon system ∼5.8 ft above the berm. During three establishment years (2008–10), all flowers and fruit were removed to encourage robust vine development, and graft unions were provided winter protection via an application of municipal compost (City of Monett, MO, USA) and/or mushroom compost (J-M Farms, Inc., Miami, OK, USA) held in place around and above the unions with a cylindrical hardware mesh. An equivalent amount of compost was applied to nongrafted vines but without a hardware mesh. Vines were individually balance-pruned (Jordan et al. 1981) following the 2011 growing season and throughout the duration of the study. Specifically, at 10- to 12-inch shoot length, all noncount shoots and double shoots were removed. Cluster thinning was performed after berry set using the “two-one-none” rule adjusted for large-clustered cultivars such as Chambourcin: shoots greater than 36 inches: retain two clusters per shoot; shoots 18 to 36 inches: retain one cluster per shoot; and shoots less than 18 inches: remove all clusters from the shoot. Shoot positioning was performed two to three times per season as needed to orient shoots in a downward position. Leaf removal was performed after berry set within the fruiting zone on the north side of the canopy.

The experiment was established in eight parallel east-west vineyard rows that were 370 ft long and 9.7 ft apart. Nine four-vine plots were established per row (72 total plots), with each vine planted 10 ft apart in-row. Each of the 10 root systems (own-rooted or rootstock cultivar) was randomly assigned to six plots (six root system replications), totaling 60 plots [note that 12 plots were randomly planted to other experimental rootstocks (all grafted to ‘Chambourcin’) that are not included in this report]. Nonexperimental ‘Chambourcin’ guard vines were similarly established and managed in rows directly north and south of the study and at both row ends to ensure that all study plants grew in similar environments.

Fruit was harvested and data collected throughout four production years (2011–14) for analysis. Bird netting (AviGard; Plantra, Inc., Eagan, MN, USA) was deployed to reduce bird predation of fruit as they ripened. Vines within plots were individually harvested, with single-vine production data later pooled by plot for analysis according to the experimental design. Number of fruit clusters and total fruit yield per vine were determined, and mean cluster weight calculated. Production yield (megagrams per hectare) was extrapolated based on the equivalent number of vines per hectare at the stated spacing. Dormant pruning weights were collected in late winter, and crop load ratio (kilograms fruit per kilogram dormant cane prunings) calculated.

A random 100-berry sample was collected at harvest from among the four vines within each plot to determine berry weight and fruit characteristics [soluble solids concentration (SSC), juice pH, titratable acidity (TA) in 2011–14], and juice K (in 2011 and 2014). SSC, pH, and TA of juice were measured as previously described (Iland et al. 2004; Thomas et al. 2017), with TA expressed as grams per liter tartaric acid. For K analysis, samples were prepared by mixing juice with 2.5% hydrochloric acid (HCl) at a 1:20 ratio (HCl:juice) before analysis at the University of Arkansas’ Agricultural Diagnostic Laboratory at Fayetteville, AR, USA, using inductively coupled plasma-optical emission spectrometry [ICP-OES (Jones and Case 1990)]. In addition, in 2011, 2013, and 2014, leaf petiole samples were collected at veraison and analyzed for nitrogen (N), P, K, calcium (Ca), magnesium (Mg), sulfur (S), manganese (Mn), Zn, iron (Fe), boron (B), copper (Cu), and molybdenum (Mo) by the Soil and Plant Testing Laboratory at the University of Missouri (Columbia, MO, USA), also using ICP-OES technology (Munter and Grande 1981).

In 2011 and 2014, additional samples of 200 random berries per plot were collected at harvest and frozen at −20 °C for subsequent analysis of anthocyanins, tannins, and total phenolics. Significant details on these sample preparations and analyses are described in Jogaiah et al. (2012), where the protocols developed by the Australian Wine Research Institute (2015, 2017) were used. Briefly, samples for anthocyanins and total phenolics were prepared simultaneously, then color measured with a spectrophotometer (Spectronic Genesys 2 ultraviolet-Vis; Thermo Scientific, Waltham, MA, USA) at 520 and 280 nm, respectively. Samples for tannins analysis were prepared separately, then similarly measured at 280 nm. Catechin standard was used at equivalent epicatechin concentrations of 0, 50, 100, 150, 200, and 250 ppm, and measured at 280 nm to create a calibration curve.

The same samples were also analyzed for tartaric, malic, and citric acid by high-performance liquid chromatography (HPLC). Briefly, 3 g polyvinylpolypyrrolidone was added to 5 mL of juice sample, mixed for 20 s to achieve decolorization, centrifuged, then filtered. Chromatography was performed on an HPLC system (ProStar 410; Varian, Inc., Palo Alto, CA, USA) with a 335 LC dual-path diode array ultraviolet-visible detector, and operated with Galaxie software (version 1.9.302.530; Agilent Technologies, Inc., Santa Clara, CA, USA). A Zorbax SB-Aq column (Agilent Technologies, Inc.) was operated at 35 °C. The aqueous mobile phase was isocratic with 99% 20 mm sodium phosphate and 1% acetonitrile at a flow rate of 1 mL·min⁻¹ (total run time 10 min). The ultraviolet-visible diode array operated at 210 nm to detect the acids. Each acid was identified by reference standards and quantified with a five-point calibration curve ranging from 0.25 to 10 g·L⁻¹. Sample injection volume was 10 µL.

The experiment was established as a completely randomized factorial study with six replications of 10 root system treatments, and with a repeated measure over time (years). While vineyard production data were determined on a per-vine basis, the experimental unit was the four-vine plot and all production data were appropriately analyzed by plot. Experimental data were compared 1) between own-rooted and grafted vines, 2) among the 10 root systems, and 3) among 4 production years. Data were statistically evaluated using statistical software (SAS version 9.4; SAS Institute Inc., Cary, NC, USA). Analysis of variance was conducted, with means separated using Fisher’s least significant difference test or the Student-Newman-Keuls/Duncan test, both at a significance level of P ≤ 0.05.

Results and discussion

Significant differences were detected in several of the yield, fruit component, and vegetative growth variables between grafted and own-rooted vines, among root systems, and among production years (Table 1). When simply comparing grafted vs. own-rooted vines across the 4 production years, fruit yield, number of clusters per vine, cluster weight, and berry size were not influenced by grafting, whereas pruning weight was greater in grafted vines. However, when the 10 root systems were compared, fruit yield, berry weight, and vine performance components varied significantly. The consideration of 4 years’ data indicated that vines grafted to 3309C had higher yields, producing an average of 21.2 Mg·ha⁻¹ compared with vines grafted to six of the other rootstocks or own-rooted (16.1 Mg·ha⁻¹ for the latter). Although vines grafted to 3309C produced numerically more clusters per vine (83.3) and among the highest cluster weights (246 g) compared with other root systems, these values were not statistically different from other root systems. Across the 4 years, the rootstock ‘Kingfisher’ produced among the largest berries (2.24 g), with statistically smaller berries produced on six of the root systems. Although numerous interacting viticultural factors can certainly influence berry weight, such factors may be further impacted by rootstocks due to their inherent abilities to uptake water and nutrients at different levels (e.g., Gautier et al. 2020; Rühl et al. 1988; Striegler et al. 1993). Many winemakers may prefer smaller berries because of their higher skin-to-pulp ratio, and the concomitant increase in aromatic and phenolic compounds (Ribéreau-Gayon et al. 2006).

Table 1.

Fruit yield and vine performance of ‘Chambourcin’ winegrape as affected by grafting, 10 root systems, and year at Mt. Vernon, MO, USA in 2011–14.ⁱ

View Table

For crop load ratio (kilograms fruit per kilogram dormant cane prunings) on high-yielding hybrid cultivars such as Chambourcin, a range of 9–14 is generally considered “balanced,” with higher values considered “excessive” (Dami et al. 2005b; Ferree et al. 2004). Across all root systems, the crop loads were nearly balanced during 2011, 2013, and 2014 (ratios of 15.0, 13.2, and 14.3, respectively). However, in 2012 (a severe heat and drought year), the crop load may have been excessive (26.5). This high ratio in 1 year skewed (increased) the mean crop load ratio data among root systems across the study. Among the 10 root systems across 4 years, crop load ratios ranged from 15.0 (1103P) to 19.8 (1616C), with most root systems being cropped just above the recommended range. Because ‘Chambourcin’ produces a large number of fruitful shoots from noncount positions (Ferree et al. 2003), the fruitfulness of noncount buds is a primary cause of overcropping, making balanced pruning less effective in some instances (Ferree et al. 2003, Pool et al. 1978). All the yield and vine vigor factors studied varied significantly among the 4 years, with 2012 producing the greatest overall yields across all root systems, including the largest number of clusters per vine but lowest berry weight and lowest pruning weight. These results support the significant interactions detected between root system and year for fruit yield and crop load ratio.

Differences in juice SSC, pH, and TA were not detected between own-rooted and grafted vines as a whole, but pH and TA varied among root systems and years (Table 2). SSCs were not different across root systems, but did vary among growing seasons. Among individual root systems, berries from Riparia Gloire and SO4 did not differ in pH from own-rooted vines, whereas the other root systems produced fruit with slightly higher juice pH. Although some adjustments for pH can be made in the winery, production of higher-pH juice from grafted vines may be an important consideration if shown to be consistent through additional research.

Table 2.

Characteristics of ‘Chambourcin’ winegrape juice as affected by grafting, 10 root systems, and year at Mt. Vernon, MO, USA in 2011–14.

View Table

No differences between grafted and own-rooted vines were found in the levels of tannins, anthocyanins, and total phenolics in the fruit; however, some differences among root systems were observed (Table 3). Numerically higher tannin content was detected in ‘Chambourcin’ grapes grafted to SO4 and 1616C, but the levels were not statistically significant. Blank et al. (2022) observed higher tannin levels in berry seed and skin of ‘Pinot Noir’ (Vitis vinifera) grafted to SO4 compared with Riparia Gloire, concluding that final tannin concentration in wine depended on the rootstock. The rootstocks Riparia Gloire, 3309C, 1616C, and 101-14 produced juice with higher concentrations of anthocyanins, whereas 1103P and ‘Kingfisher’ had the lowest. The latter two rootstocks also produced fruit with among the lower levels of total phenolics in the juice, whereas Riparia Gloire produced some of the highest total phenolics among root systems. Because tannin levels were not influenced by root system, and anthocyanins and phenolics were, it follows that complex root system × vintage interactions were detected for the latter two factors. Differences were probably not of enological significance and not great enough to impact wine composition.

Table 3.

Tannin, anthocyanin, total phenolics, potassium, and organic acids concentrations in ‘Chambourcin’ winegrape fruit as affected by grafting, 10 root systems, and year at Mt. Vernon, MO, USA in 2011 and 2014.

View Table

Vines grafted to 420A and 1103P rootstocks produced fruit with higher concentrations of K in the juice compared with own-rooted vines and vines on ‘Matador’, Riparia Gloire, SO4, and 1616C. Vanden Heuvel et al. (2004) found that rootstock can significantly influence winegrape juice pH, which is often associated with the high K uptake ability of some rootstocks, such as 420A and 1103P, resulting in higher K concentration in the juice. Differences among rootstocks in the concentration of K in the berry juice are most likely due to differences in their ability to uptake K from soil, and root-to-shoot transport of minerals (Kodur et al. 2010; Lambert et al. 2008; Rühl 1989). Potassium is the major cation taken up by grapevines (Boulton et al. 1996), and there is a close correlation between K concentration and the pH of grape juice and wine (Boulton 1980; Hale 1977; Iland 1987; Somers 1977). From an enological perspective, several minerals in the grape juice are critical for their effect on wine making processes, but K concentration can influence the winemaking process more than other minerals. High K levels could result in the precipitation of potassium hydrogen tartrate, causing the juice to have a higher-than-desirable pH, and leading to wine instability and color problems in red wines. The high K uptake ability seen in 420A and 1103P, coupled with the associated higher pH of juice, might be considered when balancing other attributes of these and other rootstocks.

Tartaric, malic, and citric acid composition did not differ between grafted and ungrafted vines, but did vary among individual root systems and across years (Table 3). Interestingly, the rootstock 1103P produced fruit with lower amounts of tartaric but higher amounts of malic acid compared with most rootstocks, whereas SO4 produced fruit with higher levels of tartaric acid but relatively low levels of malic acid. Riparia Gloire and ‘Matador’ produced fruit with the lowest levels of malic acid. Tartaric acid is the strongest acid in grapes, the most resistant to oxidation and microbial spoilage, and tends to comprise more than half of the TA (Amerine and Ough 1980). Malic acid is the second most common acid in grapes, although it is much less stable than tartaric. In terms of sourness at the same concentration, malic is the highest followed by tartaric, citric, and lactic (Amerine et al. 1965). Significant root system × vintage interactions were shown for SSC, K, malic acid, and citric acid, suggesting that the expression of these characteristics may be more complex than other fruit traits.

Rootstocks have been shown to influence 17 of the 20 elements in grape leaves (Harris et al. 2022), with the greatest impact on Mg concentrations. In the current study, the use of rootstocks significantly affected petiole mineral concentration of some minerals compared with own-rooted vines (Table 4). Specifically, N and S petiole levels were higher in own-rooted vines, whereas P, Mn, and Zn levels were lower. Among the 10 root systems across the four growing seasons, significant differences among petiole mineral levels were detected for all minerals studied. Nitrogen concentrations in the petiole were significantly higher in ‘Chambourcin’ grafted to 1103P (1.47%), 3309C, and 420A, with the lowest levels found in ‘Kingfisher’ and ‘Matador’ (both 1.25%). The highest P concentration of 0.28% was found in vines grafted to 1103P, with the lowest P uptake seen on Riparia Gloire (0.15%). Higher concentrations of petiole K were found in vines grafted to 1103P (2.57%) and SO4, compared with own-rooted vines that had the lowest K concentration (2.06%). SO4 induced a higher concentration of Ca in the petiole, whereas there was a higher uptake level of several minerals (Mg, Mn, Zn, Fe, and B) by 101-14. Vines grafted to 101-14, along with own-rooted vines, also produced the highest levels of petiole Fe and S. Although statistically significant differences were observed, levels were generally within recommended ranges with the exception of N, K, Fe, and B; for all rootstocks, N and K were higher than recommended ranges, whereas Fe and B were lower (Wolf 2008). As with other analyses in this study, petiole mineral differences varied significantly among growing years; patterns are difficult to discern but in general, macronutrient levels were higher in 2014.

Table 4.

Petiole mineral concentration of ‘Chambourcin’ winegrape as affected by grafting, 10 root systems, and year at Mt. Vernon, MO, USA in 2011, 2013–14.

View Table

Because petioles were sampled across three distinct growing seasons, it seems evident that the rootstocks studied have differing capabilities to extract different minerals from soil and transport them throughout the vine; this was also shown by Gautier et al. (2020) and Grant and Matthews (1996). With the present study in southern Missouri, USA, rootstocks 420A and 1103P showed the greatest propensity to extract and distribute needed macronutrients such as P, K, and Mg. Although the uptake of these macronutrients by 101-14 was at an average rate, it induced a higher uptake of several critical micronutrients, including Zn, Fe, and B. Riparia Gloire did not excel in mineral uptake of either macro- or micronutrients, P in particular; its performance in this aspect was comparable with own-rooted vines. Grant and Matthews (1996) demonstrated that different rootstocks might have different abilities to uptake P. Availability and uptake of some soil minerals have relationships that encompass synergy, inhibition, or antagonism (Fageria 2001; Ranade-Malvi 2011). Potassium, Ca, and Mg are macronutrients that can sometimes be antagonistic to each other in terms of ion uptake by plants (Xie et al. 2021). Their absorption depends on the capability of the plants based on the root system to uptake a given element rather than its antagonist. A rootstock that specifically influences a given mineral’s uptake can favor it over its antagonist, inducing a higher concentration of the element in the petiole and/or in the fruit. Rootstock 3309C showed a good balance in absorption of all three of these macronutrients, whereas 101-14 and Riparia Gloire had a higher uptake of K compared with P, Ca, and Mg. The same type of antagonistic relationship is present with P, Mn, and Zn in the soil. In this case, 101-14 showed a tendency to prefer uptake of Mn and Zn over P.

The environmental conditions of each growing year in this study (2011–14) had a significant influence on most viticultural production factors, often apparently exerting a stronger influence than did the root system. This is consistent with the study by Migicovsky et al. (2021), in which vintage was determined to be the largest source of variation (>40%) among root systems for pruning weight, yield, and crop load ratio. Indeed, the root system × year effect in the present study was significant for many (but not all) experimental factors (see Table 1). Variables such as yield, cluster weight, number of clusters per vine, and berry weight showed significant differences from one year to the next, clearly demonstrating the strong influence of environmental and weather conditions (Corino and Castino 1990; Hidalgo 2002; Main et al. 2002; Santiago et al. 2007). In 2011 (the first production year), there were lower production and cluster weights probably because the vineyard was not yet fully mature, but fruit that year had higher SSC. In 2012 (an excessively hot and dry summer for southern Missouri), the vineyard was more mature but was likely overcropped. The yield, number of clusters per vine, and crop load ratio that year were higher compared with the other years of study, whereas berry size and SSC were the lowest. In 2013, most experimental variables were better balanced overall.

Conclusions

During the selection of rootstocks for a given location and a given scion cultivar, compromises among vigor, yield, soil mineral uptake, and grape characteristics may be necessary. This study suggests that, among the rootstocks evaluated for ‘Chambourcin’, 3309C is highly suited to the edaphoclimatic conditions in the southern Missouri region, and it also showed distinct advantages over own-rooted vines. When grafted to ‘Chambourcin’, 3309C had a balanced performance with respect to the parameters analyzed across multiple growing seasons. It induced a good mineral uptake (including N), high yield, and produced excellent fruit with appropriate pH and acidity levels, as well as high anthocyanin content. Potassium levels in the juice were slightly higher than optimum but within acceptable range for winemaking purposes. Excellent production of winegrapes grafted to 3309C has also been observed with other cultivars in nearby environments [e.g., Sunbelt in Arkansas, USA (Morris et al. 2007)]. Other rootstocks that performed well from a yield and vine size perspective in this study were 420A, 101-14, and 1103P, but K concentrations were also somewhat higher than desired for the best wine quality. Although the new USDA nematode-resistant rootstocks ‘Matador’ and ‘Kingfisher’ performed well in this study, the production and viticultural factors evaluated suggest that their use in the region might best be reserved for locations where nematodes are a concern.

This study underscores the hypothesis that root system and rootstock can influence some important viticultural and enological production metrics, including in the midwestern United States. Regarding the viticultural performance of ‘Chambourcin’ in southern Missouri, the study suggests that the use of rootstocks may increase juice pH. This effect could be attributed to a higher root system absorption capability of K as shown by a higher petiole K concentration across 3 years of analysis. Higher concentrations of P and Mn were also present in the petioles of grafted vines compared with own-rooted vines. In this study, grafted vines also had a more balanced reproductive/vegetative ratio with a higher weight of cane prunings and a lower crop load compared with own-rooted vines. With additional research, the use of winegrape rootstocks in the midwestern United States can be fine-tuned and further improved, resulting in more targeted production goals and improved wine quality outcomes.

Units

View Table