Grapevine red blotch virus (GRBV) was identified at the University of California Davis Research Station in Oakville, California in 2008 (Calvi, 2011). The symptoms of GRBV in grapevines were similar to those of grapevine leafroll disease, which is also caused by a virus. However, leafroll virus strains were not detected in plants with GRBV, suggesting GRBV was caused by a different virus (Sudarshana et al., 2015). Archival leaf samples of grapevine ‘Early Burgundy' collected in Sonoma County, California, in 1940 and stored at the UC Davis herbarium, were found to have GRBV (Al Rwahnih et al., 2015). These findings revealed that the newly discovered GRBV was present in vineyards long before the disease was recognized in 2008 and the virus genome characterized in 2011.

The virus's most significant economic effect is diminished fruit quality (Sudarshana et al., 2015). Grapevine Red Blotch Disease (GRBD) inhibits ripening, particularly lower sugar content in berry juice and anthocyanin concentrations in berry skin, which negatively affects fruit composition and production value (Martínez-Lüscher et al., 2019b). Low sugar and anthocyanin content were suggested to be due to decreased carbon translocation; however, the exact mechanism resulting in those effects was unclear (Bowen et al., 2020). The severity of symptoms and their onset have been observed to vary with cultivar, vineyard location, and growing season (Cieniewicz et al., 2017).

The suppression of virus inoculum and control of possible vectors in the vineyard are the primary objectives of grapevine virus disease management (Maliogka et al., 2015). Despite effective virus elimination strategies employed in the laboratory, there is currently no vineyard treatments to cure grapevines from GRBV, or other viruses, emphasizing the importance of GRBV prevention and management after infection in vineyards. Viruses may be excluded from new vineyards by planting certified virus-free plants (Cieniewicz et al., 2017; Golino et al., 2017). The possible role of GRBV vectors and reservoirs, and determination of differences across viral isolates and strains were all active areas of research (Bahder et al., 2016a) and three-cornered ee-cornered alfalfa leaf hopper has been confirmed to be the vector of GRBV both in cage studies and in vineyard studies (Rumbaugh et al., 2021). Studies of the effects of GRBV on the transcriptional and hormonal regulation of grape ripening identified several pathways that were abnormally activated or repressed by GRBV infection during the late stages of fruit ripening and were linked to inadequate carbon translocation into berries and delayed maturity resulting in poor fruit composition (Blanco-Ulate et al., 2017; Martínez-Lüscher et al., 2019b). When the prevalence of GRBV in a vineyard is > 25 % it is considered significant, because this disease has a direct impact on fruit composition at the farm gate (Ricketts et al., 2017; Wallis and Sudarshana, 2016).

In healthy grapevines, the ratio of leaf area to fruit mass is strongly correlated with the amount of carbohydrates accumulated in the juice (Naor et al., 2002). As a result, an excessive crop load or smaller than ideal canopy size may delay ripening (Geller and Kurtural, 2013). Conversely, if under cropped, i.e. reduced crop mass to leaf area ratio, pruning mass is increased without affecting soluble solids accumulation (Terry and Kurtural, 2011). As a result, yield is sometimes limited in order to balance the source:sink ratio to ensure timely fruit ripening (Terry and Kurtural, 2011). The amount of carbon assimilated by leaves is allocated to many organs of the grapevine, and a fraction of the carbon pool is translocated to the fruit (Martínez-Lüscher and Kurtural, 2021). Martínez-Lüscher and Kurtural (2021) reported several compensatory processes in response to over and under cropping were such as changes in berry ripening, berry size, pruning wood and root reserves and development. Components of yield, such as clusters per vine, berries per cluster, berry mass, and berry soluble solids, are interdependent (Palliotti and Cartechini, 2000), and cluster thinning may advance the beginning of véraison, fruit ripening and anthocyanin synthesis depending on timing and intensity of the cluster thinning imposed (Guidoni et al., 2008).

Inhibition of berry ripening was reported as the most common response to GRBV infection. Components of yield which include clusters per vine, berry mass, and soluble solids, are susceptible to change together with berry ripening in compensation with each other. Although grapevine source: sink manipulation is the most frequently reported case study, most studies have not considered the impediment of virus infection to this and its subsequent effects on secondary metabolites. The objective of this work was to determine if an increase in source:sink ratio through cluster thinning can alleviate the deleterious effect of GRBV infection on total soluble sugar translocation, plant water status, leaf gas exchange and berry primary and secondary metabolites to understand the utility of cluster thinning as a management strategy for vineyards impacted by GRBV.

The experiment was conducted during the consecutive growing seasons of 2020 and 2021 at the University of California Davis, Oakville Experimental Vineyard (38.428 N, 122.409 W; Oakville, CA, USA). The site soil was a Bale-silt loam characterized as deep and gravely clay-loam having a slope less than 2 % (Yu and Kurtural, 2020). The site’s climate is considered temperate Mediterranean with a high diurnal temperature range and little to no rain during the growing season (Yu and Kurtural, 2020). The experiment was conducted on Vitis vinifera L. cv. ‘Cabernet-Sauvignon’ Clone FPS08 grafted on 110 Richter (V. rupestris Scheele x V. berlandieri Planch.) rootstock planted in 2012. The rows were oriented NW-SE and spaced 2 m x 2.4 m (vine x row). Plants were pruned to one-bud spurs and trained to a bilateral cordon on a vertical shoot positioned trellis with a cordon height of 96 cm above the ground level. The plants were shoot thinned to a density of 30 shoots⸳vine^-1 at Eichhorn-Lorenz scale stage 17 (Coombe, 1995). No additional hedging or leaf removal was performed. All other cultural practices, such as vineyard fertilization, pest management and soil management, were carried out according to University of California Cooperative Extension guidelines of the region. Flowering of the plants occurred on 10 May in both 2020 and 2021. Fruit was harvested on 14 September 2020, 127 days after flowering (DAF), and 13 September 2021 (126 DAF).

The vineyard had naturally spread distribution of GRBV. One hundred plants were initially selected based on historical visual assessments and tested for virus status using qPCR on fully expanded leaves collected from shoot bases from plants after véraison at the end of each preceding growing seasons (September to October) in 2019, 2020 and 2021 in accordance with previous GRBV studies (Bahder et al., 2016). In 2019, plants suspected of having GRBV were selected for testing based on visual leaf symptoms of red blotches while plants suspected to be free of the virus (no leaf symptoms observed) were tested to confirm GRBV(-) status. In subsequent years (2020 and 2021), virus status was reconfirmed by qPCR tests. Briefly, total nucleic acid was extracted from leaf petiole tissue using the MagMax 96 Viral RNA isolation kit with the MagMax Express-96 magnetic particle processor (Thermo Fisher Scientific, USA). No new presence of GRBV were identified in the GRBV(-) plants in the second year of study. Additionally, the samples were screened for the following viruses: grapevine leafroll-associated viruses, including Grapevine leafroll-associated virus-1 (GLRaV-1), Grapevine leafroll-associated virus-3 (GLRaV-3), and Grapevine leafroll-associated virus-4 (GLRaV-4, plus strains 9 and Car) (genus Ampelovirus); Grapevine leafroll-associated virus-2 (GLRaV-2, plus strain 2RG); (genus Closterovirus); Grapevine leafroll-associated virus-7 (GLRaV-7); (genus Velarivirus); and Grapevine rupestris stem-pitting associated virus (GRSPaV); (genus Foveavirus); as described by (Al Rwahnih et al., 2015).

To assess the effects of source-sink manipulation on carbohydrate translocation of GRBV (+) plants, two cropping levels were applied to two virus status levels (GRBV (+) and GRBV(-)) using cluster thinning in both years. At the phenological stage of pea size, about 7 mm berry diameter, clusters were removed to homogenize the number of clusters to approximately 45 clusters per plant within the experiment. The untreated (UNT) plants consisted of 45 clusters per vine. At the same time, cluster thinned (CT) treatment group were manually thinned to 10 clusters per plant. The treatments were arranged in a 2 × 2 factorial experiment in a completely randomized design resulting in four treatment combination groups: UNT GRBV (-), CT GRBV(-), UNT GRBV(+), and CT GRBV(+). Each treatment combination was replicated on four plants for a total of 16 plants in both years (N = 16) where each plant corresponded to one treatment-replicate.

The meteorological data for the site was sourced from the California Irrigation Management Information System, CIMIS, station (#77, Oakville, CA, USA) for 2019-2020 and 2020-2021 growing seasons. The station is located at the study site, 160 meters away from the experimental plot. Precipitation and Growing Degree Day accumulation (GDD) for the study site that occurred during the trial is shown in Figure 1. Precipitation was measured from October of the preceding season to October of the current season. GDD measurements began on the 1 April and finished on 30 September of the current season for each year. GDD was calculated for each day with the equation and then summed up for each season starting at budbreak:

G D D = D a i l y   m a x i m u m   t e m p e r a t u r e + D a i l y   m i n i m u m   t e m p e r a t u r e 2 - 10

CIMIS weather data provided a reference evapotranspiration rate. A crop coefficient was determined based on shade cast on the vineyard floor at the study site and used to calculate evapotranspiration by multiplying the reference evapotranspiration and the crop coefficient (Torres et al., 2021b). Irrigation was scheduled weekly based on the calculated evapotranspiration to replenish 50 % crop evapotranspiration from fruit set to harvest every year (Torres et al., 2021b).

One fully-expanded and sun-exposed leaf was selected from the main shoot axis on each grapevine and measurements of plant water status. Leaves were covered in Mylar zip-top bags 2 hours before and the assessment of stem water potential (Ψ_stem) was performed at solar noon with a pressure chamber (Model 615 Pressure Chamber Instrument., PMS Instrument Co., Corvallis, OR, USA). Ψ_stem was determined approximately every two weeks from vérasion to harvest in both years of study. In 2020, Ψ_stem was measured on 23 July (75 DAF), 5 August (87 DAF), 21 August (103 DAF) and 14 September (128 DAF). In 2021, Ψ_stem was measured on 27 July (78 DAF), 10 August (93 DAF) and 31 August (114 DAF).

An infrared gas analyzer (CIRAS-3 PP Systems, Amesbury, MA, USA) with a broad leaf chamber (4.5 cm² window) was used to measure leaf gas exchange of net carbon assimilation (A _N) and stomatal conductance (g _s). Three measurements from each treatment-replicate were taken around solar noon from an undamaged and mature leaf on the main shoot axis from each data vine for each sampling date. These three measurements were then recorded and averaged to constitute a treatment replicate. Data collection was performed on sunny days with natural sun light (photosynthetically active radiation > 1,500 μmol m^-2 s^-1) and the leaf chamber was positioned perpendicularly to sunlight. Chamber conditions were set to 40 % relative humidity, temperature to follow ambient temperature, a reference CO₂ concentration of 400 μmol CO₂ mol^-1, and a flow rate of 300 ml min^-1 into the chamber. During the 2020 growing season, gas exchange was measured on 23 July^{(75 DAF), 5 August (87 DAF) and 14 September (128 DAF). In 2021, measurements occurred on 27 July (78 DAF), 10 August (93 DAF), and 31 August (114 DAF).}

To express the season-long response of A_N , and g _s, their integrals were calculated by using natural cubic splines for plant water status and gas exchange measurements to assess the cumulative values for these variables over the whole experiment period during each growing season. Then, these cumulative values were normalized as divided by the number of days elapsed between the first measurement date and the last measurement date to make the data comparable to each individual measurement.

During the 2021 growing season leaves were sampled for sugar content on three dates; 14 July (65 DAF, immediate pre- véraison), 27 July (78 DAF, véraison) and 23 August; (105 DAF, mid- véraison) to quantify the concentration of sucrose, fructose, glucose, raffinose, and total soluble sugars as described previously (Torres et al., 2021b). Leaf samples from 4-6 mature leaves from each treatment-replicate collected and immediately weighed then oven-dried at 70 °C until weight remained constant. A tissue lyser (MM400, Retsch, Germany) was used to grind the dried leaves into a homogeneous powder. Thirty mg of the resulting powder were extracted in a 75:25 ethanol: water solution at 90 °C for 10 minutes. Immediately after extraction, samples were centrifuged for one minute at 10,000 rpm, and the supernatant was collected and immediately analyzed for the content of sugars. In leaf ethanolic extracts, total and individual sugars were measured, according to Torres et al. (2021b). Extracted samples were filtered using PTFE membrane filters (diameter: 13 mm; 0.45 μm; CELLTREAT Scientific Products, Pepperell, MA, USA) and moved into high performance liquid chromatography (HPLC) vials and subjected to reversed-phase HPLC analysis. An Agilent 1100 system along with a diode array detector (DAD) and an Infinity Refractive Index Detector (RID) (Agilent Technologies Inc., Santa Clara, CA, USA) were used. Luna Omega Sugar (150 A 4.6 mm, 3 μm particle size, Phenomenex Inc., Torrance, CA, USA) was used as the reversed-phase column, with a 5 mm guard column. The column compartment was kept at 40 °C, while the RID flow cell was kept at 35 °C. Isocratic elution with acetonitrile: water (75:25, V/V) at a flow rate of 1.0 ml/min and run time of 22 minutes was used in the mobile phase. To determine the retention time for each compound, standard solutions of 10 mg/L D-glucose, D-fructose, D-sucrose, and D-raffinose were injected, and RID was used for their detection. VWR International provided the sugar standards (Radnor, PA, USA). By comparing the peak area and retention time of each sample to standard sample curves, the sugar concentration of each sample was determined.

For the analysis of shoot sugars in phloem sap, a single shoot was collected from each vine on 27 July (78 DAF). A 7 cm stem segment was cut from near the shoot base which had the cluster node located at its center and was used for sap collection. To control for diurnal variation between samples, all shoots were collected at the same time. Sap was extracted placing shoot segments in a 15 ml conical tubes (base downwards) and centrifugating at 4500 rpm for 15 minutes at 4⁰ C. After centrifuging, extracted sap fluid was collected from the bottom of the tube and weighed. Extracted sap was then diluted by bringing volume up to 1 ml using deionized water and mixed on a vortex mixer. Diluted sap solutions were then filtered using PTFE membrane filters (diameter: 13 mm; 0.45 μm; CELLTREAT Scientific Products, Pepperell, MA, USA) into HPLC vials and immediately subjected to reversed-phase HPLC analysis using the method mentioned above for leaf sugar analysis.

For the analysis of shoot sugars in phloem sap, a single shoot was collected from each vine on 27 July (78 DAF). A 7 cm stem segment was cut from near the shoot base which had the cluster node located at its center and was used for sap collection. To control for diurnal variation between samples, all shoots were collected at the same time. Sap was extracted placing shoot segments in a 15 ml conical tubes (base downwards) and centrifugating at 4500 rpm for 15 minutes at 4 ⁰C. After centrifuging, extracted sap fluid was collected from the bottom of the tube and weighed. Extracted sap was then diluted by bringing volume up to 1 ml using deionized water and mixed on a vortex mixer. Diluted sap solutions were then filtered using PTFE membrane filters (diameter: 13 mm; 0.45 μm; CELLTREAT Scientific Products, Pepperell, MA, USA) into HPLC vials and immediately subjected to reversed-phase HPLC analysis using the method mentioned above for leaf sugar analysis.

Leaf area of each treatment-replicate was determined using a smart-phone based program as reported elsewhere (Torres et al., 2021a; Yu and Kurtural, 2020). Briefly, Leaf area index (LAI) was measured to characterize grapevine canopy growth and converted into leaf area at véraison in each year by a smartphone-based program, VitiCanopy, coupled with an iOS system (Torres et al., 2021a). The gap fraction threshold was set to 0.75, the extinction coefficient was set to 0.7, and sub-divisions were 25. A ‘selfie-stick' was used to place the device 75 cm underneath the canopy. The device was positioned with the maximum length of the screen being perpendicular to the cordon, and the cordon being at the middle of the screen according to the user's manual (De Bei et al., 2016). In each experimental unit, three images were taken to capture half the canopy of each plant and then analyzed by the software. Total leaf areas were calculated based on both LAI values and unit ground area in each experimental unit, and then the leaf area to fruit ratio was calculated. For yield per vine, clusters were removed from plants, counted, and weighed on a top-loading balance when the control treatment (GRBV (-) plants) reached 25°Brix as is a common production practice in the region. The amount of sugars per berry was calculated by multiplying the percent total soluble solids measured at harvest by the berry weight.

Immediately at harvest for both years, 20 berries were collected from each vine, berry skins were peeled by hand and placed into 15 mL centrifuge tubes with perforated lids. The berry skins were freeze-dried (Cold Trap 7385020, Labconco, Kansas City, MO, USA) until no change in mass could be detected. A tissue lyser (MM400, Retsch, Germany) was used to grind dried tissues into a fine, homogenous powder to determine the presence and quantity of low molecular weight flavonoid compounds present in the sample. The method described by Martínez-Lüscher et al. (2019a) was followed. For each sample, 50 mg of berry skin powder was extracted in methanol: water: 7 M hydrochloric acid (70:29:1, V/V/V), and held at 4⁰C overnight. After the extraction, samples were centrifuged for 10 minutes at 4,000 rpm, and supernatant was collected. Extracts were filtered into HPLC vials by using PTFE membrane filters (diameter: 13 mm; 0.45 μm; CELLTREAT Scientific Products, Pepperell, MA, USA), and analyzed on an Agilent 1260 series reversed phase HPLC system (Agilent 1260, Santa Clara, CA, USA) with a DAD. At 25 °C with flow at 0.5 mL per minute, separation was performed on a reversed-phase C18 column LiChrospher ® 100, 250 mm, 4 mm with a 5 µm particle size and a 4 mm guard column of the same material. The mobile phase, which was designed to avoid co-elution of anthocyanins and flavonols (Martínez-Lüscher et al., 2019a), consisted of a constant 5.5 % of formic acid and the following gradient (V/V) of acetonitrile in water: 0 minute 8 %, at 25 minutes 12.2 %, at 35 minutes 16.9%, at 70 minutes 35.7 %, 65% between 70 and 75 minutes, and 8 % between 80 and 90 minutes. The peak area of absorbance at 520 nm was used to identify anthocyanins, and absorbance at 365 was used for identification of flavonols. Flavonols and anthocyanins were quantified determining the peak area of the absorbance at 365 nm and 520 nm, respectively, and using quercetin 3-O-glucoside and malvidin-3-O-glucoside chloride (Extrasynthèse, Genay, France) as quantitative standards.

All data was analyzed using SAS (v. 9.4 SAS Institute, Cary, NC). The same plants were measured in both years during the execution of the experiment. The data was subjected to three-way analysis of variance for a year × virus status × crop level using the PROC GLM procedure of SAS. Whenever the year effect was significant, the analysis was conducted separately for each year. Post-hoc analyses were conducted using Duncan’s new multiple range test at p < 0.05.

The authors acknowledge funding provided by the USDA-NIFA 2019-51181-30020 for the execution of the project