Plant material.
We obtained callused, unrooted cuttings of ‘Chardonnay’ (V. vinifera; FPS selection 79.1) and ‘Merlot’ (V. vinifera; FPS selection 15) from Inland Desert Nursery (Benton City, WA, USA). We chose these cultivars because they are widely grown and popular in local vineyards; in addition, most of the V. vinifera acreage in Washington is planted to own-rooted, nongrafted vines. The callused, unrooted cuttings were rooted in situ to their experimental pots; this ensures roots were not precolonized with AM fungi before the experiment. Plants were grown in 4-L pots on greenhouse benches [Washington State University (WSU), Richland, WA, USA] in a randomized complete block design for 5 months. The experiment was established on 2 Oct 2019, and the vines were destructively removed from pots on 2 Mar 2020. There were a total of 15 vine replicates of each treatment (2 cultivars × 4 treatments × 15 replicates = 120 experimental units). Each pot contained a single plant.
Soil.
We collected field soil (Warden silt loam) from a local agricultural field (46.2544118, −119.7283880) near existing research vineyards at the WSU Irrigated Agricultural Research and Extension Center (Prosser, WA, USA). Properties and nutrients for the field soil included the following: pH 7.9; 4 ppm nitrate; 1.5 ppm ammonium; 31 ppm Olsen P; and 214 ppm potassium (Soiltest Farm Consultants, Inc., Moses Lake, WA, USA) (Supplemental Table S1). We mixed the field soil 1:1 (by volume) with medium-course landscaping sand (Beaver Bark, Richland, WA, USA) to improve drainage and autoclaved it twice (121 °C for 2 h, rest for 24 h) to eliminate resident soil organisms, including pests and pathogens. Properties and nutrients for the autoclaved sand:soil mix included the following: pH 8.0; 2.6 ppm nitrate; 7.1 ppm ammonium; 10 ppm Olsen P; and 271 ppm potassium (Soiltest Farm Consultants, Inc.) (Supplemental Table S1). All pots in the experiment contained the same autoclaved sand:soil substrate to which the AM fungal inoculant and/or P fertilizer was added, which is a common method for assessing the plant growth response to mycorrhizal fungal inoculants (Cheeke et al. 2019; Reynolds et al. 2006).
Arbuscular mycorrhizal fungal inoculant product.
We used MycoBloom as the mycorrhizal inoculant product for the AM fungal inoculation treatment (MycoBloom LLC), which included the following AM fungal species: Acaulospora spinosa; Cetraspora pellucida; Claroideoglomus claroideum; Claroideoglomus lamellosum; Entrophospora infrequens; Funneliformis mosseae; and Racocetra fulgida. MycoBloom was used as the inocula for this experiment for the following reasons: it is widely available for purchase from online retailers; it contains seven different AM fungal species that have been shown to improve the growth of a variety of perennial plant species (Bauer et al. 2017; Cheeke et al. 2019), including those in perennial agroecosystems (Koziol et al. 2019); the fungi in the inocula were isolated from perennial plants and thus may associate well with perennial grapevines; some of the fungal species in the inoculant are also found in vineyards, including E. infrequens, C. claroideum, and F. mosseae (Cheng and Baumgartner 2004); and the AM fungi in this inoculant product have been used successfully to promote plant growth in more than 15 peer-reviewed scientific publications (Cheeke et al. 2019; Koziol and Bever 2016), demonstrating the viability of the product. This is important because some commercial mycorrhizal inoculant products fail to establish, even when added to sterilized soil/substrate, thus yielding no experimental treatment or effects (Salomon et al. 2022). Each inoculated pot in our experiment had 400 cm3 of the fungal inoculant (10% inoculation rate by volume) added to the rooting zone of each vine, and the remainder of the pot contained autoclaved sand:soil mix (Supplemental Fig. S1). Although this is higher than the 2.5% inoculation rate suggested for container gardening on the MycoBloom website, we used a 10% inoculation rate for this study for the following reasons: it is within the range of effective inoculation rates reported for greenhouse studies that tested the perennial plant growth response to AM fungal inoculation (Cheeke et al. 2019); other studies have shown that higher application rates of inoculant products are often needed to be effective (Koziol et al. 2019); and not all spores/propagules in formulated products may be viable at the time of application. By adding the same amount of inocula to the rooting zone of each vine, we were able to reduce the variation in the amount of AM fungi that each pot received because heterogeneity and patchiness can occur when trying to mix large batches of soil with inocula before assembling the experimental units. Moreover, because the inocula was added at the rooting zone at planting and covered with autoclaved sand:soil mix, the chance of the fungal treatment contaminating other pots in the greenhouse was reduced (e.g., via splashing, dust, or cross-contamination while measuring). The symbiotic relationship between plant roots and AM fungi is dynamic, and the fungal hyphae will continue to grow and colonize roots over the course of the experiment through the production of fungal hyphae. The uninoculated control pots contained only autoclaved sand:soil mix. We did not add autoclaved MycoBloom to the control pots because autoclaving is known to release a flush of nutrients from killed organisms and through chemical changes to substrates under high temperature and pressure (Anderson and Magdoff 2005; Berns et al. 2008; Skipper and Westermann 1973). Although calcined clay was included as a filler in MycoBloom (calcined clay is sometimes added to inoculant products to maintain moisture in greenhouse pots), our plants were watered daily; therefore, the small amount of calcined clay in the inoculated pots was unlikely to affect plant growth over the course of the experiment. To account for potential effects of nonmycorrhizal microbes present in MycoBloom, each pot received 50 mL of a microbial filtrate prepared from the fungal inocula, which was filtered through a sieve (38 μm) and then through filter paper (5–10 μm), allowing bacteria to pass through, but not fungal spores, roots, or larger organisms.
Potted plant management.
The daily temperature in the greenhouse was recorded digitally by Hobo data loggers (Onset, Cape Cod, MA, USA), and temperatures and humidity levels on each bench were monitored daily with benchtop digital recorders (AcuRite, Lake Geneva, WI, USA). The average greenhouse daily low temperature was 16 °C, and the average daily high was 28 °C. Humidity ranged from 20% to 40% during the growing period. Vines received additional lighting beginning at 76 d after planting to achieve a total photoperiod of 16 h from 1000-W high-pressure sodium bulbs. Vines were watered daily for 3 to 4 min using an automated drip irrigation system (drip rate: 25 mL/min). We managed powdery mildew (foliar and fruit fungal disease caused by Erysiphe necator) using a mineral oil spray per the manufacturer’s guidelines for grapes (12 to 28 mL/L; 1.2%–2.8%; PureSpray GREEN, Intelligro, Canada) every 10 d, starting at the first sign of infection (∼80 d after planting). The dilute mineral oil spray was added directly to the surface of the leaves, and there was no direct contact between the foliar mineral oil spray and mycorrhizal fungi in the roots.
When most vines had three to four true leaves (∼20 d after planting), we added a P fertilizer used by local growers (NUE 0–30–0; 8.6 g P2O5/pot; BioGro, Mabton, WA, USA) to the P fertilizer treatments. Each vine was also fertilized with a P-free (15–0–15) fertilizer (0.08 g/pot; Simple Lawn Solutions, Lake Panasoffkee, FL, USA) 111 d after planting, which added N and K to the soil to reduce the potential for macronutrient deficiencies during the experiment. To reduce the potential for micronutrient deficiencies, we applied a foliar micronutrient treatment (∼1% each B, Cu, Fe, Mg, Mn, Zn; 0.26 mL/vine; BioGro, Mabton, WA, USA) twice thereafter using a hand-pumped sprayer.
Plant growth measurements.
We recorded the initial bud number for each cutting at the time of planting to account for variations in size before treatment. This number was used as a covariate in the statistical analysis. We began collecting shoot length data 60 d after planting using a flexible measuring tape, and we measured from the base of the shoot to the tip of the apical meristem. We continued to record shoot length every 30 d for the remainder of the experiment. At 60 d, vines were pruned to one primary shoot that was trained onto a 1.2-m-high bamboo stake (Schreiner 2007). When the primary shoot reached the top of the stake, it was cut off to encourage lateral shoot growth instead of vertical shoot growth, and the following length measurements included lengths from lateral shoots (Schreiner 2007).
At 5 months after planting, we destructively harvested the vines and separated the roots from the shoots for biomass. The roots were thoroughly washed to remove all traces of soil and sand from the root system and blotted dry with a paper towel before recording fresh weights of the whole root system. Subsamples of fine roots were collected from multiple parts of the root system to ensure a representative subsample from each vine was collected to assess the percentage of mycorrhizal colonization of the roots. The fresh weights of the whole root system and the root subsample were recorded separately, and a dry weight conversion was used to add back the weight of the subsampled roots to obtain the total root biomass (g, dry weight) before the analysis. Shoots and roots were dried for 48 h at 70 °C to collect aboveground and belowground biomass data (g, dry weight). Plant tissue (leaf, petiole, green stem) was collected from all vines to determine nutrients, including N, P, K, and Ca (KUO Testing Laboratories, Pasco, WA, USA).
Mycorrhizal colonization.
Roots were cleared and stained to observe fungal structures (modified from Phillips and Hayman 1970). Briefly, roots were cleared using 10% KOH (simmered for 20 min), soaked at room temperature in 5% bleach solution for 2 min to lighten the roots, acidified at room temperature in 5% lactic acid, and then stained at room temperature with 0.05% trypan blue in lactoglycerol. Root subsamples from each plant were cut into ∼1-cm fragments and mounted onto microscope slides (one slide per plant; 120 slides in total). Roots were assessed for the presence or absence of AM fungal structures, including the hyphae, arbuscules, and vesicles, out of 100 intersections per sample using a compound microscope at 200× total magnification with a vertical crosshair in the eye piece for scoring each intersection (McGonigle et al. 1990). The percentage of colonization of AM fungi was determined as the number of intersections containing AM fungal structures out of 100 total root intersections analyzed.
Statistical analysis.
Data were analyzed using R (R Core Team 2021). To calculate the mycorrhizal growth response of each cultivar, we used the following formula: [(BiomassInoculated − BiomassUninoculated)/BiomassUninoculated]*100 for each P fertilizer treatment. The mycorrhizal growth response represents the relative change in aboveground biomass caused by the addition of AM fungal inocula in the different P fertilizer treatments. This resulted in a total of the following four mycorrhizal growth response values: ‘Merlot’ to AM fungi with no P fertilizer added; ‘Merlot’ to AM fungi with P fertilizer added; ‘Chardonnay’ to AM fungi with no P fertilizer added; and ‘Chardonnay’ to AM fungi with P fertilizer added. For example, a mycorrhizal growth response of 23 represents a 23% increase in the aboveground biomass of the inoculated vines compared with the uninoculated vines in a particular P fertilizer treatment.
We used linear mixed effects models in the “lme4” package (Bates et al. 2015) to test the effects of cultivar, mycorrhizal inocula, P fertilizer, and their interactions on vine shoot length, shoot biomass (leaves, petioles, stem), root biomass, root-to-shoot ratio, percentage of mycorrhizal colonization of the roots, and nutrients of vine tissue. Because ‘Merlot’ and ‘Chardonnay’ varied in their growth rates and their responses to the experimental treatments (Supplemental Table S2), and because we could not meet the assumption of equal variance within each group to include both cultivars in the same model, the final statistical analyses were conducted for each cultivar separately. To determine the effect of AM fungal inocula on vine growth in each P fertilizer treatment, we focused primarily on the AM fungal × P fertilizer interaction term in the model and the results of the a priori contrasts comparing growth response of the inoculated vines and uninoculated vines within each P fertilizer treatment.
In each linear mixed effects model, the initial bud number was included as a covariate to account for potential variations in the initial size of the vine cuttings before treatment, and the greenhouse block was included as a random effect to account for potential environmental variations. Because uninoculated vines contained no fungal structures in roots, only the inoculated vines were included in the models testing the effects of P fertilizer on AM fungal colonization for each cultivar. We tested the significance of terms in our linear mixed effects models using an analysisi of variance with the R package car (Fox and Weisberg 2011). Assumptions for each model were checked using a visual inspection of residuals from quantile–quantile plots. The effects of AM fungal inocula on vine growth and/or tissue nutrients were then decomposed into a priori contrasts that separately tested the average growth response of each cultivar to AM fungal inocula in each P fertilizer treatment using the estimated marginal means ‘emmeans’ package of R. Tukey adjustments were performed to limit the type 1 error rate caused by α-inflation during the contrast comparisons.