Introduction

Gray mold, caused by Botrytis cinerea, is among the most serious diseases in vineyards, and its control has been almost exclusively based on the application of chemical fungicides. The use of such phytochemicals is not considered to be a long-term solution because they are potentially harmful for the environment and human health and contribute to the development of resistant pathogen strains. So far, considerable interest has been dedicated to alternative strategies including activation of plant defenses by non-pathogenic microorganisms. This biocontrol strategy can be defined as the mechanisms by which plants exhibit increased levels of resistance to a broad spectrum of pathogens by the prior activation of genetically programmed defense pathways. The most extensively studied type of resistance is systemic acquired resistance (SAR) (Durrant and Dong 2004), which is expressed locally and systemically after a localized infection by a necrotizing pathogen or the application of some chemicals such as benzothiadiazole (Conrath et al. 2002). SAR is characterized by the accumulation of salicylic acid and pathogenesis-related (PR) proteins. The colonization of roots with beneficial rhizospheric or endophytic bacteria (Van Loon et al. 1998; Compant et al. 2005; De Vleesschauwer et al. 2006; Bakker et al. 2007; Verhagen et al. 2010) can also lead to a type of systemic resistance, commonly known as induced systemic resistance (ISR). This kind of resistance has been demonstrated with different plant species against several pathogens under various conditions (Pieterse et al. 1996; Van Loon et al. 1998; Iavicoli et al. 2003; Meziane et al. 2005).

The bacteria-mediated ISR has recently been described as a result of bacteria-mediated priming of rapid molecular and cellular defense responses, such as strong and fast transcription of defense-related genes, callose deposition, accumulation of active oxygen species, papillae formation and accumulation of phytoalexins (Ahn et al. 2007; Verhagen et al. 2004, 2010). Although the priming and induction of direct defenses constitute different mechanisms of protection, both result in a phenotypically similar resistance to pathogens. Furthermore, some induced resistance phenomena are based on a combination of both mechanisms (Verhagen et al. 2011).

The accumulation of stilbenic phytoalexins in leaves and grape skins has been reported in different conditions including fungal infection (Langcake and Pryce 1976; Jeandet et al. 2010), treatment with UV-irradiation (Langcake and Pryce 1976; Jeandet et al. 1991), heavy metals (Adrian et al. 1996), but also treatment with various microbe-associated molecular patterns (MAMPs) that enhance resistance against Botrytis cinerea and Plasmopara viticola (Aziz et al. 2003, 2006). The main stilbenes in grapevine are resveratrol and its oligomeric derivative ɛ-viniferin, a glycosylated form of resveratrol known as piceid (5,4′- dihydroxystilbene-3-O-β-glucopyranoside), or a dimethylated derivative, pterostilbene (3,5-dimethoxy-4′-hydroxystilbene) (Jeandet et al. 2010, 2014). It was also shown that an isomer of ɛ-viniferin, δ-viniferin, is one of the major stilbenes produced as a result from resveratrol oxidation in grapevine leaves infected by P. viticola (Pezet et al. 2003). Stilbenic phytoalexins have been shown to possess biological activity against a wide range of pathogens and can be considered as markers for plant disease resistance in grapevine (Coutos-Thevenot et al. 2001; Pezet et al. 2003; Aziz et al. 2006). Recently, we showed that phytoalexin accumulation is induced in grapevine plantlets by rhizobacteria or bacterial MAMPs that could be linked to the development of ISR (Verhagen et al. 2011).

Triggering of defenses has mainly been investigated using one single bacterium, but different bacteria are present as complex mixtures in plants or in the rhizosphere. In practical agriculture, the use of single beneficial microorganisms often exhibits a partial performance (Magnin-Robert et al. 2007). Thus, more emphasis was laid on the combined use of two or more compatible strains of bacteria, which turned out to be more successful than individual treatments (Nandakumar et al. 2001; Bharathi et al. 2004; Thilagavathi et al. 2007; Magnin-Robert et al. 2013). Though several researchers have reported the increased efficiency of biological control with a mixture of fluorescent Pseudomonas spp, relatively few mechanisms have been proposed to explain the obtained increased resistance.

It is generally believed that energy costs can arise following pathogen attacks from the allocation of resources to defense and away from plant growth and development. However, induced resistance by beneficials evolved to save energy by priming plants for enhanced defense upon pathogen challenge (Van Hulten et al. 2006; Verhagen et al. 2011). Priming is a phenomenon that boosts plants for fast and strong activation of cellular defense responses only once primed plants are attacked by a pathogen or an insect. As a consequence, priming is less costly for the plant than induced defenses (Verhagen et al. 2004; Conrath et al. 2006; Van Hulten et al. 2006; Verhagen et al. 2010). Different bacteria isolated from healthy tissues and the rhizosphere of grapevine plants (Trotel-Aziz et al. 2008) have been shown to enhance disease resistance in grapevine leaves when applied at the root level under in vitro conditions (Verhagen et al. 2011). Among them, Pseudomonas fluorescens PTA-CT2 (CT2), Pantoea agglomerans PTA-AF2 (AF2) and Bacillus subtilis PTA-271 (271) can trigger immune responses and enhanced local and systemic resistance against B. cinerea when they are sprayed on the aerial parts of the plant or applied to the plant roots (Magnin-Robert et al. 2007; Verhagen et al. 2011; Gruau et al. 2015). These bacteria triggered early oxidative burst and primed plant cells and leaves of in vitro plantlets for enhanced stilbenic phytoalexin production to various extents. They also triggered an increased activity of chitinase and β-1,3-glucanase in systemic leaves and berries in vineyards when used as a soil drench (Magnin-Robert et al. 2007, 2013). Although these bacteria are able to enhance resistance to control B. cinerea on grapevine plants in vineyard conditions, relationships between bacteria-induced resistance and phytoalexin accumulation as well as grape yield remained to be deciphered.

This study was undertaken to determine whether induced resistance in grapevine with single or binary combinations of beneficial bacteria is associated to enhanced systemic production of phytoalexins and whether these treatments can impact grape yield under vineyard conditions. For this purpose, bacterial suspensions were applied by drenching soil of grapevine plants and resistance to gray mold as well as phytoalexin accumulation were evaluated in both leaves and berries. Grape yield has also been assessed at full ripening.

Materials and methods

Cultivation of bacterial strains

The bacterial strains used in this study were isolated from grapevine plants and rhizospheric soils of vineyards in Champagne (France). Bacillus subtilis PTA-271 (271) has been isolated from the rhizosphere and Pseudomonas fluorescens PTA-CT2 (CT2) and Pantoea agglomerans PTA-AF2 (AF2) were isolated form tissues of healthy grapevine plants (Trotel-Aziz et al. 2008). Each bacterium was grown separately in liquid King B (KB) medium (King et al. 1954) at 28 °C with continuous shaking (100 rpm) for 24 h before use. Bacteria were collected by centrifugation and resuspended in 10 mM MgSO4 at 1 × 108 CFU/ml before binary mixtures were made (with half dose of each isolate) for applications in vineyards. All treatments with single or mixed strains consisted of 1 × 108 CFU/ml as a final concentration.

Fungal pathogen

The highly virulent Botrytis cinerea strain 630 (kind gift of Y. Brygoo, INRA, Versailles, France) was grown and maintained on potato dextrose agar (PDA, Sigma) in 200-ml Erlenmeyer flasks at 22 °C with a 16/8 h photoperiod (cool white neon tubes 3 000 lx). Conidia were harvested from 14- to 20-day-old cultures and collected by rubbing with 10 ml of sterile distilled water, filtered on sterile glass beads filter to remove mycelia, and the concentration was determined using Thoma slide and photonic microscope, and adjusted with sterile distilled water to 1 × 106 conidia per ml.

Field experiments

Fifteen-year-old grapevine plants (Vitis vinifera cv. Chardonnay, on 41B rootstock) were treated during 2006 in two vineyards with similar environmental and agronomic conditions, Nogent L’abbesse (49.255°N latitude and 4.156°E longitude) and Cernay-les-Reims (49.2667°N latitude and 4.100°E longitude) in the Champagne area (France). Vine spacing was 1.05 m within the row and 1.20 m between rows in both vineyards. Fungicides needed for the control of downy and powdery mildews that have no effect on B. cinerea, as well as canopy management were applied in both vineyards, but no botryticides were used during growing season. Treatments consisted of soil drenches with 250 ml of 1 × 108 CFU/ml bacterial suspension per plant, which flooded the first 10 cm of the soil in contact with the rootstock. Bacterial suspensions were applied twice (June 7th and July 13th, 2006), both individually (CT2, AF2, 271) and in mixtures of CT2+AF2, CT2+ 271 and AF2+271. Control plants were treated with 250 ml MgSO4. The experimental design (given in Supplementary Material S1) was similar in both vineyards, as randomized complete blocks with twelve plants per plot and three replications. Both leaves from the top of shoots (30 per treatment) and berries (10 bunches per treatment) without visible disease symptoms were harvested at different dates during the growing season and stored at −80 °C before analysis of the produced phytoalexins.

Challenge inoculation and disease assessment

Young fully expanded leaves were excised from the top of shoots (30 leaves per treatment) 2 months after the second application of bacteria, immediately placed into moistened plastic bags and taken to the laboratory. Leaves were then rinsed with sterile distilled water, placed on sterile wet absorbing paper in Petri dishes. One needle-prick wound was applied on the middle of the abaxial surface of each leaf, and the fresh wounds were covered with 10-μl drops of a conidial suspension of B. cinerea (1 × 106 conidia per ml). The Petri dishes were then placed at 22 °C with a photoperiod of 16 h of light. Disease development in leaves was measured as average diameter of lesions formed 4, 7 and 12 days post-challenge with B. cinerea, and disease rating was expressed as percentage of lesion diameter relative to control. For grapevine berries, disease severity was evaluated at full ripening (September 25, 2006) in vineyards by measuring the percentage of naturally infected berries per cluster (disease intensity) based on actual berry count, and the percentage of infected clusters per plant (disease frequency). A berry or cluster was considered as infected if it carried the typical sporulation of B. cinerea. Measurements were performed on 50–60 clusters.

Phytoalexin analysis

Stilbenes were extracted with pestle and mortar from leaves and berries in 2 ml of 85 and 100 % methanol, respectively. Tubes were placed overnight at 4 °C and then centrifuged for 10 min at 5000 g. Supernatants were evaporated under nitrogen stream and residues were solubilized in 1 ml of 100 % methanol. All extraction steps were done in subdued light conditions. The extracts were clarified by filtration through a Millex-GN 0.22-μm filter (Millipore, St-Quentin en Yvelines, France) before HPLC analysis. Samples were loaded onto a Lichrocart C-18 reversed phase column (250 × 4 mm, 5 μm, Waters, St-Quentin en Yvelines, France) equilibrated with a 90:10 (v/v) H2O/acetonitrile mobile phase. Phytoalexins were eluted with a linear gradient of 10 to 85 % acetonitrile at a flow rate of 1 ml/min. The phytoalexins, trans-resveratrol and ε-viniferin were detected with a photodiode array detector coupled to a fluorometer (λex = 330 nm, λem = 374 nm) and quantified on the basis of standard calibration curves (Hatmi et al. 2015).

Grape yield evaluation

At harvest, grape yield was determined for each treatment. The total number of healthy clusters was counted on each plant of all treatments and control. Each healthy cluster was weighted separately and the average weight was determined for each treatment.

Data analysis

Each treatment consisted of three replications and each replication consisted of 12 plants per block. The effects of selected bacteria on disease development, evaluated on 30 leaves per treatment and 50–60 clusters per treatment and per block, were performed by using analysis of variance (ANOVA), and Duncan’s multiple range test (P ≤ 0.05) was used for post-hoc comparison of means. Statistica software (Statsoft Inc., Tulsa, USA) was used for statistical data analysis. Defense reactions were determined on 10 leaves and 10 berries (from 3 to 4 clusters per treatment) per site all in triplicates. Data are means ± standard deviation, and Duncan’s multiple range test (P ≤ 0.05) was used for post-hoc comparison of means calculated from pooled-data from each vineyard sites.

Results

Protective effects of individual or binary mixtures of bacteria against B. cinerea challenge on grapevine leaves

The development of gray mold disease was followed on the leaves collected from the two vineyards to evaluate the effect of bacteria on grapevine resistance against B. cinerea. As shown in Fig. 1, gray mold lesions were widely developed 4 days after challenge (d.a.c.) in leaves of the control plants, whereas the sizes of the lesions were significantly reduced in the inoculated plants (Fig. 1A). At 7 (Fig. 1B) and 12 d.a.c. (Fig. 1C), the lesion sizes in the control plants were larger, but progression of the lesions remained limited in bacterized plants. Among the single bacteria, B. subtilis PTA-271 (271) showed the highest protective effect, with about 75 % reduction of the symptoms of B. cinerea compared to the control. P. fluorescens PTA-CT2 (CT2) and P. agglomerans PTA-AF2 (AF2) provided a lower level of protection, with approximately 30 % reduction of Botrytis symptoms compared to the control. Regarding binary mixtures, treatment with CT2+AF2 was the most protective against B. cinerea, with 80 to 90 % reduction of Botrytis symptoms. The two other combinations CT2+271 and AF2+271 showed protecting levels similar to those observed with single isolates CT2 and AF2 (with approximately 35 to 40 % reduction of the symptoms of B. cinerea). Thus combination of the two bacteria that are the least effective when they are alone (CT2 and AF2) provided the best protection. However, adding CT2 or AF2 to strain 271 resulted in a significant reduction of 271 efficacy against B. cinerea on leaves, especially at 7 and 12 d.a.c. The variability between the two vineyards remained comparable to that observed for replicates within vineyards (not shown).

Fig. 1
figure 1

Protection of grapevine leaves against Botrytis cinerea after root treatment with single and mixtures of beneficial bacteria. Bacteria were applied on June 7 and July 13, 2006, by soil drenching at a final density of 1 × 108 CFU/ml. Treatments consisted of control (without bacteria), P. fluorescens PTA-CT2 (CT2), P. agglomerans PTA-AF2 (AF2), B. subtilis PTA-271 (271), CT2+AF2, CT2+271, and AF2+271. Two months after the second treatment (September 14), expanded young leaves from control and treated plants were detached and challenged with B. cinerea. Disease was evaluated as percent of necrotic lesion area compared to control (disease index) on 30 leaves per treatment recorded 4 (A), 7 (B), and 12 (C) days after challenge inoculation (d.a.c.). Different letters indicate statistically significant differences between treatments (Duncan’s multiple range test; P < 0.05)

Full size image

Protective effects of individual or binary mixtures of bacteria against B. cinerea on grape berries

Disease severity (Fig. 2A) and frequency (Fig. 2B) in the control plants were about 10 and 27 % infection, respectively. Berries from bacteria-inoculated plants showed significant reduction of disease symptoms under natural contamination. The individual isolate 271 offered the best efficiency in terms of disease intensity reduction (about 87 %) when compared to the control. CT2 and AF2-treated plants also exhibited a significant reduction of disease intensity reaching about 78 and 72 %, respectively (Fig. 2A). Disease frequency was also reduced to a lower extent by individual CT2 and 271, while AF2 had no significant effect (Fig. 2B). Interestingly, the highest efficiency was observed with the CT2+AF2 binary mixture, which resulted in a 93 % reduction of disease intensity. This reduction was also very important with the other 2 binary mixtures; it was 87 % with CT2+271 and 78 % with AF2+271. Treatment with CT2+271 was the most efficient in reducing disease frequency (by approximately 65 %). Combinations CT2+AF2 and AF2+271 also showed significant reduction of disease frequency, which remained similar to those observed with individual isolates (Fig. 2B). As for the leaves, the variability observed on berries between the two vineyards remained comparable to that reached for replicates within vineyards.

Fig. 2
figure 2

Protection of grapevine berries against Botrytis cinerea after root treatment with single and mixtures of beneficial bacteria. Bacteria were applied on June 7 and July 13, 2006, by soil drenching at a final density of live bacteria of 1 × 108 CFU/ml. Treatments consisted of control (without bacteria), P. fluorescens PTA-CT2 (CT2), P. agglomerans PTA-AF2 (AF2), B. subtilis PTA-271 (271), CT2+AF2, CT2+271, and AF2+271. Disease intensity (A) and frequency (B) were measured in both vineyards from all treated-plants at full ripening (September 25) as percent of infected berries per cluster and percent of infected clusters per plant, respectively. Different letters indicate statistically significant differences between treatments (Duncan’s multiple range test; P < 0.05)

Full size image

Phytoalexin production in grapevine leaves and berries after treatment with individual or binary mixtures of beneficial bacteria

To provide insights into defense reactions associated with bacteria-induced resistance, stilbenic phytoalexins were monitored in both leaves and berries of control and treated plants from the two vineyards. Following the first bacterial treatment, leaves collected 12 days post treatment (dpt) showed an increased resveratrol level (Fig. 3A) as well as total stilbenes (Fig. 5A) only in CT2-treated plants. At 26 dpt, resveratrol amount continued to increase (Fig. 3A) as did total stilbenes (Fig. 5A) in response to all bacterial treatments compared to the control. At 26 dpt with CT2, 271 or CT2+AF2, leaves showed a significant increase in ε-viniferin compared to the control (Fig. 3B), and those treated with AF2 showed a slight but not significant increase in δ-viniferin (Fig. 3C). Interestingly, compared to other bacterial treatments at this date, treatment with CT2+AF2 induced the highest level of total phytoalexins in leaves (Fig. 5A), which reflected the highest level of resveratrol produced (Fig. 3A). Thirty five dpt, total stilbene levels in grapevine leaves drastically dropped by half compared to data from the previous date (Fig. 5A), although treatment with CT2+271 showed a significant increase of ε-viniferin and total phytoalexins (Figs. 3B and 5A). Forty days after the second bacterial treatment, that is at 75 dpt, resveratrol (Fig. 3A) and ε-viniferin (Fig. 3B) levels increased again to reach their highest levels whatever the bacterial treatment. Bacterial treatments showing the highest phytoalexin-inducing effects in leaves were CT2 and CT2+AF2 for total stilbenes and resveratrol content (Figs. 3A and 5A), and CT2, CT2+AF2 and CT2+271 for ε-viniferin (Fig. 3B). Leaf level of δ-viniferin also increased (unsignificant) but to a lesser extent than that of ε-viniferin (Fig. 3C).

Fig. 3
figure 3

Production of phytoalexins by grapevine leaves after root treatment with single and mixtures of live bacteria. Live bacteria were applied June 7 and July 13, 2006, by soil drenching at a final density of live bacteria of 1 × 108 CFU/ml. Treatments consisted of control (MgSO4 without bacteria), P. fluorescens (CT2), P. agglomerans (AF2), B. subtilis (271), CT2+AF2, CT2+271, and AF2+271. Phytoalexins, resveratrol (A), ε-viniferin (B) and δ-viniferin (C) were quantified on various days after first treatment (dpt) in fully expended young leaves from control and treated plants. Data are means from three replicates ± SD. The dotted bar indicates the time of the second treatment, and different letters indicate statistically significant differences between treatments (Duncan’s multiple range test; P < 0.05)

Full size image

In berries, stilbene levels were significantly increased at 61 dpt by some bacterial treatments (Figs. 4 and 5B). At that time, bacterial treatments showing the highest resveratrol-inducing effect in berries were CT2 and CT2+AF2, as in leaves collected at 26 and 75 dpt. However, it can be observed that at 61 dpt only individual bacteria induced a significant increase of resveratrol dehydrodimers ε-viniferin (Fig. 4B) and its isomer δ-viniferin (Fig. 4C). Over time, the highest stilbene level produced in berries was observed with CT2 with a peak at 75 dpt (Fig. 5B) which mainly corresponded to the accumulation of ε-viniferin (Fig. 4B) and δ-viniferin (Fig. 4C). At 75 dpt, the mixture CT2+AF2 also induced a strong accumulation of ε-viniferin in berries as did strains CT2, AF2 and to a lesser extent 271 (Fig. 4B). After 75 dpt, the grapevine stilbene levels slightly decreased, but remained significantly higher with CT2, AF2 and CT2+AF2 than the control at 94 and 99 dpt (Fig. 5B). Consequently, CT2 and CT2+AF2 treatments induced the highest total phytoalexin levels in both leaves (Fig. 5A) and berries (Fig. 5B). At 99 dpt, almost all treatments induced a significant accumulation of resveratrol in berries (except 271 strain, Fig. 4A), while ε-viniferin was only significantly increased by CT2, AF2, and CT2+AF2 (Fig. 4C). CT2 alone induced a fast accumulation of resveratrol in leaves (Fig. 3A), while CT2+AF2 induced a later but stronger increase of resveratrol both in leaves and berries (Figs. 3A and 4A). Surprisingly, 271 which strongly protected grapevine leaves and berries against B. cinerea (Figs. 1 and 2A), was found to induce only small amounts of phytoalexins both in leaves and berries compared to the other bacterial treatments.

Fig. 4
figure 4

Production of phytoalexins by grapevine berries after root treatment with single and mixtures of live bacteria. Live bacteria were applied June 7 and July 13, 2006, by soil drenching at a final density of live bacteria of 1 × 108 CFU/ml. Treatments consisted of control (MgSO4 without bacteria), P. fluorescens (CT2), P. agglomerans (AF2), B. subtilis (271), CT2+AF2, CT2+271, and AF2+271. Various days after first treatment (dpt), phytoalexins, resveratrol (A), ε-viniferin (B), and δ-viniferin (C) were quantified at different stages of berry development from control and treated plants. Data are means from three replicates±SD and different letters indicate statistically significant differences between treatments (Duncan’s multiple range test; P < 0.05)

Full size image
Fig. 5
figure 5

Total phytoalexins in grapevine leaves and berries after root treatment with single and mixtures of bacteria. Treatments consisted of control (MgSO4 without bacteria), P. fluorescens (CT2), P. agglomerans (AF2), B. subtilis (271), CT2+AF2, CT2+271, and AF2+271. Phytoalexins were quantified at various days after first treatment (dpt) in leaves (A) and berries (B). Data are means from three replicates±SD and different letters indicate statistically significant differences between treatments (Duncan’s multiple range test; P < 0.05). The dotted bar indicates the time of the second treatment

Full size image

Bacterial treatments did not affect grape yield

Soil applications of individual bacteria or their binary mixtures did not affect negatively the number of healthy clusters per plant at full ripening as compared with untreated plants (Fig. 6A). On the contrary, regarding the number of clusters grape yield was slightly but significantly increased after soil treatments with CT2, 271, CT2+271, and to a lesser extent with the CT2+AF2 mixture (Fig. 6A). Interestingly, the weight of clusters was also significantly increased following treatments with CT2, CT2+271, and CT2+AF2 (Fig. 6B), and to a lesser extent with the 271 strain (Fig. 6B). Treatments with the AF2 strain or with the AF2+271 mixture did not affect the number of healthy clusters per plant nor the average weight of clusters per plant.

Fig. 6
figure 6

Grape yield at harvest after treatment of grapevine plants with single and mixtures of beneficial bacteria. Treatments consisted of control (MgSO4 without bacteria), P. fluorescens (CT2), P. agglomerans (AF2), B. subtilis (271), CT2+AF2, CT2+271, and AF2+271. Yield was evaluated as the number of healthy clusters per plant (A) and average weight of clusters in g per plant (B). Data are means from three replicates of 12 plants ± SD and different letters indicate statistically significant differences between treatments (Duncan’s multiple range test; P < 0.05)

Full size image

Discussion

This study is a continuation of our previous work on the induced systemic resistance of grapevine by beneficial bacteria isolated from healthy grapevine plants in vineyards. Although induced systemic resistance (ISR) by some beneficial bacteria was reported to be associated to a priming state of enhanced phytoalexin accumulation in grapevine plantlets (Gruau et al. 2015) or cell suspensions (Verhagen et al. 2011), relationships between ISR and the responsiveness of stilbenic phytoalexins in both leaves and berries under vineyard conditions remained unknown. Moreover, the effects of bacterial mixtures on this defense response have not been reported so far. In this study, we compared the effectiveness of single or binary mixtures of selected strains Pseudomonas fluorescens PTA-CT2 (CT2), Pantoea agglomerans PTA-AF2 (AF2) and Bacillus subtilis PTA-271 (271) to induce resistance against B. cinerea and its potential connection with phytoalexin accumulation in leaves and berries and grape yield in vineyard conditions.

The present study demonstrated that both single and mixtures of selected bacteria can protect grapevine leaves and berries to various extent against gray mold disease in field-grown grapevine plants. This protective effect mainly involves an activation of ISR (Magnin-Robert et al. 2007, 2013), accompanied in most cases by a systemic accumulation of resveratrol or viniferins, especially following treatment with CT2 alone or in mixture with AF2 (CT2+AF2). Although selected strains did not antagonize one another when co-cultured in vitro (Suppl Fig. S2), synergistic protective effects were not observed when applied as mixtures. These effects may be explained by a low microbial density and/or a low host tissue colonization or a low expression of defense responses according to bacterial treatments in vineyard conditions. The great protective efficacy of CT2 alone or in combination with the other strains could probably be due, at least in part, to its dual action as an inducer of ISR (Verhagen et al. 2011) and an antagonist of B. cinerea as shown in vitro (Trotel-Aziz et al. 2008). Pseudomonas spp. and Bacillus spp. were often shown to induce systemic resistance against a variety of pathogens (van Loon et al. 1998; De Vleesschauwer and Höfte 2009; Lugtenberg and Kamilova 2009; Van der Ent et al. 2009; Pineda et al. 2010). Similarly, Pantoea spp. was able to protect grapevine against gray mold (Trotel-Aziz et al. 2008) or apple and pear against fire blight disease (Stockwell et al. 2010). The latter authors have observed that the biocontrol of fire blight on pear and apple trees by P. fluorescens A506 and P. agglomerans C9-1 was less effective with mixed inocula than with individual strains. This is consistent with our results indicating that adding CT2 or AF2 to strain 271 resulted in a significant reduction of the 271 efficacy against B. cinerea on leaves and to some extent in berries. We suggest that the low efficacy of certain bacterial mixtures could result from some competitiveness that might occur between bacterial populations in the field, or from negative interactions between defense signaling pathways induced by each bacterium. In contrast, combining the two bacteria that are the least effective when they are alone (CT2 and AF2) provided the best protection. This protecting effect may result from cooperative signaling pathways or potentiation of induced resistance by bacteria in vineyard.

Improved resistance against pathogen infections by beneficial bacteria is frequently associated to priming the plants for activation of various cellular defense responses that are subsequently induced upon pathogen attack (Conrath et al. 2002). The potentiated responses include oxidative burst (Verhagen et al. 2011), cell wall reinforcement (Benhamou and Bélanger 1998), and the production of secondary metabolites (Yedidia et al. 2003; Verhagen et al. 2010). Here, we showed that some bacteria applied in single or in binary mixtures can induce resveratrol and viniferin accumulation in leaves and berries and that renewing treatments during the growing season primed the accumulation of resveratrol and ε-viniferin in leaves and also δ-viniferin in berries. P. fluorescens CT2 triggered high accumulation of resveratrol, ε-viniferin and δ-viniferin in leaves and berries, while P. agglomerans AF2 led to a high accumulation of ε-viniferin with few amounts of δ-viniferin, and B. subtilis 271 only induced small amounts of these phytoalexins. Accordingly, the CT2+AF2 mixture triggered the highest accumulation of resveratrol and ε-viniferin in grapevine leaves and berries. However, although 271 showed a high protective effect on leaves and berries against B. cinerea, it induced lower levels of ε-viniferin and δ-viniferin than CT2 and AF2. Similar results were observed with 271+CT2 and 271+AF2 mixtures in berries. These results support the view that selected bacteria may use distinct signaling pathways to trigger ISR as reported by Verhagen et al. (2010). This could explain the discrepancies between the level of induced resistance and phytoalexin response, but cannot rule out the potential role of resveratrol and its derivatives in induced resistance (Aziz et al. 2006; Gruau et al. 2015). Phytoalexin synthesis in grapevine was also shown to be upregulated by abiotic stress (Hatmi et al. 2015).

The increased accumulation of stilbenic phytoalexins in protected leaves and berries from CT2 or CT2+AF2-treated plants strongly indicated the possible active role of these compounds in the ISR mechanisms. Accumulation of stilbenes was frequently associated with an enhanced grapevine resistance to both biotrophic and necrotrophic pathogens (Chong et al. 2009; Jeandet et al. 2014). The remarkable increase of ε-viniferin and δ-viniferin production in berries in some bacteria-treated plants (which also showed enhanced resistance to B. cinerea) indicated that these oligomers are possible markers for induced resistance to gray mold. Using in vitro plantlets, we showed that CT2, AF2 or 271 applied at the root level induced ISR against B. cinerea by priming aboveground plant parts for enhanced stilbene accumulation, which could explain their resistance level to the pathogen (Verhagen et al. 2011; Gruau et al. 2015). Stilbenes have antimicrobial activities (Adrian et al. 1997) and they also accumulated in response to different beneficial bacteria or MAMPs including oligosaccharides and the SA analogue, BTH (Aziz et al. 2006; Verhagen et al. 2010; Dufour et al. 2013). The overexpression of stilbene synthase, the key enzyme responsible for the synthesis of these phytoalexins, also conferred improved resistance to pathogens (Coutos-Thevenot et al. 2001; Jeandet et al. 2014).

The induction of defenses based on secondary metabolite production represents a significant metabolic cost for a plant since carbon and nitrogen resources devoted to plant growth can partially be used in the production of these molecules during plant-microbe interactions. However, the extent of such a “cost” related to the plant defense responses induced or primed by beneficial bacteria in vineyards remained unknown. In this study, the negative effect of B. cinerea on the grape harvested volume was evident whereby non-treated-plants showed higher percentage of Botrytis-infected grape bunches than treated plants (Fig. 2b). Application of CT2 or 271 alone or of CT2 in combination with the other strains showed no negative impact on grape productivity, defined as grape yield per plant (number of healthy bunches per plant multiplied by average weight of bunches). Therefore, increased resistance levels in AF2- and/or AF2+271-treated plants do not affect the grape yield. These results clearly indicate that it is possible to potentiate the plant resistance by beneficial bacteria without compromising grape yield in the vineyard. Different authors indicated that beneficial bacteria can induce ISR by priming aboveground plant parts for enhanced defense against pathogens without significant energy costs to plant metabolism and growth (Conrath et al. 2002; Verhagen et al. 2004; Walters and Heil 2007; Van der Ent et al. 2009). Priming has already been described in grapevine for resistance to Plasmopara viticola after treatment with BABA (Hamiduzzaman et al. 2005; Dubreuil-Maurizi et al. 2010), sulfated laminarin (Trouvelot et al. 2008) or beneficial bacteria (Verhagen et al. 2011). Perazzolli et al. (2011) also underlined that the activation of such a priming state in Trichoderma harzianum T39-treated grapevines did not require any notable costs, even after repetitive applications of the resistance inducer. Indeed, T39 did not affect shoot growth, weight of leaves, shoots and roots, leaf dimensions nor chlorophyll content.

Our results provide some evidence for the efficiency of individual bacteria or bacterial mixtures to induce systemic resistance in vineyard conditions. However, the relationship between induced resistance and phytoalexin accumulation seemed to be dependent on bacterial strain, bacterial combination and plant organ. This also suggests the existence of different pathways in grapevine leading to induced resistance to B. cinerea. These findings, together with enhanced accumulation of phytoalexins and beneficial effect of P. fluorescens, P. agglomerans and B. subtilis on grape systemic resistance against B. cinerea without compromising grape production, make the use of beneficial bacteria a promising low-impact tool in modern disease management in vineyards. Further investigations are needed to decipher mechanisms and factors underlying bacteria-induced priming of defense reactions as related to ISR in vineyards, and whether such beneficial bacteria can interact with the rhizospheric or endophytic microbiome.