New Potential Biological Limiters of the Main Esca-Associated Fungi in Grapevine

The strains Trichoderma harzianum TH07.1-NC (TH), Aphanocladium album MX95 (AA), Pleurotus eryngii AL142PE (PE) and Pleurotus ostreatus ALPO (PO) were tested as biological limiters against Fomitiporia mediterranea Fme22.12 (FM), Phaeoacremonium minimum Pm22.53 (PM) and Phaeomoniella chlamydospora Pc22.65 (PC). Pathogens were obtained from naturally Esca-affected ‘Nero di Troia’ vines cropped in Grumo Appula (Puglia region, Southern Italy). The antagonistic activity of each challenge organism was verified in a dual culture. TH and PO completely overgrew the three pathogens. Partial replacement characterized PE-FM, PE-PM, PE-PC and AA-PC interactions. Deadlock at mycelial contact was observed in AA-FM and AA-PM cultures. The calculated antagonism index (AI) indicated TH and PE as moderately active antagonists (10 < AI < 15), while AA and PO were weakly active (AI < 10). The maximum value of the re-isolation index (s) was associated with deadlock among AA-PM, AA-PC and PE-FM dual cultures. The tested biological limiters were always re-isolated when PO and TH completely replaced the three tested pathogens. TH and AA confirmed their efficiencies as biological limiters when inoculated on detached canes of ‘Nero di Troia’ in dual combination with FM, PC and PM. Nevertheless, additional experiments should be performed for a solid conclusion, along with validation experiments in the field.


Introduction
The disease known as 'Esca' is one of the longest-recognized and most devastating threats to grapevines (Vitis vinifera) and viticulture around the world [1]. The most recent studies consider this mycopathy as different diseases that overlap in the same vine or develop at different stages of plant life [1][2][3][4][5]. Cross-sections of rooted cuttings, trunk, branches, and shoots show brown to black spots often accompanied by a dark, viscous exudate ('black goo'). Longitudinally, xylem necrosis extends in columnar strips, called 'brown wood streaking' [1][2][3][4][5]. Brown wood streaking affects rooted cuttings (also named brown wood streaking of grapevine cuttings) and young (2-7 years) vines (defined as Petri disease, formerly also known as black goo, slow dieback, and Phaeoacremonium grapevine decline) [1][2][3]. Inside the trunk or the branches, the wood shows 'white rot' (which is the origin of the name Esca), as well as cracking or fissuring of the bark and wood (known as 'mal dello spacco'-cracking disease-in Italy) [1,3,6]. On leaves, small pale green or chlorotic zones, roundish or irregular, occur dispersed to the veins or adjacent the margin. Progressively, these areas increase, merge, and, in part, necrotize, and at last leave only a narrow strip of unaffected green tissue along the main veins. Depending on the cultivar, dead tissues appear dark brown to red-brown and the diseased leaves develop a 'tiger stripes' appearance. Occasionally, the necrotic zones of the lamina desiccate and detach and leaf margins become irregular [7]. These symptoms characterize grapevine leaf stripe disease, previously named 'young Esca' [1][2][3][4]. Brown wood streaking of rooted cuttings, from a forest soil sample, and Pseudomonas protegens strain DSM 19095T significantly inhibited P. chlamydospora and P. minimum in vitro growth [15,25]. The volatile compound produced by Pseudomonas isolate S45, Stenotrophomonas isolate S180, and Novosphingobium isolate S112 induces deleterious effects on F. mediterranea mycelia, while Enterobacter isolate S11, Paenibacillus isolates S150 and S270, Weeksellaceae S259, and Bacillus strain S5 even promoted F. mediterranea growth in a dual culture [26]. Strains of the cellulolytic and xylanolytic Paenibacillus sp. display a synergistic interaction with F. mediterranea and enhance wood degradation [26].
Strains of Actinobacteria and Streptomyces isolates E1 and R4 induced a significant reduction in infection rates at the lower end of the rootstock in the context of Petri disease [15,27,28].
The application of Pythium oligandrum at the root level reduced P. chlamydospora and P. minimum necroses in the stem and triggered the plant defense pathways, including PR proteins, phenylpropanoid, oxylipin, and oxidoreduction systems [15,30,31].
Epicoccum layuense, an ascomycete often associated with the mycobiomes of grapevines, through in vitro dual culture, inhibited the growth of P. chlamydospora, F. mediterranea, and some P. minimum strains without colony contact, suggesting the production of inhibitory compounds [32]. E. layuense strain E24 colonized the rooted grapevine cuttings of Cabernet Sauvignon and Touriga Nacional cultivars under greenhouse conditions and decreased (31-82%) the wood symptomatology via chemical interaction and competition for space, depending on the pathogen and grapevine cultivar [32]. Epicoccum mezzettii E17 overgrew F. mediterranea in vitro and competed for space and nutrients [32].
Strains of Clonostachys rosea, in vitro, overgrew and inhibited P. chlamydospora and P. minimum through antibiosis and mycoparasitism [15,39]. Under greenhouse conditions, amended soil with the endophytic C. rosea strain 19B/1 significantly decreased the length of the necrotic lesions caused by P. chlamydospora [39]. Lecanicillium lecanii (ATCC 46578) reduced in vitro the growth of P. chlamydospora and P. minimum by carbon and nitrogen competition [34] or by the production of antibiotic compounds [39]. Fusarium oxysporum strain F2 reduced P. chlamydospora in vitro growth by 43% and sporulation by 90%; nonetheless, the strain showed no reduction in the discoloration length inside the trunk, despite an 82% reduction in the P. chlamydospora DNA amount [40].
This work aimed to assess the antagonistic activity of four new biological limiters against the three most important Esca-associated fungi of grapevine. The challenge organisms applied included Trichoderma harzianum, the chitinolytic fungus Aphanocladium album, and two xylotrophic fungi belonging to the genus Pleurotus. Radial growth inhibition, the morphology of the interaction, and the antagonism index were used to evaluate the in vitro dual interactions. The number of positive re-isolations after pathogen and biological limiter interaction was calculated. T. harzianum and A. album were additionally evaluated on V. vinifera cv Nero di Troia detached canes against the three main fungal species associated with the Esca complex.

Strains, Media, and Growth Conditions
Target organisms (Table 1) were isolated from wood samples of 12-year-old cv Nero di Troia vines, grafted on 157-11 located in the countryside of Grumo Appula (Puglia region, southern Italy) and trained to a 'tendone' not irrigated. In the 2020 and 2021 growing seasons, the vineyard was surveyed, and vines with typical Esca symptoms were selected ( Figure 1). During the winter pruning operations (January 2022), samples of branches were taken from five symptomatic vines. Internal symptoms through the wood, revealed after cutting, included vascular streaking, light brown to black discoloration, and white rot ( Figure 1). Wood samples were wetted with denatured ethanol and surface-sterilized by a flame. From each sample, 15 wood chips (2 mm × 2 mm × 2 mm) were aseptically cut from the margins of diseased wood and plated (5 chips per dish) either on potato dextrose agar (PDA; from Oxoid Part of Thermo Fisher Scientific-Microbiology, Hampshire, UK) amended with 100 mgL −1 of each streptomycin and ampicillin (PDA-SA) or PDA modified by adding 1 mL L −1 of thiabendazole lactate (PDA-T: 2.3 g of thiabendazole in 10 mL of lactic acid). The incubation of Petri dishes occurred at 24 ± 1 • C in the dark for up to 28 days. Emerging colonies were hyphal-tippurified. Isolates were identified based on their microscopic morphological characteristics. Strains CBS 229.95, CBS 631.94, and DBPV-1 of P. chlamydospora, P. minimum, and F. mediterranea, respectively [41], were used as a reference for morphological features. The obtained strains were stored on PDA slants in Wheaton bottles at 4 ± 1 • C in the fungal collection of the Department of Soil, Plant and Food Sciences (Di.S.S.P.A.), Plant Pathology Section, University of Bari.

Growth Rate
Potential biological limiters and target organisms were singly grown in a 90 m dish containing 18 mL of PDA medium. A plug (3 mm in diameter) of each str lected from the margins of actively grown cultures, was placed 1 cm from the b the plate on the line of the dish diameter. Inoculated plates were sealed with Par and incubated in darkness at 25 ± 1 °C. Radius measurements were performed eve hours following the line of the dish diameter. The average of the daily radius in was calculated. All strains were tested in triplicate and the experiment was rep least two times.

In Vitro Dual Culture Interactions
Dual interactions between the tested strains (Table 1) were performed in plas dishes (diameter 90 mm, height 15 mm) containing PDA (18 mL per plate). M plugs (3 mm in diameter) were cut from the edge side of the actively growing pure and used as an inoculant. The target organism and potential biological limiter plu placed together on the same plate on opposite sides, 1 cm from the border of ( Figure 2a). As a control, the target organism and potential biological limiter plu placed alone. Inoculate plates were sealed with Parafilm M and incubated at 25 ± the dark.
Based on the differences in the growth rates of each fungal strain, the slower placed 8 days in advance of the faster-growing TH.

Growth Rate
Potential biological limiters and target organisms were singly grown in a 90 mm Petri dish containing 18 mL of PDA medium. A plug (3 mm in diameter) of each strain, collected from the margins of actively grown cultures, was placed 1 cm from the border of the plate on the line of the dish diameter. Inoculated plates were sealed with Parafilm M and incubated in darkness at 25 ± 1 • C. Radius measurements were performed every eight hours following the line of the dish diameter. The average of the daily radius increment was calculated. All strains were tested in triplicate and the experiment was repeated at least two times.

In Vitro Dual Culture Interactions
Dual interactions between the tested strains (Table 1) were performed in plastic Petri dishes (diameter 90 mm, height 15 mm) containing PDA (18 mL per plate). Mycelium plugs (3 mm in diameter) were cut from the edge side of the actively growing pure culture and used as an inoculant. The target organism and potential biological limiter plugs were placed together on the same plate on opposite sides, 1 cm from the border of the dish (Figure 2a). As a control, the target organism and potential biological limiter plugs were placed alone. Inoculate plates were sealed with Parafilm M and incubated at 25 ± 1 • C, in the dark.
Based on the differences in the growth rates of each fungal strain, the slower PC was placed 8 days in advance of the faster-growing TH.
Three replicates were maintained for each treatment. The experiment was conducted twice for reproducibility. Plates were observed every eight hours to record the time of the first contact between the two mycelia.
in Figure 2a). Discs were placed on PDA plates and incubated at 25 ± 1 °C in the dark. The score of the re-isolation success of potential biocontrol agents against each pathogen was quantified as ln[(A + 1)/(P + 1)], where A and P are the mean numbers of the successfully re-isolated biological limiter and target organism, respectively (n = max 5). The estimated value of s ranges between +1.79 and −1.79. Higher values of s indicate greater re-isolation success [45,46].

Interaction on Detached Grapevine Canes
From 20 healthy 'Nero di Troia' dormant vines, canes (diameter 10-12 mm) were collected in January 2023. In the laboratory, the canes were cut into one node segment (80 mm length) each 2 cm above the bud to simulate a fresh pruning wound. Cane segments were autoclaved (121 °C, 40 min) and distributed, respecting the polarity, in sterile 50 mL Falcon-type plastic tubes (one per tube) containing 20 mL of sterile distilled water.
For AA, TH, PM, and PC, conidia suspensions were prepared in 0.2% agar water as a bio-adhesive. For FM, a mycelia homogenate was obtained in 0.2% agar water using a Sorvall model 17105 Omni-Mixer homogenizer (DuPont de Nemours, Inc., Wilmington, DE, USA). The viability of conidia and mycelia fragments was assessed on PDA by the decimal dilutions method, adjusted to 10 6 colony-forming units per milliliter, and used (10 μ) to inoculate the apical end of each cane. AA, TH, PC, PM, and FM were assayed singly and in dual pathogen-limiter arrangements. In dual combinations, target organisms were inoculated 5 days before or after the biological limiters. Canes treated with 0.2% agar water and untreated canes were used as controls. A total of 20 canes per treatment were used. All tubes were incubated at 25 ± 1 °C, in the dark, for 12 weeks. Re-isolations were performed to verify the vitality of the fungal strains inoculated and their progression inside the cane tissue. Each apical end was gently pressed on the PDASA medium surface Every day, the colony radii of the target organism were assessed in dual culture and on the control plate. In plates with dual culture, radii were measured in the direction of the potential biological limiter. The measurements evaluated 10 days after inoculation were used to determine the percentage of radial growth inhibition as 100 × [(C − T)/C], where C and T are the mean radii (mm) of the pathogen in the absence of a biological limiter (control) and dual culture, respectively. Based on the radial growth inhibition rates, antagonist activity was considered [42] very high (>75%), high (61-75%), moderate (51-60%), or low (<50%).
Mycelial interactions in dual culture were examined daily and scored with the rating proposed by Badalyan et al. [43,44]. The antagonism index, which defines the ability of a fungus to compete with another species, was calculated as ∑ (n × i), where n is the number of each type or subtype of interaction and i is the corresponding score [43,44]. Biological limiters were considered active (antagonism index > 15), moderately active (antagonism index = 10-15), or weak (antagonism index < 10) antagonists [43,44].
Twenty days after mycelia interaction, the observed results of the competition were quantified after fungal re-isolation. From each plate, seven agar discs (5 mm in diameter) were cut ( Figure 2a): five from the interaction zone and two (one per strain) from areas with the presumed growth of only one individual fungus (positions labeled "L" and "P" in Figure 2a). Discs were placed on PDA plates and incubated at 25 ± 1 • C in the dark. The score of the re-isolation success of potential biocontrol agents against each pathogen was quantified as ln[(A + 1)/(P + 1)], where A and P are the mean numbers of the successfully re-isolated biological limiter and target organism, respectively (n = max 5). The estimated value of s ranges between +1.79 and −1.79. Higher values of s indicate greater re-isolation success [45,46].

Interaction on Detached Grapevine Canes
From 20 healthy 'Nero di Troia' dormant vines, canes (diameter 10-12 mm) were collected in January 2023. In the laboratory, the canes were cut into one node segment (80 mm length) each 2 cm above the bud to simulate a fresh pruning wound. Cane segments were autoclaved (121 • C, 40 min) and distributed, respecting the polarity, in sterile 50 mL Falcon-type plastic tubes (one per tube) containing 20 mL of sterile distilled water.
For AA, TH, PM, and PC, conidia suspensions were prepared in 0.2% agar water as a bio-adhesive. For FM, a mycelia homogenate was obtained in 0.2% agar water using a Sorvall model 17105 Omni-Mixer homogenizer (DuPont de Nemours, Inc., Wilmington, DE, USA). The viability of conidia and mycelia fragments was assessed on PDA by the decimal dilutions method, adjusted to 10 6 colony-forming units per milliliter, and used (10 µ) to inoculate the apical end of each cane. AA, TH, PC, PM, and FM were assayed singly and in dual pathogen-limiter arrangements. In dual combinations, target organisms were inoculated 5 days before or after the biological limiters. Canes treated with 0.2% agar water and untreated canes were used as controls. A total of 20 canes per treatment were used. All tubes were incubated at 25 ± 1 • C, in the dark, for 12 weeks. Re-isolations were performed to verify the vitality of the fungal strains inoculated and their progression inside the cane tissue. Each apical end was gently pressed on the PDASA medium surface to create 5 different imprints. Then, every cane was divided longitudinally. Chips (2 mm × 2 mm × 2 mm) were aseptically taken at approximately 5, 10, 15, and 20 mm (Figure 2b) from the upper end and plated (5 per dish) on PDASA. Inoculated dishes were incubated at 25 ± 1 • C, in the dark, for 28-45 d. The developed fungal colonies were identified based on their macroand microscopic morphological characters. The percentage of colonization frequency of re-isolated fungi was estimated as 100 × (N × n), where N is the number of colonies of each species developed from each chip, and n is the total number of plated fragments (n = 15).

Statistical Analysis
Plates and tubes were allotted in a randomized design. Normality and homogeneity of variances were verified with Shapiro-Wilk's test and Levene's test, respectively. The standard deviations (sd) were calculated for all mean values. The experimental data obtained were compared using an analysis of variance (ANOVA), followed by the Fisher least significant difference or Kruskal-Wallis test (p = 0.05). The data of inhibition radial growth were analyzed as radius values and expressed as a percentage. The frequency of re-isolated fungi percentage was transformed to arcsine before analysis.

Growth Rate
All the tested strains developed a different growth rate ( Figure 3). to create 5 different imprints. Then, every cane was divided longitudinally. Chips (2 mm × 2 mm × 2 mm) were aseptically taken at approximately 5, 10, 15, and 20 mm (Figure 2b) from the upper end and plated (5 per dish) on PDASA. Inoculated dishes were incubated at 25 ± 1 °C, in the dark, for 28-45 d. The developed fungal colonies were identified based on their macro-and microscopic morphological characters. The percentage of colonization frequency of re-isolated fungi was estimated as 100 × (N × n), where N is the number of colonies of each species developed from each chip, and n is the total number of plated fragments (n = 15).

Statistical Analysis
Plates and tubes were allotted in a randomized design. Normality and homogeneity of variances were verified with Shapiro-Wilk's test and Levene's test, respectively. The standard deviations (sd) were calculated for all mean values. The experimental data obtained were compared using an analysis of variance (ANOVA), followed by the Fisher least significant difference or Kruskal-Wallis test (p = 0.05). The data of inhibition radial growth were analyzed as radius values and expressed as a percentage. The frequency of re-isolated fungi percentage was transformed to arcsine before analysis.

Growth Rate
All the tested strains developed a different growth rate (Figure 3).    (Figure 3a). Among the tested pathogens (Figure 3b), PC (5.7 mm d −1 ) was the slowest, while the fastest growth rate was recorded for PM (17.7 mm d −1 ) and FM (18.7 mm d −1 ).

In Vitro Dual Culture Interactions
In the dual cultures, pathogens and potential biological limiters required 56 to 96 h for the first colony contact (Table 2). Short times were associated with organisms showing fast mycelial growth (e.g., TH against FM, PM, or PC). On the contrary, long times affected the interactions between slow-growth organisms (e.g., AA against PM and PC).

In Vitro Dual Culture Interactions
In the dual cultures, pathogens and potential biological limiters required 56 to 96 h for the first colony contact (Table 2). Short times were associated with organisms showing fast mycelial growth (e.g., TH against FM, PM, or PC). On the contrary, long times affected the interactions between slow-growth organisms (e.g., AA against PM and PC). The percentage of radial growth inhibition (Figure 4) ranged from 48 to 87% in the PO-FM and TH-PC8 interactions, respectively. The percentage of radial growth inhibition (Figure 4) ranged from 48 to 87% in t PO-FM and TH-PC8 interactions, respectively. The dual culture assays displayed a miscellaneous pattern of mycelial interacti ( Figure 5). Following the definition of mycelial interaction proposed by Badalyan et [43,44], deadlock (mutual inhibition in which neither strain was able to overgrow t other) at mycelial contact, replacement (overgrowth without initial deadlock), and part replacement after an initial deadlock with mycelial contact occurred among the pairin tested.
AA showed deadlock at mycelial contact (interaction type A by Badalyan et [43,44]) during the interactions with FM and PM. Partial replacement after initial deadlo with mycelial contact (interaction type CA1 as suggested by Badalyan et al. [43,44]) w exhibited during AA-PC dual cultures and PE towards FM, PC, and PM. PO and TH com pletely replaced (interaction type C as indicated by Badalyan et al. [43,44]) the three test phytopathogenic organisms, including PC inoculated 8 days in advance of TH. Amo the 78 pairings tested, replacement of the pathogen by the potential biological limiter w The dual culture assays displayed a miscellaneous pattern of mycelial interaction ( Figure 5). Following the definition of mycelial interaction proposed by Badalyan et al. [43,44], deadlock (mutual inhibition in which neither strain was able to overgrow the other) at mycelial contact, replacement (overgrowth without initial deadlock), and partial replacement after an initial deadlock with mycelial contact occurred among the pairings tested.
AA showed deadlock at mycelial contact (interaction type A by Badalyan et al. [43,44]) during the interactions with FM and PM. Partial replacement after initial deadlock with mycelial contact (interaction type C A1 as suggested by Badalyan et al. [43,44]) was exhibited during AA-PC dual cultures and PE towards FM, PC, and PM. PO and TH completely replaced (interaction type C as indicated by Badalyan et al. [43,44]) the three tested phytopathogenic organisms, including PC inoculated 8 days in advance of TH. Among the 78 pairings tested, replacement of the pathogen by the potential biological limiter was more frequent (53.8%) than partial replacement after initial deadlock with mycelial contact (30.8%) and deadlock (15.4%).
Microorganisms 2023, 11, x FOR PEER REVIEW 9 of 18 more frequent (53.8%) than partial replacement after initial deadlock with mycelial contact (30.8%) and deadlock (15.4%).  Table  1. Red letters show the type of interaction as proposed by Badalyan et al. [43,44]: A = deadlock (mutual inhibition in which neither strain was able to overgrow the other) at mycelial contact, C = replacement (overgrowth without initial deadlock), CA1 = partial replacement after an initial deadlock with mycelial contact. In brackets is the percentage of overgrowth 20 days after inoculation. NT = not tested. Red and green arrows indicate the sites where biological limiters and pathogens were inoculated, respectively.
Based on the calculated antagonism index values, AA and PO were weak antagonists, while PE and TH were moderately active antagonists, reaching antagonism indexes of 5.5, 9.0, 10.5, and 12.0, respectively.
The outcomes of re-isolation success in the dual cultures are shown in Table 3. No target pathogens were re-isolated after the interactions PO-FM, PO-PM, PO-PC, TH-FM, TH-PM, TH-PC, and TH-PC8. In all the other combinations, the pathogen was always reisolated from the agar disc cut at the P position (Figure 2a). FM was re-isolated from the connective line with AA, while PM and PC were obtained from the deadlock lines with PO.  Table 1. Red letters show the type of interaction as proposed by Badalyan et al. [43,44]: A = deadlock (mutual inhibition in which neither strain was able to overgrow the other) at mycelial contact, C = replacement (overgrowth without initial deadlock), C A1 = partial replacement after an initial deadlock with mycelial contact. In brackets is the percentage of overgrowth 20 days after inoculation. NT = not tested. Red and green arrows indicate the sites where biological limiters and pathogens were inoculated, respectively.
Based on the calculated antagonism index values, AA and PO were weak antagonists, while PE and TH were moderately active antagonists, reaching antagonism indexes of 5.5, 9.0, 10.5, and 12.0, respectively.
The outcomes of re-isolation success in the dual cultures are shown in Table 3. No target pathogens were re-isolated after the interactions PO-FM, PO-PM, PO-PC, TH-FM, TH-PM, TH-PC, and TH-PC8. In all the other combinations, the pathogen was always re-isolated from the agar disc cut at the P position (Figure 2a). FM was re-isolated from the connective line with AA, while PM and PC were obtained from the deadlock lines with PO.   The score of re-isolation success was quantified as ln[(A + 1)/(P + 1)], where A and P are the mean numbers of the successfully re-isolated biological limiter and target organism, respectively (n = max 5). 2 For species acronym definitions, see Table 1. 3 The data are the means of six replicates ± sd. Within each row, values with different letters are significantly different according to the Kruskal-Wallis test at p ≤ 0.05. 4 PC was inoculated concomitantly (PC) or 8 days in advance of TH (PC8). 5 NT = not tested. * = re-isolated with only the biological limiter.
A similar pattern was observed from the re-isolations performed with the fragments cut at 5 mm below the upper end of the cutting surface (Table 4). From the single inoculations with AA, TH, PM, PC, or FM, the inoculated fungi were recovered from all the imprints (Table 4). In dual inoculations AA5-FM, AA5-PM, AA5-PC, FM5-AA, PM5-AA, and PC5-AA, only AA was re-isolated, whereas TH was obtained from the TH5-FM, TH5-PM, TH5-PC, FM5-TH, PM5-TH, and PC5-TH combinations. Table 4. Colonization frequency of re-isolated fungi (%) 1 from 'Nero di Troia' detached canes  inoculated with AA, TH, FM, PM, PC, and their dual combination with pathogens deposited 5  days before (FM5-AA, PM5-AA, PC5-AA, FM5-TH, PM5-TH, PC5-TH) or after (AA5-FM, AA5-PM,  AA5-PC, TH5-FM, TH5-PM, TH5-PC) 1 Colonization frequency of re-isolated fungi was estimated as 100 × (N × n), where N is the number of colonies of each species developed from each chip, and n is the total number of plated fragments (n = 15). 2 For acronym definitions, see Table 1. 3 The data are the means of fifteen imprints per treatment. 4 Data are the means of fifteen chips for treatment. 5 CN = non-inoculated control. 6 WA = 0.2% agar in water. 7 NFI = no fungal isolation.
A similar pattern was observed from the re-isolations performed with the fragments cut at 5 mm below the upper end of the cutting surface (Table 4).
Negative was the isolation from the longitudinal sections of the canes performed at approximately 10, 15, and 20 mm below the upper end of the cutting surface for all the tested detached canes (Table 4). No organisms were obtained from the re-isolations performed with the imprints and the chips cut at 5, 10, 15, and 20 mm below the apical end of the detached cane used as a non-inoculated control (CN) and the control inoculated with the adhesive suspension WA (Table 4).

Discussion
The use of microbial antagonists to manage disease agents is one of the current frontiers in reducing the worrying dependence on pesticides, ensuring food safety and security, and protecting the environment and consumers [47][48][49].
T. harzianum TH1, A. album MX95, P. ostreatus ALPO, and P. eryngii AL142PE were tested as potential biological limiters against F. mediterranea, P. minimum, and P. chlamydospora, the main pathogens worldwide associated with grapevine Esca complex.
A dual culture evaluates the antagonistic properties of microorganisms. This technique shows rapidly and clearly the mutual effects of the paired organisms and their interactions. However, the method excludes the host plant and cannot be used for biocontrol agents that allow disease control through tissue colonization, systemic resistance induction, and/or niche competition [43][44][45]51,52].
The growth rate on the PDA medium differentiates the behavior of tested organisms and could support potential biological limiters and pathogens during dual culture interaction.
Following the suggestions of Bell et al. [53], TH and PO are valid antagonists, because they covered completely the surfaces of the three tested pathogens (type C interaction following the Badalyan et al. [43,44] score rating).
This was supported by the antagonism index and the quantification of the re-isolation success. Low values of the antagonism index indicate a weak response of the strain in terms of inhibition, while high antagonism indices are associated with highly competitive and inhibitory properties in the antagonist [43,44]. The quantification of the re-isolation success of a potential biological control agent against a pathogen is a reliable tool to confirm the success of an antagonist in the interaction with the mycelium of a competing pathogen and represents an important parameter in dual culture assays [45]. The re-isolation of both the antagonist and pathogen from the interaction zone allowed us to assume that the challenge strains were not highly active competitors. In contrast, the re-isolation of only the antagonist from the interaction zone suggested the elimination of the challenger organism and obviously higher success in competition [45]. During the interactions labeled as "type c" (replacement), shown by TH and PO against the three tested pathogens, the biological limiter was always re-isolated from the overlapping zone and the area of pathogen presence (e.g., letter P in Figure 2a). This confirms the antagonistic aptitude of TH and PO and their ability to stop or inhibit the growth of challenged organisms [43,44,53]. Re-isolations of pathogens and biological limiters, even from the same discs, taken from the interface zone after type A and type C A1 interactions, were obtained in AA-FM, PE-PM, and PE-PC. In the other combinations, the pathogenic organism was replaced by the biological limiter. The outcomes of these experiments allowed us to assume the weak competitive capacity of FM, PM, and PC, as hypothesized for Scleroconidioma sphagnicola [45] and Eutypella parasitica [46]. A similar re-isolation result after deadlock at a distance (interaction type B) was reported by Koukol et al. [45], which was not recorded in our study.
The detached cane assay, a modification of the single node cutting technique [54], assesses cane fruitfulness, reduces the time needed to evaluate pruning wound treatments and screenings of vine cultivars for grapevine trunk disease development [55], and allows potential biocontrol activity characterization [56]. Here, we adapted the technique as a model for the study of Esca-associated fungi in wood colonization and the interactions with AA and TH, two probable biological limiters. Due to the results of the study of the interaction among P. chlamydospora, P. minumum, and F. mediterranea performed in vitro and on grapevines cv. Italia and Matilde [57], single interaction pathogens vs. biological limiters were here evaluated. During previous in vitro and in planta interactions, P. chlamydospora and P. minumum competed for the substratum, albeit not directly challenging: F. mediterranea overgrew P. chlamydospora, and the interaction of P. minumum vs. F. mediterranea showed a deadlock at mycelia contact. In triple cultures, P. minumum in some way "prevented" F. mediterranea from overgrowing P. chlamydospora [57]. A similar pattern was seen on the woody tissue of both tested grapevine cultivars [57]. Furthermore, in naturally infected vine material, F. mediterranea is commonly associated with decayed wood, while P. chlamydospora and P. minimum are linked to brown wood streaking [1,3,5,[7][8][9][10]. Inside the trunks of Esca-affected vines, rotted tissues are often bordered by a brown line that separates rotted from nondecayed wood [2]. It is probable that when colonizing a vacant resource, such as a fresh wound, each invading fungus became competitive and could exclude other microorganisms. Another theory of wood colonization during Esca development considers P. chlamydospora able to reduce plant resistance due to its toxic activity, P. minimum affects cell wall integrity through its enzymatic activity, and F. mediterranea enables the complete degradation of wood tissues, resulting in white rot formation [1][2][3]6,15,57].
To avoid a competitive association among P. chlamydospora, P. minumum, and F. mediterranea, single pathogen inoculation was here considered.
The detached cane assay confirmed the ability of PM, PC, and FM to colonize pruning wounds and to use this as a penetration method to colonize the wood tissues of the vine [1][2][3]49,50,57]. The experiments showed that the pathogens and potential biological limiters were effective in colonizing the apical end of the cane and progressing inside the woody tissue when singly inoculated. TH and AA also effectively protected pruning wounds against FM, PM, and PC for at least 3 months after treatment and confirmed the antagonistic effect exhibited in the dual cultures.
The four species used in the present study are well-known antagonists of plantpathogenic fungi and nematodes.
Trichoderma is a genus of fungi from the family Hypocreaceae, commonly associated with the rhizosphere [64]. Several species of this genus and their Hypocrea teleomorphs are used in agriculture as bio-regulators of plant growth and biocontrol agents for the management of nematodes and plant diseases [64][65][66]. Among the Trichoderma species useful as biocontrol agents, T. harzianum is widely used in plant disease control [67]. The species is proposed as a complex of species based on secondary metabolite production, the target pathogen, the host range, and the distribution area [66][67][68][69]. Morphological and molecular characterizations of commercial T. harzianum biological control products show the presence of nine new species [67]. Different Trichoderma species, including T. atroviride, T. harzianum, T. asperellum, T. gamsii, and T. longibrachiatum, have been widely used against fungi in the Esca complex during grapevine nursery propagation processes and for pruning wound protection [1,12,15,19,20,33,[36][37][38]. Different species of Trichoderma provide efficacy against fungi in the Botryosphaeriaceae and the Diatrypaceae associated with grapevine trunk disease [56]. Antibiosis, mycoparasitism, and nutrient and/or space competition are mechanisms of action associated with Trichoderma's antifungal activity [15,19,20,69]. Therefore, effectiveness in pruning wound protection requires the establishment of a biological agent. For this, Trichoderma-based commercial products give the greatest protective action several days after application.
In the present study, TH and AA effectively prevented pruning wound colonization by F. mediterranea, P. minimum, and P. chlamydospora mycelia and conidia spread. They effectively also stopped the colonization developed by each pathogen.
P. chlamydospora, P. minimum, and F. mediterranea, the main fungal species involved in the Esca complex of grapevine, were successfully added to the target organisms of these interesting biological limiters.
Among the four species here tested, T. harzianum and A. album demonstrate a long history as fungal antagonists. Strains of T. harzianum, including the commercial one (Trichodex ® ), effectively controlled P. chlamydospora and P. minimum in vitro and under greenhouse, field, and nursery conditions [1,15,22,34]. During antagonistic interaction, T. harzianum explores different modes of action: it competes for nutrients, utilizing carbon and nitrogen sources, induces direct antagonism, overgrows its "hosts", induces growth reductions [15,22,35], and reduces infection [36][37][38] and the necrosis length on rootstock and nursery material [15,[36][37][38]. The strain DiSSPA TH07.1-NC, tested in this study, was a moderately active antagonist (antagonism index of 12.0), showed overgrowth, inhibited radial growth, and quickly interacted with the tested Esca-associated pathogens. A direct effect of this strain on PC, PM, and FM can be assumed from the score of the re-isolation success and the colonization frequency. No target pathogens were re-isolated after the interactions with TH on PDA plates and from dual-inoculated detached 'Nero di Troia' canes. Furthermore, the strain DiSSPA TH07.1-NC was able to colonize the apical end of each treated cane, to produce abundant mycelial efflorescence on the cutting surfaces and penetrate the first 5 mm of canes during 30 days.
Strains of the necrotrophic mycoparasite A. album inhibit the sporulation and the growth of several powdery mildew and rust agents, soil-borne plant pathogens, and nematodes [59][60][61][62]. This antagonistic efficiency is supported by the production of hydrolytic enzymes such as protease, glucanase, and several chitinases involved in the cell wall degradation of many phytopathogenic fungi [58].
The strain DiSSPA MX95, tested in this study, was a weak antagonist (antagonism index of 5.5), showing deadlock (mutual inhibition in which neither strain was able to overgrow the other) at mycelial contact against FM and PM and partial replacement after an initial deadlock with mycelial contact against PC. Despite its slow growth rate, DiSSPA MX95 inhibits radial growth and greater re-isolation success on PDA plates. During the experiments on detached 'Nero di Troia' canes, the strain DiSSPA MX95 was able to colonize the apical end of each treated cane, to produce abundant mycelial efflorescence on the cutting surfaces and penetrate the first 5 mm of canes.
This was the first application of P. ostreatus and P. eryngii strains against fungal species involved in the Esca complex of grapevine. The strains DiSSPA ALPO and DiSSPA AL142, tested in this study, were moderately active antagonists (antagonism index of 9.0 and 10.5 for PO and PE, respectively), showed overgrowth, inhibited radial growth, and quickly interacted with the tested Esca-associated pathogens. A direct effect of these strains on PC, PM, and FM can be assumed from the score of the re-isolation success: no target pathogens were re-isolated after the interactions with PO and PE on PDA plates.
Further screening of the strains used in this study should be conducted to assess their efficacy against other grapevine trunk disease fungi, including Ilyonectria spp., Cadophora luteo-olivacea, Diplodia seriata, and Neofusicoccum parvum, commonly found in nurserypropagated material and responsible for young grapevines' decline and death.
A shortcoming of our study is related to the slight repetitions performed. Additional experiments, testing a broader range of P. chlamydospora, P. minimum, and F. mediterranea strains, are needed for the confirmation and clarification of our results and a solid conclusion. However, further testing should be performed to optimize the application methods (e.g., dose and time), together with validation experiments in the field. Moreover, combinations of AA, PO, PM, and TH will be tested.

Conclusions
The interactions between the three most important species isolated from the wood of Esca-complex-affected vines and the four strains tested as biological limiters in the dual cultures, here analyzed in terms of the first contact between the two mycelia, the percentage of radial growth inhibition, the type of mycelial interaction, the antagonism index, and the quantification of fungal re-isolation, suggested that the tested strains P. ostreatus ALPO, P. eryngii AL142, A. album MX-95, and T. harzianum TH07.1-NC were competitive and caused the greatest inhibition of the three pathogens. Interactions on detached grapevine canes among TH and AA against FM, PC, and PM confirmed the in vitro effects. In particular, A. album MX-95 and T. harzianum TH07.1-NC are the most promising strains to be used as biological control agents against the three main pathogens associated with the Esca complex of grapevine.