The potential role of Aureobasidium pullulans in the development

Esca belongs to the group of grapevine trunk diseases - fungal diseases present worldwide in all wine-growing regions. Some aspects of the disease, like the development of external symptoms, have still not been completely discovered and are believed to be affected by several factors, including interactions within the vine microbiome. The examination of the occurrence of the yeast-like fungus Aureobasidium pullulans in the healthy wood of Esca-diseased grapevines via both isolation and qPCR measurements showed a positive correlation between its abundance and the severity of foliar symptoms, suggesting the contribution of this fungus to Esca pathogenesis via an indirect action. In vitro confrontation tests revealed antagonistic interaction between A. pullulans and the Esca pathogen Phaeomoniella chlamydospora . Mutual growth inhibition and the induction of asexual sporogenesis were observed for both fungi without cytotoxic effects. In planta confrontation tests revealed that A. pullulans in combination with P. chlamydospora can lead to severe foliar damage in a strain-dependent manner. This phenomenon could be explained by the altered metabolism of the Esca pathogen in the presence of A. pullulans , or by the cumulative/synergistic effects of the secreted polysaccharides and/or proteins of the two fungi. The present study shows the importance of microbial interactions in the development of plant diseases, highlighting that even a non-pathogenic microorganism can act as a disease-enhancer.


INTRODUCTION
Esca disease belongs to the group of grapevine trunk diseases (GTDs) which are devastating in vineyards worldwide. The syndromes of GTDs (Esca complex, Eutypa-and Botryosphaeria-diebacks, Phomopsis disease and Black foot disease; for a review see Gramaje et al. (2018)) are caused by filamentous fungi of different genera (Phaeoacremonium, Phaeomoniella, Diplodia, Neofusicoccum, Lasiodiplodia, Eutypa, Ilyonectria, etc.), often leading to the death of the affected plants (Surico et al., 2006;Mugnai et al., 1999;Mondello et al., 2018). GTD pathogens infect vines, especially through wounds, and they colonise woody tissues, developing internal wood discolorations and necrosis (Bertsch et al., 2013). Even though the pathogens grow inside the trunk, there are typical GTD external symptoms that develop on the green organs of the vine; i.e., on the shoots, leaves and, in some cases, the berries. Shoots and leaves can dry rapidly, often affecting the whole plant (apoplexy). On leaves, chlorotic and necrotic areas appear, while berries can develop small black spots (Black measles) or can become completely dry, depending on the GTD (Bertsch et al., 2013). To date, following the ban of sodium-arsenite and some benzimidazoles, there is no simple and reliable GTD control method in neither the vineyard nor the nursery (Gramaje et al., 2018;Mondello et al., 2018).
The main agents of the Esca disease complex are the ascomycetous fungi Phaeoacremonium minimum (Phm) and Phaeomoniella chlamydospora (Pch), sometimes associated with Fomitiporia mediterranea (Fomed), and other basidiomycetous fungi of the genus Fomitiporia (Moretti et al., 2021), which cause white rot in the Esca itself (Graniti et al., 2000a;Fischer, 2002;Cloete et al., 2014). The typical Esca internal symptoms are vascular discoloration with the production of dark tarry drops and wood necrosis. In leaves, both chlorotic and necrotic areas have typically been described as tiger-striped leaves (Mugnai et al., 1999), and later as Grapevine Leaf Strip Disease (GLSD -Surico, 2009). The absence of the pathogens in symptomatic leaves, their presence as endophytes in healthy vines (Bruez et al., 2014;Del Frari et al., 2019;Mondello et al., 2022) and the erratic symptoms expression on Esca-affected plants clearly indicate that there is no direct and simple relationship between the presence of pathogens and the development of the Esca external symptoms. Therefore, Esca disease expression (and those of other GTDs, in general) is believed to be strongly affected by other factors, such as abiotic stresses, weather conditions and microbial interactions (Claverie et al., 2020;Fischer and Peighami-Ashnaei, 2019;Songy et al., 2019). Thus, the main hypothesis for explaining the complexity of foliar symptoms expression in Esca-affected vines is based on the toxicity of some pathogens' secreted metabolites. These phytotoxins, whose production can be modulated by internal (fungus) and /or external (environmental) factors, can spread throughout plants as a result of the sap flux reaching the leaves, thus determining metabolic disorders and, finally, the symptoms (Mugnai et al., 1999;Claverie et al., 2020;Trotel-Aziz et al., 2019).
The study of this molecular background has led to the identification of several toxic effector molecules of differing chemical nature. The results of studies have suggested that secreted polypeptides of Pch and Phm have a direct toxic effect on plant grapevine cells (Luini et al., 2010) or act as degrading enzymes, damaging the colonised vascular tissues (Bruno and Sparapano, 2006b). It has also been suggested that secondary metabolites secreted by Esca pathogens take part in symptom development. According to the results of analyses of naturally infected grapevines and phytotoxicity tests, two Esca-associated phytotoxic secondary metabolites have been found to be scytalone and isosclerone (Bruno and Sparapano, 2006a;Bruno et al., 2007). Polysaccharides have also been found to contribute to Esca foliar symptom development, especially pullulan (Bruno and Sparapano, 2006a;Bruno et al., 2007), which is produced by Pch and Phm. Its phytotoxicity has also been hypothesised in the cases of other fungal plant diseases, like chestnut blight caused by Cryphonectria parasitica (Forabosco et al., 2006). This suggests that pullulan may act as a non-host-specific phytotoxin. Even if secreted proteins, secondary metabolites (e.g., scytalone, isosclerone, naphthalenone, and mellein) and extracellular polysaccharides (e.g., pullulan) have proved to be toxic to grapevines (Bruno et al., 2007;Burruano et al., 2016;Graniti et al., 2000b;Andolfi et al., 2011;Abou-Mansour et al., 2015;Trotel-Aziz et al., 2019), their mode of action and their role in the development of the symptoms in natural conditions are not completely clear. The fact that GLSD foliar symptoms have never been observed in grapevine artificially infected with Phm, Pch and Fomed, either alone or in combination, raises questions about the real impact of their effector molecules in Esca disease, confirming that other factors could promote or block foliar symptoms expression.
Recent studies indicate a possible role of the vine microbiome in GTD symptoms expression (Bettenfeld et al., 2020;Niem et al., 2020). According to this hypothesis, healthy vines are the result of a balance of the different elements of the vine microbiome. Therefore, a diseased vine could be the result of an unbalanced microbiome, in which interactions between microorganisms have negative effects for plant health, especially in the case of pathogen attack.
Aureobasidium pullulans (Apu) is a cosmopolitan dimorphic fungus and a relevant member of the epiphytic (Barata et al., 2012) and endophytic (Knapp et al., 2021) microbiome of the aboveground parts of grapevines. Apu is mostly known for its biocontrol capacities and its ability to produce industrially important enzymes and the polysaccharide pullulan (Prasongsuk et al., 2018). The effects of this fungus on the grapevine host are also beneficial according to previous studies. Apu has proved to be a potential biocontrol agent against Botrytis cinerea, the causal agent of grapevine bunch-rot (Galli et al., 2021), and Greeneria uvicola, the pathogen of bitter rot in grapes (Rathnayake et al., 2018). Apu is an effective inhibitor of Diplodia seriata (Botryosphaeria dieback), directly inhibiting and possibly elicitating defense responses in the host, as has been suggested by the elevated expression of PR6 pathogenesis-related gene in Apu-treated plants . In contrast to the abovementioned beneficial effects of Apu on plant hosts, its phytopathogenicity has also been reported in certain cases (Guo et al., 2020;Lee et al., 2019;Prashanthi and Kulkarni, 2005;Wu et al., 2017;Xie et al., 2022). Moreover, the fact that Apu is a potent producer of the aforementioned phytotoxin pullulan raises further doubts about whether it is generally a beneficial microorganism.
Our initial investigations on Esca-affected grapevines in the field produced surprising results: the abundance of Apu was found to positively correlate with the severity of the foliar symptoms ( Figure 1B). Since Apu is a well-characterised inhabitant of the grapevine microbiome that does not have any known damaging effects on the host, we hypothesised that the disease-promoting effect of this fungus is based on its interaction with the known Esca pathogens Pch and Phm. Interactions between the fungal species were studied in vitro confrontation assays, revealing a modulatory effect of Apu and Pch on each other's development. Their direct and indirect effects on the development of foliar symptoms were examined by fungal artificial inoculation. Strain-dependent differences in the combined pathogenicity of Apu and Pch strains are suggested to be related to the measured differences in the ability of strains to produce known virulence factors of Esca disease like exopolysaccharides (EPS) or extracellular proteins.

Isolation and identification of fungal strains
The wood endophytic fungal population of Esca-affected vines was isolated in 2017 from Vitis vinifera cv. Cabernet-Sauvignon plants in the vineyard of Eszterházy Károly Catholic University. To isolate the wood-colonising fungi small pieces of the vine trunk were cut out. Since there is a large variance in the severity of foliar symptoms between the canes of a plant, the wood trunks were sampled just below the canes monitored for foliar symptom severity. The obtained wood chips were placed on Potato Dextrose Agar (PDA) plates supplemented with 10 µg/ml oxytetracycline to prevent bacterial growth. The plates were incubated at 25 °C for 4 weeks. The emerging mycelia were first placed individually on PDA plates, then pure cultures were obtained by subculturing a small portion of actively growing mycelia or by streaking conidia when available. The strains were stored in 50 %v/v glycerol at -80 °C. A total of 43 trunks with the foliar symptoms of Esca disease were sampled.
Total DNA was extracted from the fungal biomass using DNeasy Plant Mini Kit (Qiagen) for molecular identification. Internal transcribed spacer regions were amplified using ITS1-f (5'-CTTGGTCATTTAGAGGAAGTAA-3') and ITS4 (5'-TCCTCCGCTTATTGATATGC-3') primers. Reaction products were checked by agarose gel electrophoresis. The obtained amplicons were sequenced (BaseClear B.V. -Netherlands) and the sequences were compared in online databases using BLAST (www.ncbi.nlm.nih.gov). Fungal strains were identified based on sequence similarity and morphological characteristics.

Measurement of Apu DNA in grapevines by Quantitative Real Time PCR
In September 2018, to quantify the presence of Apu in the wood of Esca-disease-expressing vines, 10 cv Cabernet-Sauvignon grapevines showing foliar symptoms of disease and 10 asymptomatic plants were sampled in the same vineyard of 2017; this was done by drilling the trunk at the base of canes being examined for symptom severity. The samples were lyophilised and grounded in a tissue-lyser (50 Hz, 2×5 min) with the aid of a stainless-steel ball. DNA was extracted with DNeasy Plant Mini Kit (Qiagen) according to the manufacturer's instructions and measured by Nanodrop 2000 (Thermo Fisher Scientific) spectrophotometer. The amount of Apu DNA was determined by using species-specific primers as described by Martini et al. (2009).

Investigation of the severity of foliar symptoms on Esca-affected grapevines
The symptomatic leaves of Esca-diseased vines (3 per each symptomatic cane) were removed ans photographed using a table scanner. The symptomatic areas were analysed by image analysis software (Photoshop CS5) to calculate symptom severity. For each cane, the symptom severity was calculated as the mean of 3 leaves.

Direct assay: dual in vitro culture assays
To verify any possible interactions between Apu (8 isolates) and Pch (6 isolates) or Phm (3 isolates), the fungi isolated from Esca-diseased vines were inoculated in dual culture assay. Mycelial plugs (diameter of 3 mm) from the margins of 1-week-old colonies were placed on a PDA plate at a distance of 5 cm from each other. The controls consisted in plates individually inoculated with the used strains (3 replicates per combination). The petri dishes were incubated at 25 °C in the dark for two weeks.

Indirect assay: cultural filtrate assays
To verify the nature of the observed antagonism, the strains of the pathogen-Apu combinations showing interaction were used in the culture filtrate assay. Apu strain Y60 and Pch strain P3 were inoculated separately in 3×100 ml liquid YG medium (2 % m/v glucose, 1 % m/v yeast extract) at 25 °C and 180 rpm for 7 days. Afterwards, the fungal mycelia were removed by filtration, and the cultural filtrate (CF) was sterilised using a membrane with a 0.2 µm pore size. CFs were mixed with PDA medium to obtain plates of CF at concentrations of 50, 25, 12.5, and 0 % v/v. PDA plates with 50 % v/v YG medium were also prepared as a control. Mycelial plugs (diameter of 3 mm) were inoculated on the plates for the measurement of the effect of culture filtrates on growth rate and sporulation, while 100 conidia were streaked on the plates for the assessment of germination rates. Fungal cultures were incubated at 25 °C in the dark for 8 days. Mycelial growth and conidia germination were monitored in order to calculate the respective growth and germination rate. The number of formed conidia for Pch and of single-cell forms (conidia and yeast) for Apu was obtained using a Burker-chamber. Three isolates of Apu and three isolates of Pch were used for the indirect assays. The experimental design is summarised in Supplementary Table S1. All tests were done in triplicates.

Artificial inoculation of grapevine detached green shoots
To also test the interaction between Apu and Pch in planta, artificial inoculations were performed on detached green shoots. Three strains of Apu (Y37, Y43 and Y60) and three of Pch (P3, P37, and P46) were used in this test. In detail, green canes of cv. Cabernet-Sauvignon were collected at the beginning of July and cut into small sections containing a node, an internodal section and a leaf. The cuttings were surface sterilised using 70 %v/v ethanol and then wounded with a sterile scalpel in the upper third part. Apu and Pch mycelial plugs (diameter of 3 mm) from the margin of a 1-week-old colony were placed within the wound. Each strain was inoculated alone and in combination with a strain of the other fungus. In the latter case, strains were inoculated at 1 cm distance from each other. Each of these inoculation conditions was replicated 6 times. The controls were inoculated with sterile PDA plugs. The cuttings were placed in water and incubated without artificial lighting and with full exposure to sunlight for 15 days in July. The relative area of the leaves in all conditions was measured after photographing as described in Section 3 of Materials and methods. The experiment was repeated three times.

Analytical methods
The production capacity of the EPS and extracellular proteins (potential virulence factors) of Apu and Pch strains was determined to explain the strain-dependent differences in pathogenicity. To further investigate the possible interaction of Apu and Esca pathogens in disease expression, we also studied the possible role of Apu and Pch exopolysaccharides (EPS) in the development of Esca foliar symptoms.

Purification of exopolysaccharides from fungal culture filtrates
To obtain EPS, culture filtrates of Apu (Y37, Y43 and Y60) and Pch (P3, P37 and P46) strains were added to two volumes of 96 % v/v ethanol and incubated at 4 °C overnight. EPS were pelleted by centrifugation (4000 rpm, 30 min) and washed twice with 70 %v/v ethanol. EPS were dried and weighed to prepare 5 mg/ml stock solutions.

Proteins and polysaccharides quantification
The concentrations of proteins were assayed by a dye-binding method (Bradford, 1976) using BSA (bovines serum albumin) as a standard. Polysaccharide content was evaluated by the phenol-sulfuric acid method (Dubois et al., 1956) using purified EPS of the Y60 Apu strain (described above) as a standard. Absorbance was measured by Nanodrop 2000 (Thermo Fisher Scientific). All measurements were done in triplicate.

Measurement of reducing sugars
The reducing sugar content of purified fungal EPS was determined using the neocuproine method (Dygert et al., 1965). Absorbance was measured by Nanodrop 2000 (Thermo Fisher Scientific) using glucose as a standard. All measurements were done in triplicate.

Foliar disc assays
The effect of the interaction of Apu and Pch EPS, alone or in combination, on foliar symptom development was tested using foliar disc assays. Foliar discs (diameter: 1 cm) were obtained from leaves of cv Chardonnay and Cabernet-Sauvignon using a cork-borer and put in aqueous solution containing single (Pch or Apu) or coupled (Pch+Apu) EPSs at 50, 250 and 500 µg/mL each. The controls comprised water without any EPS and, as a positive control, water with mellein (same concentrations), which is a secondary metabolite with ascertained phytotoxic effect on grapevine leaves. In each of these conditions three discs were used. The foliar discs were observed at 24, 48 and 96h for necrosis development.
To obtain quantitative phytotoxicity data, image analysis was conducted on photographs of the foliar discs, using Fiji software (Schindelin et al., 2012). The images were converted to black and white, processing only red and yellow pixels (reflecting on necrosis and discoloration) and omitting the green ones. Afterwards, pixel density was determined.

Statistical analysis
Significances of differences were determined by ANOVA analysis with Tukey post-test, while correlations were assessed by applying Spearman's test. In both tests, results with p < 0.05 were considered to be significant.

The presence of A. pullulans in the endophytic microbiome is linked to increased severity of foliar symptoms on Esca-affected grapevines
A total of 43 Esca-symptomatic grapevines was sampled to evaluate their endophytic fungal population. Apu and Pch were isolated at frequencies of 16 % and 14 % respectively. Interestingly, Apu seemed to be associated with vines with higher (46-81 %) foliar symptom severity ( Figure 1A), and in accordance with this observation, grapevines with isolated Apu showed significantly higher severity values ( Figure 1B).
The measurement of Apu DNA by qRT-PCR showed a significantly increased value in the diseased plants ( Figure 2A). This increase was due to the occurrence of high values in the diseased plants rather than a general increase in Esca-affected ones. Moreover, the severity of the foliar symptoms showed a correlation with Apu DNA content according to Spearman's test (p < 0.05) in diseased plants ( Figure 2B).    Growth rate (A), sporulation (B) and germination rate (C) of P3, P37 and P46 Pch strains growing on PDA medium supplemented with different concentrations of the Apu culture filtrate Y60 strain or with 50 % v/v YG medium. Asterisks mark significant differences (ANOVA test) relative to the untreated (0%) controls. *p < 0.05; **p < 0.005; ***p < 0.0005.

FIGURE 5. Effects of Pch secreted molecules on Apu development.
Growth rate (A), formation of yeast and conidia (B) and germination rate (C) of Y37, Y43 and Y60 Apu strains growing on PDA medium supplemented with different concentrations of the Pch strain P3 culture filtrate or with 50 %v/v YG medium. Asterisks mark significant differences (ANOVA test) relative to the untreated (0 %) controls. *p < 0.05; **p < 0.005; ***p < 0.0005.

In vitro interaction of Apu with Pch and Phm, causal agents of Esca disease
Antagonistic activity of Phm against Apu was observed in the in vitro confrontation tests. While the growth of all Phm isolates was stimulated by the Apu colonies, the latter showed death of cells (condensation of cytoplasm and outflow at the mycelial tips) in the confrontation zones. Because of this inhibition of Apu growth by Phm (not related to the observed Apu disease-promoting effect), the interactions between these fungi were excluded from further analysis. In the Apu-Pch confrontation tests, both species showed a decreasing growth rate close to the confrontation zone ( Figure 3A). However, no signs of cell death (e.g., cytoplasm condensation and cytoplasm outflow) were detected at the microscopic level. By contrast, all the tested Apu and Pch strains showed increased sporulation close to the confrontation areas ( Figure 3B). For further investigation of the interaction between Apu and Pch, three isolates of each species were selected (Supplementary Table S2). The Pch strains formed white colonies when cultured on PDA plates at 25 °C in the dark, and showed different rates of greenish melanisation when grown at room temperature (21 °C) in ambient light, P37 having the lowest pigmentation. Differences were also observed in the sporulation of Pch strains with the highest value for P46, followed by P3 and P37. All the strains produced unicellular ellipsoidal conidia. Not only did the observed morphological characteristics agree with the description of Pch (Gams and Crous, 2000), but the ITS sequences were highly similar to the CBS 229.95 ex-type strain. All the Apu strains formed white-pinkish colonies with even margins on PDA medium at 25 °C. After two weeks of incubation at room temperature in ambient light conditions, strong black melanisation was observed in Y60, and a lesser pigmentation in Y43, while Y37 showed brownish coloration after about four weeks. Microscopic examinations revealed hyaline mycelia and the presence of yeast forms, conidia and chlamydospores in agreement with the general characteristic of the species (Ciferri et al., 1957). The ITS sequences of all Apu isolates showed a high degree of similarity to the CBS 584.75 ex-type strain. Morphological characteristics of the selected Apu and Pch strains showed considerable differences, suggesting that they represent both physiological and genetical variants of the same species.

Study of the crossed antibiosis between Apu and Pch using culture filtrates
The Y60 Apu culture filtrate reduced the growth rate and increased the sporulation of the tested Pch strains without inhibiting their germination ( Figures 4A, B, C). At values of between 12.5 and 50 % v/v concentrations of Apu culture filtrate, the number of Pch conidia significantly increased ( Figure 4B).
The same effects were observed for Apu strains growing on medium supplemented with the culture filtrate of Pch strain P3. The Apu germination rate was not affected ( Figure 5C) while the growth decreased ( Figure 5A) and the production of conidia and yeast cells increased ( Figure 5B) in the presence of P3 Pch culture filtrate.

Effects of co-inoculation of Apu and Pch on green canes of grapevine
After 15 dpi, the leaves on the artificially-infected green canes showed leaf damages ( Figure 6) that differed in terms of severity and type depending on the inoculated strain (Figure 7). No or very slight symptoms were observed on mock-inoculated canes, which remained mostly asymptomatic ( Figure 7A).  Two out of the three inoculated Apu strains (Y37, Y43) caused mild damage, significantly reducing the healthy leaf area in case of Y43 ( Figure 7B). Both the occurrence and severity of foliar damage increased in the Apu+Pch inoculated canes in a strain-dependent manner. The highest severity was observed in the Y37 Apu and P46 Pch co-inoculation ( Figure 7B), which resulted in the death of most of the leaves. A significant decline was also observed in the shoots inoculated with Y43 Apu and P46 Pch strains, with uniform chlorotic, necrotic and apoplectic areas appearing ( Figure 7B).
In single inoculated canes, all Pch strains and none of the Apu strains caused necrosis at the inoculation point. Interestingly, canes with apoplectic leaves in combined Y37 Apu and P46 Pch inoculations showed massive necrosis ( Figure 8A), resulting in significantly higher necrosis length in this inoculation than in the other groups ( Figure 8B).
The results of the artificial infections agree with the hypothesis that Apu could promote the development of foliar symptoms of Esca disease, highlighting the strain-dependent nature of this interaction. Moreover, it seems that Apu may also contribute to the development of local vascular necrosis, as was seen in the case of Y37 Apu and P46 Pch combined inoculations.

Production of potential virulence factors by Apu and Pch strains
The secretion of proteins by the Apu strains showed significant differences among strains. Y37, Y43 and Y60 produced 26.4, 20, and 42.7 µg/ml respectively ( Figure 9A). Of the Pch strains, P46 was the greatest protein producer, with 11 µg/ml, while P3 and P37 secreted 3 and 2.3 µg/ml of proteins respectively ( Figure 9A). The EPS produced by Apu strains were greater than Pch ones ( Figure 9B). Y37, Y43 and Y60 Apu strains secreted 4.4, 3.5 and 3.0 mg/ml EPS, while P3, P37 and P46 Pch strains secreted only 0.7, 0.4 and 1.3 mg/ml. The average molecular weight of EPS produced by the examined fungal strains was calculated as the rate of measured EPS mass, and the reducing sugar molarity was expressed as a glucose equivalent. Differences between the tested strains were observed in this aspect ( Figure 9C). Of the Apu strains, Y60 produced EPS with the highest molecular weight (183 kDa) followed by Y43 (84.3 kDa) and Y37 (33.4 kDa). In the Pch strains, the highest EPS molecular weight was measured in P46 (15.2 kDa), followed by P3 (11.2 kDa) and P37 (4.4 kDa).

Effect of Apu and Pch EPS on grapevine foliar discs
No necrotic areas were observed in any of the foliar discs treated with single EPS after 96 hours. By contrast, the combination of P46 EPS and Y37, Y43 and Y60 EPS induced edge necrosis on Cabernet-Sauvignon foliar discs after 48h ( Figure 10A). No effect was observed on Chardonnay, while the positive control based on mellein content induced necrotic areas for both cultivars at all tested concentrations. The image analysis of the photographed foliar disks revealed higher red and yellow pixel densities on the EPS-treated samples than on the untreated control ( Figure 10B), indicating the negative effect of fungal EPS on the leaves. This difference was significant when 50 µg/ml P46 EPS was combined with 50 or 500 µg/ml EPS of Y43 Apu strain, or 50 µg/ml EPS of Y60 strain ( Figure 10B).

DISCUSSION
In the present study, our results indicate that there is a positive relationship between the presence of Apu in the vine microbiome at the base of the analysed canes and the severity of Esca disease foliar symptoms. This is a surprising result, since Apu is a well-characterised fungus to date with very limited information on its phytopathogenicity. Apu has been reported to be a pathogen on some hosts and to cause spot blight on pear (Xie et al., 2022), anthracnose on Paeonia suffruticosa (Lee et al., 2019), and stem and fruit spot on pitaya (Wu et al., 2017). Apu is also a rare but possible causal agent of post-harvest rot of fruits, like sweet cherries (Kim, 2014), pears (Dobra et al., 2014), or even grapes (Lorenzini and Zapparoli, 2015). In addition to cultivated plants, Apu has also showed pathogenicity against some weeds (Prashanthi and Kulkarni, 2005;Guo et al., 2020). It is worth noting that all the abovementioned diseases are the result of the infection of green and annual plant parts, unlike in the case of GTDs. Despite its low relevance as a phytopathogen, Apu is considered to have potential as a biocontrol agent against plant pathogenic fungi (Prasongsuk et al., 2018). Data on the occurrence of Apu in Esca-affected grapevines are very controversial. Some studies did not find relevant incidences of Apu in symptomatic grapevines (Bruez et al., 2014), others reported similar incidences in both symptomatic and asymptomatic plants (Hofstetter et al., 2012) or considered its presence as typical on healthy wood in symptomatic plants (Rusjan et al., 2017). To date, the only known relationship between Apu and Esca pathogens is the production of pullulan. This non-toxic and widely used polysaccharide (in pharmaceutical and food industries) is believed to participate in the pathogenesis of Esca disease (Bruno et al., 2007;Graniti et al., 2000b). Some researchers suggest that pullulan blocks the water and nutrient flows in grapevine by obstructing the vascular tissues, but this theory needs to be validated.
Regarding the possible role of Apu in Esca disease expression, two possible modes of action were examined: i) the direct interaction of Apu with the Esca pathogens, increasing their virulence in grapevine; and ii) the production of effector molecules (e.g. pullulan) by Apu that may contribute, in synergy with Esca pathogens effectors, to increasing the severity of the symptoms.
In the in vitro confrontation tests, Pch and Apu mutually inhibited each other's growth, increased sporulation and did not secrete molecules that were toxic to one another. In addition to the direct confrontation tests, these results were further confirmed in the culture filtrate tests. For these latter experiments, three strains of Apu and Pch fungal species were chosen from our collection. The characterisation of these strains in terms of traits like growth rate, sporulation and pigmentation suggests that they represent different phenoand genotypes of their species. This allowed us to identify some strain-dependent differences, as well as possible general attributes of these fungal species. It seems that these two fungi can mutually modulate each other's development causing a shift to the stationary phase, possibly resulting in an altered metabolism. This may be achieved by the so called "quorum sensing" mechanism, which allows both bacteria (Miller and Bassler, 2001) and fungi (Albuquerque and Casadevall, 2012) to change their development and metabolism according to their cell density. The biosynthesis of fungal phytotoxins (as secondary metabolites) generally occurs during the stationary growth phase (Calvo et al., 2002); therefore, the modulatory effect of Apu on Pch development may enhance the pathogenicity of the latter. In contrast to Pch, Phm greatly inhibited Apu growth. Due to the molecules it secretes, Phm is also able to kill Apu cells, while the latter have no effect on Phm.
Our artificial co-inoculations allowed us to study possible interactions between Pch and Apu that affect the severity of Esca disease. As expected, in the artificial inoculations, Pch strains alone developed necrosis at the inoculation point, but neither foliar symptoms nor damages. Apu strains did not cause necrosis at the inoculation point or any foliar symptoms, except for the Y37 and Y43 strains, which caused mild foliar symptoms (partial necrosis and chlorosis) in about 50 % of inoculated plants. This result suggests that, depending on the strain, Apu can cause damage to grapevine leaves. The discolorations on Apu-treated leaves were yellow, suggesting a decrease in chlorophyll content, rather than the stress response of the plants that results in an accumulation of red anthocyanine pigments (Kaur et al., 2022). This phenomenon is also in accordance with the decreased chlorophyll content in Esca-affected leaves reported previously (Petit et al., 2006). The Apu/Pch co-inoculation revealed other pathogenesis-related effects. The Apu strains Y43 and Y60, in combination with all the tested Pch strains, showed the same symptoms as single Apu inoculations (chlorosis and partial necrosis for Y43; no symptoms for Y60) with the exception of Y43/P46 co-inoculations, which caused low but significant foliar damage. In Y37/P46 coinoculations, about 80 % of shoots were apoplectic, with a significant necrosis of both leaves and canes. These results suggest that the hypothesised effect of the Apu/Pch interaction on Esca foliar and vascular symptom expression could be strain-dependent. The possible strain-specificity of the Apu/ Pch interaction was evaluated by studying the phytotoxicity of EPS. Phytotoxicity tests with leaf discs showed that the Apu and Pch EPS fractions are not toxic at the concentrations tested. Only the combination of the Apu EPS+P46 EPS determined the development of necrosis on the leaf discs. This result suggests that some molecules in EPS derived from Apu and Pch strains may interact synergistically to damage grapevine leaves. The phytotoxicity of the combined EPS of the two species further reinforces the importance of pullulan in Esca disease development and may explain the contribution of Apu to this process (Bruno et al., 2007;Graniti et al., 2000b).
The production of extracellular polysaccharides and proteins by Apu and Pch strains in liquid cultures was also investigated. The characteristic feature of the Y37 strain among the tested Apu strains was that it produced the highest amount of EPS and had the lowest average molecular weight. The latter trait may facilitate the transport of Y37-derived EPS from the wood to the leaves, allowing it to cause foliar symptoms. The negative correlation between the molecular weight of a microbial EPS and its ability to affect a plant host has already been reported in a previous study (Zou et al., 2018). Of the Pch strains, P46 produced the highest amount of extracellular proteins, about 4 times more than both the P3 and P37 strains. Since we know that the secreted proteins of Pch are involved in Esca symptom expression (Luini et al., 2010), the latter result may explain the high pathogenicity of P46 among the examined Pch strains. These results suggest that the secreted polysaccharides of Apu and the secreted proteins of Pch are responsible for the severe foliar damage observed in the Y37-P46 co-inoculations. To confirm the role of these molecules in the development of foliar symptoms, further experiments should be carried out by measuring their phytotoxicity before and after specific enzymatic decomposition. This should be followed by the identification of individual phytotoxic proteins and polysaccharides and their subsequent characterisation.
Our surprising finding that Apu may promote Esca disease development is not unprecedented. Regarding their complex nature, GTDs are believed to be partly the result of microbial interactions in the vine microbiome. The interactions of known GTD pathogens have been previously discussed in several studies. Possibe interactions between the Esca pathogens Pch, Phm, and Fmed (Bruno and Sparapano, 2006b;Sparapano et al., 2001) were examined in great detail, while the disease-promoting effects of non-Esca GTD pathogens on Esca pathogenesis have also been suggested (Martín et al., 2022). Recent have suggested that microorganisms that frequently occur in grapevine without remarkable pathogenicity also contribute to GTD pathogenesis. DNAbased analyses of the grapevine microbiome have found a significantly high abundance of the fungus Acremonium alternatum (Bekris et al., 2021), as well as the bacterial taxa Xanthomonadales, Rhizobiales and Enterobacteriales (Bruez et al., 2015), in symptomatic plants. While disease-promoting interactions have been suggested by several of the above studies, only a limited number of papers have focused on the validation and characterisation of such interactions: Haidar et al. (2021) confirmed that the synergistic interaction between a Paenibacillus sp. isolate and Fmed was associated with the degradation of grapevine wood in in vitro experiments; in addition, a synergistic interaction between the virulent GTD pathogen Neofusicoccum parvum and the bacterial taxa Bacillus pumilus and Xanthomonas sp. was found to be linked to the development of vascular necrosis in artificially inoculated grapevine cuttings (Haidar et al., 2016).

CONCLUSIONS
According to the results of this study, even fungal species, which are generally considered to be beneficial members of the plant microbiome, may promote the pathogenesis of Esca disease. The fungal species Apu showed a strain-dependent disease-promoting effect, probably based on the secretion of polysaccharides. The strain-dependence of pathogenesis in the case of Pch strains suggests that the proteins secreted by this fungus may also play a role in symptom development. In addition to the possible synergistic action of Pch proteins and Apu polysaccharides, direct interaction between these two fungi may also contribute to the development of Esca disease.