The origin of grape berry-associated microbes is so far not fully understood (Martins et al., 2013). The microbial community of the fruit surface is thought to originate from the vineyard environment, with each locality having its own embedded microbial signature underlying regional wine characteristics (Barata et al., 2012). While soil, grapevine bark, leaves and other nearby plants are natural reservoirs for the microbial community associated with the grape berry, microbial transmission is largely induced by humans, insects and weather events (e.g., wind and rain) during the different developmental stages (Ramírez et al., 2020). The composition of these microbes on the grape surface change with fruit developmental stage, climate and sanitary conditions, and a core microbiota comprising the yeast genera Aureobasidium, Hanseniaspora and Metschnikowia is ubiquitously present regardless of environmental factors and plant genotype. A catalogue of cultivable yeasts isolated from grape berries and wines from the Douro wine region is shown in Figure 1 (Martins et al., 2014a).

Table

For each species, images show the colony morphology (left) and the cell architecture observed by DIC microscopy (right). Original photographs by Viviana Martins.

Culture-dependent approaches can be used to study the cultivable fraction of grape yeasts by growing and isolating them on selective and non-selective media. The macroscopic identification of a genus or species is mainly based on the observation of elevation, opacity, production of polysaccharide capsular material, pigmentation, pastiness, consistency of cream, smoothness, dryness and texture of the yeast colony (Pallmann et al., 2001). Microscopic identification is carried out via the observation of structures in culture, such as shape of cells, buds, pseudohyphae and mature hyphae. Other detailed structures, such as organelles, bud scar and fermentation capacity, can be visualised by transmission electron microscopy (Osumi, 2012). In addition, biochemical phenotypic assays are also used to screen for certain groups of the microbiome; for example, enzymatic tests (for example, urease, glycosidase, aminopeptidase and phosphatase), cycloheximide resistance, phenoloxidase production and esculin hydrolysis. In addition, it is possible to determine the potential of different isolates to ferment and assimilate in microtubes containing different colorimetric substrates (carbohydrates, organic acids, nitrogen and KNO₃) in the medium (Bascomb and Manafi, 1998; Hafeez and Aslanzadeh, 2018). Substrate utilisation or hydrolysis can be determined from increased turbidity, the generation of coloured products or the generation of fluorescent products after incubation. Biochemical tests can be used to identify different taxa and at different levels of performance, based on the type of biochemical substrates used (Pincus et al., 2007). Drumonde-Neves et al. (2016) cultivated yeasts on differential Wallerstein Nutrient Media as part of microbiota analyses carried out in vineyards in Europe (during the 2012 and 2013 to 2016 vintages) with the aim of identifying and characterising yeasts from grapes which can contribute to wine typicity when used in winemaking. By sequencing 26S rDNA D1/D2 domains, 5.8S rRNA gene and ITS region, the authors were able to identify Saccharomyces cerevisiae and non-Saccharomyces species: Candida boidinii, Candida zemplinina, Hanseniaspora guilliermondii, Issatchenkia terricola, Zygosaccharomyces bailii, Hanseniaspora uvarum, Zygoascus hellenicus and Hanseniaspora opuntiae. Thus, after a pre-selection of yeast isolates, their identity at species and strain level can be routinely confirmed by carrying out ITS/LSU barcoding, following the PCR amplification and Sanger sequencing of these regions (Martins et al., 2022). At advanced level, microorganisms can also be identified using a character strain approach in which microbial strains are analysed in more detail, taking into account factors such as genomic variations, metabolic pathways and physiological traits specific to each strain (Ghazi et al., 2022). This approach enables a more accurate and granular classification of microorganisms, leading to a deeper understanding of microbial diversity. Such recognition and differentiation of microbial strains can provide insight into the functional capabilities and ecological roles of specific strains – information that is necessary for targeting applications and ensuring their efficiency in various scientific and industrial contexts (Van Rossum et al., 2020). The incorporation of molecular tools, such as ITS ribotyping and bioinformatic analysis, has enhanced the study of the cultivable yeast occurrence and diversity in grape berries.

Recently, studies are increasingly applying DNA-based community fingerprinting methods, such as denaturing gradient gel electrophoresis (DGGE), single-strand conformational polymorphisms (SSCP), terminal restriction fragment length polymorphisms (T-RFLP), and automated ribosomal intergenic spacer analysis (ARISA), which can be used for the identification and analysis of grape berry microbiota without cultivation (Laforgue et al., 2009; Morgan etal ., 2017). DGGE coupled to PCR can be used for evaluating microbiota evolution during the ripening and fermentation stages, but this method is culture-dependent and is not highly sensitive (Laforgue et al., 2009). On the other hand, SSCP, T-RFLP and ARISA have been found to be useful tools for characterising the microbial diversity of wine (Martins et al., 2014a). However, these methods do not always provide a complete inventory of the microbial consortium living on grapes, as they sometimes do not reveal species of low abundance.

More recently, metagenomics involving high-throughput techniques is changing the paradigm in this line of research, laying the foundation for the characterisation of microbial diversity and composition, microbial functional analyses and potential biocontrol applications (Urich et al., 2008; Mackelprang et al., 2011; Zhou et al., 2022). These approaches are preferred to culture-dependent methods, because they are more accurate and less labour-intensive; they provide extensive microbial profiling of a sample, as both cultivable and non-cultivable taxa are detected. In metabarcoding approaches, microbial DNA is extracted from grape samples using commercial kits, and amplicon library paired-end sequencing can be carried out using primers for specific regions, such as ITS1; the generated raw sequence data can be deposited in public repository databases. The development of whole metagenome sequencing has further provided excellent conditions for detecting even species of the lowest abundance, making it one of the most preferred microbiota fingerprinting methods (Constantini et al., 2022).

Analysis of grape berry microbiota using culture-dependent and -independent methods (such as Next Generation Sequencing approaches) has made it possible to not only identify the yeast fingerprints of grape berries and other vine organs in many wine regions, but also to assess important aspects of plant-microbe interactions (such as the impact of organic and conventional fungicide treatments on grape berry yeast community) and the evolution of yeast biodiversity during spontaneous fermentations (Martins et al., 2014a; Agarbati et al., 2019). Although the majority of taxa in a sample can be detected using culture-independent methods, some species, including highly fermentative yeasts, can only be detected using culture-dependent methods involving enrichment procedures; these also enable the utilisation of isolates in subsequent experiments, such as inoculated fermentations. Studies on grape microbiota in which a combination of culture-dependent and -independent methods have been used are scarce (Martins et al., 2014a; Agarbati et al., 2019); however, the use of both methods is essential for ensuring the detection of the highest possible number of taxa, and thus accurately reflecting yeast diversity in each sample.

Over the years, it has been acknowledged that farming systems and vineyard soil management strategies can have an indelible impact on the grape microbial community. The low abundance of beneficial fungal diversity in grapes grown using common chemical fertilisers and pesticides is an indicator of toxicity of the chemical compounds to organisms (Wu et al., 2021). Increasing research shows that the population and composition of yeast communities resident on grape berries are modified as a result of antifungal treatments (Rogiers et al., 2005; Martins et al., 2014a; Yao et al., 2022). The foliar application of the conventional fungicides spiroxamina, copper-oxychloride and sulfur copper sulfate on cv. Montepulciano vines has been found to cause a reduction in yeast biodiversity, negatively affecting the fermenting yeasts, except for oxidative yeasts (such as Aureobasidium pullulans) at harvest time (Agarbati et al., 2019). In another study, the application of iprodione, pyrimethanil and fludioxonil plus cyprodinil on cv. Rebula vines had only a minor impact on the overall composition of grape berry microbial communities, but affected specific yeast species at harvest time (Čadež et al., 2010). Meanwhile, Setati et al. (2013) also reported higher yeast species diversity in the biodynamic vineyard and the lower diversity in the conventional vineyard where chemical fungicides like folpan^® (N-(trichloromethylthio)-phthalimide), rootex^™ (phosphorous acid), kumulus^® (sulphur), dithane^® (mancozeb), acrobat^® (dimethomorph), talendo^® (proquinazid), cungfu^® (copper hydroxide) and topaz^® (mono- and di-potassium salts of phosphorous acid) had been applied. In spite of their pest control benefits, the intensive application of copper- and sulphur-based fungicides in organic agriculture can negatively affect the epiphytic yeasts and yeast-like community on grape berry surfaces due to their wide fungicidal spectrum, causing sluggish or stuck fermentations (Martins et al., 2014a; Martins et al., 2014b). Meanwhile, other studies involving the application of synthetic-organic fertilisers and fungicides (e.g., cyprodinil, fludioxonil, metiram, penconazole and pyraclostrobin) used in conventional agriculture showed an increase in fungal diversity, as well as in the relative abundance of beneficial fungi on grape berry surfaces, thereby improving grape quality and storability (Čadež et al., 2010; Wu et al., 2021). However, many active components of conventional fungicides, including penconazole, spiroxamina, quinoxyfen, myclobutanil, among others, can have negative effects on the grape berry yeast biodiversity (Agarbati et al., 2019).

The preharvest field application of fungicides like potassium phosphate + sulphur + copper hydroxide for the control of vineyard diseases has also been shown to modulate the abundance of yeast microbiota on grapes and in musts (Grangeteau et al., 2016; Grangeteau et al., 2017); for instance, H. uvarum and P. kudriavzevii were predominant during spontaneous winery fermentations and S. cerevisiae was predominant during inoculated winery fermentations. The lower abundance of A. pullulans with potassium phosphonate (K-Pho) treatment against powdery mildew in conventional vineyards and its predominance in berries treated with potassium bicarbonate + copper hydroxide (K-Bic+C-h) and sulphur + metiram (S+Me) supports the above findings (Englezos et al., 2022), further suggesting that this yeast-like fungus could contribute to the effective control of nonbeneficial microbes. The impact of fungicides on yeast microbiota thus likely depends on other variables, since studies have shown either a reduction in biodiversity in both organic and conventional regimes or no significant effects of fungicide treatments.

Biocontrol agents for plant disease management are receiving increasing attention, and thus the focus of research is now on indigenous yeast strains in vineyards that exhibit antagonistic activity against pathogenic filamentous fungi. At primary level, to assess the biocontrol properties of the yeasts, researchers rely on methodologies involving a wide range of in vitro assays (e.g., the dual culture assay, the mouth-to-mouth dual assay, the agar plug diffusion assay and the agar disk diffusion assay), as well as in vivo investigations conducted in controlled laboratory, greenhouse or field conditions (Table 1).

Table

titre du tableau
Approach	Methodology	Advantages	Shortcomings	References
Agar surface dual culture assay	Co-inoculation of suspensions of antagonist and pathogen spores on the same Petri dish containing appropriate solid growth media, 2-3 cm apart; measurement of the inhibition zone around pathogen following incubation.	Ease of execution and broad applicability; allows the simultaneous screening of several yeast species towards the same pathogen; provides information on the effect of antagonist on pathogen spore germination and growth.	Result less reliable than in vivo assays; inhibition zone diameter depends on the initial distance between the inoculation spots of the antagonist and the pathogen.	Ayogu et al. (2023)
Dual culture assay on liquid media	Coinoculation of fungal spores or mycelia into liquid media in test tubes containing the yeast; microscopic examination of preparation for antagonism features after incubation.	Ease of implementation.	Supports the evaluation of antifungal activity of only one yeast at a time.	Serna-Díaz et al. (2020)
Agar plug diffusion assay	Transferring an agar plug (about 0.5 cm2) of filamentous fungi onto the middle of another agar plate previously inoculated with antagonistic yeast; measuring mycelia growth after incubation.	Gives idea of inhibitory effect of antagonists over hyphae proliferation; offers opportunity for testing processed antagonist products (heat denatured yeasts cells, cell-free supernatants).	No information provided on the effect of antagonist on spore germination; determination of antifungal activity of only one yeast per assay; agar plugs may have different pathogen loads.	Zepeda-Giraud et al. (2016)
Agar disk diffusion assay	Overlay of agar plates with the fungal suspension; placement of the yeast-laddened filter paper discs evenly on the agar surface; measurement of diameters of inhibition growth zones around the disks after incubation.	Simple, low cost; offers opportunity to test greater number of yeasts against one mould; ease of results interpretation.	The antagonistic activity of yeast may be affected by the permeability (quality) of the paper used in the assay.	Liu et al. (2017)
Mouth-to-mouth dual assay	Inoculation of pathogen suspension or agar plug in solid growth media and replacement of the Petri dish lid by an inverted open Petri dish containing yeast solid agar culture, allowing diffusion of yeast VOCs to the Petri dishes' headspace; measurement of mycelia growth of mould after incubation.	Identification of yeasts producing VOCs with antagonistic action.	Less reliable than the in vivo assay; does not provide information about antagonistic mechanisms involving production of non-VOCs.	Parafati et al. (2015)
In vivo dual assay	Inoculation of purposely wounded healthy berries, detached leaves, young vines, canes or root stocks with the yeast suspension, followed by a mould spore suspension at given concentrations; measuring antagonism by severity or disease incidence after incubation.	Controlled in planta approach; relatively cheap, simple and quick characterisation of antagonistic potential in vivo.	Antagonist activity is often lower than in in vitro approaches; does not yet contemplate the influence of environmental factors on antagonist efficiency; possible risk of affecting non-target microbial residents, especially in field setting.	Parafati et al. (2015); Maluleke et al. (2022); Pollard-Flamand et al. (2022); Ayogu et al. (2023)

In vitro dual assays are of primary importance when screening for biocontrol activity, since they are easy to perform and broadly applicable to the identification of bacterial, yeast, fungal, oomycete and viral biocontrol agents. Dual culture assays can be used to detect the release of antifungal compounds by yeasts that are in the proximity of filamentous fungi, and to study antagonist-pathogen interactions. They have been extensively used for determining the antifungal activity of native grape yeasts (Cordero-Bueso et al., 2017; Ayogu et al., 2023).

The agar plug diffusion method can be used not only for determining the inhibitory effects of antagonists, but also for testing processed antagonist products, such as heat-denatured yeast cells and cell-free supernatants (Zepeda-Giraud et al., 2016). However, this method still needs to be refined and complemented with other in vitro or in vivo assays. In addition, a major limitation is that the concentration of the pathogen may not be known, as a cut out plug of an actively growing mycelia of fungi is mainly used for inoculation.

The agar disk diffusion method - also known as the NCCLS (National Committee for Clinical Laboratory Standards) or Kirby-Bauer method and which was primarily developed for routine antimicrobial susceptibility testing - has found use in some yeast antagonism assays (Balouiri et al., 2016). It is a simple, low-cost method that can be used to test high numbers of microorganisms and antimicrobial agents, and the results can be interpreted easily.

The production of volatile organic compounds (VOCs) (e.g., ethyl acetate, isoamyl alcohol and 2-phenethyl ethanol) in the presence of a pathogen is a very common mechanism of antagonistic activity of many yeasts. The mouth-to-mouth dual assay can be carried out to assess strains exhibiting this type of behaviour (Parafati et al. 2015; Zhao et al., 2023).

The in vivo dual assay has been widely reported for its effectiveness in evaluating the biocontrol potential of various yeast strains against serious phytopathogens (Parafati et al., 2015; Agarbati et al., 2022). It is often performed on intentionally wounded plant materials, including healthy berries. Other in vivo and in planta approaches involving the use of sterile detached leaflets, young vines, canes and root stocks can also be applied to evaluate the biocontrol activity of various microbial species without going to the field (Agarbati et al., 2022). In particular, non-destructive detached leaf and young vine experiments can be carried out to determine the interactions between plants and phytopathogens, as well as to quickly characterise the infectivity of pathogens and plant resistance to them. In a recent study, native yeasts of grape berry, A. pullulans, W. anomalus, M. pulcherrima, C. intermedia, and R. glutinis, were identified as potential biocontrol yeasts for the control of A. niger, B. cinerea, C. acutatum, C. nymphaeae, P. expansum and P. glabrum both in vitro and in vivo using detached leaves and in vitro-grown grapevine plantlets (Figure 2) (Ayogu et al., 2023).

Table

Inhibition of A. niger infection by W. anomalus (a), with images showing a non-inoculated control leaf (C), a leaf inoculated with only the mould (M), a leaf inoculated with only the yeast (Y) and a leaf inoculated with both yeast and mould (Y+M). Heatmap with antagonistic activity of yeasts inoculated at 10⁸ CFU/mL towards fungi pathogens inoculated at 10⁷ spores/mL (b); values were centred and scaled in the row direction to form virtual colours, as presented in the colour key, and indicate the disease incidence index, as follows: 0 = no visible symptoms (0-5 % leaf area infected); 1 = trace symptoms (6-25 % leaf area); 2 = slight lesions (26-50 % leaf area); 3 = pale/yellowing (51-75 % leaf area); 4 = extensive deformation/loss of integrity (> 75 % leaf area). Data was retrieved from Ayogu et al. (2023).

While pathogenic fungi, such as Aspergillus niger, Botrytis cinerea, Colletothricum spp. and Penicillium spp., represent the most studied grapevine spoilage moulds that cause pre- and postharvest losses worldwide (Di Canito et al., 2021), the obligate biotrophs Erysiphe necator and Plasmopara viticola are also of grave concern, causing highly devastating mildews (Koledenkova et al., 2022). Compared to other fungi, the initial screening of these two biotrophs in vivo at laboratory scaleis more challenging (performed mostly on plant tissues), as they are both difficult to grow on culture media (Colcol and Baudoin, 2016). To date, their complete eradication is still difficult to achieve, as the traditional practices of pruning, defoliation, vineyard cover management, hot water treatment, irradiation, as well as chemical treatments, will not always be completelyeffective, and the outcome will depend on the product, method and organism involved (Table 2). Thus, the FAO estimates that plant diseases still cost the global economy $220 billion/year, roughly estimated as 20-40 % annual productivity losses (FAO, 2019). Of the grapevine fungal diseases, grey mould caused by B. cinerea and downy mildew caused by E. necator represent the endemic diseases responsible for economic losses for grape growers and wine producers worldwide (Rienth et al., 2021).

Table

titre du tableau
Disease	Characterisation	Causal agents	Active ingredients approved in EU	References
Bunch rot/Grey rot	Brown rot on bunches which can also affect other plant parts.	Botrytis cinerea	Copper (II) sulphate	Nally et al. (2012); Di Canito et al. (2021)
Anthracnose/Ripe rot	Circular, reddish-brown spots, which enlargen to cover the whole fruit. Leaves and branches become covered with brown spots and wither.	Colletotrichum acutatum,C. nymphaeae , C. gloeosporioides, C. musae, C. fructicola, C. aenigma, C. pseudoacutatum	Fludioxonil, folpet, captan	Ishii et al. (2022)
Aspergillus rot	Grape clusters lighten and acquire a dark colour. Berries turn into a bluish-brown mass following shrinkage and cracking.	Aspergillus carbonarius, A. niger, A. Alliaceus, A. ochraceus	Copper (II) sulphate, fludioxonil, cyprodynil, captan	Di Canito et al. (2021)
Fruit rot	Light-brown soft, watery rot followed by the appearance of blue‐green conidia on the fruit surface at advanced stages of decay.	Penicllium expansum, P. glabrum. P. purpurescens, P. decumbens, P. chrysogenum, P. crustosum, P. aurantiogriseum	Thiabendazole, imazalil, carbendazim	Di Canito et al. (2021)
Powdery mildew	Mycelial growth and sporulation occur on the surface of leaves and stems, resulting in a white fuzzy mildew appearance.	Erysiphe necator	Copper (II) sulphate, azoxystrobin, trifloxystrobin	Lukšić et al. (2022)
Downy mildew	Light green to yellow spots on upper leaf surface which eventually turn brown, berries turn a mottled dull green or reddish purple and readily fall from the cluster.	Plasmopara viticola	Copper oxychloride, copper hydroxide, captan	Koledenkova et al. (2022)
Black spot	A leaf spot that is characterised by small, water-soaked lesions.	Alternaria alternata, A. brassicicola	Copper (II) sulphate, fludioxinil	Ji et al. (2023)

With the aim of reducing by 50 % the use of harmful chemical pesticides by 2030, new EU rules encourage the use of microorganisms as biocontrol agents (European Commission, 2022). The results of a large number of studies support the use of native grape yeasts in the management of vineyard diseases, showing that various species and strains have differing efficacy and pathogen specificity. A recent review explored the distinct mechanisms involved in yeast-mediated biocontrol (Freimoser et al., 2019).

Although an assortment of native yeasts is continuously being evaluated for biocontrol potential and plant growth enhancement, little information is available regarding confirmation via studies in the field of the various promising results obtained in vitro; this is partially due to the variable conditions that characterise disease incidence and vine response across regions (Droby et al., 2016). A summary of successful biocontrol studies in field conditions is provided in Table 3. The most efficient formulations were found to be those for the control of rots caused by Botrytis and Aspergillus spp., and new information has also been obtained regarding the mechanisms underlying the antagonistic activity of yeasts for the biocontrol of grapevine mildewsin field trials.

Table

titre du tableau
Yeast strain	Action mechanism	Target pathogen	Crop	Localisation	Vintage	References
Candida sake CPA-1	Antifungal toxins	B. cinerea, P. expansum	Grape	Lleida, Spain	2005 - 2006, 2010 - 2011	Calvo-Garrido et al. (2013)
Candida oleophila 182	Induction of systemic resistance	P. digitatum	Grapefruit	Dagan, Israel	NA	Droby et al. (2002)
Cryptococcus laurentii LS-28	Competition for nutrients and space, secretion of hydrolytic enzymes (β-1,3- glucanase), antibiotic	B. cinerea, P. expansum, A. niger, R. stolonifer, Geotrichum citri-aurantii	Grape, citrus, apple, pear, strawberry, kiwi, peach, plum, olive	NA	NA	Liu et al. (2013)
Metschnikowia fruticola NRRL Y-27328, and CBS 8853	Competition for nutrients and space	cinerea, niger, Alternaria spp.	Grape, strawberry	California, United States	2000 - 2002	Karabulut et al. (2003)
Aureobasidium pullulans AP159-18	Competition for nutrients and space	B. cinerea	Grape	Siena, Italy	2015 - 2018	Galli et al. (2021)
A. pullulans Y-1	Competition for space	A. carbonarius	Grape	Corinthos County and Rhodes Island, Greece	2004 - 2006	Dimakopoulou et al. (2008)
A. pullulans 533 and 547	Competition for nutrients	B. cinerea	Table grape, sweet cherry	Southern Italy	1998 - 2001	Schena et al. (2003)
A. pullulans SCSUI8, SCSUH15 and STR10	Competition for nutrients	Greeneria uvicola	Grape berry, leaf	Tumbarumba, Australia	2011 - 2014	Rathnayake et al. (2018)
pullulans LS30	NA	Aspergillus carbonarius.	Grape	Abruzzo, Italy	2006 - 2007	de Felice et al. (2008)
Candida guilliermondii A42	Competition for nutrients	B. cinerea, A. niger, R. stolonifer	Grape	Beijin, China	1996 - 1998	Qing et al. (2000)
Rhodotorula glutitnis LS-11	Competition for nutrients and space, secretion of antifungal toxins	expansum	Grape, apple	Campobasso, Italy	NA	Lima et al. (1998)
M. pulcherrimaNCAIM Y0177071, NCAIM Y0177072, NCAIM Y0177071	Iron depletion	A. clavatus, A. niger, B. cinerea, Mucor piriformis	Grape	Tokaj, Hungary	2002-2003	Sipiczki (2006)
Debaryomyces hansenii 18858	Competition for nutrients	digitatum	Grapefruit	Dagan, Israel	NA	Droby et al. (1989)
Pseudozyma flocculosa	Secretion of antifungal unsaturated fatty acids	Erysiphe necator	Grape	Patras, Greece	1999 - 2001	Konstantinidou-Doltsinis et al. (2007)
Unassigned yeast taxa (ANYI-2, SATYI-1 and NYI-1)	Induction of phytoparasitism	E. necator	Grape	Maharashtra, India	NA	Raghuwanshi et al. (2017)
Aurereobasidium pullulans DSM 14940 and DSM 14941	NA	Plasmopara viticola	Grape	NA	NA	Harm et al. (2011)
Trichoderma harzianum T39144	NA	P. viticola	Grape	San Michele all’Adige, Italy	NA	Vecchione et al. (2007)

NA: Not available.

In a Greek vineyard containing cv. Grenache Rouge and Agiorgitiko, the grape berry-isolated yeast A. pullulans Y-1 (at 10⁷ cells/mL) was field-tested for use against A. carbonarius (at 10⁴ spores/mL): the yeast efficiently reduced the infection symptoms in berries at harvest and ochratoxin A contamination in must, showing a similar pattern to the commercial fungicide fludioxonil + cyprodinil (Dimakopoulou et al., 2008). Another field trial with A. pullulans strains L47 and 547 on table grapes showed a significant reduction of incidence and severity of grey mould (Schena et al., 2003): treatment with strain L547 significantly reduced postharvest rots in sweet cherries (by 47 %) and table grapes (by 38 %), whereas the application of strain L47 alone or in combination with calcium chloride and sodium bicarbonate effectively controlled rot incidence caused by B. cinerea in wounded sweet cherries. A strain of the yeast-like fungi A. pullulans LS30 field-tested for its biocontrol activity against Aspergillus spp. on wine grapes in the vineyard was also demonstrated to significantly reduce the severity of grape rots and OTA contamination by 81.9 % (de Felice et al. 2008).

Other in vivo field trials have demonstrated the efficacy of A. pullulans Y-1 in the control of A. carbonarius, whichcauses sour rot in grapevines (Dimakopoulou et al., 2008). Remarkably, in a commercial organic vineyard in Catalonia, Spain, when applied without the addition of biodegradable coatings, Candida sake CPA-1 significantly reduced sour rot incidence and severity, while novel film-forming formulations and the liquid (potato-starch) formulations of C. sake CPA-1 controlled B. cinerea by reducing disease severity from 56 % (CS) to 84 % (Carbó et al., 2019).

The seasonal application of the vineyard isolates Candida guilliermondii A42 and Acremonium cephalosporium B11 to table (Thompson Seedless and Superior Seedless) and wine (Sauvignon blanc) grapes in Israeli vineyards showed C.guilliermondii A42 to be effective in reducing decay due to moulds in both types of grapes in the given seasons (Zahavi et al. 2000). In this study, although A.cephalosporium B11 did not effectively reduce the sour rot of grape caused by A.niger , the findings indicate that both native yeasts can prevent disease progression in the vineyard and reduce postharvest decay.

Other studies, like that of Calvo-Garrido et al. (2017), have shown that formulations of C.sake CPA-1 combined with Fungicover to reduce Botrytis incidence and severity by 35 - 85 %. The overall results show a synergy between antagonistic yeasts and most phytosanitary products commonly used in viticulture, with mixed formulations being the most effective for biocontrol (Calvo-Garrido et al., 2017). Furthermore, the efficiency of the same formulations has been found to be comparable to that of commercialised bio-based products, such as BOTRY-Zen ^® and ARMOUR-Zen ^®, when used against Botrytis infection in organic vineyards (Calvo-Garrido et al., 2013).

Field experiments on the biocontrol of powdery mildew have demonstrated that formulations based on Pseudozyma flocculosa (Sporodex^® L) reduce disease incidence, thus representing promising biocontrol strategies (Konstantinidou-Doltsinis et al., 2007). The antagonistic action of P. flocculosa is linked to the secretion of extracellular fatty acids like flocculosin and cis-9-heptadecenoic acid, which disrupt the cytoplasmic membrane in a variety of fungi (Konstantinidou-Doltsinis et al., 2007). Under field conditions, other native grapevine leaf yeast isolates with no assigned taxa (ANYI-2, SATYI-1 and NYI-1) have been shown to reduce powdery mildew incidence by ≈80 % (Raghuwanshi et al., 2017): the yeast treatment caused a shift in the leaf microflora, reducing phyllosphere microbial biodiversity while increasing the abundance of yeast population, and was suggested to induce phytoparasitism via adherence of yeast spores to the fungal spore epidermis, leading to its shrinkage. Assays conducted in greenhouse conditions on potted vines have shown that a treatment based on A.pullulans (Aureo) provides some protection against P.viticola; after the treatments, the activity of the pathogenesis-related (PR) protein glucanase slightly increased, but not enough to consider this microorganism as a resistance inducer (Harm et al., 2011). However, the study highlights its potential to partly induce natural resistance metabolites that enhance the tolerance of grapevines to P.viticola through synergistic effects with conventional fungicides, thereby reducing their ecological burden, while not being sufficiently effective to entirely replace them. Similarly, under greenhouse conditions using potted vines of cv. Pinot gris and cv. Chardonnay, microbial formulations comprising Trichoderma harzianum T39144 (Trichodex) and a combination of Clonostachis rosea and T. harzianum demonstrated good ability to reduce the disease incidence of P.viticola (Vecchione et al., 2007). The performance of microbial formulations lasts for only a short time after inoculation and it is possible that the disease will not be completely controlled (Vecchione et al., 2007); therefore, for effective disease management, a strategy needs to be found for controlling the fungi at multiple site and stages of its life cycle.

As mentioned previously, the outcomes of vineyard foliar spraying of yeast formulations do not always align with those obtained in laboratory scale experiments. For the optimal performance of an antagonistic yeast, the cells must colonise the plant surface and consequently remain active during periods conducive to plant infection (Dimakopoulou et al., 2008); this requires the optimisation of several variables: the composition of the yeast formulations, the treatment timings and the operative methods for applying them in the vineyard.

The use of yeast isolates from wine grape microbiota for developing biofungicides is favored by their environmental friendliness, ease of manipulation and GRAS status (Freimoser et al., 2019; Gonzalez and Morales, 2022). The grape berry-associated yeasts play an essential role in wine fermentations and co-fermentations, as the metabolic interactions that occur after controlled inoculations using Saccharomyces and non-Saccharomyces yeasts can enhance the complexity of wine flavour, contributing to the production of innovative and distinct wines (Álvarez et al., 2023). These yeasts are considered safe, but their potential risks to consumers should not be excluded (Riesute et al., 2021). In terms of cost, the use of yeasts seems to be cheaper than using conventional pesticides. Apart from products such as Botector^®, Noli^®, Romeo^® and Shemer^®, relatively few bioproducts are commercially available for use with grapes (Table 4). Certain microbes that have been studied in the laboratory for biocontrol may not be accepted for commercialisation due to inconsistent results in the field (Droby et al., 2016).

Table

titre du tableau
Product	Yeast species	Target pathogens	Crops	References
Blossom Protect	Aureobasidium pullulans DSM 14940 and DSM 14941	E. amylovora, B. cinerea, C. gloeosporioides	Apple	Kołodziejska et al. (2021)
Botector	Aureobasidium pullulans DSM 14940 and DSM 14941	B. cinerea	Apple, pear, grape, strawberry	Schilder (2013)
BoniProtect	Aureobasidium pullulans	Pezicula sp., Nectria sp., B. cinerea, M. fructigena, P. expansum	Apple, pear	Lima et al. (2005)
Noli	Metschnikowia fruticola NRRL Y-27328 KM1110 WDG	B. cinerea, Monilinia spp.	Table and wine grape, strawberry	Dukare et al. (2018)
Romeo	Saccharomyces cerevisiae cell walls	B. cinerea,Erysiphales, Plasmopara	Grape, tomato, strawberry, lettuce and cucumber	Dukare et al. (2018)
Aspire	Candida oleophila I-182	P. digitatum, Rhizopus stolonifer, Fusarium sp., B. cinerea,Sclerotinia sclerotium	Pome fruit, citrus, banana	Droby et al. (1998)
Nexy	Candida oleophila strain O	B. cinerea, Penicillium sp., Colletotrichum sp.	Pome, banana, citrus, sweet potato, peach, pepper, carrot	Droby et al. (2002)
CandifruitTM	Candida sake CPA-1	P. expansum, B. cinerea, Rhizopus nigricans	Apple, pear	Teixidó et al. (2011)
Yieldplus	Cryptococcus albidus	Botrytis sp., Penicillium sp., Mucor sp.	Pome fruit	Teixidó et al. (2011)
Shemer	Metschnikowia frutícola	Botrytis sp., Penicillium sp., Rhizopus sp., Aspergillus sp.	Grape, apricot, citrus, peach, pepper, strawberry, sweet potato	Droby et al. (2002)

Botector acts competitively and does not attack the metabolism of fungal pathogens, indicating that the product could be used to combat Botrytis cinerea in grape berries and other crops without the risk of resistance developing after several applications (Bio-Ferm, 2017). More research is needed to identify other microbial species, including fungi and bacteria, that could be marketed as novel agricultural bioproducts for other crops. It should be noted that some of the earlier commercialised biocontrol yeast formulations, such as Aspire and Yieldplus, were eventually withdrawn from distribution due to inconsistent efficacy, as well as low marketing success and profit (Spadaro and Droby, 2016). Most commercialised bioproducts are still subject to regulation by local authorities (Table 4). In this regard, 65 strains of microorganisms have to date been approved for plant protection in the European Union: 36 strains of fungi (e.g., Trichoderma harzianum strains ICC012, T25 and TV1, T. afroharzianum T-22, Isaria fumosorosea 97), yeasts (e.g., Aureobasidium pullulans DSM 14940 and DSM 14941, Metschnikowia fructicola NRRL Y-27328, Saccharomyces cerevisiae LAS02, Candida oleophila strain O) and oomycetes (e.g., Pythium oligandrum M1); 21 strains of bacteria (e.g., Bacillus amyloliquefaciens AH2, Bacillus thuringiensis subsp. kurstaki SA-12 ); and 7 viruses (e.g., Cydia pomonella Granulovirus (CpGV), Adoxophyes orana granulovirus). Of these agents, 32 strains have been recognised as fungicidal agents and can be used as such (see: ec.europa.eu, June 2020 data).

The shortcomings of using biocontrol agents are that they may not eradicate all the pests if they are not applied at the right time (i.e., before the pathogen’s establishment in the host), and they might not provide a high level of protection when used on a large scale (Spadaro and Droby, 2016). In addition, there is need for skillful management of the process; the farmers need to have an understanding of the biology of both the target pests and their natural enemies before they can use the agents. However, despite these shortcomings, the use of biological control agents can majorly contribute to safer food production by eliminating the occurrence and accumulation of pesticides within the global food chain and food delivery system.