Biocontrol Effects of Paecilomyces variotii against Fungal Plant Diseases

The genus Paecilomyces is known for its potential application in the control of pests and diseases; however, its use in agriculture is limited to few species. Research interest in new formulations based on microorganisms for the control of pathogens is growing exponentially; therefore, it is necessary to study new isolates, which may help control diseases effectively, and to examine their compatibility with established agricultural control methods. We analysed in vitro and in vivo the antagonistic capacity of Paecilomyces variotii against seven phytopathogens with a high incidence in different crops, and we examined its compatibility with 24 commercial fungicides. P. variotii was applied in the following pathosystems: B. cinereal—melon, Sclerotinia sclerotiorum—pepper, R. solani—tomato, F. solani—zucchini, P. aphanidermatum—melon, M. melonis—melon, and P. xanthii—zucchini. The results showed strong control effects on M. melonis and P. xanthii, reducing the disease severity index by 78% and 76%, respectively. The reduction in disease severity in the other pathosystems ranged from 29% to 44%. However, application of metabolites alone did not cause any significant effect on mycelial growth of phytopathogens, apart from F. solani, in which up to 12% inhibition was observed in vitro when the extract was applied at a concentration of 15% in the medium. P. variotii was compatible with most of the tested fungicides, and of the 24 fungicides tested at the maximum authorised dose, 6 acted as fungicides, 4 as fungistatics, and the remaining showed inhibition rates ranging from 18.2% to 95.8%. These results indicate that P. variotii is a potential biological control agent to be used against several aerial and soil diseases, thus it should be integrated into modern pest management strategies.


Introduction
Various pests and diseases cause significant annual losses in global agriculture by reducing yield and crop quality [1,2] and such losses are estimated at around 20-40%, which represents a significant economic impact valued at hundreds of millions of dollars per year [3,4]. Common practice for controlling such pathogens is the indiscriminate application of agrochemicals; however, biological control by endophytic microorganisms is a more environmentally friendly alternative [5][6][7].
The endophytic fungus Paecilomyces variotii is an ascomycete that is morphologically characterised by the development of verticillate conidiophores with irregular branches in which cylindrical phialides are inserted with a pronounced narrowing at the neck where yellow or pale brown conidia are located [8,9]. This fungus produces chlamydospores with walls that are smooth or slightly rough, depending on the strain. P. variotii is thermophilic, ubiquitous, and highly versatile, and its spores can withstand temperatures above 80 • C for 15 min [10]; furthermore, it grows rapidly, can develop under oxygen-limited conditions, and produces a wide range of metabolites, therefore this species stands out among decomposing fungi [11]. It can adapt to a large variety of habitats (in air, edaphic, and aquatic environments) and can be found in plant debris and compost, inside plants as an 2. Materials and Methods 2.1. Fungal Isolates P. variotii CDG33 was obtained from autochthonous plants grown at the Cabo de Gata Natural Park, Almería, Spain [51]. This isolate has been tested to determine its potential biocontrol effect against Botrytis cinerea Pers, Sclerotinia sclerotiorum (Lib.) de Bary, Fusarium solani (Mart.) Sacc, Pythium aphanidermatum (Edson) Fitzp, Rhizoctonia solani J.G. Kühn, Mycosphaerella melonis (Pass.) and Podosphaera xanthii (Castagne) U. Braun & Shishkoff.
Different crop plants infected with pathogens were collected in the province of Almería, Spain, during 2017-2019. Selected stem and leaf sections with active lesions were cut, and tissue sections from the boundaries between healthy and discoloured areas were placed in potato dextrose agar (PDA; Bioxon, Becton Dickinson, Mexico). Pathogenicity tests were carried out to confirm pathogenic capacity.
The fungal isolate was grown and maintened on PDA in the dark at 25 -27 ± 2 • C for 5 or 10 days. Spore suspensions of P. variotii and phytopathogen isolates (except for R. solani, S. sclerotiorum, and P. xanthii) were prepared by flooding plates of 10-day-old cultures with sterile distilled water, scraping the surface using a sterile glass rod, and subsequent filtering of the suspensions, after which they were adjusted to a concentration of 1 × 10 6 spores/mL using a Neubauer haemocytometer. In the case of R. solani and S. sclerotiorum, inoculation was performed using mycelium adjusted to a concentration of approximately 10 4 colony forming units/mL (CFU·mL −1 ). To produce mycelium, two agar plugs were cut from a colony and were placed in a flask containing PDB. The flasks were placed on a shaker for four days at 25 • C, and the content was homogenised for subsequent inoculation. The inoculum of P. xanthii was directly collected from affected leaves.

Dual Culture Bioassays
P. variotii was screened for in vitro antagonism against six phytopathogens using the confrontation assay of Santos et al. [52]. Petri dishes (9 cm diameter) containing 17 mL PDA were prepared, and 0.5 cm plugs of mycelium of all fungi were cut from the growing edge of 7-days-old cultures with active growth of each isolate. The plugs were placed at the margin of Petri dishes with a distance of 8 cm between the antagonist and phytopathogen fungi. A plug of PDA medium was used as a control treatment while the pathogen plug was placed at the other side. All plates were sealed using parafilm and were incubated in the dark at 25 • C for different periods of time, depending on the growth rate of the respective pathogen (3-10 days).
Radial fungal colony growth was measured daily. Results were transformed into percentages of mycelium growth inhibition (PIRM). These tests were carried out using five replicates. The areas where antagonist and pathogen merged were also observed using a dissection microscope and a compound light microscope at 100-fold and 400-fold magnification.

Effects of Non-Volatile Metabolites of P. variotii on Fungal Pathogens
To determine the effects of non-volatile metabolites of P. variotii, 300 mL PDB medium was inoculated with two 0.5 cm plugs of mycelium and incubated without shaking at 25 • C in the dark for 7, 14, 21, 28, 58, 88, and 118 days. Subsequently, the medium was filtered through double layered muslin cloth to remove mycelium and was sterilized by microfiltration through sterilized membranes of 0.22 µm pore size (Merck Millipore, Burlington, MA, USA). Extracts were then incorporated and mixed with cooled PDA at 5, 10, and 15% (v/v) and were immediately poured into Petri dishes (50 mm). In the control treatments, distilled water was used instead of extract. Agar plugs (0.5 cm diameter) covered with actively growing mycelia of the different pathogens were individually inoculated on the media and were incubated in the dark at 25 • C for 10 days, apart from P. aphanidermatum which were incubated for 3 days, and S. sclerotiorum and B. cinerea which were incubated for 5 days under the same conditions. After incubation, mycelial growth of each fungus was measured as the average of two perpendicular diameters of the colony, and inhibitor effects were calculated in comparison to the controls. Five replicates were used for each incubation time and pathogen.

Antifungal Volatile Organic Compounds Bioassay
In vitro inhibition effects of volatile organic compounds produced by P. variotii against different phytopathogenic fungi was determined. For this, a plug (0.5 cm diameter) was cut from the actively growing edges of P. variotii plates and was placed at the centre of 5.5 cm Petri dishes containing PDA medium. After two days of incubation of the antagonist, the same procedure was repeated for each of the pathogens, and a Cartesian axis was drawn from the centre of each plate to facilitate measurements. Both open plates were placed on the base of a square 12 cm Petri dish on which a further inverted base was placed after which they were sealed using parafilm. The air volume was approximately 403.20 cm 3 , preventing direct contact between the two fungi. The control plates were prepared following the same procedure but without the antagonist plug. The plates were incubated at 25 ± 1 • C, and mycelial growth of the pathogen was measured in both directions, which was performed daily for fungi with rapid growth (B. cinerea, S. sclerotiorum, P. aphanidermatum and R. solani) and every two days for those growing more slowly (M. melonis and F. solani), until the edge of the plate was reached. Three replicates were examined each day.

Detached Leaf Assay
Before the in vivo assay, suppressive effects of P. variotii on B. cinerea, S. sclerotiorum, and M. melonis were assessed using a detached leaf assay as described by Novak et al. [53] and Patial et al. [54]. Leaves of cucumber (variety Marketmore), pepper (var. red cherry) and tomato (var. Largo de Reus) seedlings were disinfected using 3% sodium hypochlorite for 30 s and were washed twice to remove residues. The leaves were immersed for 3 min in a solution containing P. variotii spores at a dose of 10 6 spores·mL −1 or were immersed in an aqueous solution containing the fungicide Switch (cyprodinil 37.5% and fludioxonil 25% (WG) w/w; Syngenta, Basel, Switzerland) at 600 ppm, to compare efficacy. Whole leaves or fragments were then placed on wet filter paper in Petri dishes, and the centre of the leaf or leaf fragment was carefully punctured using a sterilised needle. A 0.5 cm plug containing the corresponding pathogen was placed at the puncture site. Petri dishes were then incubated at 25 • C, and number of leaves with symptoms were counted and photographed 72 h after inoculation. This experiment was repeated twice.

Compatibility of P. variotii with Fungicides
In vitro compatibility of P. variotii with different selected fungicides for mycelial growth inhibition was established using the poisoned food technique [55,56]. The minimum recommended dose (D2), the maximum recommended dose (D3), 0.5 × D2, and 1.5 × D3 (Table 1) of each fungicide were tested. Using a sterile cork borer, mycelial plugs (0.5 cm diameter) were cut from actively growing seven-day-old fungal cultures and were placed in the centre of a Petri dish containing PDA supplemented with various pesticides. Five replicates were used per treatment. Fungicidal or fungistatic effects of each fungicide was determined by transferring the initial plug of P. variotii which did not grow with fungicide, to a PDA plate without fungicide.

Greenhouse Evaluation of P. variotii on Disease Severity of Seven Phytopathogens
Biocontrol effects of P. variotii on different pathosystems was determined: B. cinerealmelon (var. Piel de Sapo), S. sclerotiorum-pepper (var. Red Cherry), R. solani-tomato (var. Largo de Reus), F. solani-zucchini (P. aphanidermatum-melon, M. melonis-melon, and P. xanthii-zucchini (var. Milenio). Seeds were disinfected using 2% hypochlorite for 3 min and were washed thoroughly with tap water to eliminate residues. Subsequently, the seeds were sown in 500 mL pots containing a commercial peat mix at one seed per pot. A 5 mL spore suspension of each isolate of P. variotii (PAE) was added to each pot at 1 × 10 6 propagules/plant; the control treatments received 5 mL water. To assess diseases of aerial parts, foliar spraying was carried out at the same dose to wet the whole plant with the P. variotii solution. Spraying was carried out three days before applying conidia/mycelium of the respective pathogen.
To prepare P. aphanidermatum inocula the procedure described by Marin et al. [58] was followed. Inocula of the other phytopathogens were prepared by scraping and subsequent filtration as previously described, apart from R. solani, S. sclerotiorum and P. xanthii. Phytopathogens were inoculated when the plants showed a second true leaf, and using a sterile micropipette, inoculation was performed by uniformly applying the zoospores/conidia suspension (5 mL) at a concentration of 10 4 CFU·mL −1 uniformly to the peat surface. In the case of B. cinerea and M. melonis, the pathogen was applied by spraying the plant five times at the same concentration (6 mL). Before this, the first true leaf had been cut to facilitate pathogen entry. Inoculation with R. solani was carried out by mixing mycelium into the substrate; inoculation with S. sclerotiorum was performed using the spray mycelium method as described by Chen and Wang [59]. The stem was wounded to facilitate pathogen entry. Symptom severity was recorded continuously, and 30-60 days after inoculation, a final disease severity index was estimated according to the following scale: 0 = healthy plant; 1 = initial symptoms; 2 = moderate symptoms (25%); 3 = affected plant (50%); 4 = severely affected plant (75%); and 5 = dead plant ( Figure 1).

Greenhouse Evaluation of P. Variotii on Disease Severity of Seven Phytopathogens
Biocontrol effects of P. variotii on different pathosystems was determined: B. cinereal-melon (var. Piel de Sapo), S. sclerotiorum-pepper (var. Red Cherry), R. solanitomato (var. Largo de Reus), F. solani-zucchini (P. aphanidermatum-melon, M. melonismelon, and P. xanthii-zucchini (var. Milenio). Seeds were disinfected using 2% hypochlorite for 3 min and were washed thoroughly with tap water to eliminate residues. Subsequently, the seeds were sown in 500 mL pots containing a commercial peat mix at one seed per pot. A 5 mL spore suspension of each isolate of P. variotii (PAE) was added to each pot at 1•10 6 propagules/plant; the control treatments received 5 mL water. To assess diseases of aerial parts, foliar spraying was carried out at the same dose to wet the whole plant with the P. variotii solution. Spraying was carried out three days before applying conidia/mycelium of the respective pathogen.
To prepare P. aphanidermatum inocula the procedure described by Marin et al. [58] was followed. Inocula of the other phytopathogens were prepared by scraping and subsequent filtration as previously described, apart from R. solani, S. sclerotiorum and P. xanthii. Phytopathogens were inoculated when the plants showed a second true leaf, and using a sterile micropipette, inoculation was performed by uniformly applying the zoospores/conidia suspension (5 mL) at a concentration of 10 4 CFU•mL −1 uniformly to the peat surface. In the case of B. cinerea and M. melonis, the pathogen was applied by spraying the plant five times at the same concentration (6 mL). Before this, the first true leaf had been cut to facilitate pathogen entry. Inoculation with R. solani was carried out by mixing mycelium into the substrate; inoculation with S. sclerotiorum was performed using the spray mycelium method as described by Chen and Wang [59]. The stem was wounded to facilitate pathogen entry. Symptom severity was recorded continuously, and 30-60 days after inoculation, a final disease severity index was estimated according to the following scale: 0 = healthy plant; 1 = initial symptoms; 2 = moderate symptoms (25%); 3 = affected plant (50%); 4 = severely affected plant (75%); and 5 = dead plant ( Figure 1). To determine the control effect of P. variotii on powdery mildew, an inoculum of P. xanthii was prepared from field-collected zucchini leaves affected by cucurbit powdery mildew (Figure 2). Using a sprayer, sterile distilled water was sprayed under pressure to rinse the leaves of fungal conidia. The suspension was collected, and experimental plants were immediately inoculated at a concentration of 10 4 CFU•mL −1 . To determine the suppressive effect of P. variotii on the disease, the leaf area ( Figure 2) affected by powdery mildew was determined using WinDIAS 3.1.lnk software (Dynamax. Fresno. CA. USA) to calculate the proportion of affected leaf area with respect to total leaf area. Additionally, the numbers of affected leaves and petioles per plant were counted. To determine the control effect of P. variotii on powdery mildew, an inoculum of P. xanthii was prepared from field-collected zucchini leaves affected by cucurbit powdery mildew (Figure 2). Using a sprayer, sterile distilled water was sprayed under pressure to rinse the leaves of fungal conidia. The suspension was collected, and experimental plants were immediately inoculated at a concentration of 10 4 CFU·mL −1 . To determine the suppressive effect of P. variotii on the disease, the leaf area ( Figure 2) affected by powdery mildew was determined using WinDIAS 3.1.lnk software (Dynamax, Fresno, CA, USA) to calculate the proportion of affected leaf area with respect to total leaf area. Additionally, the numbers of affected leaves and petioles per plant were counted. All pathogenicity tests were performed under greenhouse conditions and in different seasons to provide optimal environmental conditions for each pathogen. Experimental units consisted of four repetitions with 24 plants per pathosystem. Two experiments were conducted using a completely randomized block design.

Statistical Analyses
Data were analysed using analysis of variance with Statgraphics Centurion version XVI software. A preliminary test was performed to assess the normal distribution of the data. Mean separation was carried out using Fisher's least significant difference test. Data were tested using a one-way analysis of variance or Student's t-test; statistical significance is reported at p < 0.05. All pathogenicity tests were performed under greenhouse conditions and in different seasons to provide optimal environmental conditions for each pathogen. Experimental units consisted of four repetitions with 24 plants per pathosystem. Two experiments were conducted using a completely randomized block design.

Statistical Analyses
Data were analysed using analysis of variance with Statgraphics Centurion version XVI software. A preliminary test was performed to assess the normal distribution of the data. Mean separation was carried out using Fisher's least significant difference test. Data were tested using a one-way analysis of variance or Student's t-test; statistical significance is reported at p < 0.05.

In Vitro Antagonism of P. variotii
The results of the dual culture assay, expressed as inhibition percentages, are shown in Figure 3. Colony growth of P. variotii occurred rapidly (48-72 h) with isolated colonies covering the plate, demonstrating the inhibitory effect of the antagonist (Figure 4). The highest percentages of inhibition were detected in P. aphanidermatum and F. solani. Moreover, P. variotii initially showed inhibition of approximately 2 mm (Figure 4), which subsequently disappeared under direct contact between the two colonies. This inhibition zone indicates the presence of fungistatic metabolites secreted by P. variotii. Parasitism was not observed, even though hyphal destruction and vacuole formation were observed.

In Vitro Antagonism of P. Variotii
The results of the dual culture assay, expressed as inhibition percentages, are shown in Figure 3. Colony growth of P. variotii occurred rapidly (48-72 h) with isolated colonies covering the plate, demonstrating the inhibitory effect of the antagonist (Figure 4). The highest percentages of inhibition were detected in P. aphanidermatum and F. solani. Moreover, P. variotii initially showed inhibition of approximately 2 mm (Figure 4), which subsequently disappeared under direct contact between the two colonies. This inhibition zone indicates the presence of fungistatic metabolites secreted by P. variotii. Parasitism was not observed, even though hyphal destruction and vacuole formation were observed.

In Vitro Effects of Cell-Free Filtrate and Antifungal Volatile Compounds of P. variotii on Mycelial Growth of Phytopathogens
Cell-free filtrate of P. variotii did not show any inhibitory effect on mycelial growth of the analysed fungi in the different tested conditions, except for F. solani, which showed inhibition ranging from 1.7% to 12% at concentrations of 10% and 15%, regardless of the incubation time ( Figure 5).

In Vitro Effects of Cell-Free Filtrate and Antifungal Volatile Compounds of P. Variotii on Mycelial Growth of Phytopathogens
Cell-free filtrate of P. variotii did not show any inhibitory effect on mycelial growth of the analysed fungi in the different tested conditions, except for F. solani, which showed inhibition ranging from 1.7% to 12% at concentrations of 10% and 15%, regardless of the incubation time ( Figure 5).

In Vitro Effects of Cell-Free Filtrate and Antifungal Volatile Compounds of P. Variotii on Mycelial Growth of Phytopathogens
Cell-free filtrate of P. variotii did not show any inhibitory effect on mycelial growth of the analysed fungi in the different tested conditions, except for F. solani, which showed inhibition ranging from 1.7% to 12% at concentrations of 10% and 15%, regardless of the incubation time ( Figure 5).  By contrast, volatile metabolites of P. variotii showed no significant inhibition of the tested fungi. In the case of M. melonis, low inhibition of 5% and 7% were observed for the L1 and L2 readings (L), respectively. For F. solani, 3.2% inhibition was detected in the L3 reading ( Figure 6). By contrast, volatile metabolites of P. variotii showed no significant inhibition of the tested fungi. In the case of M. melonis, low inhibition of 5% and 7% were observed for the L1 and L2 readings (L), respectively. For F. solani, 3.2% inhibition was detected in the L3 reading ( Figure 6).

Compatibility for Use of P. Variotii with Fungicides
The observed effects of 24 fungicides on growth and development of P. variotii are presented in Table 1. The results indicate that mycelial growth of P. variotii was affected by the different doses of each of the fungicides tested in vitro, compared to the control. According to the OILB scale [57], the compatibility of the 24 fungicides tested for minimum (D2) and maximum (D3) recommended doses in horticultural crops was as follows: 2 were harmless (Pencycuron and Propamocarb; inhibition < 30%), 6 were slightly toxic (diethofencarb, triadimenol, myclobutanil, azoxystrobin, kresoxim-methyl, and Fosetyl-AL; 30-75%), 4 were moderately toxic (thiophanate-methyl, flutriafol, Chlorothalonil, and Cymoxanil; 75-90%), and 10 were toxic (iprodione, tetraconazole, pyrimethanil, etridiazole, copper oxychloride, mancozeb, cyprodinil + fludioxonil, Folpet + metalaxyl-M, dimethomorph + mancozeb, and benalaxil + cymoxanil + mancozeb; >90%). Only two fungicides (fenhexamid and copper hydroxide) showed different behaviour according to the scale at maximum and minimum doses; both were toxic at the maximum doses (D3) and slightly toxic and moderately toxic, respectively, at the minimum doses (D2). The four fungicides formulated with two or three active ingredients were toxic, with 100% mycelial growth inhibition at all tested doses; one mixture (cyprodinil + fludioxonil) was fungistatic, and the other three were fungicidal. Of the remaining six toxic fungicides, total growth inhibition was observed only in three and five at doses D2 and D3, respectively, with fungicidal effects only using pyrimethanil at the minimum dose (D2) and with iprodione, pyrimethanil, and etridiazole at the maximum dose (D3).

Detached Leaves
Inoculation of detached cucumber, pepper, and tomato leaves with mycelium of B. cinerea, S. sclerotiorum, and M. melonis resulted in marked inhibition of disease development in leaves inoculated with P. variotii, at the same efficacy as the control fungicide. Leaves not inoculated with P. variotii showed mycelial growth and obvious symptoms of rotting (Figure 7). Means with * denote significant difference between both treatments according a pairwise comparison using Student's t-test (p < 0.05).

Compatibility for Use of P. variotii with Fungicides
The observed effects of 24 fungicides on growth and development of P. variotii are presented in Table 1. The results indicate that mycelial growth of P. variotii was affected by the different doses of each of the fungicides tested in vitro, compared to the control. According to the OILB scale [57], the compatibility of the 24 fungicides tested for minimum (D2) and maximum (D3) recommended doses in horticultural crops was as follows: 2 were harmless (Pencycuron and Propamocarb; inhibition < 30%), 6 were slightly toxic (diethofencarb, triadimenol, myclobutanil, azoxystrobin, kresoxim-methyl, and Fosetyl-AL; 30-75%), 4 were moderately toxic (thiophanate-methyl, flutriafol, Chlorothalonil, and Cymoxanil; 75-90%), and 10 were toxic (iprodione, tetraconazole, pyrimethanil, etridiazole, copper oxychloride, mancozeb, cyprodinil + fludioxonil, Folpet + metalaxyl-M, dimethomorph + mancozeb, and benalaxil + cymoxanil + mancozeb; >90%). Only two fungicides (fenhexamid and copper hydroxide) showed different behaviour according to the scale at maximum and minimum doses; both were toxic at the maximum doses (D3) and slightly toxic and moderately toxic, respectively, at the minimum doses (D2). The four fungicides formulated with two or three active ingredients were toxic, with 100% mycelial growth inhibition at all tested doses; one mixture (cyprodinil + fludioxonil) was fungistatic, and the other three were fungicidal. Of the remaining six toxic fungicides, total growth inhibition was observed only in three and five at doses D2 and D3, respectively, with fungicidal effects only using pyrimethanil at the minimum dose (D2) and with iprodione, pyrimethanil, and etridiazole at the maximum dose (D3).

Detached Leaves
Inoculation of detached cucumber, pepper, and tomato leaves with mycelium of B. cinerea, S. sclerotiorum, and M. melonis resulted in marked inhibition of disease development in leaves inoculated with P. variotii, at the same efficacy as the control fungicide. Leaves not inoculated with P. variotii showed mycelial growth and obvious symptoms of rotting (Figure 7). posed to the in vitro results, with a reduction of 31.11% and 28.57%, respectively. In the case of F. solani, all plants treated with P. variotii reached disease grade 5, showing a delay of approximately 10 days.
Foliar application of P. variotii considerably suppressed the disease caused by P. xanthii (Figure 9), with a reduction of 75.80% in the proportion of leaf area showing symptoms. Similarly, the number of leaves and petioles with powdery mildew was reduced by 30.4% and 17.4%, respectively.

In Vivo Suppressive Effects of P. variotii on Diseases Caused by Phytopathogenic Fungi
In general, the severity of disease caused by different soil and aerial phytopathogens decreased due to application of P. variotii (Figure 8). The strongest biocontrol effect occurred against stem gummosis, with a reduction of 78% in disease severity after infection with by M. melonis after spraying with P. variotii. Lower values occurred in the other phytopathogen experiments. Reductions of 44%, 38.88%, and 37.20% in disease severity after infection with B. cinerea, S. sclerotiorum, and R. solani, respectively, were observed. Due to the random arrangement of treatments, disease was also observed in the uninoculated controls (T0) of some of the pathosystems. The highest disease severity indices following application of P. variotii occurred in the P. aphanidermatum and F. solani treatments, as opposed to the in vitro results, with a reduction of 31.11% and 28.57%, respectively. In the case of F. solani, all plants treated with P. variotii reached disease grade 5, showing a delay of approximately 10 days.
Foliar application of P. variotii considerably suppressed the disease caused by P. xanthii (Figure 9), with a reduction of 75.80% in the proportion of leaf area showing symptoms. Similarly, the number of leaves and petioles with powdery mildew was reduced by 30.4% and 17.4%, respectively.

Discussion
P. variotii has been described as a biological control agent of fungal and bacterial diseases. Its biocontrol effect is mostly due to its rapid growth and sporulation, resulting in higher competition success for space and available resources compared to other species [60]. However, few studies have conducted in vivo experiments on disease control by P. variotii [19].
The current study analysed in vitro and in vivo suppressive effects of P. variotii CDG33 on several phytopathogenic fungi that cause diseases of aerial and soil parts of crop plants.
In vitro antagonism results based on the dual culture technique showed very heterogeneous values in terms of inhibition percentages, ranging from 25% to 82.0%. Results obtained in in vitro experiments can vary substantially depending on experimental conditions and may differ even between isolates of the same species. Perveen et al. [33] reported similar results from the dual culture of P. variotii and P. aphanidermatum, with an inhibition percentage of 67-72%, whereas Rodrigo et al. [23] observed lower values that did not exceed 40%. In the present study, inhibition of P. aphanidermatum exceeded 80%, which is mainly due to the rapid dispersal of P. variotii spores during the experiment. Moreover, substantial inhibition of F. solani was observed, which is in line with the results of Ramzan et al. [27]. However, no direct relationship between in vitro and in vivo antagonism values was found, especially regarding P. aphanidermatum and F. solani. The same was observed in R. solani and M. melonis, which, despite showing the lowest in vitro antagonism values of no more than 40%, showed the highest disease reduction values in the plant experiments. This has been commonly observed [61,62] and can be caused, among other factors, by differences in techniques and components, such as the culture medium, presence of soil or substrate, and the plant itself. Even so, it is one of the most commonly used techniques and is considered the basis for screening microorganisms as a preliminary step to in vivo experiments, and to produce effective biological control agents in agriculture.
P. variotii produces a large number of volatile and non-volatile compounds [19,63,64]; however, little or no inhibition effect of the metabolites produced by isolate CDG33 on the tested phytopathogens was observed. This may be because the effect of volatile and non-volatile metabolites on the tested phytopathogens decreases over time within the first 24-48 h, and the diameter of the fungal colony in the treatments with filtrate was similar to that of the control. Only a mild inhibitory effect of non-volatile metabolites on F. solani was observed. Rodrigo et al. [23] obtained similar results, with a significant reduction of F. moniliforme with P. variotii filtrates. Such a time-dependent decrease in the effect of metabolites on Biscogniauxia mediterranea, the causal agent of charcoal disease in cork oak, was described previously [65].
It is important to assess compatibility of biological control agents and agrochemicals by using a combination of chemical and biological tools. Chemical treatments should be based on selection of agrochemicals with the lowest impact on biological control agents. Many studies, even though in clinical settings, confirmed the incompatibility of P. variotii with several fungicides, since different species of the genus Paecilomyces cause a variety of infections in immunocompromised patients [66][67][68]. Furthermore, previous studies indicate that Paecilomyces isolates from agricultural environments are more resistant to common fungicides. This is the case with P. niveus isolates from naturally occurring infections in apples, where it can cause post-harvest spoilage and patulin accumulation [69]. Species of the genus Paecilomyces may differ regarding their compatibility with fungicides, which may, however, also be the case regarding different strains of the same species due to genetic differences [70,71]. Even though the results obtained in the laboratory may not confirm those of the field experiments due to the higher severity of in vitro assays, our results indicate that P. variotii may be compatible with many of the tested fungicides regarding fungistatic effects or reduction of growth rates. Of the 24 tested fungicides, only 4 were incompatible at the minimum recommended doses, and 6 at the maximum doses. Particularly when fungicide mixtures with different modes of action are used, these products may thus interfere with successful biological control by P. variotii, which showed 100% fungal inhibition with fungicidal effects on P. variotii in three of the tested formulations. Regarding fungicides with one active ingredient, only iprodione, pyrimethanil, and etridiazole acted as fungicides in the authorised dose range. Other products that completely inhibited growth, such as tetraconazole, copper hydroxide, and copper oxychloride, were not biocidal.
Similar results were obtained by Er and Gökçe [72] who applied iprodione at 500 ppm active ingredient (ai) to an isolate of Paecilomyces fumosoroseus, even though a reduction of 51% in mycelial growth was observed using a different isolate, Loureiro et al. [73] classified the effect of iprodione on P. fumosoroseus as toxic at 500 ppm ai and very toxic at 750 ppm ai; Biango-Daniels et al. [69] analysed the compatibility of 30 isolates of P. niveus with the fungicide pyrimethanil to determine the effective fungicide dose at which growth was inhibited by 50% (E C 50). The E C 50% values ranged from 29 to 109 and were higher in isolates from crops. In the current study, the lowest tested dose (D1; 100 ppm ai) completely inhibited growth, although without a fungicidal effect which only occurred using higher doses. In contrast, the study of Gallego-Velásquez et al. [74] showed that the application of ethidiazole at 350 ppm ai completely inhibited growth of P. lilacinum. In our study, complete inhibition of P. variotii growth occurred at dose D3 (1440 ppm ai), and growth inhibition above 92% was observed only at 960 and 480 ppm ai. Other studies showed that copper hydroxide at 1660 and 2592 ppp ai used on P. fumosoroseus resulted in growth inhibition of approximately 9% and 82%, respectively [75], while Dermici and Denihan [76] observed a minimum inhibitory concentration of >5000 ppm ai for P. lilacinus. Copper oxychloride at 1500 ppm ai completely inhibits growth of P. lilacinus [76], however others found inhibition of only 44.1% using copper oxychloride at 4200 ppm ai. In the current study, copper oxychloride completely inhibited growth of P. variotii at all doses, and copper hydroxide did the same at the highest dose, thus both compounds were fungistatic. By contrast, Ondráčková et al. [76] observed growth inhibition of 79.6% in P. lilacinus treated with mancozeb at 1500 ppm ai; the lowest dose tested in our study was 800 ppm ai, which resulted in >90% inhibition, and it was not fungicidal at any of the tested doses. The fungicides pencycuron and propamocarb were compatible with P. variotii at all tested doses. These results corroborate those of previous studies on the effects of pencycuron at 540 ppm ai [77] and propamocarb at 1042.5 ppm ai [74] on P. lilacinus.
The present study demonstrated a high capacity of P. variotii CDG33 to reduce the severity of diseases caused by different aerial and soil phytopathogens in vivo, even though some pathogens showed very limited inhibition in vitro, according to a previous study [78]. The genus Paecilomyces is best known for its role as an entomopathogen and nematophagus fungus; however, other research findings also support the use of Paecilomyces fungi to prevent or treat fungal diseases [79]. In this regard, P. variotii has been previously described as a BCA in in vivo experiments in which the application of this microbial antagonist reduced colonisation of sunflower seedlings with M. phaseolina [28]. Similarly, Maitlo et al. [26] showed a significant reduction in the severity of disease caused by F. oxysporum f. sp. ciceris from 80% (control) to 6% and 36% following the application of P. variotii and P. lilacinus, respectively. By contrast, in vitro inhibition percentages of P. tenuis were described previously [80], which are similar to those observed in our study, particularly regarding R. solani (28%) and F. oxysporum (51%), even though lower concentrations of Pythium sp. (24%) were used.
No inhibitory effect on phytopathogens through volatile metabolites was observed. A previous study showed in vitro inhibition of B. cinerea by approximately 43% due to P. variotii [21] which is lower than the inhibition effect exerted by the P. variotii isolate CDG33, and this study suggested a significant reduction in rotting of kiwifruit caused by Botrytis, which was more effective than the conventional fungicide vinclozolin. Similar in vitro antagonism values (44%) of Paecilomyces sp. were observed in a study conducted by Marín-Chacón et al. [81].

Conclusions
According to our results, the application of the ascomycete P. variotii CDG33 as a preventive and/or control of fungal plant diseases may be a viable alternative to the use of conventional synthetic compounds. These insights can help reduce the amount of agrochemicals used which may leak into other ecosystems, and will therefore help reduce the environmental burden of agriculture.

Conflicts of Interest:
The authors declare that there is no conflict of interests regarding the publication of this manuscript.