Biocontrol agents and resistance inducers reduce Phytophthora crown rot (Phytophthora capsici) of sweet pepper in closed soilless culture

Copyright: © 2021 G. Gilardi, A. Vasileiadou, A. Garibaldi, M. L. Gullino. This is an open access, peer-reviewed article published by Firenze University Press (http://www.fupress.com/pm) and distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.


INTRODUCTION
Sweet pepper (Capsicum annuum) is a popular vegetable crop, which is greatly appreciated for its quality and taste for fresh consumption and for processing, and is increasingly grown in greenhouse environments. The commercial greenhouse production of this vegetable is extensive in European countries, including Spain (12,420 ha), Italy (2,370 ha), Poland (1,830 ha) and The Netherlands (1,320 ha) (2017 data; FAOSTAT 2019).
These diseases have been known in Italy since 1949 (Sibilia, 1952), and are still widespread, as a result of a lack of effective fumigants, no commercially acceptable resistant cultivars, and limited availability of resistant rootstocks which often only have partial resistance to the pathogen . Although several sweet pepper cultivars show full or partial resistance to P. capsici, stem blight, root rot and foliar blight of these cultivars are probably reduced by separate genetic systems, so that cultivars resistant to crown rot may not be resistant to foliar blight (Foster and Hausbeck, 2010). This complicates the use of genetic disease resistance in sweet pepper, increasing the need to improve marketable yield potential and good fruit quality (Barchenger et al., 2018;Acquadro et al., 2020). Development of pathogen resistance to fungicides, including the commonly used mefenoxam and metalaxyl (Bower and Coffey, 1985;Lamour and Hausbeck 2001;Parra and Ristaino, 2001;Tamietti and Valentino, 2001), and alternatives to phenylamides, such as cyazofamid, dimethomorph, flumorph and oxathiapoprolin (Bower and Coffey 1985;Kousik et al., 2008;Bi et al., 2014;Miao et al., 2016), has complicated management of these diseases.
Soilless cultivation is increasingly being adopted for many vegetable crops Lee, 2015. Sambo et al., 2019), as this is a good option when crop rotation is not feasible and resistant cultivars are not available, and because of increasing limitations in the use of fumigants and fungicides . Host species, plant growth substrate type and the target pathogen all play important roles in management of diseases in soilless cultivation (Zheng et al., 2000;van der Gaag and Wever, 2005;Khalil et al., 2009;Lee and Lee, 2015), and specific evaluations are required for new crops. The effects of soilless cultivation on sweet pepper have not been fully exploited, particularly in terms of effects on the reduction of soil-borne diseases, such as Phytophthora crown rot.
Apart from their intrinsic capability of reducing problems caused by soil-borne pathogens, soilless cultivation systems offer good opportunities for exploiting new disease management strategies, including the use of resistance inducers and biocontrol agents, where increasing limitations apply to use of traditional fungicides.
The present study was carried out in an experimental closed soilless system, under controlled conditions, to evaluate the efficacy of host resistance inducers, based on K-phosphite and K-silicate, used alone or combined, and of experimental biocontrol agents, against P. capsici on sweet pepper. These treatments were compared with a commercial formulation of Trichoderma gamsii + T. asperellum. Different treatments and types and timing of application were tested, aiming to develop appropriate strategies for their application in practical soilless production systems.

Soilless growing system and experimental layout
Twelve trials were carried out in a glasshouse at the AGROINNOVA Centre of Competence of the University of Torino, in Grugliasco, Torino, Italy, at temperatures ranging from 22 to 30°C, using fully automated closed soilless system. This is a small-scale hydroponic system, with recirculating nutrient solution. Each hydroponic unit consisted of one channel (6 m long and 25 cm wide) connected to a storage tank (300 L capacity) filled with nutrient solution, which was automatically delivered to the plants using an electronic control unit (Idromat2, Calpeda S.p.a.). Nutrient solution (at 1.5-1.6 mS cm -1 ) was pumped by emitters (one per pot) at a flow rate of 6 L h -1 from the storage tank, and fed to plants through drip emitters, and was left to drain back into the storage tank by gravity. The plants were fertigated with 5:3:8 N:P:K fertilizer (nutrient solution containing 120 mg L −1 total N, 30 mg L -1 P and 150 mg L -1 ; conductivity 1.5-1.6 mS cm -1 ). Nutrient solutions contained: 11.24 mM NO 3 , 4.8 mM NH 4 , 0.75 mM KH 2 PO 4 , 0.75 mM K 2 SO 4 , 0.012 mM Iron chelate EDTA, 2 mM MgO, 2 mM SO 3 , 0.2 mM B, 0.001 mM Mo, 0.15 mM Zn, 3.1 mM CaO, 0.05 mM Cu ++ , 0.25 mM Mn, and 12.2 mM K. The pH and conductivity of the nutrient solutions were checked regularly by using a pH meter and a SevenGo DUO TM SG23 conductivity meter (Tettler). The irrigation programme (three to six times per day) was revised according to the environmental conditions, particularly temperature. Each experimental unit consisted of six pots replicated five times (n = 30 pots). Two plants were planted in each pot, and six pots were each sub-replicate of 12 plants each. Five replicates were used per treatment (60 plants per treatment).

Plant material, and tested products and protocols
Fifteen-day-old plants of sweet pepper 'Corno di Toro' (Furia Sementi), which is susceptible to Phytophthora crown rot, were transplanted into 3 L capacity plastic pots filled with a growing medium containing: black peat soil (Brill Type 5: 15% of blond peat, 85% of black peat; pH 5.5-6.0, 1,100 g m -3 of N:P:K and traces of molybdenum, Georgsdorf) in all the trials.
Each set of trials included one untreated and inoculated experimental control, and different treatments with salt products or biocontrol agents tested alone or in combinations, which were applied according to three different protocols (Table 1). These were: Protocols I and II. The fertilizer-based phosphite (Alexine 95PS: 52% P 2 O 5 + 42% K 2 O, Massò), which is labelled as a phosphorus supplement for soilless application, and potassium silicate (K 2 SiO 3 , 33.7 to 34.7%, Andrea Gallo S.r.l.) were tested (Trials 1 to 6) in 2018 and 2019, to select the optimal rates and types of application for potassium phosphite and potassium silicate, used alone or in combinations. K 2 SiO 3 at 200 mg L -1 and K-phosphite were added directly to the nutrient solutions (NS) at the standard rate of 1.30 g L -1 P 2 O 5 + 1.05 g L -1 K 2 O, or at half this rate (0.65 P 2 O 5 + 0.52 g L -1 K 2 O + 0.525 g L -1 ) in experimental Protocol I. The same products, alone or in combinations, were applied to each pot around the crown of the seedlings using 100 mL per pot of the suspension prepared according to Protocol II. The treatments were carried out three times every 7 d (Table 1).
Protocol III. The following BCAs, isolated from suppressive composts, were tested: Pseudomonas sp. PB26 (Pugliese et al., 2008), Fusarium solani FUS25 (Gullino and Pugliese, 2011), Trichoderma sp. TW2 (Cucu et al., 2020), and a 1:1:1 mixture of three Pseudomonas spp. strains: Pseudomonas sp. FC 7B (EU836174), Pseudomonas putida FC 8B (EU836171), and Pseudomonas sp. FC 9B (EU836172). These strains were previously isolated from a suppressive rockwool substrate in a soilless system (Clematis et al., 2009;Srinivasan et al., 2009) (Table 1). The bacterium strains were maintained at 4°C in Luria Bertani (LB) slants throughout the study. The fresh bacterium suspensions were each prepared by inoculating a loop-full of cells into 30 mL of LB medium in 100 ml capacity Erlenmeyer flasks, and then incubating the suspension on a rotary shaker at 600 rpm. The cell suspension was centrifuged and resulting pellets were resuspended in sterile deionized water. The optical densities (OD 600 ) of cultures incubated for 48 h at 23°C were checked immediately before application, and was adjusted with sterile deionized water to 1 × 10 8 cells mL -1 before application. Trichoderma sp. TW2 was grown in a 1000 mL capacity flask containing 200 mL of potato dextrose broth (Sigma-Aldrich) and maintained under static conditions at 23°C. After 13 to 15 d, the produced conidia and mycelium were transferred to 200 mL of sterile distilled water and homogenized using a handheld rotary mixer. The conidium suspension obtained for the Trichoderma sp. TW2 isolate was standardized to 1 × 10 7 cfu mL -l .
The antagonist Fusarium solani FUS25 was propagated in 1000 mL capacity-flasks each containing 250 mL of potato dextrose broth maintained on a rotary shaker for 8 to 10 d at 160 rpm. The cultures were centrifuged at 8,000 g for 20 min at 4°C. Conidia and mycelium pellets were each transferred into 200 mL of sterile distilled water and homogenized using a rotary mixer. The inoculum suspensions were adjusted with sterile deionized water to 1 ×10 7 conidia mL -1 before application.
Each BCA suspension was applied at a final concentration of 1×10 7 cfu mL -1 to each pot and around the base of 15-d-old seedlings immediately after planting.
The BCAs were applied to the growing medium five times at 7 d intervals using 100 mL per pot of the suspension, following experimental Protocol III (Table 1). The experimental BCAs were compared with a commercial formulation of Trichoderma asperellum + T. gamsii (10% a.i.;'Remedier',Isagro), applied at the label rate of 0.25 g per liter of peat substrate (Table 1).

Phytophthora capsici strain and inoculation
A highly virulent strain of Phytophthora capsici (coded PHC 6/16; AGROINNOVA culture collection), originally isolated from sweet pepper, was cultured on selective oomycete medium (Masago et al., 1977) at 20°C for 1 week. Zoospores were produced according to a modified protocol of that described by Kim et al., (1997). One mycelium/agar plug (5 mm in diam.), taken from an actively growing colony, was transferred to a 1000 mL capacity flask containing a wheat-hempseed medium (200 g wheat kernels, 100 g hempseeds, 320 mL water, sterilized at 121°C for 30 min), and then incubated at 22°C in a growth chamber under continuous light for 7 d, followed by 3 d at 15°C in darkness. Zoospore suspension was prepared by mixing 80 g of the oomycete biomass in 1 L of distilled water, mixing the suspension for 10 min, then removing the aqueous extract from solid sediment by filtering through two layers of cheesecloth and vigorously mixing it. Zoospores were released by chilling the liquid cultures at 4°C for 1 h, followed by 1 h at room temperature (25°C). The zoospore concentration was adjusted to 1 × 10 5 zoospores mL -1 using a haemocytometer. A 10 mL aliquot of the zoospore suspension was then pipetted onto the peat media around the base of each plant.
In trials 1 to 6, the first treatment with salts was carried out 48 h before infestation of the peat substrate, while in trials 7 to 12, the first treatment with BCAs was carried out 48-72 h before infestation of the peat substrate (Table 1).

Disease assessments, and statistical analyses
The sweet pepper plants were assessed starting from the first appearance of wilt caused by P. capsici root and crown infections. Disease incidence (DI) was evaluated at intervals of 3 to 7, d by counting and removing the dead plants with symptoms of Phytophthora root, crown and stem rot. At the final assessment, the remaining plants were removed by each pot and the total number of dead plants including those wilted, due to severe root Areas under the disease-progress curves (AUDPCs) were calculated using the formula of Shaner and Finney (1977), for a total number of observations per trial of three to five disease incidence assessments. The biomass of healthy plants (fresh weight from 1-2 cm above soil surface) in each experimental unit (six pots), was measured at the end of trials 7 to 12 in Protocol III. The experimental units each consisted of a 3 L capacity pot with two plants (six pots), and sub-replicates contained 12 plants/treatment. Trials carried out using each of the three protocols were repeated at least three times (Tables 1 and 1S), and they were combined when the 'trial' factor was not statistically significant (P > 0.05).
Analysis of variance (ANOVA) with SPSS software (Version 26) was used to determine the effects of trial, treatments and their interactions, on disease incidence (DI) calculated at the end of each trial, AUDPC and fresh weight data. Prior to ANOVA, homogeneity of variances was evaluated and arcosin transformation of the percentage data was applied when necessary to normalize variances. When a statistically significant F test was obtained for treatments (P ≤ 0.05), the data were subjected to mean separation using Tukey's test at P ≤ 0.05.

Efficacy of resistance inducers (dosage, type of application) against Phytophthora crown rot
First symptoms of Phytophthora root and crown rot and sometimes black lesion on stems started to be visible between 4 to 7 d after inoculation during the trials carried out using Protocols I and II. These symptoms then developed quickly (average air temperature ranging from 24 to 28°C), with the final assessments carried out 27 to 33 d after planting (Protocol I), or 17 to 28 d after planting (Protocol II).
The DI and AUDPC data from each set of trials carried out according to Protocol I (Trials 1 to 3) or Protocol II (trials 4 to 6) were analyzed together for each experimental run (Tables 3 and 4), because statistically significant differences were not observed among trials (P > 0.05). The average incidence of Phytophthora root and crown rots was 46% in the untreated controls in trials 1, 2 and 3, and was greater (DI = 93%) in trials 4, 5 and 6 ( Table 3).
The most efficacious control of the pathogen was observed from K-phosphite at the standard dosage (80% efficacy) in all the experiments for both types of application (Table 3), while efficacy the half rate treatments also reduced percentage of dead plants (compared to the untreated controls) by 47%, when applied via the nutrient solution, or by 62% when applied directly to the soil substrate.
K-silicate alone reduced the percentage of dead plants with 20-23% efficacy. No improvement in disease control was observed when K-silicate was applied in combination with K-phosphite for either of the tested doses. Slightly less disease control was observed for K-phosphite, at the standard dosage, combined with K-silicate, when applied via nutrient solution (68% efficacy) or to the substrate (74% efficacy) (Table 3). Similar effects were observed for K-phosphite, at a reduced dosage, combined with K-silicate, when applied via nutrient solution (53% efficacy) or to the substrate (63% efficacy) (Table 3).
For AUDPC values, K-silicate reduced disease development, compared to the untreated control, for both types of application. K-phosphite, at both tested dosages, applied alone or combined with K-silicate, affected AUDPC compared to the untreated control, and differences among these treatments (Table 4).
K-phosphite and K-silicate either alone or combined, applied via nutrient solution or when distributed to the growing medium, did not reduce the development of sweet pepper plants (data not shown).

Effects of biocontrol agents
First symptoms of Phytophthora root and crown rots started to be visible at 5 to 13 d after inoculations with P. capsici carried out 48-72 h after planting during the tri-als carried out using Protocol III. These symptoms rapidly under the experimental conditions used (average air temperature 24 to 28°C), with the final assessment carried out between 34 to 37 d after planting trials 7-9, and 23 to 26 d in trials 10-12 carried out with Protocol III.
The data from trials 7, 8 and 9 (DI, df = 2, F = 0.369, P = 0.545; AUDPC, df = 2, F = 0.073, P = 0.930), and from trials 10, 11 and 12 (DI, df = 2, F = 2.252, P = 0.111; AUDPC, df = 2, F = 0.781, P = 0.461), were combined when there was heterogeneity between the trial runs (Tables 5 and 6). Inoculation with the pathogen led to high disease incidence in the untreated controls (78% in trials 7, 8 and 9 and 63% in trials 10, 11 and 12), permitting evaluation of the different BCAs under study (Table 5). All the BCAs reduced DI in the trials 7, 8 and 9 (df = 5, F = 18.021, P < 0.0001) and in trials 10, 11 and 12 (df = 5, F = 15.538, P < 0.0001), compared to the untreated controls. The experimental biocontrol agents generally gave more disease control than the commercial formulation of Trichoderma asperellum + T. gamsii Values from trials 1, 2 and 3 or trials 4, 5 or 6, each with five replicates per treatment, were combined when statistically significant differences were not observed (Trial P > 0.05). Significant according to the F tests and degrees of freedom (df) used in its calculation. Each mean is associated with its standard error (± SE). c Means in the same column, followed by the same letter, do not differ according to Tukey's Test (P ≤ 0.05). Each mean is associated with its standard error (± SE). d E%: percentage reduction of disease incidence of wilted and dead plants, compared to the untreated controls, at the end of trials 1, 2 or 3, corresponding to, respectively, 30, 25 or 24 d after inoculation (Protocol I), and at the end of trials 4, 5 or 6, corresponding to, respectively, 35, 21 or 17 d after inoculation (Protocol II). .525 g L -1 ) or their combinations, added directly to the nutrient solution NS (Protocol I), or to the substrate in each pot using 100 mL per pot of the suspension prepared (Protocol II). The treatments were carried out three times every 7 days. b AUDPC values were calculated for three to four assessments at 6-7 d intervals during trials 1, 2 and 3 (Protocol I), and four or five assessments at 3-6 d intervals during trials 4, 5 and 6 (Protocol II). c Values from trials 1, 2 and 3 or trials 4, 5 and 6, each with five replicates per treatment, are combined when statistically significant differences (trial P > 0.05; F tests) were not observed. d Means in each column followed by the same letter are not different (P ≤ 0.05; Tukey's Test). Each mean is associated with its standard error (± SE). Table 5. Mean Phytophthora crown rot incidences after applications of experimental BCA treatments, applied according to Protocol III, on for soilless grown sweet pepper 'Corno di Toro' . The data are average disease incidence (DI) at the end of trials 7, 8 and 9 and 10, 11 and 12. Disease reductions were observed for all the tested BCAs, and a similar trend was observed for the different disease amounts in trials 7, 8 and 9 (df = 5, F = 27.655, P < 0.0001) and trials 10, 11 and 12 (df = 5, F = 33.733, P < 0.0001). The least AUDPC values, compared to the non-treated controls, were from Fusarium solani FUS25, followed by the tested Pseudomonas strains and Trichoderma TW2 ( Table 6).
Applications of the BCAs to the growing media generally increased fresh plant biomass at the end of both sets of trials, compared to the untreated controls, although inconsistent results were observed for Pseudomonas PB26 and Trichoderma TW2 (Figure 1).

DISCUSSION
Phytophthora capsici is a broad host range pathogen (Granke et al., 2012), important in most sweet pepper growing areas. There are few chemical options for management of the diseases caused by this pathogen, due to the lack of effective soil fumigants, its resistance to some effective fungicides (Bower and Coffey, 1985;Hwang et al., 1996;Lamour and Hausbeck, 2001; Parra and Table 6. Mean areas under disease progress curves after applications of different experimental BCA treatments, applied according to Protocol III, for Phytophthora crown rot caused by P. capsici on soilless grown sweet peppers 'Corno di Toro' . The data are expressed as average area under disease progress (AUDPC) at the end of trials 7-9 and 10-12. a The BCA suspensions (1 ×10 7 cfu mL -1 ) were applied five times after planting at 7 d intervals to individual pots around the base of 15 d-old seedlings, using 100 ml of suspension per pot. The experimental BCAs were compared with a commercial formulation of Trichoderma asperellum + T. gamsii (10% a.i.; 'Remedier' , Isagro), applied at the label rate of 0.25 g L -1 of peat substrate. b AUDPC values were calculated considering four to five assessments at 4 to 9 d intervals during trials 7, 8 and 9, and four assessments at 4 to 8 d intervals during trials 10, 11 and 12. c Values from trials 7, 8 and 9 or trials 10, 11 and 12, each with five replicates per treatment, are combined when statistically significant differences (P > 0.05; F tests with and degrees of freedom (df) indicated. d Means each column accompanied by the same letter are not different (P ≤ 0.05; Tukey's Test). Each mean is accompanied by its standard error (± SE). , and trials 10, 11 and 12 (21 to 23 d after inoculation). a Mean fresh weight of plants per treatment from trials 7, 8 and 9 (df = 5; F = 30.446; P < 0.0001) or trials 10, 11 and 12 (df = 5; F = 34.384; P < 0.0001), each with five replicates per treatment, are combined because significant differences were not observed for trials 7, 8 and 9 (df = 2; F = 2.684 P = 0.055) or trials 10, 11 and 12 (df = 2; F = 1.701 P = 0.190), as indicated by F tests. b Means accompanied by the same letter do not differ (P ≤ 0.05; Tukey's Test). Each mean is accompanied by associated with its standard error (± SE). Ristaino, 2001;Tamietti and Valentino, 2001;Kousik et al., 2008;Miao et al., 2016;Barchenger et al., 2018;Hunter et al., 2018), and the complexity for host breeding to select cultivars resistant to local pathogen isolates (Acquadro et al., 2020). Soilless crop cultivation is expanding throughout the world, but by itself this approach cannot solve the problems of managing soil-borne pathogens (Jenkins and Averre, 1983), since many of them, including P. capsici, may be introduced through infected planting material (Granke et al., 2012) or irrigation water (Ristaino et al., 1993;Hong and Moorman, 2005;Gevens et al., 2007). Growth and spread of zoosporic pathogens are particularly severe problems in soilless systems (Stanghellini and Rasmussen, 1994). Management strategies that minimize inoculum dispersal have, therefore, considerable potential for disease reduction. Increased understanding of the impacts of crop management measures in soilless systems is needed for the sweet pepper-P. capsici pathosystem.
The present study examined the efficacy of resistance inducers, based on K-phosphite and K-silicate, used alone or in combinations, and of experimental biocontrol agents under conditions of high disease pressure, in a closed soilless system. K-silicate only partially reduced the incidence of P. capsici (20-23% efficacy) and its development, also when combined with K-phosphite, at the tested standard or reduced dosages. However, phosphite combinations could possibly reduce the risks of selection of Phytophthora isolates resistant to phosphite (Hunter et al., 2018), and address control of more than one pathogen in integrated pest management. Addition of 75 or 100 mg L -1 of silicon to the hydroponic nutrient solution reduced the severity of anthracnose caused by Colletotrichum capsici or C. gloeosporioides in Capsicum annuum (Jayawardana et al., 2015).
The mode of action of silicates and their functioning in several pathosystems are not yet fully understood. French-Monar et al., (2010) reported that bell pepper plants treated with calcium silicate had reduced root lesions and crown and stem necrosis caused by P. capsici, although at lower levels of reduction than found in other hosts. Together with possible formation of physical barriers to pathogens, from silicon accumulation in plants with differences depending on species, this element can influence plant defense responses and interact with key components of plant stress signalling systems, thus leading to induced resistance to pathogens (Liang et al., 2006;Wang et al., 2017).
Of the products tested in the present study, potassium phosphite provided the best control of P. capsici, when applied to pepper plants by adding soluble forms to nutrient solution of the closed hydroponic system, and when added to the growing medium. No effects on pepper development were detected after addition of this compound. However, the possible phytotoxic effects of phosphite, applied in soilless systems via a nutrient solution, should be considered. These compounds may precipitate and accumulate, as has been described for several fertilizers (Sambo et al., 2019). For example, tomato and pepper plants treated with either commercial or technical formulations of phosphite in a soilless system exhibited reduced growth (Forster et al., 1998).
The results obtained in the present study are similar to those of Förster et al. (1998), who observed reduced symptoms caused by P. capsici on tomato and pepper plants grown in a greenhouse hydroponic system treated with phosphite. Several factors are associated with the inhibition of pathogen growth as a result of phosphite treatments. These include phosphite concentration, the nature of the salts, the acidification of the plant growth medium and the pathogen life cycle (Guest and Bompeix, 1990;Smillie et al., 1989;Khalil et al., 2009;Khalil and Alsanius, 2011;Sambo et al., 2019). The results obtained in the present study show that the concentration of applied K-phosphite affected control of Phytophthora root and crown rot of sweet pepper, with greatest control provided by the standard dosages here tested.
Phosphite may act directly on pathogens, by inhibiting their growth (Fenn andCoffey, 1984, Smillie et al., 1989;Grant et al., 1990;Smillie et al., 1989), and possibly by priming the host defense in several pathosystems during pre-infection or post-inoculation stages of the pathogens. Liu et al. (2016) reported that phosphite, at >5 μg mL -1 , had a direct effect on mycelium growth and zoospore production in the sweet pepper-P. capsici pathosystem. Moreover, this compound increased transcription of antioxidant enzyme genes, and those involved in ethylene and abscisic acid biosynthesis, which mediated control of the pathogen at a higher phosphite concentration (1 g L -1 ) .
Biocontrol agents may be worthwhile for disease management in soilless production systems (Paulitz, 1997;Postma, 2009;Vallance et al., 2011;Lee and Lee, 2015). In the past 20 years, several studies have demonstrated reduced disease and practical implementation (Lamichhane et al., 2017;Villeneuve, 2017;Barratt et al., 2018;Raymaekers et al., 2020). In the present study, the Pseudomonas putida isolate mixture (FC 7B, FC 8B, FC 9B), Pseudomonas sp. PB26, F. solani FUS25 and Trichoderma sp. TW2, introduced into the soilless system 48-72 h before the pathogen inoculation, followed by four applications at 7 d intervals, reduced P. capsici on pepper by 45 to 64%. This result indicates further research is warranted. The most consistent results in control of Phytophthora crown rot were provided by F. solani FUS25 (60-64%% efficacy). The same biocontrol strain sometimes provided good, but variable results (8-54% efficacy) on zucchini, when applied to growing medium immediately at inoculation with P. capsici and 5-6 d before planting . Although the mechanism of action of F. solani FUS25 is not known, the most likely strategy for its use should be as a protectant or in preventative treatments. Establishment of BCAs in host root systems can also vary according to the host. The root systems of sweet pepper have greater surface areas than root systems of zucchini, which instead develops roots with few branches.
The good results obtained in the present study add evidence for applying BCAs in soilless cultivation systems against diseases caused by oomycetes. Pseudomonas sp. is known to be effective in reducing cucumber root colonization by Pythium aphanidermatum (Moulin et al., 1994;Chatterton et al., 2004), and Pythium disease on cucumber grown in a closed rockwool system (Postma et al., 2000). Streptomyces griseoviride ('Mycostop') is effective against Pythium ultimum on cucumber (Wolfhechel and Funck Jensen, 1991), while Trichoderma virens ('Soilgard') and Gliocladium catenulatum ('Prestop') are active against Pythium aphanidermatum, the causes of the damping-off of cucumbers grown in rockwool (Punja and Yip, 2003).
Increased understanding of BCA modes of action is needed to achieve a widespread application of these agents. Disease suppression could be related to different mechanisms: including production of antibiotics, secondary metabolites, lytic enzymes, phytohormones, siderophores, volatiles, and induction of host resistance (Köhl et al., 2019). For example, pepper plants inoculated with Fusarium oxysporum f. sp. lycopersici developed local and systemic resistance against P. capsici (Silvar et al., 2009), and the foliar pathogen Botrytis cinerea (Dı´az et al., 2005). Endophytic Trichoderma isolates induced resistance in hot pepper to P. capsici (Bae et al., 2011). Pseudomonas induced motility inhibition of P. capsici zoospores (Zohara et al., 2016), which are the only mobile propagules found in recirculating nutrient solutions (Stanghellini et al., 1996). Iron competition was important in the antagonistic activity of Trichoderma asperellum against F. oxysporum f. sp. lycopersici of tomato grown in a soilless medium based on perlite (Segarra et al., 2010).
Little is known about the capability of biocontrol agents to suppress Phytophthora blight on peppers grown in soilless systems.
Selected strains of the Pseudomonas, Bacillus, and Trichoderma have long been known for ability to improve plant growth and induce host systemic resistance against diseases and pests in different ecosystems, including soilless systems (Paulitz, 1997;Domenech et al., 2006;Gravel et al., 2007;Berg, 2009;Lee and Lee, 2015.;Sambo et al., 2019). Several biocontrol agents introduced into hydroponically grown fruit and vegetables have provided positive effects on yields and quality of horticultural products. In the present study, significant reduction in Phytophthora crown rot observed after treatment with Fusarium solani FUS25 was confirmed, by increased biomass fresh weight, with similar or greater increases than those from the commercial mixture of Trichoderma asperellum + T. gamsii.
In general, if a biocontrol agent is given time to become well-established in plant growth media, before pathogens are introduced (often through planting material), it can prevent infections or can interfere with pathogen inoculum production that may spread throughout the systems (Fry, 1982). Pseudomonas chlororaphis, Bacillus cereus, and B. gladioli strains, applied in smallscale hydroponic units, suppressed root colonization of chrysanthemum by Pythium aphanidermatum when applied 14-7 d before pathogen inoculation, rather than at the same time as inoculation (Liu et al., 2007). Selected antagonistic Pseudomonas, Fusarium or Trichoderma strains, previously tested in pot trials in peat medium against Fusarium wilt agents of lettuce and wild rocket (Gilardi et al., 2019;Srinivasan et al., 2009), or in naturally infested soil in a zucchini-P. capsici pathosystem (Cucu et al., 2020), were here used in preventative treatments in a closed soilless system, before inoculation of peat substrate with P. capsici. Optimizing the conditions in the soilless environment to which biocontrol agents were introduced, resulted in improved disease management consistency.
The results obtained in this study provide evidence for using phosphite-based products and biocontrol agents against Phytophthora crown and root rot of pepper, grown in soilless systems. They also show that there is not one solution for the management of these diseases. Different options should be considered and adapted to the different situations, relying on good extension services. Soilless cultivation provides good opportunities for exploitation and practical application of new disease management tools, such as resistance inducers and biocontrol agents. These have been intensively studied in recent years, expanding the integrated disease management options for intensive vegetable production (Paulitz, 1997;Lamichhane et al., 2017;Messelink et al., 2020).