Selection of Trichoderma spp . isolates antagonistic to Rosellinia necatrix

Fifty-six bulk isolates of Trichoderma spp. from avocado (Persea americana Mill.), carnation (Dianthus caryophyllus L.), litchi (Litchi chinensis Sonn), rice (Oryza sativa L.) and sugar beet (Beta vulgaris L.) crops located in southern Spain were evaluated for antagonism against one isolate of Rosellinia necatrix Prill. Isolates of both types of fungi were tested in dual and cellophane culture. The origin, cultural characteristics, overgrowth sporulation and staining of growth medium were recorded. As a result, 21 Trichoderma bulk isolates were selected and their corresponding monoconidial isolates were evaluated as above. Next eight monoconidial Trichoderma isolates with the largest in vitro antagonism were selected, and they were additionally tested against nine representative isolates of R. necatrix from nine virulence groups. These were established after pathogenicity tests on 57 isolates of R. necatrix from diseased avocado orchards in southern Spain. These monoconidial Trichoderma isolates were considered as potential biological control agents with a high potential for effective control of white root rot of avocado. Additional key words: avocado white root rot; biocontrol; cellophane culture; dual culture.

Antibiotics produced by Trichoderma species have been shown to inhibit fungal pathogens.Trichoderma species are often able to suppress the growth of endogenous fungi on an agar medium, and this mechanism has been observed in T. virens suppressing Macrophomina phaseolina, which causes charcoal rot in a range of crops (Howell, 2003).Various authors have demonstrated that mycoparasitism is not the main mechanism by which Trichoderma controls fungal pathogens.
The aim of this study was to identify isolates of Trichoderma with efficient in vitro antagonism to isolates of R. necatrix obtained from diseased avocado orchards in southern Spain, and select these isolates as biological control agents (BCAs) against WRR.

Material and methods
The experiments of this work were carried out in the period 2001-2006.In vitro evaluation of Trichoderma spp.against R. necatrix A total of 56 bulk Trichoderma isolates were collected from avocado, carnation, garlic, litchi, rice and sugar beet crops in southern Spain (Table 1), and evaluated for in vitro antagonistic activity to R. necatrix isolate Rn 400.

Antagonism of Trichoderma spp. bulk isolates in dual culture
Experiment 1 Forty-eight bulks isolates of Trichoderma spp.were evaluated for in vitro antagonism to Rn 400 in dual cultures (Royse and Ries, 1978).Petri dishes (90-mm diam.) containing 20 mL of potato dextrose agar (PDA; Difco Laboratories, Detroit, MI, USA) were each inoculated with a 5-mm diam.mycelial disc of a 7-dayold culture of R. necatrix grown under chamber conditions (25°C in darkness).Three days later, Petri dishes were co-inoculated with a 5-mm diam.mycelial disc of a three-day-old culture of Trichoderma spp.isolate at a distance of 5 cm from the R. necatrix mycelial disc.R. necatrix colony growth was measured along two radii: R1 (between the sowing point and farthest point of the colony) and R2 (between the sowing point and the edge of the colony) from where R. necatrix and Trichoderma mycelia came into contact.Percentage of radial growth inhibition (%RGI) was calculated as %RGI = [(R1 -R2) / R1]*100 (Royse and Ries, 1978).Controls comprised R. necatrix cultures without Trichoderma; and were conducted in duplicate.The profusion of growth over opposite microorganisms (i.e., overgrowth) such as R. necatrix growth over Trichoderma or vice versa) was examined, as were sporulation and staining of growth medium.Thus, 25 bulk Trichoderma isolates with efficient antagonistic properties to R. necatrix were selected for further evaluation.

Experiment 2
The 25 selected Trichoderma isolates plus 7 new isolates were re-evaluated using the same dual culture techniques as used in experiment 1, except that the cultural characteristics were not recorded.Isolates were replicated 10 times.The 7 new isolates of Trichoderma from escape trees (i.e.healthy trees in an orchard affected by WRR) were included based on promising results from preliminary experiments in our laboratory.
The experiments were complete randomised designs and an analysis of variance (ANOVA) was applied to the average of arcsin-transformed data of 3 and 10 replicates per isolate for experiment 1 and 2, respectively.The isolate means were compared by Fisher's Least Significant Difference (LSD) test (p < 0.05) to separate the means (Steel and Torrie, 1985).All statistical

Antagonism of bulk Trichoderma isolates in cellophane culture
Bulk Trichoderma spp.isolates (32 in total = 25 from Experiment 1 + 7 new isolates from Experiment 2) with an additional isolate previously tested in our laboratory were evaluated in cellophane culture (Dennis and Webster, 1971) against one isolate of Rn 400.Sterile cellophane film was transferred to Petri dishes containing 20 mL PDA.Each Petri dish was inoculated with a 5-mm diam.mycelial disc from different isolates of Trichoderma, one isolate per Petri dish in the centre of the dishes and incubated in chamber conditions for 2 days.The cellophane f ilm was removed, and a 5-mm diam.mycelial disc of Rn 400 transferred to the growth medium at a rate of 1 disc per Petri dish.When control mycelia of R. necatrix (no previous Trichoderma exposure) covered the Petri dish, two colony diameters of all treatments were measured; isolates were replicated five times.Thus, 21 isolates with efficient antagonism in vitro to R. necatrix were finally selected for additional evaluation.
The experiment was a complete randomised design and an ANOVA was applied to the average data from five dishes per isolate.Isolate means were compared by Fisher's LSD test (p < 0.05) (Steel and Torrie, 1985).

Preparation of Trichoderma monoconidial isolates
In order to minimise variability in the antagonistic response observed with bulk isolates of Trichoderma, monoconidial isolates were obtained for evaluation in dual and cellophane culture.Selected Trichoderma bulk isolates were grown on PDA for 7 days in chamber conditions.Spores were then removed by scraping the colony surface with 5 mL of sterile deionised water, and 0.1 mL of the spore preparation transferred to Petri dishes containing water agar and was incubated for 1 day at room temperature (RT = 24°C).A single conidium per isolate was recovered using an optical microscope, and transferred to dishes containing PDA, except for isolate CH 304 for which two conidia were recovered (described as CH 304 • 1-mc and CH 304 • 2-mc).Monoconidial cultures were maintained by serial transfer of bulk mycelium on PDA and stored at 4°C for future study.

Selection of monoconidial isolates of Trichoderma in dual and cellophane culture
In two experiments, 22 monoconidial isolates of Trichoderma were evaluated against Rn 400 in dual and cellophane culture as described previously; there were five replicates per isolate.Eight Trichoderma isolates were then selected based on the degree of antagonism to Rn 400 as measured in the same way as «Antagonism of Trichoderma spp.bulk isolates in dual culture» section for dual culture and «Antagonism of bulk Trichoderma isolates in cellophane culture» section for cellophane culture.
Experiments were complete randomised designs and an ANOVA was applied to the average of arcsintransformed data (dual culture) and data of five replicates per isolate (cellophane culture).Isolate means were compared by Fisher's LSD test (p < 0.05) to separate the means.

In vitro antagonism of eight Trichoderma monoconidial isolates to nine isolates of R. necatrix
To obtain representative virulent isolates of R. necatrix, the pathogenicity of 57 isolates (Table 2) from infested avocado orchards of southern Spain were tested on 12-month-old avocado plants from germinated seeds of cv.Topa-Topa.The plants were grown in 1.15 L pots containing a Laura substrate consisting of peat, coconut fibre, and perlite at a ratio of 6:1:0.6 v/v/v, respectively.Inoculations were as described by Sztejnberg and Madar (1980).Specifically, 3.75 g of wheat seed per L of substrate colonised by R. necatrix isolates was added at different depths and distances from the plant stem.The experiment was conducted under greenhouse conditions (18-26°C and relative humidity (RH) of 56-88%).
Five replicate pots per treatment were used.Aerial symptoms were evaluated every three days on the following scale of 1-5: 1 = healthy plant; 2 = plant with first symptoms of wilt; 3 = plant wilted; 4 = plant wilted with first symptoms of leaf desiccation; and 5 = plant completely desiccated and dead.Data were calculated as the area under the disease progress curve (AUDPC) Selection of Trichoderma spp.isolates antagonistc to Rosellinia necatrix (Campbell and Madden, 1990).These AUDPC data were statistically analysed as a complete randomised design comparing the means by Fisher's LSD test (p < 0.05).Nine bulk representative Rosellinia necatrix isolates, each from a different virulence group, were selected.The activity of eight isolates of Trichoderma, selected for their antagonism to R. necatrix, was again tested in cellophane and dual culture over the nine isolates of R. necatrix.Data describing in vitro antagonisms were statistically analysed using a factorial design where the main factor was the R. necatrix isolates and the sub-factors were the Trichoderma isolates.The means were compared by Fisher's LSD test (p < 0.05).

Antagonism of bulk Trichoderma isolates in cellophane culture
Of the 33 Trichoderma bulk isolates tested, 16 showed an antagonistic effect significantly different (p < 0.05) Thus, 21 bulk isolates of Trichoderma were selected based on high or medium %RGI in dual culture or cellophane culture, in combination with culture characteristics of overgrowth, staining and sporulation (Table 6).7).

Evaluation of monoconidal isolates of Trichoderma in cellophane culture
Total inhibition of Rn 400 growth occurred with Trichoderma isolates CH 252, CH 273, CH 303 and CH 316; inhibition was less with CH 296 although the difference was not significant (p < 0.05) (Table 7).Trichoderma isolates CH 256,CH 230,CH 242,CH 220,CH 222,CH 254,CH 251 and CH 215 had little effect on the growth of Rn 400.Isolates CH 304•1 and CH 304•2, from a same bulk isolate, had different effects in cellophane culture and variability was similar to that of the bulk isolates.
Hence, isolates CH 273, CH 296 and CH 303 were selected on the basis of high or total inhibition of Rn 400 growth in both cellophane and dual culture (Ta-ble 6).In addition, isolates CH 101, CH 304•1 and CH 314 were selected for a high %RGI in dual culture, and isolates CH 252 and CH 316 were selected for their inhibition in cellophane culture (Table 7).

In vitro evaluation of eight monoconidial Trichoderma isolates to nine R. necatrix isolates in different virulence groups
The study of pathogenicity of 57 isolates of R. necatrix was finished 24 days after inoculation when all control plants were dead.During the experiment, plants inoculated with the isolates Rn 12 and Rn 29 did not show symptoms of wilt, while all the remaining inoculated plants showed symptoms of wilt or were dead.Nine significantly different groups of virulence were established and the isolates Rn 320, Rn 400, Rn 10, Rn 17, Rn 50, Rn 33, Rn 30, Rn 49 and Rn 12 selected (Table 8).
The average effect of each monoconidial isolate of Trichoderma over the nine isolates of R. necatrix is shown in

Discussion
The presence of Trichoderma spp. in the root systems of plants that survive disease may be evidence of biological control.The best method for obtaining potential BCAs might be where candidate Trichoderma are isolated from plants and soils in situations where they are thought to function in disease control (Howell, 1998).It is often recommended that potential BCAs should be sought from healthy plants in fields of di-seased plants of the same species (Linderman et al., 1983).The antagonistic organisms would be expected to be able to function in the same environmental niche as the target pathogen (Knudsen et al., 1997).Therefore, isolates of Trichoderma with high antagonistic activity but from different crops might not be expected to be as effective in the rhizosphere of avocado trees.For our selection process, the origin of isolates from avocado was considered and the majority of the Trichoderma isolates selected were from the rhizosphere of healthy avocado trees adjacent to diseased avocado trees.In addition, some Trichoderma isolates were obtained from different hosts (i.e., carnation) where they had been proven effective in our laboratory for controlling other soil-borne fungi (i.e., Fusarium opxysporum f. sp.dianthi) (López Herrera et al., 2008).Elad et al. (1982), established a positive correlation between in vitro degradation of cell wall lytic activity  due to large number of T. harzianum isolates and the degree of biological control against the pathogens Sclerotium rolfsii, Rhizoctonia solani and Pythium aphanidermatum in vivo.Knudsen et al. (1997) also reported a positive correlation between fungi antagonistic isolates in vitro and biological control of Drechslera teres and Tilletia caries on cereals fields.The high variability observed within data in dual culture and cellophane experiments with the same bulk isolates of Trichoderma and R. necatrix, and also among replicates in the same experiment, was possibly associated with the bulk origin of the fungal isolates used.Bulk isolates constituted a pool of several genotypes with differing activity levels in essential aspects such as growth or sporulation, as well as in more specific factors, such as biocontrol.This high diversity may be related to activity of the different genotypes giving rise to synergisms between different isolates with variable genotypes and coexisting in bulk Trichoderma isolates (Harman et al., 1998).
Variability within strains may be due to differences in the degree of genetic variation within the genus Trichoderma (Hjeljord and Tronsmo, 1998).This may give strains within the genus Trichoderma a high The area under the disease progress curve (AUDPC), obtained from data of aerial symptoms, was evaluated every 3 days on the following scale of 1-5: 1: plant healthy; 2: plant with first symptoms of wilt; 3: plant wilted; 4: plant wilted with first symptoms of leaf desiccation; 5: plant completely desiccated and dead.Data were standardised means of five replicates, which values were compared by Fisher's protected LSD test (p < 0.05).LSD value = 0.489.In each column, numbers followed by the same letter are not significantly different according to the LSD test.
degree of ecological adaptability as reflected in its worldwide presence in soils under a range of differing environmental conditions (Harman et al., 1998).There is also intra-and inter-specific variability in the intensity of Trichoderma response against other microorganisms (Hjeljord and Tronsmo, 1998;Clarkson et al., 2004).Thalli of wild or successively transferred strains of Trichoderma are likely to comprise complex heterokaryons (i.e., individual nuclei may differ).Therefore, the thallus of the genus Trichoderma may be considered a complex community of nuclei, some differing subtly and others differing markedly from their neighbours (Harman et al., 1998).To reduce such variability, we prepared monoconidial isolates of Trichoderma with conidia receiving a single nucleus from the phiallide and hence being homokaryotic.The monoconidial generation comprises individuals that each possess a single genotype that may differ to a greater or lesser extent from that of other individuals of the same population (Harman et al., 1998).This variation may explain differences in the responses between bulk and the corresponding monoconidial isolates.Nevertheless, Worasatit et al. (1994) observed considerable variation among single spore isolates of T. koningii and their capacity to inhibit R. solani growth on agar.A similar variation occurred with isolates of T. harzianum inhibiting the growth of Gaeumannomices graminis var.triciti (Ghisalberti et al., 1990).Variability observed in our experiments did not disappear with mono-conidial isolates of Trichoderma, perhaps because R. necatrix is a dikaryotic homokaryon (Kanda et al., 2003) and a bulk isolate providing a new source of variation.
We observed that isolates with high %RGI in dual culture did not show the same response in cellophane culture (e.g., isolates CH 238 and CH 300).This may be attributable to the lack of evidence of stimulation of control activity due to a lack of direct contact between pathogen and antagonist in cellophane cultures.Kubicek et al. (2001) reported high-levels of induction of extracellular chitinolytic enzymes when Trichoderma was grown on purified chitin, fungal cell walls or mycelia as the exclusive source of carbon.Similar behaviour was reported by Inbar and Chet (1995) who studied the role of recognition in the induction of specific chitinases during mycoparasitism by T. harzianum of Sclerotium rolfsii.
In our dual culture experiments, RGI did not exceed 40%, which is in contrast to an RGI of 70% reported by Dubey et al. (2007) for Trichoderma spp.against Fusarium oxysporum f. sp.ciceris and by Royse and Ries (1978) for Cytospora cincta against Alternaria alternata, Epicoccum purpurascens, Coniothyrium olivaceum and Aureobasidium pullulans.This analysis of inhibition is based mainly on the competition for space in Petri dishes.For this reason, when two fungi with very different growth rates (e.g., R. necatrix and Trichoderma) are confronted, the %RGI is not high, but is still useful for selecting isolates with the potential for biocontrol.
We have identified isolates with significantly higher rates of sporulation, although high sporulation is a common characteristic of this genus (Gams and Bissett, 1998).High sporulation would favour rapid colonisation of substrate, and hence would be of valuable property as BCA due to high reproductive activity (Benítez et al., 2004).
We selected antagonistic isolates that overgrew the pathogen and rejected isolates that were overgrown by the pathogen.Haran et al. (1996) reported dual culture experiments in which T. harzianum was overgrown by R. solani but hardly overgrown by S. rolfsii under the same conditions, thus demonstrating differential chitinolytic activity.Similar results were obtained by Limón et al. (2004) for transformed isolates of T. harzianum 2413 that overgrew R. solani and prevented overgrowth of pathogen-antagonism.Nevertheless, Mukherjee and Raghu (1997) reported a direct relationship between overgrowth and BCA of Trichoderma spp.and S. rolfsii.
Although there may be a relationship between staining of growth medium and antibiosis (Rey et al., 2001), it was not always observed with our isolates.Isolates CH 300 and CH 221 showed considerable staining of growth medium, which did not correspond with antibiotic activity in cellophane cultures.This lack of corresponding activity may be due to specificity between antagonist and pathogen (Gams and Bissett, 1998), or indicative of its diversity and variability similar to Streptomyces spp.(Ndonde and Semu, 2001).
The high variability that exists in the virulence of isolates found in avocado orchards of southern Spain, corresponds with the high genetic diversity observed by Pérez Jiménez et al., (2002) and the existence of a somatic incompatibility system in isolates of R. necatrix of the same origin.The response differences in antagonism observed among the eight monoconidial Trichoderma isolates to nine R. necatrix isolates representative of different virulence groups in dual and cellophane cultures, suggests the existence of different antagonistic modes of action to pathogens.
In conclusion, nine groups of virulence were established for a collection of 57 isolates of R. necatrix, which demonstrates high variability among isolates of the pathogen infesting avocado orchards in southern Spain.The in vitro experiments with a high number of bulk isolates of Trichoderma from different hosts lead to the final selection of eight Trichoderma monoconidial isolates being tested over nine isolates of R. necatrix.These eight Trichoderma monoconidial isolates can be considered as BCAs with high potential for effective control of R. necatrix.These eight monoconidial isolates have been later evaluated in new experiments, not included in this study, as biocontrol agents against avocado white root rot, and have provided high levels of WRR control when two of them were tested singly or combined (Ruano-Rosa and López-Herrera, 2009).

Table 3 .
Selection of Trichoderma spp.isolates antagonistc to Rosellinia necatrix 1089 Effect of different Trichoderma bulk isolates on radial growth inhibition (%RGI) of R. necatrix (Rn 400).«Dual» culture technique; Experiment 1 a Least significant difference (LSD) (p < 0.05) = 9.79; Multiple comparisons between means are based on arcsin-transformed values.However, mean percentages are shown.The data were means of three replicates, which were compared by Fisher's protected LSD test (p < 0.05).In each column, numbers followed by the same letter are not significantly different according to the LSD test.b S: profuse sporulation.OT: overgrown of Trichoderma spp. with sporulation over R. necatrix.A: staining of growth medium, possible antibiosis.OR: overgrowth of R. necatrix over Trichoderma.-: without important characteristic.monstrated high inhibition in dual culture but not in cellophane culture.Variability in isolate responses was high but less so than in dual culture.

Table 5 .
Effect on growth of R. necatrix (Rn 400) of different Trichoderma spp bulk isolates in cellophane culture

Table 6 .
Level of antagonism to R. necatrix and cultural characteristics of selected Trichoderma bulk isolates a A: staining of growth medium, possible antibiosis; S: profuse sporulation; OT: overgrown of Trichoderma spp. with sporulation over R. necatrix; -: without important characteristics Avocado «escape»: healthy avocado tree with diseased neighbours.b

Table 7 .
Effect of different Trichoderma monoconidial isolates on radial growth inhibition (%RGI) of R. necatrix (Rn 400) in dual and cellophane culture b LSD (p < 0.05) = 2.04.The data were means of five replicates, which values were compared by Fisher's protected LSD test (p < 0.05).In each column, numbers followed by the same letter are not significantly different according to the LSD test.

Table 9 .
Mean effect of monoconidial Trichoderma isolates on nine representative R. necatrix isolates of virulence groups Nine significantly different groups of virulence: Rn 320, Rn 400, Rn 10, Rn 17, Rn 50, Rn 33, Rn 30, Rn 49 and Rn 12. b LSD (p < 0.05) = 5.01 Multiple comparisons means are based on arcsintransformed values.However, mean percentages values are shown.c LSD (p < 0.05) = 1.34.The data were means of three replicates, which values were compared by Fisher's protected LSD test (p < 0.05).In each column, numbers followed by the same letter are not significantly different according to LSD test.