Selection of Trichoderma stromaticum isolates for efficient biological control of witches’ broom disease in cacao

Biological Control 51 (2009) 130–139

Contents lists available at ScienceDirect

Biological Control journal homepage: www.elsevier.com/locate/ybcon

Selection of Trichoderma stromaticum isolates for efficient biological control of witches’ broom disease in cacao Leandro Lopes Loguercio a,*, Aítala Carvalho de Carvalho a,b, Givaldo Rocha Niella c, Jorge Teodoro De Souza d, Alan W. Villela Pomella b,1 a

Dept. de Ciências Biológicas (DCB), Universidade Estadual de Santa Cruz (UESC), Rod. BR 415, Km 16, Ilhéus-BA 45662-000, Brazil Almirante Cacau Ltda, Fazenda Almirante, BR 101 p/Barro Preto, Km 2, Cx. Postal 55, Itajuípe-BA 45630-000, Brazil c Centro de Pesquisas do Cacau (CEPEC/CEPLAC), Rod. BR 415, Km 22, Cx. Postal 07, Ilhéus-BA 45600-970, Brazil d Centro de Ciências Agrárias Biológicas e Ambientais (CCABA), Universidade Federal do Recôncavo da Bahia (UFRB), Centro, Cruz das Almas-BA 44380-000, Brazil b

a r t i c l e

i n f o

Article history: Received 5 February 2009 Accepted 10 June 2009 Available online 14 June 2009 Keywords: Moniliophthora Biocontrol Screening Theobroma cacao Antagonistic interaction Mycoparasitism Hemibiotrophic pathogen

a b s t r a c t Witches’ broom is the most devastating disease of cacao in Brazil, and losses to it entail serious socioeconomical and environmental problems. Biological control of the causal agent Moniliophthora perniciosa (Mp) using the antagonistic fungus Trichoderma stromaticum (Ts) is promising, although the identification of superior isolates is necessary. Here, we report a study on the selection of more effective Ts isolates based on field experiments. Sixty-three Ts isolates from a local collection were applied on brooms and placed under typical conditions of shaded-cacao plantations in southeastern Bahia State (Brazil), during two periods of three months each. The percentages of Ts sporulation and incidence and severity of Mp were the parameters used for biocontrol assessments. The results from both experiments were very distinct, indicating a high phenotypic variation in this collection and suggesting a significant effect of the environment in the Ts–Mp interaction. Ts-sporulation rates were negatively correlated with the presence of Mp in the brooms and a number of isolates reduced Mp incidence more efficiently than the reference isolate. Contrasting isolates in their efficiency of reducing Mp incidence were selected and further tested in four subsequent field trials for validation purposes. The results partially confirmed their biocontrol phenotypes but also suggested isolate-specific responses to environmental variations. Inhibition of Mpbasidiospore germination by total protein secreted in culture supernatants of Ts isolates correlated well with field results and revealed a potentially useful procedure for pre-screening of large collections towards selection of better biological control isolates. The characteristics and efficiency of the method as a reliable protocol for identification of superior BCAs in the witches’ broom—cacao pathosystem is discussed. Ó 2009 Elsevier Inc. All rights reserved.

1. Introduction The witches’ broom of cacao, caused by the fungus Moniliophthora perniciosa (Stahel) Aime & Phillips-Mora (ex Crinipellis—Aime and Philips-Mora, 2005), has become the most devastating disease on this crop in the Brazilian States of Amazonas, Bahia, and Espírito Santo. For the last 17 years, losses in cacao production in Bahia have averaged 60%, leading to severe economical, social and environmental problems (Donald, 2004; Griffith et al., 2003; Anderbrhan et al., 1999; Pereira et al., 1996). M. perniciosa (Mp) is a hemibiotrophic pathogen, whose basidiospores infect plant meri-

* Corresponding author. Address: Bio-Protection Research Centre, Lincoln University, P.O. Box 84, Canterbury 7647, New Zealand. Fax: +64 3 325 3864. E-mail addresses: [email protected], [email protected] (L.L. Loguercio). 1 Present address: Sementes Farroupilha, Rua Major Gote, No. 585, 8° Andar, Cx. Postal 90, Patos de Minas-MG 38702-054, Brazil. 1049-9644/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.biocontrol.2009.06.005

stematic tissues and develop intercellularly, causing loss of apical dominance, hypertrophy, and hyperplasia (biotrophic phase); later, tissue necrosis of the broom is induced, whose exploitation (saprophytic phase) leads to formation of basidiocarps, the source of the infectious propagules (Silva et al., 2002; Griffith et al., 2003; Meinhardt et al., 2008). Phytosanitation and protective fungicide applications are common practices to control Mp (Purdy and Schmidt, 1996; Costa et al., 2006). The use of copper-based fungicides has shown to be economically nonviable and technically ineffective in some cases (e.g. Krauss and Soberanis, 2002), whereas systemic fungicides are not routinely used in cacao farming, due to high costs and risks of contaminating cacao beans and the environment (Meinhardt et al., 2008). The replacement of diseased trees with genetically resistant clonal varieties is considered the most viable alternative (Lopes et al., 2003). However, this results in a delay for production to restart due to the long life cycle of cacao, and the possibility of

L.L. Loguercio et al. / Biological Control 51 (2009) 130–139

resistance breakdown threatens the long-term sustainability of this approach. Integrated pest management (IPM) is, therefore, a complementary and necessary strategy (Costa et al., 2006). Biological control is an important component of IPM programs as it helps both in reducing diseases to economically viable levels and decreasing the use of chemical fungicides (Krauss and Soberanis, 2001; Bajwa and Kogan, 2004). Increasing public concerns about pesticide residues in food products and the degrading effects of fungicides on soil biota that negatively impacts soil health (Van-Zwieten, 2004; Van-Zwieten et al., 2004), indicate that an environmentally sustainable method for disease control in cacao is necessary. In areas of organic certified production of cocoa with restrictions to use of chemicals, biologically based management appears to be one of the few viable approaches (Krauss et al., 2006). Antagonistic fungi such as Trichoderma spp. have shown significant success against plant diseases in several agricultural systems (reviewed by Benitez et al., 2004; Kubicek et al., 2001; Hjeljord and Tronsmo, 1998). Studies on the potential of Trichoderma-based biological control against pod diseases in cocoa-producing regions of Peru (Krauss and Soberanis, 2002), Cameroon (Tondje et al., 2007; Deberdt et al., 2008) and Brazil (Pomella et al., 2007; Hanada et al., 2008, 2009) indicate promise for this approach. The most likely mode of action of Trichoderma in these cases is parasitism of the pathogen. Significant diversity in species of Trichoderma has been described (Druzhinina and Kubicek, 2005; Samuels et al., 2006), thereby providing the possibility of finding distinct biocontrol mechanisms and host specificities (Hjeljord and Tronsmo, 1998; Lorito, 1998). In addition, antagonists naturally found in the areas of application tend to be more easily and widely acceptable by the public, with potentially faster benefits. Trichoderma stromaticum Samuels & Pardo-Schultheiss (Samuels et al., 2000) is a mycoparasite of Mp that is promising as a biological control agent (BCA) for the witches’ broom disease of cacao, as it colonizes the necrotic tissue of the plant and prevents basidiocarp production by the pathogen (Pomella et al., 2007; Hjorth et al., 2003; Bastos, 1996, 2000; Costa et al., 1996). However, the isolate currently used in the cacao-growing area of southeastern Bahia was introduced from the Amazon region and its field performance is greatly influenced by environmental factors (Santos, 2005; Sanogo et al., 2002). T. stromaticum (Ts) isolates are naturally widespread in Bahia and comprise two distinct genetic groups, I and II, with the latter showing a higher level of variability within the group, based on AFLP analysis (De Souza et al., 2006). Moreover, isolates from group II tend to persist longer as endophytes in cacao trees than those of group I (De Souza et al., 2008). Since the teleomorphic form of the genus, Hypocrea stromatica (Bezerra et al., 2003), was also identified in this cacao region, genotypic variability of Ts is thereby expected to be found. Hence, the search for locally adapted isolates that may be more efficient as BCAs against Mp is likely to be worthwhile (Costa et al., 2006). In addition, the ideal situation of achieving a ‘‘biological broom pruning” by treating the cacao canopy with Ts would avoid the need to remove dead brooms, one the most expensive practices on cocoa farming for control of witches’ broom (Pomella et al., 2007). Due to the complexity of variables involved in the phenotype of a BCA (Rajkumar et al., 2005; Elad, 2003; Hermosa et al., 2000; Knudsen et al., 1997), field experiments are more likely to render a methodology capable of selecting superior isolates because all genetic and environmental factors are considered in an integrative manner. Nevertheless, this strategy might be hampered if an excessive number of isolates are to be screened, due to potential space or time constraints, or resource limitation. In this study, 63 Ts isolates were assessed for their antagonistic capabilities against the witches’ broom pathogen, under usual field conditions of cacao farming in southeastern Bahia. The objectives were to establish a reliable strategy for identifying and selecting superior strains for

131

improvement of this method for disease control. Additional studies were also performed, to investigate the potential of using in vitro methods for a first round of isolates selection, prior to screening experiments in the field.

2. Materials and methods 2.1. Isolates and culture conditions Sixty-three isolates of T. stromaticum (Ts) were provided by Almirante Cacau Ltda (Itajuípe-BA, Brazil) for use in this study. They were previously collected in the tropical cacao-growing region of southeastern Bahia (Brazil) and classified into genetic groups I and II based on AFLP analysis (De Souza et al., 2006), as shown in Table 1. The abbreviations ‘GI’ and ‘GII’ are used throughout this text. Of the 63 isolates available, 54 were from GI and 9 from GII. An isolate designated ‘TVC’ (AM7, GII) was kindly provided by the Centro de Pesquisas do Cacau (CEPEC/CEPLAC). Because this isolate has been made available to producers since 1999 as the basis of the product ‘Tricovab’ for biological control of M. perniciosa (Mp), it was used as the Ts isolate of reference. Pieces of 0.25-cm2 filter paper containing spores from Ts isolates stored at 4 °C (Dhingra and Sinclair, 1985) were placed in Petri dishes containing potato-dextrose-agar medium for incubation at room temperature (25 °C) for 14 days until sporulation (Sanogo et al., 2002). The conidia that formed were collected and suspended in distilled water, and their concentrations adjusted with a haemocytometer to 107 sporesmL 1. These suspensions were applied to the dead Mp-infected cacao-stem segments (brooms) used for the biocontrol field experiments. For the experiments involving extraction of total secreted protein, the Ts spore-containing filter papers were transferred into liquid TLE medium (0.1% Bacto peptone, 0.03% urea, 0.2% KH2PO4, 0.14% (NH4)2SO4, 0.03% MgSO4, 0.03% CaCl2, and 0.1% trace elements, composed of 2.5% citric acid, 2.5% ZnSO4, 0.5% Fe(NH4)2(SO4)26H2O, 0.125% CuSO45H2O, 0.025% MnSO4, 0.025% NaNO3, 0.025% H3BO3). Two alternative carbon sources, 0.5% glucose or 0.5% dried mycelia of Mp + plus 0.03% glucose, were used. Ts isolates were cultured in 1-L flasks, containing 500 mL of medium, under 110 rpm constant agitation for 6 days at 25 °C, with the supernatants being processed afterwards as described below. 2.2. Broom collection and preparation Brooms of several lengths, averaging 0.75 cm in diameter, were collected from plants located in areas heavily infested with Mp. To select those for use in the experiments, brooms were tested for the presence of the pathogen and for incidental, spontaneous Ts colonization. Segments of 2 cm in length were cut from the ends of each broom and placed inside humid chambers of small plastic bags containing wet filter paper. These segments were incubated at 25 °C for 7 days; appearance of the characteristic white, cotton-like mycelia, and white-to-green sporulation were the indicators of Mp and Ts, respectively. Brooms were standardized to lengths of 20 cm and only those containing Mp and free of Ts were used. 2.3. Biological control field experiments Two screening experiments were conducted with the 63-isolates collection, the first extending from 9 May until 14 August, 2003, and the second from 31 July until 6 October, 2003. Rainfall and temperature data were collected from the station ‘591’ at the Almirante Cacau Farm. For these field trials, the 63 Ts isolates and the control treatment without Ts were applied on the 20

132

L.L. Loguercio et al. / Biological Control 51 (2009) 130–139

Table 1 Description of the T. stromaticum isolates used in this study. Isolate IDa

Local of originb

Sourcec

Genetic groupd

AM6 AM7* BA1 BA2 BA4 BA6 BA7 BA8 BA10 BA11 BA12 BA13 BA14 BA15 BA17 BA19 BA20 BA21 BA22 BA23 BA24 BA25 BA26 BA27 BA28 BA29 BA30 BA31 BA32 BA33 BA34 BA35 BA36 BA37 BA38 BA39 BA40 BA41 BA42 BA43 BA44 BA45 BA46 BA47 BA48 BA51 BA52 BA53 BA60 BA64 BA65 BA66 BA67 BA68 BA69 BA70

Medicilandia Belém Ubaitaba (plant 1) Ubaitaba (plant 1) Ubaitaba (plant 2) Aurelino Leal Ibirataia - Ipiaú Ituberá (plant 1) Ituberá (plant 2) Ituberá (plant 2) Ituberá (plant 3) Ituberá (plant 4) Ituberá (plant 5) Ituberá (plant 6) Ituberá (plant 7) Ubatã - Ibirapitanga (1) Ubatã - Ibirapitanga (1) Ubatã - Ibirapitanga (1) Ubatã - Ibirapitanga (2) Ibirapitanga Buerarema (plant 2) Buerarema (plant 1) Arataca Camacã (2) Camacã (2) Camacã (1) Itajuipe Itajuipe Itajuipe Itajuipe Inema-Ilhéus Inema-Ilhéus Inema-Ilhéus Inema-Ilhéus Inema-Ilhéus Coaraci Coaraci Coaraci Ibicaraí Ibicaraí Ibicaraí Ibicaraí Uruçuca (4) Uruçuca (1) Uruçuca (2) Uruçuca (3) Uruçuca (1) Uruçuca (1) Una Lomanto Junior Lomanto Junior Lomanto Junior SAP.4 NAP.25.12 Itabuna Mascote

– Broom Fallen broom 1 Fallen broom 2 Broom Broom Fallen broom Broom Broom 1 Broom 2 Fallen broom Broom Broom Broom Fallen broom Broom 1 Broom 2 Broom 3 Broom Broom Pod Fallen broom Pod Pod 1 Pod 2 – Pod Broom 1 Broom 2 Broom 3 Broom 1 Broom 2 Broom 3 Broom 4 Broom 5 Broom 1 Broom 2 Broom 3 Pod Broom 1 Broom 2 Broom 3 Broom Pod – Pod Branch cortex Pod seeds – – Broom T. grandiflora plant – – Broom –

II II I I I I I I I I I I I I I I I II II I I I I I I II I I I I I I I I I I I I I I I I I I I I II II I I II II I I I I

a

ID system based on previous study (De Souza et al., 2006); AM,isolate from the Amazon region of Brazil (State of Pará); BA, isolate from the State of Bahia. Seven other isolates tested in the present work (labeled ALF-540, -541, -557, -1092 to -1095) were not previously studied, therefore, they were not converted to the ‘BA/ AM’ system. *‘TVC’ isolate used in the marketed biocontrol product ‘Tricovab’. b Municipality names. Numbers in parenthesis indicate different locations in the same Municipality. When two names are presented in an entry, the first is from the District/Village and the second is from the Municipality. c Broom, dead branch/twig from the cacao-tree canopy; Fallen broom, dead broom from the ground; Pod, cacao fruit from the canopy; dashes indicate unknown source. Numbers on right indicate different positions in the same plant. T. grandiflora, Theobroma grandiflora. d De Souza et al. (2006).

cm-length brooms, which were hung at an average height of 1.50 ± 0.3 m (above ground, below the canopy) inside a cacao plantation at the Almirante Farm. The cacao trees were planted in rows

spaced 3  3 m. Spore suspensions of the Ts isolates were amended with 2% sucrose and 2% OPPAÒ (mineral oil—Petrobrás, Macaé-RJ, Brazil) for application onto the brooms by spraying until near run off. This formulation is standard for application of Tricovab (Pomella et al., 2007) and is known to not affect disease development (Pomella, unpublished data). Each Ts isolate was sprayed onto a set of 21 brooms, which were randomly distributed in the field in three groups of seven brooms to facilitate data recording. Brooms treated only with the formulation (2% sucrose + 2% OPPAÒ) served as controls. From the results of each field experiment, the Ts isolates were ranked according to their biological control activity, based on the levels of Mp incidence. Isolates with the same Mp incidence were further ranked based on Mp severity. After these biocontrol ranks were established, the resulting lists of 63 ordered isolates were equally divided into three categories or classes. Each category (‘high’, ‘medium’ and ‘low’) corresponded to a third of the isolates containing the highest, intermediate and lowest biocontrol-activity phenotypes, respectively. For the validation experiments, four selected isolates (BA4 and 8 from GI; BA29 and 66 from GII) were studied in subsequent years for their activity against Mp, in a cacao plantation located at an experimental area of CEPEC/CEPLAC. The TVC isolate (AM7, GII) used in the Tricovab product was included as a reference for comparison purposes. Two 3-month experiments and two 7-month experiments were performed. Five brooms sprayed with an individual Ts isolate comprised a treatment; treatments were distributed in a randomized-block design, with eight replicates (blocks) in a single experiment, totaling of 40 brooms per treatment. Controls were sprayed with distilled water. For both screening and validation experiments, only a singleapplication treatment was used for testing the Ts isolates. The observation and recording of Ts-sporulating brooms during the field trials were done two to three times a week. Afterwards, all the brooms (sporulated and non-sporulated) were brought to the laboratory, cut into six (for screening experiments) and four (for validation experiments) equal-length pieces and incubated in humid chambers at room temperature for 7–10 days. The biological variables assessed were the accumulated percentage of Ts sporulation at the end of the field trial and the residual presence of Mp after humid-chamber incubation. The latter was evaluated on the basis of the percentages of brooms (Mp incidence) and broom pieces (Mp severity) per treatment showing the white mycelia of Mp. During the field trial, the growth of Mp is typically not observable as it is the basidiocarp production, requiring incubation at humid-chamber in the lab for mycelia detection. 2.4. Inhibition of Mp-basidiospore germination For supernatant processing and extraction of total secreted protein, Ts isolates were grown for 6 days in TLE liquid media containing either glucose or dried Mp mycelia as carbon sources and the cultures were filtered using 3 MM paper to obtain the supernatant. The total protein secreted in 600-mL of supernatant was extracted on ice as previously described (Loguercio et al., 2002), with modifications. Ammonium sulphate for a final concentration of 75% (w/ v) was slowly added under gentle stirring over 40–50 min and keeping on ice for at least 1 h to complete precipitation. Each mixture was centrifuged at 30,500 g for 20 min in 50-mL screw-cap plastic tubes to obtain the total protein pellet. The supernatant was discarded and the pellet dried at room temperature (15 min). Pellets from a single isolate were pooled and resuspended in 2.5 mL of PBS buffer (pH 7.2), thereby allowing the protein content of a culture supernatant to be 240 times concentrated. To remove the remaining salt, the suspension was dialyzed overnight at 4 °C against 2 L of water. The dialysis membrane had a

L.L. Loguercio et al. / Biological Control 51 (2009) 130–139

133

cut-off of 4000–8000 Da. The protein samples were quantified by the dye-binding method of Bradford (1976). Each of five different quantities (0.1, 0.5, 1.0, 5.0 and 10.0 lg) of total protein, extracted from the culture supernatant of each Ts isolate, were mixed with 2  104 basidiospores of Mp and spread onto a Petri dish containing 2% agar media. After 4.5 h of incubation at room temperature, a drop of cotton-blue stain in 20% lactophenol was added to stop the Mp-basidiospore germination. The number of germinated spores was counted directly on the Petri dish, using a light microscope at 400 magnification. Spores were considered as germinated when the length of the germ tube exceeded the spore length. The germination percentages were the average of readings from three different Petri plates per treatment. 2.5. Statistical analyses The results from the two field screening trials were subjected to correlation-regression analysis with data pairs corresponding to Ts sporulation in the field as the independent variable and residual Mp presence (incidence and severity) after humid chamber as the dependent variable. For the correlation-regression analysis related to field and in vitro results, the above Ts sporulation, Mp incidence and severity results were the dependent variables, whereas the independent variables were the percentages of Mp-basidiospore germination after incubation with total secreted protein from Ts isolates. Results from the same 10 isolates were used in each regression analysis, being four isolates from the ‘high’, two from the ‘medium’ and four from the ‘low’ biocontrol-activity class. Statistical significance of regression was assessed by ANOVA. The results from the four subsequent validation experiments, with the five Ts isolates selected for comparison, were first tested for normal distribution of data by the Lilliefors test (Lilliefors, 1967), with the percentage data of Mp incidence previously transformed by the square root of ‘x + 1’ (Steel and Torry, 1980). Since some samples had the null hypothesis for normality rejected (p < 0.05) in each of the four experiments, the non-parametric Kruskall–Wallis test was employed separately for each experiment. When the H value was significant (p < 0.05), means comparison was performed by the Student–Newman–Keuls test (p < 0.10). The normal distribution of the Mp-basidiospore germination data was confirmed by the Lilliefors test, so ANOVA was performed with amounts of protein and biocontrol activity classes as the sources of variation (significance was considered at the p < 0.10 level). Since media with either carbon sources can be used to compare phenotypic classes of Ts, the statistical analysis was performed for each medium independently. For all the ANOVA analyses, means comparisons were performed by the Student’s ttest. All statistical analyses were performed using the open-access statistical software ‘BioEstat 5.0’ (Ayres et al., 2007).

3. Results 3.1. Assessment of M. perniciosa control and field sporulation of T. stromaticum A collection of 63 Ts isolates, most of which collected within the main cacao-producing region in southeastern Bahia (De Souza et al., 2006), was evaluated for antagonism against Mp. Two experiments were performed in which brooms were sprayed with the isolates in a single-application treatment and set above ground, under usual field conditions of cacao plantations (Fig. 1). The percentage of brooms showing Ts sporulation at the end of the field trial (Fig 1a) indicated colonization by the isolates, and the percentage of these brooms showing growth of the phytopathogen mycelia after a humid-chamber incubation (Fig 1b) indicated the

Fig. 1. Aspects of the two major parameters used to assess biocontrol efficiency of T. stromaticum isolates under usual conditions of cacao plantations in southeastern Bahia. (a) Aspect of T. stromaticum sporulation on brooms during the field tests (white bar = 2 cm); (b) Aspect of the residual white mycelium of M. perniciosa after humid-chamber incubation of brooms previously tested in the field (black bar = 2 cm); (c) levels of rainfall and temperature variation during field trials (indicated over the bars).

residual presence of Mp (incidence and severity), after the antagonistic interaction took place. The weather conditions were different for the two experiments. The first occurred in a period of a higher rainfall and higher maximum and minimum temperatures, which showed an overall descending pattern as the season progressed (Fig. 1c). The second was performed in a period of lower rainfall and lower temperatures, although the ascending period for temperature had started. It is noteworthy that in May, 52% of total rainfall (146.4 mm) occurred during 4 days at the end of the month (Fig. 1c). Since a decrease in the pathogen survival is expected after the antagonistic interaction in the field, the presence of Mp remaining in the brooms was used as a ranking parameter to identify superior BCAs. After ranking, three basic phenotypic categories resulted from equally dividing the total number of isolates, i.e. high, medium and low antagonistic activity (Table 2). In both experiments, the control treatment with no Ts application showed low average levels of spontaneous sporulation in the field (6%), and the residual incidence and severity of Mp was about 100% and 60%, respectively.

134

L.L. Loguercio et al. / Biological Control 51 (2009) 130–139

Table 2 Classification of 63 T. stromaticum isolates based on the biocontrol phenotypes observed in the two field experiments. Phenotypea

Residual M. perniciosa (%)

Exp. 1

Exp. 2

Exp. 1

Exp. 2

Exp. 1

Exp. 2

GI

G II

79.1 66.9 56.5 67.50

32.9 15.6 12.0 20.18

5.8 31.6 60.3 32.58

64.0 80.0 92.6 78.89

1.9 16.8 34.1 17.61

37.3 55.0 65.2 52.51

Alf-557; BA2, 8, 23, 11, 26, 32 BA69, 28, 40, 46 Alf-541,-1092; BA4, 43, 44, 51, 60

BA29 BA66 AM6, 7; BA53

Incidence

High Medium Low Total mean

Representative isolatesb

Ts field sporulation (%)

Severity

Italic values indicate the averages. a The isolates were ranked as a function of the Mp incidence and severity values observed after field trials; an independent ranking was performed for each experiment, with the resulting list of 63 ordered isolates being equally divided into the three ‘high’, ‘medium’ and ‘low’ biocontrol-activity classes indicated (see Section 2). Values on the table are the averages of all isolates belonging to the corresponding class. b Only the isolates that ranked consistently on the same third (phenotype) on the two experiments are indicated; G I and G II refers to the genetic groups described by AFLP analysis in a previous work (De Souza et al., 2006); out of the 63 isolates tested, 54 were G I and 9 were G II.

The split of 63 Ts isolates into three groups revealed a concentration of Mp incidences around an average value of 5.8% for the class with high biocontrol activity in the first experiment, whereas in the second, the average Mp incidence was 64% for the same class (Table 2). Isolates belonging to the genetic groups GI and GII (De Souza et al., 2006), were found in all three classes in both experiments. From a total of 1323 brooms evaluated in each experiment, 894 (67.5%) showed Ts sporulation in the first experiment, whereas only 267 (20.2%) sporulated in the second, which indicated a higher discriminative power among isolates for experiment 1. In terms of Mp incidences on a per isolate basis, an overview of observations for the 63 isolates showed remarkably contrasting outcomes from both experiments (Fig. 2). Although some isolates were consistently ranked, being classified in the same category in both experiments (Table 2), most changed their position among classes between experiments (i.e. ranked high in the first experiment and medium or low in the second, and vice-versa—Fig. 2). The extent of phenotypic variation among the tested Ts isolates was much higher under the conditions of the first experiment, with individual Mp incidence ranging from total absence (0%), found in 12 isolates (BA2, 8, 10, 23, 29, 32 35, 39, 47 and Alf-557, -1094, -1095), up to 83.3% (isolate ALF-541). For experiment 2, Mp incidence ranged from 41.6% to 100% (Fig. 2). In the second exper-

iment, a generally poorer biocontrol performance for the isolates was observed, as shown by an overall higher incidence and severity of Mp (Table 2, Fig. 2). In both experiments, very few brooms presented production of Mp basidiocarps, which occurred erratically and, for this reason, was not considered in our analyses. Upon inspection of the Ts sporulation and Mp incidence data from both experiments, a possible inverse relationship between these variables was noticed, so that a correlation–regression analysis was performed to test this hypothesis. A negative correlation between these variables was apparent (Fig. 3), although a high experimental variation was generally observed, as evidenced by the scattering of the plotted data pairs and the relatively low values of the coefficients of determination (R2). 3.2. Performance of selected T. stromaticum isolates from GI and GII in the field Aimed at verifying the consistency against Mp of Ts isolates screened under those two different environmental conditions (Figs. 1 and 2, Table 2), a series of similar biological control experiments were established in subsequent years, in another experimental area of shaded cacao plantation. In addition, as indicated by some preliminary work (Santos, 2005; Carvalho, 2006), the

Fig. 2. Comparative biocontrol activity (Mp incidence) for all 63 Ts isolates. Results from the first and second experiments are shown by black and grey bars, respectively. The order of isolates along the x-axis is presented based on the rank from the experiment 1, considering a decreasing antagonism activity (increasing Mp incidence). Results were based on the percentage of 21 brooms treated with each isolate that showed the growth of pathogen’s white mycelia in humid chamber, after field trials. The ‘star’ indicate the TVC reference isolate.

135

L.L. Loguercio et al. / Biological Control 51 (2009) 130–139

Fig. 3. Regression analysis between Ts sporulation in the field and remaining incidence and severity of Mp in the brooms. Data-pairs for all 63 Ts isolates were plotted. (a and b) Correspond to the first and second experiment, respectively. Based on an ANOVA for the regressions, all results were statistically significant (p < 0.05).

hypothesis that Ts isolates from GI tend to perform better as antagonists to Mp than those from GII was also investigated. Two contrasting representatives (i.e. from different biocontrol classes— Table 2) from GI, two from GII and the reference isolate were used for these validation experiments. Based on the Kruskall–Wallis test (p < 0.05), the isolates responded differently to each other, not only in different experiments (Table 3) but also from what was previously observed (Table 2). For instance, GI isolates BA8 and BA4 (high and low classes on the screening trials, respectively—Table 2) inverted their behaviour relatively to each other on experiments B and D (Table 3). A similar situation occurred for GII isolates BA29 and BA66 in experiments A, B and D, where no difference in biocontrol activity was detected (Table 3). An evaluation of the combined results from the four experiments, which accounted for a total period of more than 1.5 years in the field, indicated that the GI isolates were more effective in reducing survival of Mp in the hanging brooms than GII isolates (Fig. 4). 3.3. Effects of the total protein secreted by T. stromaticum in liquid culture on M. perniciosa-basidiospore germination As an attempt to associate field results with an in vitro parameter that could allow pre-screening of isolates to be further tested in field trials, the effects of total protein secreted

in the supernatant of Ts cultures upon germination of Mp basidiospores were assessed. Based on the ranking from the more discriminative screening experiment 1, Ts isolates representing the highest and lowest activity against Mp were compared (Fig. 5). Considering the effects of increasing amounts of total secreted protein on the decrease of Mp-basidiospore germination, ANOVA revealed statistically significant differences among protein amounts for both the high-(p < 0.05) and low-class (p < 0.10) series. The interaction ‘biocontrol class  protein amounts’ was not significant, either on media with glucose (p = 0.19) or with dried Mp mycelia (p = 0.67). Comparing the results between the contrasting biocontrol classes, inhibition of Mp-basidiospore germination correlated well with field results, as the group of isolates with highest biocontrol activity in the field also showed higher inhibition of basidiospore germination at all levels of total protein tested (Fig. 5). However, when assessing pairwise comparisons at different amounts of total protein, not all differences between the high and low classes in both types of carbon sources were statistically significant according to the t test (p < 0.10, Fig. 5). When assessing the effects of total secreted proteins of BA8 and BA4 isolates (GI) on Mp-basidiospore germination, the results were compatible to those of the high class (data not shown), thereby correlating well with results from the validation field experiments (Fig. 4).

Table 3 Mp incidencea of five selected Ts isolates after four biocontrol field experiments. Experim.

Control

Ts isolates G II

A B C D avg

9.8 ± 0.23 ab 9.0 ± 0.04 a 10.0 ± 0.00 a 9.2 ± 0.28 9.51

GI

TVC

BA29

BA66

BA8

BA4

9.6 ± 0.29 a 8.4 ± 0.73 a 9.4 ± 0.33 ab 7.4 ± 1.08 8.70

9.5 ± 0.21 a 7.6 ± 1.04 a 7.5 ± 1.00 c 7.5 ± 1.14 8.03

9.3 ± 0.36 a 7.8 ± 0.84 a 9.7 ± 0.19 abd 6.5 ± 1.14 8.33

5.3 ± 1.05 b 3.9 ± 0.98 b 5.9 ± 1.50 ce 7.4 ± 1.13 5.63

7.5 ± 0.63 b 3.2 ± 1.09 b 8.2 ± 1.07 bcde 5.2 ± 1.39 6.03

Italic values indicate the averages. a Values on the Table are the square root-transformed mean ± SE percentages of Mp incidence (see Section 2). The two extreme values of 10.0 and 3.2 in the table correspond to 100.0% and 9.2%, respectively. b Statistics performed by a non-parametric analysis of variance (Kruskall–Wallis test, p < 0.05) on a per experiment basis; H value was not significant in experiment ‘D’. Means with same letters do not significantly differ by the Student–Newman–Keuls test (p < 0.10).

136

L.L. Loguercio et al. / Biological Control 51 (2009) 130–139

Fig. 4. Summary of results from biocontrol experiments performed to validate the screening of Ts isolates. The average and standard error bars represent data from all four experiments combined (see Table 3 for results from individual experiments).

In order to validate the association observed between in vitro and field data, a correlation–regression analysis was performed. Results only for those amounts of total protein that showed a larger contrast between classes in terms of inhibitory effects on Mpbasidiospore germination (Fig. 5) are shown in Table 4. Despite both media tested showing differences between contrasting biocontrol classes (Fig. 5), ANOVA of the regression showed statistical significance (p < 0.10) only for those data-pairs in which supernatant proteins were secreted in medium with glucose (Table 4). Mp severity and Ts sporulation parameters showed a significant correlation with Mp-basidiospore germination for both protein amounts tested, whereas Mp incidence did so only for 10 lg of total protein (Table 4). 4. Discussion The establishment of a sufficiently reliable strategy for identifying superior biological control isolates must consider the local biotic and abiotic conditions where the BCA will be applied (Howell, 2003). An assessment of local T. stromaticum (Ts) germplasm was conducted in the most important cocoa-producing region of Brazil, the tropical southeastern Bahia State, with isolates characterized genetically and phenotypically (De Souza et al., 2006; Carvalho, 2006). Ts sporulation and Mp incidence and severity (Fig. 1) were reliable biological parameters for assessment of the Ts genotype– environment interaction and were useful in establishing biological control capabilities (Fig. 2, Table 2). Ts sporulation can be considered a measure of both the antagonistic capacity and environmental adaptability of the tested isolates, since it represents the completion of a life cycle at the expense of the mycoparasitized host. However, not all brooms colonized by Ts (a characteristic assessed by visual inspection of sporulation) became Mp-free, although most tend to be (Bastos, 2000; Hjorth et al., 2003). Since the remaining incidence of Mp in the brooms (Fig 1b) was a direct way to assess the final antagonistic results of Ts isolates, this parameter was considered to be more reliable as a basis to rank the 63 Ts isolates tested (Fig. 2). Although, the lack of assessable basidiocarp production in any experiment prevented a further indepth description of Ts isolates biocontrol capabilities, it is logic to assume that only those brooms in which Mp remained alive are the ones with potential to produce basidiospores.

Based on a separate rank per screening experiment, a significant variation within and between experiments for the same set of isolates was observed with regard to Ts–Mp antagonism (Fig. 2). These overall contrasting results suggest that the Ts response to the environment (Fig. 1c) is likely to be isolate-specific. The genetic variation of these isolates has already been established (De Souza et al., 2006) and the teleomorphs of Ts have been found in the collection area (Bezerra et al., 2003). Therefore, it is likely that such a variability within Ts populations provides enough opportunity for the occurrence of distinct Ts genotypes with specific responses to the environment (Costa et al., 2006), so that a change in their relative biocontrol potentials would not be unexpected. The differences in results between experiments also suggest that, in terms of survival inside brooms, Mp can cope better with its mycoparasite when weather conditions are more like those in experiment 2 (Figs. 1c and 2). The extent to which Ts spore production and biocontrol effects are inversely related can be seen from the negative correlation between these parameters in both experiments (Table 2, Fig. 3). This concurs with previous work showing that higher levels of Ts sporulation tend to yield lower levels of Mp incidences after field trials (Costa et al., 2006; Hjorth et al., 2003). In addition, the level of variation observed from the correlation–regression analysis (Fig. 3) suggests that the magnitude of this association varies according to isolate-specific interactions with the environment. From a screening perspective, Ts field-sporulation data could provide information earlier in the research process, perhaps serving as pre-screening method for larger collections of isolates. However, since the ‘high Ts sporulation/low Mp incidence’ association was not strictly observed for several isolates (Fig. 3), the assessment of both parameters simultaneously provides a more precise characterization of the isolates, without too much greater labor intensity. The results from the screening experiments were further explored by four subsequent validation experiments that included the same set of five Ts isolates, selected from both genetic groups and with contrasting biocontrol phenotypes. The tendency for varying results in Mp incidences among and within isolates (Table 3 and Fig. 2) confirmed the strong influence of a changing environment from one experiment to another (e.g. see Fig. 1c). Nevertheless, an analysis of the combined results from all four experiments,

137

L.L. Loguercio et al. / Biological Control 51 (2009) 130–139

totaling >1.5 years in the field (Fig. 4), strongly suggests that the isolates from GI tend to be more effective BCAs than GII isolates, even when considering a wider range of environments. Only in certain conditions, such as those of experiment C (Table 3), the GII-

Fig. 5. Inhibition of Mp basidiospores germination by total protein secreted by Ts isolates representing the contrasting biocontrol classes from experiment 1 (‘high’ and ‘low’). The error bars correspond to five replicates, i.e. five Ts representatives randomly chosen among the 21 isolates of a class. The protein secreted in the supernatants was obtained from cultures grown in TLE medium with two types of carbon sources (top and bottom graphs) for 6 days. Differences in Mp germination for increasing amounts of protein were statistically significant by ANOVA (p < 0.05) only within the high class series, in both carbon sources (not shown). Probability values (p) from t tests comparing high and low classes pairwise, within each protein amount, are shown above error bars. This experiment was repeated once with proteins extracted from repeated cultures for the isolates and showed similar results.

isolate BA29 showed a better performance than others from its group, not being significantly different than the GI isolates tested. Although this confirmed previous non-systematic observations (Pomella, unpublished data), the number of isolates tested here is rather small for a more general conclusion. These results also suggest that the overall conditions in which the screening experiment 1 was performed was exceptionally favorable to the antagonistic interaction Ts–Mp. The reference isolate TVC, for instance, showed a Mp incidence of 33.3% (Fig. 2), its best performance when compared to the screening experiment 2 (Mp incidence of 75.0%) and to all four validation experiments, whose average Mp incidences for TVC ranged from 91.2% in experiment A to 53.8% in experiment D (Table 3). A more detailed discussion about the screening/validation results is needed at this point. As stated above, the genetic and phenotypic variability among Ts isolates within a collection (Fig. 2) allowed the identification of Ts isolates with a better potential as a BCA (Table 2, Fig. 4). However, the magnitude of environmental influence in field trials suggests that an individual analysis of any given new Ts isolate may be misleading from one environment to another, if other reference isolates (‘controls’) are not simultaneously available for comparison. Class- or group-based analyses such as those performed in the screening (ranking and classification of isolates) and validation experiments (systematic long-term evaluation of the same set of isolates) seem to be required for effective characterisation of Ts phenotypes. Additional work indicates that assessment of isolates individually may only be possible if the responses to the variation in microclimatic factors are monitored, so that isolate-specific environmental requirements for better biocontrol results in the field can be gauged (Loguercio et al., 2009). Although advantages of using the reference isolate TVC as a BCA to reduce Mp-basidiocarp production were reported both on and above ground (Costa et al., 2006), conditions for the antagonism Ts–Mp are generally more suitable at the leaf-litter environment in cacao-growing areas of Bahia (Bastos, 2000; Santos, 2005). This helps to explain the overall poor performance of the TVC in our experiments (Fig. 2; Table 3). Further support to this is added by the current recommendation for Tricovab (TVC) use by CEPLAC, i.e. four applications upon canopy-removed infected brooms placed on the ground, every 30 days on the period of higher humidity (Costa et al., 2006). With the objective of screening for efficient ‘‘biological broom pruners” to reduce costs of disease management (Pomella et al., 2007), our experiments focused on brooms placed above ground, simulating those infected brooms located on the branches and twigs of the cacao tree. Under these circumstances, the lack of significance observed for TVC in relation to the control treatment in various different conditions (Table 3) indicates this isolate is not a good option for use directly on the cacao canopy,

Table 4 Analysis of correlationa between the field-related parameters and the inhibition of Mp-basidiospore germination by two amounts of total protein secreted in the culture supernatants (SN) of Ts isolates. Ts field-related parametersb

Carbon sources of Ts culture mediac Glucose

Mp-dried mycelia

5 lg SN prot. R Mp incidence Mp severity Ts sporulation

10 lg SN prot. p

0.50 0.59 0.56

0.145 0.071 0.089

R

5 lg SN prot. p

0.61 0.72 0.65

0.059 0.019 0.040

R

10 lg SN prot. p

0.16 0.14 0.11

0.665 0.703 0.761

R

p 0,40 0,35 0,42

0.255 0.327 0.221

a The correlation coefficients (R) between the parameters were estimated by the square-root of the linear regression coefficient (R2) obtained from plotted data pairs. The significance probability for each pairwise analysis (p) was calculated by ANOVA of the regression. b Corresponded to the values for the Ts isolates obtained from experiment no. 1. c TLE (also see Fig. 5).

138

L.L. Loguercio et al. / Biological Control 51 (2009) 130–139

at least under the conditions of a low number of applications as suggested by this study. Nevertheless, application strategies in the Ts–Mp-cacao system that combine different isolates with distinct canopy-microclimate responses (Loguercio et al., 2009) certainly warrant further investigation. Our results clearly suggest that it may be possible to find suitable isolates that can appropriately antagonize Mp specifically during periods of the year with lower levels of humidity (this work; Loguercio et al., 2009). Field evaluations usually require sufficiently large experimental areas, a relatively long time to collect the data, and particularly favorable climatic conditions for readily discriminatory results (Figs. 1c and 2). Depending on the number of isolates to be screened, achieving these requirements may not be feasible. Establishing a first-tier screening strategy with less time and costs, and with a high probability of identifying good biocontrol activities would, therefore, be highly advantageous. The predominant mycoparasitism between Ts and Mp (Bastos, 1996; Samuels et al., 2000) suggests that the protein fraction present in Ts culture supernatants may properly indicate superior biological control isolates, as many of the secreted enzymes of a mycoparasite show hydrolytic activities against the host’s cell-wall components (Benitez et al., 2004; Steyaert et al., 2003; Kubicek et al., 2001; De Marco et al., 2000; De Marco and Felix, 2002). Our results were compelling in suggesting that inhibition of Mp-basidiospore germination by total proteins from Ts culture supernatants can work as a fast in vitro method for preliminarily discriminating between ‘good’ and ‘bad’ antagonists (Fig. 5, Table 4). It is relevant to mention, though, that such a ‘pre-field’ screening approach could only be deemed as possible because the in vitro data was correlated to results of a field trial (experiment 1) with very favorable conditions for the Ts–Mp antagonism. Nevertheless, the generally low R coefficients obtained (Table 4) and the existence of few outliers (isolates from a biocontrol class that did not show the corresponding levels of inhibition of basidiospore germination—data not shown) suggest that experimental variance can be quite high. Despite that the use of more isolates (data-pairs) might have reduced the experimental variation, relying only upon this in vitro procedure would probably be insufficient for a completely effective identification of better biocontrol isolates (Howell, 2003). On the other hand, when screening a large collection of unknown isolates, this approach would help decreasing the number of isolates with biological control potential to be further tested in the field (Rajkumar et al., 2005; Knudsen et al., 1997). Interestingly, the results from glucose medium showed a better power of discrimination than those from Mp dried-mycelia medium (Fig. 5, Table 4) which had supposedly triggered secretion of mycoparasitism-related enzymes. A possible explanation for such a phenomenon might be related to the timing of protein secretion in culture (e.g. Chet et al., 1998), which is a question being currently addressed. Alternatively, a possible effect of other metabolites of a non-protein nature (Bastos, 2006) that might have been co-precipitated from supernatants cannot be ruled out. The identification of superior Ts BCAs in the witches’ broom context of tropical southeastern Bahia (Brazil) has proved to be a complex task, with a diverse array of factors and variables to consider. The necessity of including field trials in this research was stressed here, as such an approach reproduces the real conditions of biological control applications, accounting in a simultaneous and integrated manner for the uncontrolled environmental variables that interfere with the system (Howell, 2003). More precise and consistent interpretations of isolates behavior in the field could be achieved by a systematic evaluation of local changes in meteorological factors (Santos, 2005; Loguercio et al., 2009). In addition, attempts to produce spores from group I isolates on a large scale by using the same strategy employed for Tricovab (group II), i.e. fermentation on rice grains (Niella, 2005; Pomella

et al., 2007), were not successful (Medeiros, unpublished data), thereby directing part of our current research efforts into circumventing this problem. We hope these findings can aid in the development of efficient, environmentally sustainable and cost-effective strategies for cacao disease management, besides being possibly useful in other crop systems for identification of superior BCAs involving similar antagonistic interactions. Acknowledgments The authors are grateful to Mr. Marcus Montargil, Agnaldo Guimarães and Mrs. Lívia Santana dos Santos for great technical assistance in the experiments, to Dr. Luiz Roberto M. Pinto for statistical advices, and to Dr. Robert Hill for critical review of this manuscript. The research was funded by MARS-Almirante Cacau and PADCTCNPq (Brazil). A.C. de Carvalho was supported by fellowships from CNPq and FAPESB (Brazil). References Anderbrhan, T., Figueira, A., Yamada, M.M., Cascardo, J., Furtek, D.B., 1999. Molecular fingerprinting suggests two primary outbreaks of witches’ broom disease (Crinipellis perniciosa) of Theobroma cacao in Bahia, Brazil. European Journal of Plant Pathology 105, 167–175. Ayres, M., Ayres Jr., M., Ayres, D.L., dos Santos, A.A.S., 2007. BioEstat–Aplicações estatísticas nas áreas de ciências biomédicas. Fundação Mamirauá, Belém-PA, Brazil. Available from: . Bajwa, W.I., Kogan, M., 2004. Cultural practices: springboard to IPM. In: Koul, O., Dhaliwal, G.S., Cuperus, G.W. (Eds.), Integrated Pest Management: Potential, Constraints and Challenges. CABI Publishing, Wallingford-UK, pp. 21–38. Bastos, C.N., 1996. Mycoparasitic nature of the antagonism between Trichoderma viride and Crinipellis perniciosa. Fitopatologia Brasileira 21, 50–54. Bastos, C.N., 2000. Trichoderma stromaticum sp. nov. on the production of basidiomatas and on infections of shoots and cushion flowers of cocoa by Crinipellis perniciosa. Agrotrópica 12, 59–62 (portuguese). Bastos, C.N., 2006. Biological control of Crinipellis perniciosa by an isolate of Trichoderma sp. In: Akrofi, A.Y., Baah, F. (Eds.), INCOPED Fifth Proceedings of International Seminar on Cocoa Pests and Diseases. San Jose, Costa Rica, pp. 1–7. Benitez, T., Rincon, A.M., Limon, M.C., Codon, A.C., 2004. Biocontrol mechanisms of Trichoderma strains. International Microbiology 7, 249–260. Bezerra, J.L., Costa, J.C.B., Bastos, C.N., Faleiro, F.G., 2003. Hypocrea stromatica sp. Nov. Teleomorfo de Trichoderma stromaticum. Fitopatologia Brasileira 28, 408– 412. Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry 72, 248–254. Carvalho, A.C., 2006. Bioprospection of Trichoderma stromaticum isolates for the Biological Control of the witches’ broom disease of the cocoa tree. Masters Thesis. State University of Santa Cruz/UESC, Ilhéus-BA, Brazil, 82 p. Chet, I., Benhamou, N., Haran, S., 1998. Mycoparasitism and lytic enzymes. In: Harman, G.E., Kubicek, C.P. (Eds.), Trichoderma and Gliocladium—Enzymes, Biological Control and Commercial Applications, vol. 2. Taylor & Francis, London-UK, pp. 153–172. Costa, J.C.B., Bezerra, J.L., Cazorla, I.M., 1996. Controle biológico da vassoura-debruxa do cacueiro na Bahia com Trichoderma polysporum. Fitopatologia Brasileira 21 (Suppl), 397. Costa, J.C.B., Bezerra, J.L., Veloso, J.L.M., Niella, G.R., Bastos, C.N., 2006. Controle Biológico da vassoura-de-bruxa do cacaueiro. In: Venzon, M., Paula Júnior, T.J., Pallini, A. (Eds.), Tecnologias alternativas para o controle de pragas e doenças. EPAMIG, Viçosa-MG, Brazil, pp. 25–47. Deberdt, P., Mfegue, C.V., Tondje, P.R., Bon, M.C., Ducamp, M., Hurard, C., Begoude, B.A.D., Ndoumbe-Nkeng, M., Hebbar, K.P., Cilas, C., 2008. Impact of environmental factors, chemical fungicide and biological control on cacao pod production dynamics and black pod disease (Phytophthora megakarya) in Cameroon. Biological Control 44, 149–159. De Marco, J.L., Felix, C.R., 2002. Characterization of a protease produced by a Trichoderma harzianum which controls cocoa plant witches’ broom disease. BMC Biochemistry, 3, 3. Available from: . De Marco, J.L., Lima, L.H.C., Souza, M.V., Felix, C.R., 2000. A Trichoderma harzianum chitinase destroys the cell wall of the phytopathogen Crinipellis perniciosa, the causal agent of witches’ broom disease of cocoa. World Journal of Microbiology and Biotechnology 16, 383–386. De Souza, J.T., Pomella, A.W.V., Bowers, J., Pirovani, C.P., Loguercio, L.L., Hebbar, P., 2006. Genetic and biological diversity of Trichoderma stromaticum, a mycoparasite of the cacao witches’ broom pathogen. Phytopathology 96, 61–67. De Souza, J.T., Bailey, B.A., Pomella, A.W.V., Erbe, E.F., Murphy, C.A., Bae, H., Hebbar, K.P., 2008. Colonization of cacao seedlings by Trichoderma stromaticum, a mycoparasite of the witches’ broom pathogen, and its influence on plant growth and resistance. Biological Control 46, 36–45.

L.L. Loguercio et al. / Biological Control 51 (2009) 130–139 Dhingra, O.D., Sinclair, J.B., 1985. Basic Plant Pathology Methods. CRC Press, Boca Raton-USA. 355 p. Donald, P.F., 2004. Biodiversity impacts of some agricultural commodity production systems. Conservation Biology 18, 17–37. Druzhinina, I., Kubicek, C.P., 2005. Species concepts and biodiversity in Trichoderma and Hypocrea: from aggregate species to species clusters? Journal of Zhejiang University Science 6B, 100–112. Elad, Y., 2003. Biocontrol of foliar pathogens: mechanisms and application. Communications in Agriculture and Applied Biological Sciences 68, 17–24. Griffith, G.W., Nicholson, J., Nenninger, A., Birch, R.N., Hedger, J.N., 2003. Witches’ brooms and frosty pods: two major pathogens of cacao. New Zealand Journal of Botany 41, 423–435. Hanada, R.E., De Souza, J.T., Pomella, A.W.V., Hebbar, K.P., Pereira, J.O., Ismaiel, A., Samuels, G.J., 2008. Trichoderma martiale sp. Nov., a new endophyte from sapwood of Theobroma cacao and a potential agent of biological control. Mycological Research 112, 1335–1343. Hanada, R.E., Pomella, A.W.V., Soberanis, W., Loguercio, L.L., Pereira, J.O., 2009. Biocontrol potential of Trichoderma martiale (ALF 247) against black-pod disease (Phytophthora palmivora) of cacao. Biological Control 50, 143–149. Hermosa, M.R., Grondona, I., Iturriaga, E.A., Diaz-Minguez, J.M., Castro, C., Monte, E., Garcia-Acha, I., 2000. Molecular characterization and identification of biocontrol isolates of Trichoderma spp. Applied and Environmental Microbiology 66, 1890–1898. Hjeljord, L., Tronsmo, A., 1998. Trichoderma and Gliocladium in biological control: an overview. In: Harman, G.E., Kubicek, C.P. (Eds.), Trichoderma and Gliocladium— Enzymes, Biological Control and Commercial Applications, vol. 2. Taylor & Francis, London-UK, pp. 131–151. Hjorth, S., Pomella, A.W.V., Hockenhull, J., Hebbar, K.P., 2003. Biological control of witches’ broom disease (Crinipellis perniciosa) with the co-evolved fungus Trichoderma stromaticum: testing different delivery regimes. In: 14th International Cocoa Research Conference—Towards a Sustainable Cocoa Economy: What Strategies to this End? vol. 2. COPAL, Accra-Ghana, pp. 691– 697. Howell, C.R., 2003. Mechanisms employed by Trichoderma species in the biological control of plant diseases: the history and evolution of current concepts. Plant Disease 87, 4–10. Knudsen, I.M.B., Hockenhull, J., Funck-Jensen, D., Gerhardson, B., Hökeberg, M., Tahvonen, R., Teperi, E., Sundheim, L., Henriksen, B., 1997. Selection of biological control agents for controlling soil and seed-borne diseases in the field. European Journal of Plant Pathology 103, 775–784. Krauss, U., Ten Hoopen, G.M., Hidalgo, E., Martínez, A., Stirrup, T., Arroyo, C., García, J., Palacios, M., 2006. The effect of cane molasses amendment on biocontrol of frosty pod rot (Moniliophthora roreri) and black pod (Phytophthora spp.) of cocoa (Theobroma cacao) in Panama. Biological Control 39 (23), 2–239. Krauss, U., Soberanis, W., 2001. Biocontrol of cocoa pod diseases with mycoparasite mixtures. Biological Control 22, 149–158. Krauss, U., Soberanis, W., 2002. Effect of fertilization and biocontrol application frequency on cocoa pod diseases. Biological Control 24, 82–89. Kubicek, C.P., Mach, R.L., Peterbauer, C.K., Lorito, M., 2001. Trichoderma: from genes to biocontrol. Journal of Plant Pathology 83 (special issue), 11–23. Lopes, V.L., Monteiro, W.R., Pires, J.L., da Rocha, J.B., Pinto, L.R.M., 2003. On farm selection for witches’ broom resistance in Bahia, Brazil—a historical retrospective. In: 14th International Cocoa Research Conference—Towards a Sustainable Cocoa Economy: What Strategies to this End? vol. 2. COPAL, AccraGhana, pp. 1001–1006. Lilliefors, H., 1967. On the Kolmogorov–Smirnov test for normality with mean and variance unknown. Journal of the American Statistical Association 62, 399–402. Loguercio, L.L., Barreto, M.L., Rocha, T.L., Santos, C.G., Teixeira, F.F., Paiva, E., 2002. Combined analysis of supernatant-based feeding bioassays and PCR as a firsttier screening strategy for Vip-derived activities in Bacillus thuringiensis strains effective against tropical fall armyworm. Journal of Applied Microbiology 93, 269–277.

139

Loguercio, L.L., Santos, L.S., Niella, G.R., Miranda, R.A.C., de Souza, J.T., Collins, R.T., Pomella, A.W.V., 2009. Canopy-microclimate effects on the antagonism between Trichoderma stromaticum and Moniliophthora perniciosa in shaded cacao. Plant Pathology, in press. Lorito, M., 1998. Chitinolytic enzymes and their genes. In: Harman, G.E., Kubicek, C.P. (Eds.), Trichoderma and Gliocladium—Enzymes, Biological Control and Commercial Applications, vol. 2. Taylor & Francis, London-UK, pp. 73–99. Meinhardt, L.W., Rincones, J., Bailey, B.A., Aime, M.C., Griffith, G.W., Zhang, D., Pereira, G.A.G., 2008. Moniliophthora perniciosa, the causal agent of witches’ broom disease of cacao: what’s new from this old foe? Molecular Plant Pathology 9, 577–588. Niella, G.R., 2005. Produção massal de Trichoderma stromaticum para o controle da vassoura-de-bruxa do cacaueiro. In: 9° Simpósio de Controle Biológico—Anais. Entomological Society of Brazil, Recife-PE, p. 45. Pereira, J.L., Almeida, L.C.C., Santos, S.M., 1996. Witches’ broom disease of cocoa in Bahia: attempts at eradication and containment. Crop Protection 15, 743–752. Pomella, A.W.V., de Souza, J.T., Niella, G.R., Bateman, R.P., Hebbar, K.P., Loguercio, L.L., Lumsden, R.D., 2007. The use of Trichoderma stromaticum in the management of witches’ broom disease of cacao in Bahia State, Brazil. In: Vincent, C., Goettel, M., Lazarovits, G. (Eds.), Biological Control: A Global Perspective—Case Studies from Around the World. CABI Publishing, Wallingford-UK, pp. 210–217. Purdy, L.H., Schmidt, R.A., 1996. Status of cacao witches’ broom: Biology, epidemiology, and management. Annual Review of Phytopathology 34, 573– 594. Rajkumar, M., Lee, W.H., Lee, K.J., 2005. Screening of bacterial antagonists for biological control of Phytophthora blight of pepper. Journal of Basic Microbiology 45, 55–63. Samuels, G.J., Pardo-Schultheiss, R., Hebbar, K.P., Lumsden, R.D., Bastos, C.N., Costa, J.C., Bezerra, J.L., 2000. Trichoderma stromaticum sp. Nov., a parasite of the cacao witches broom pathogen. Mycological Research 104, 760–764. Samuels, G.J., Suarez, C., Solis, K., Holmes, K.A., Thomas, S.E., Ismaiel, A., Evans, H.C., 2006. Trichoderma theobromicola and T. Paucisporum: two new species isolated from cacao in South America. Mycological Research 110, 381–392. Sanogo, S., Pomella, A., Hebbar, K.P., Bailey, B., Costa, J.C.B., Samuels, G.J., Lumsden, R.D., 2002. Production and germination of conidia of Trichoderma stromaticum, a mycoparasite of Crinipellis perniciosa on cacao. Phytopathology 92, 1032– 1037. Santos, L.S., 2005. Antagonistic interaction Trichoderma stromaticum—Crinipellis perniciosa: climate effects in the biocontrol of cacao witches’ broom and development of microssatelite markers. Masters Thesis. State University of Santa Cruz/UESC, Ilhéus-BA, Brazil. 72 p. Silva, S.D.V.M., Luz, E.D.M.N., Almeida, O.C., Gramacho, K.P., Bezerra, J.L., 2002. Redescription of the symptomatology caused by Crinipellis perniciosa in cacao. Agrotrópica 14, 1–24. portuguese. Steel, R.G., Torry, J.H., 1980. Principles and Procedures of Statistics, 2nd ed. McGrawHill Book Co, New York-USA. Steyaert, J.M., Ridgway, H.J., Elad, Y., Stewart, A., 2003. Genetic basis of mycoparasitism: a mechanism of biological control by species of Trichoderma. New Zealand Journal of Crop and Horticultural Sciences 31, 281–291. Tondje, P.R., Roberts, D.P., Bon, M.C., Widmer, T., Samuels, G.J., Ismaiel, A., Begoude, A.D., Tchana, T., Nyemb-Tshomb, E., Ndoumbe-Nkeng, M., Bateman, R., Fontem, D., Hebbar, K.P., 2007. Isolation and identification of mycoparasitic isolates of Trichoderma asperellum with potential for suppression of black pod disease of cacao in Cameroon. Biological Control 43, 202–212. Van-Zwieten, L., Merrington, G., Van-Zwieten, M., 2004. Review of impacts on soil biota caused by copper residues from fungicide application. In: SuperSoil 2004—3rd Australian New Zealand Soils Conference. University of Sydney, Australia, pp. 1–8. Van-Zwieten, L., 2004. Impact of pesticides on soil biota. In: Lines-Kelly, R. (Ed.), Soil biology in agriculture—Proceedings. Tamworth Sustainable Farming Training Centre, NSW Department of Primary Industries, Orange, pp. 72–79.