SEARCH

SEARCH BY CITATION

Keywords:

  • biological control;
  • durability;
  • experimental evolution;
  • fitness;
  • resistance;
  • selection

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The objective of this work was to estimate the risk of a decrease in the efficacy of biocontrol as a result of selection pressure exerted by biocontrol agents on Botrytis cinerea, focusing on pyrrolnitrin, an antibiotic identified in diverse biocontrol agents having an effect on B. cinerea. To evaluate a possible decrease in sensitivity to pyrrolnitrin, 10 successive generations of five isolates of B. cinerea were produced in vitro in the presence of a sublethal dose (10 μg L−1) of the antibiotic. For one isolate, a significant reduction in the sensitivity to pyrrolnitrin at the fifth generation was observed with a resistance factor of c. 11. The production of 10 additional generations for four of these isolates, with increasing doses of pyrrolnitrin (100–4000 μg L−1), resulted in the development of variants of B. cinerea with high levels of resistance to the antibiotic (RF > 1000) and a reduced sensitivity in vitro to a pyrrolnitrin-producing bacterium. Reverse adaptation of resistant variants after 10 additional generations in the absence of selection pressure was not observed, suggesting stability of the resistance. Comparison of the pyrrolnitrin-resistant generations and their sensitive parental isolates for mycelial growth, sporulation and aggressiveness on plant tissues revealed that the high level of resistance to pyrrolnitrin resulted in a high fitness cost. Mycelial growth was reduced between 1·7 to 3·6 times and sporulation reduced 3·8 to 6·6 times that of sensitive parental isolates. Similarly, aggressiveness was 7 to 10 and 3 to 10 times lower for resistant isolates on tomato and apple, respectively. This study provides evidence that a fungal plant pathogen is able to gradually build up resistance to an antibiotic produced by a biocontrol agent.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Botrytis cinerea (teleomorph Botryotinia fuckelianal) is the causal agent of grey mould, an economically important disease that generates losses on a wide range of crops (Jarvis, 1980). The control of this disease relies mainly on the use of fungicides, but the development of resistance of B. cinerea populations to the different families of fungicides complicates their use and threatens their effectiveness (Leroux, 2004). Biological control could be a good alternative or complement, and many biological control agents against B. cinerea have been described in past years (Elad & Stewart, 2004; Elmer & Reglinski, 2006). Known modes of action include hyperparasitism, induction of host-plant resistance, competition for nutrients and the production of inhibitory metabolites (Elad & Stewart, 2004).

The antibiotic pyrrolnitrin (3-chloro-4-(2′-nitro-3′-chlorophenyl)-pyrrole) is produced by various bacteria, such as Enterobacter agglomerans (Chernin et al., 1996), Serratia plymuthica (De Vleesschauwer & Höfte, 2007) and different species of Pseudomonas or related genera (de Souza & Raaijmakers, 2003) described as potential biological control agents against several plant pathogens. Several pyrrolnitrin-producing bacteria have been described as biological control agents against B. cinerea (Janisiewicz & Roitman, 1988; Chernin et al., 1996). Moreover, the antibiotic itself has been reported to inhibit the growth of bacteria (El-Banna & Winkelmann, 1998) and fungi (Levenfors et al., 2004) including B. cinerea (Janisiewicz & Roitman, 1988; Hammer et al., 1993; Schoonbeek et al., 2002). It has also received much attention in the medical field because of its activity against some opportunistic fungal or bacterial human pathogens, including the causal agent of tuberculosis (Biava et al., 2007).

Some recent studies have shown that B. cinerea can withstand a wide variety of fungitoxic compounds from different origins, including the principal families of fungicides (Leroux et al., 2002), the grapevine phytoalexin resveratrol (Schoonbeek et al., 2001) and the antibiotic 2,4-diacetylphloroglucinol (2,4-DAPG) produced by Pseudomonas spp. (Schouten et al., 2008). Thus, B. cinerea appears to be a very adaptable fungus. The multinucleate and possibly heterokaryotic nature of this fungus, the presence of transposons in the genome, the existence of virus-like particles and the possible occurrence of sexual reproduction (Beever & Weeds, 2004) may explain this high level of variability. These known biological properties of B. cinerea might, potentially, compromise the durability of biological control methods despite the commonly reported assumption that resistance of plant pathogens or pests to biological control agents will develop less frequently than that to chemical control methods (Holt & Hochberg, 1997; Duffy et al., 2003). Although information on the durability of biological control efficiency is scant, recent results concerning pest management in agricultural systems showed that this assumption may not always be justified (Asser-Kaiser et al., 2007). Moreover, inconsistent efficacy of biological control against plant pathogens still appears to be frequent in the field (Elad & Stewart, 2004). This fact is generally attributed to variable climatic conditions encountered in the field, the lack of ecological competence of the biological control agent and/or an unstable quality of the products (Elad & Stewart, 2004). However, a possible reduction in efficacy of a biological control agent could also arise if plant pathogens produce natural mutants with reduced susceptibility to the biological control agents used on crops. Knowledge on the potential development of resistance to biological control agents can thus help to devise or improve resistance management strategies. Buildup of field resistance to biological control agents could result from the selection of pre-existing plant pathogen isolates with low susceptibility in natural populations, or it could arise if plant pathogens have the ability to produce natural mutants with reduced susceptibility under the selection pressure of products used by farmers.

The main purpose of the present study was to assess the potential for buildup of resistance in B. cinerea isolates to pyrrolnitrin-producing biological control agents. To this end, successive generations were produced in vitro for five isolates of B. cinerea. The specific objectives were (i) to monitor these generations for possible changes in sensitivity to pyrrolnitrin and to a pyrrolnitrin-producing biological-control bacterium; and, for the resistant variants detected, (ii) to assess their fitness in comparison with the parental isolates and (iii) to evaluate the stability of the resistant phenotype by producing additional generations in the absence of selection pressure.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Isolates of Botrytis cinerea

Five single-spore isolates of B. cinerea originating from France were selected from a collection maintained in the laboratory. The choice was made on the basis of differences in aggressiveness to tomato plants and apple fruits and of differences in patterns of resistance to six fungicides representing four chemical groups. All selected isolates were sensitive to pyrrolnitrin (Table 1). For the duration of this work, they were maintained in stock cultures stored at −20°C in a 0·06 m phosphate buffer containing 20% (v/v) glycerol.

Table 1.   Characteristics, aggressiveness on tomato plants and apple fruits, and sensitivity to pyrrolnitrin and fungicides of Botrytis cinerea isolates used for in vitro conidial-transfer experiments
IsolateHostYear isolatedAggressivenessaPyrrolnitrinActive ingredients of fungicides (chemical family)b
Tomato plantApple fruitCarbendazim (benzimidazoles)Iprodione, procymidone, vinclozolin (dicarboximides)Fludioxonil (phenylpyrroles)Dicloran (aromatic hydrocarbons)
  1. a+++, high level of aggressiveness; ++, moderate level of aggressiveness; +, low level of aggressiveness; −, not aggressive in conditions of test.

  2. bConcentrations of fungicides used: carbendazim 1 μg mL−1, iprodione 2·5 μg mL−1, procymidone 10 μg mL−1, vinclozolin 5 μg mL−1, fludioxonil 0·1 μg mL−1, dicloran 25 μg mL−1.

  3. S, sensitive; R, resistant.

BC1Tomato1989++++++SSRSR
BC21Strawberry1991+++++SSSSS
BC25Tomato1991++SSSSS
BC26Tomato1991+++SSRSR
H6Tomato1991+SRSSS

In vitro conidial-transfer experiment

This experiment was carried out to produce 20 consecutive generations of spores on potato dextrose agar medium (PDA 39 g L−1; Difco Laboratory) amended with pyrrolnitrin (purified microbial pyrrolnitrin >95%, Sigma Chemical). The pyrrolnitrin was dissolved in methanol and a stock solution containing 1 g L−1 was prepared and stored at 4°C in the dark to preserve its activity. Stored spore suspensions of the five isolates of B. cinerea were first inoculated on PDA medium without pyrrolnitrin. After 14 days of incubation at 21°C, spores were harvested from the plates in 2 mL phosphate-glycerol solution. This spore suspension represented the initial parental generation G0. To produce the next generation, this suspension was adjusted to 106 spores per mL using a haemocytometer and 40-μL aliquots were spread on PDA medium amended with pyrrolnitrin in Petri dishes at a final concentration of 10 μg L−1. After 14 days of incubation in the dark at 21°C, spores representing generation G1 were harvested in 2 mL phosphate-glycerol solution. The spore suspension was adjusted to 106 spores per mL and 40-μL aliquots were used to produce the next generation. The remaining suspension was stored at −20°C. From G0, 10 successive generations (G1 to G10) were produced on PDA medium in the presence of 10 μg pyrrolnitrin L−1 (Fig. 1). This concentration of pyrrolnitrin was determined, in a preliminary experiment described below, as the highest dose allowing sufficient conidial production by any of the five isolates of B. cinerea after 14 days of incubation at 21°C in the dark. Using the spores of generation G10, another 10 successive generations were produced in conditions of increasing selection pressure. The concentration of pyrrolnitrin in the PDA medium was progressively increased from 10 μg L−1 (for the production of the first 10 generations) to 4000 μg L−1 for the production of generation G20 (Fig. 1). Spores from alternate generations (G12, G14, G16, G18 and G20) were preserved at −20°C for further work. As a control, 20 successive generations were also produced on a pyrrolnitrin-free PDA medium.

image

Figure 1.  Presentation of the conidial-transfer procedure for Botrytis cinerea isolates. The concentration of pyrrolnitrin in μg L−1 is indicated in the plates for each generation. For the pyrrolnitrin generations, the PDA medium was amended with 10 μg pyrrolnitrin L−1 until the 10th generation, then with increasing doses of pyrrolnitrin, reaching 4000 μg L−1 at the 20th generation. Control generations were produced on PDA medium without pyrrolnitrin. Three independent conidial-transfer procedures (lineages 1, 2 and 3) were realized for each isolate. Each generation was transferred to a new plate every 14 days, except the 11th generation, which was incubated for 30 days.

Download figure to PowerPoint

For each of the five isolates, the whole experiment was carried out three times independently, providing three lineages of 20 generations produced under selection pressure (Fig. 1) and three independent ‘control’ lineages produced on pyrrolnitrin-free PDA. Control generations produced on unamended PDA and generations produced in presence of pyrrolnitrin were designated GnC and GnP, respectively, where n indicates the generation rank in the lineage.

Determination of pyrrolnitrin sensitivity

In a preliminary experiment, the effect of pyrrolnitrin on mycelial growth and spore production by the parental generation of each isolate was assessed. This information was needed to determine the dose of pyrrolnitrin to be used in the production of successive generations under selective pressure, as described above. In a second phase, sensitivity to pyrrolnitrin was tested on generations G1P, G2P, G3P, G4P, G5P, G7P, G10P, G12P and G20P for each of the 15 lineages generated under selection pressure. In addition, these tests were also conducted for generations G10C and G20C of the lineages produced on pyrrolnitrin-free medium for comparison.

The effect on mycelial growth was assessed on PDA medium containing different concentrations of pyrrolnitrin (0, 1, 2, 5, 10, 50, 100, 500, 1000, 2000, 4000, 6000, 8000 and 13 000 μg L−1). To carry out this test, the spores of a given generation (including parental generation G0) were first transferred to unamended PDA medium and the plates were incubated for 4 days at 21°C in the dark. Mycelial plugs (5 mm in diameter) were then excised from the periphery of the growing colonies and transferred to the centre of fresh PDA Petri plates (90 mm in diameter) containing the desired concentrations of pyrrolnitrin. The plates were incubated at 21°C in the dark and colony diameter was monitored for 4 days after inoculation. For each of the three lineages per isolate, three replicate plates were used for each dose of pyrrolnitrin and the whole experiment was carried out three times independently. The 50% effective concentration (EC50) was estimated as described by Leroux et al. (1999). A resistance factor (RF) was measured as the ratio of the average EC50 value of a given generation for a given isolate (mean for the three replicate lineages) over the EC50 value of the sensitive parent of the considered isolate (generation G0). Statistical analyses were performed separately for each isolate on the EC50 values, to test the hypothesis that the level of sensitivity to pyrrolnitrin remained identical for the successive generations. These analyses used the anova module of statistica software to test both a possible lineage effect and a generation effect. When a significant effect was observed, the multiple comparison test of Student–Newman–Keuls was used to compare the means.

The effect of pyrrolnitrin on spore production was assessed for the parental generation G0 of each isolate. Plates of PDA medium amended with different concentrations of pyrrolnitrin were each inoculated in the centre with a mycelial plug (5 mm in diameter) taken from the periphery of a 4-day-old colony. After 14 days of incubation at 21°C in the dark, the spores were scraped from the medium, suspended in water and spore concentration determined using a haemocytometer. Three replicate plates were inoculated for each dose of pyrrolnitrin and the whole experiment was carried out three times independently. The data were analysed with statistica software, using the three repetitions of the experiment as blocks and the individual plates as replications.

Stability of pyrrolnitrin resistance

To test the possibility of ‘reverse adaptation’ of resistant variants in the absence of selection pressure, resistant variants obtained at the 20th generation under selection pressure (G20P) for isolates BC1, BC25, BC26 and H6 were studied. For one lineage per isolate, spores of generation G20P were used to produce 10 additional generations on PDA medium without pyrrolnitrin (G21 to G30). An incubation time of 14 days was applied between each successive generation. The EC50 and RF values of the last generation (G30) were determined for each isolate as described above.

Interaction between B. cinerea and a pyrrolnitrin-producing bacterium

The antagonistic activity of a pyrrolnitrin-producing bacterium was tested in vitro against the parental generation G0 of four isolates of B. cinerea (BC1, BC25, BC26 and H6) and against some of their generations obtained under selection pressure (G10P and G20P). For this work, one lineage was used per isolate. Strain ChPhzS24 of Pseudomonas chlororaphis (Mazurier et al., 2009) was used, which was kindly provided by P. Lemanceau (INRA, Dijon, France). Confrontation tests were conducted on TSA medium (3 g tryptic soy broth L−1 and 15 g agar L−1) in 90-mm-diameter Petri plates. In each plate, the fungus (5-mm-diameter mycelial plug) and the bacterium (20-μL droplet at 107 CFU per mL) were placed 25 mm apart. In control plates, a droplet of sterile distilled water was used instead of the bacterial suspension. Three replicate plates were inoculated for each test and incubated at 21°C in the dark. The whole experiment was conducted three times independently. Mycelial growth (radius in mm) was measured daily for 5 days after inoculation. For each test, the inhibition of mycelial growth was assessed with an index computed as [1 − (radius of mycelial growth on control plates/radius of mycelial growth in presence of the bacterium)] × 100. Additionally, the width of the inhibition zone was measured 5 days after inoculation. The levels of sensitivity to the bacterium of the pyrrolnitrin-sensitive parents (G0) and the generations (G10P and G20P) were compared using the anova module of statistica software. For a given isolate, the inhibition of mycelial growth index and the width of the inhibition zone, averaged over the three independently conducted experiments, were used as elementary replicates. When a significant effect was observed, the multiple comparison test of Student–Newman–Keuls was used to compare the means.

Estimation of fitness cost in vitro

Different parameters of fitness were studied for the pyrrolnitrin-sensitive parental generation G0 and for several generations of isolates BC1, BC25, BC26 and H6 which presented reduced sensitivity to pyrrolnitrin. These fitness parameters were assessed on PDA medium and included spore germination, mycelial growth and spore production. Spore germination was first assessed after 24 h of incubation. It was estimated by examining 100 arbitrarily selected spores from each of three replicate plates. As all tests systematically resulted in the observation of 100% germination, an attempt was made to explore possible differences in the early stage of the germination process. To this end, germination was also quantified every hour during a 6-h incubation period at 21°C. The early kinetics of germination were assessed by computing an average hourly rate of germination for the period between the third and the sixth hour after inoculation. Mycelial growth and spore production were assessed as described above for the determination of pyrrolnitrin sensitivity, except that plates were placed in a growth chamber with a photoperiod of 14 h to hasten sporulation. Mycelial growth was assessed by measuring the diameter of colonies 3 days after inoculation and by estimating the daily radial growth rate between 1 and 4 days after inoculation. Sporulation was computed as spores produced per Petri plate after 14 days. The experiments were repeated three times, each with three replicates. Statistical analyses were performed separately for each isolate and each type of fitness parameter, to test the hypothesis that fitness remained identical for successive generations. In these analyses, the anova module of statistica software was used to test both a possible lineage effect and a generation effect. When a significant effect was observed, the multiple comparison test of Student–Newman–Keuls was used to compare the means.

Estimation of fitness cost in planta

To assess possible changes in aggressiveness, several generations were tested on apple fruits and tomato plants. For each generation of a given isolate, the aggressiveness of the three lineages was tested.

The aggressiveness of generations G0, G12P and G20P obtained with isolates BC1, BC25, BC26 and H6 was assessed on apple cv. Golden Delicious. Generation G20C, produced on pyrrolnitrin-free medium, was included in all tests for comparison. In addition, generation G10P of BC26 was also tested because of its resistance to pyrrolnitrin. In these tests, three 2-mm-wounds were inflicted on the surface of the fruit and immediately inoculated with 10 μL of a spore suspension of B. cinerea containing 106 spores per mL. Apples were incubated in a growth chamber at 21°C. The diameter of lesions was monitored daily for 7 days after inoculation.

Tests on tomato cv. Monalbo plants (INRA, Avignon, France) were conducted for generations G0, G12P and G20P of isolate BC1 and for generations G0, G10P, G12P and G20P of isolate BC26. Generation G20C, produced on pyrrolnitrin-free medium, was included in all tests for comparison. Two types of bioassays were used with two independent sets of plants. To test the aggressiveness of B. cinerea on the stem, leaves were removed from the plants leaving a 5-mm petiole stub on the stems and 10 μL spore suspension deposited on the petiole stub. A spore concentration of 106 spores per mL was used for isolate BC1, but 107 spores per mL was used for BC26 because isolate BC26 was known to have a low level of aggressiveness on tomato plants (Table 1),. The lengths of resulting stem lesions were monitored daily for 7 days after inoculation. The aggressiveness of generations G0, G10P, G12P, G20P and G20C of isolate BC26 was also tested on petioles. To this end, leaves were removed from the plants leaving a 50-mm petiole stub and 10 μL of spore suspension containing 107 spores per mL deposited on the wound. The length of lesions on petiole stubs was monitored daily for 7 days after inoculation. All plants were incubated in a growth chamber with a photoperiod of 14 h and maintained at 21°C with a relative humidity above 90%.

The experiments were all repeated three times independently per lineage, each with three replicate plants or fruits. To take into account the kinetics of disease development for each isolate, the audpc was computed as described by Decognet et al. (2009). Statistical analyses were performed separately for each isolate on the audpc values, to test the hypothesis that the level of aggressiveness remained identical for the different generations. In these analyses, the anova module of statistica software was used to test both a possible lineage effect and a generation effect. When a significant effect was observed, the multiple comparison test of Student–Newman–Keuls was used to compare the means.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Sensitivity to pyrrolnitrin of the wild-type parental isolates

The mycelial growth of the five isolates used in this study (parental generation G0) was highly affected by pyrrolnitrin, with estimated values of EC50 between 1 and 7 μg L−1 (Table 2). At 10 μg pyrrolnitrin L−1, the inhibition of mycelial growth varied between 64·8, 56·3, 81·9, 71·9 and 98·5% for BC1, BC21, BC25, BC26 and H6, respectively). At a dose of pyrrolnitrin of 100 μg L−1 or greater, the mycelial growth of all five isolates was completely inhibited (Fig. 2).

Table 2.   Level and stability of resistance to pyrrolnitrin of different generations produced from Botrytis cinerea isolates BC1, BC21, BC25, BC26 and H6. Generations G0, G7P, G10P, G12P and G20P are, respectively, the wild-type parental isolates and the seventh, 10th, 12th and 20th generations produced in the presence of pyrrolnitrin in the medium. For BC21, generations were not pursued after the seventh generation. Generation G30 is the 30th generation obtained from G20 by producing 10 consecutive generations on PDA medium without pyrrolnitrin
IsolateGeneration testedEC50 (μg L−1)aRFb
  1. aAverage values obtained with three replicate lineages ± standard error of the mean.

  2. bRF = resistance factor calculated by dividing the EC50 value of each of the generations tested by the EC50 value of the sensitive parent of the considered isolate (G0).

BC1G05·1 ± 0·1 
G10P9 ± 11·7
G12P8700 ± 11001700
G20P8700 ± 11001700
G308700 ± 11001700
BC21G07 ± 1 
G7P7 ± 11
BC25G05 ± 2 
G10P9 ± 11·8
G12P9000 ± 6001800
G20P9000 ± 6001800
G309000 ± 7001800
BC26G06 ± 1 
G10P63 ± 1311
G12P7200 ± 5001200
G20P7400 ± 4001230
G307400 ± 5001230
H6G01·3 ± 0·3 
G10P7 ± 15·4
G12P4600 ± 1003540
G20P4600 ± 1003540
G304600 ± 2003540
image

Figure 2.  Colony diameter (a) and spore production (b) estimated 4 and 14 days after inoculation, respectively, for parental generation G0 of five Botrytis cinerea isolates on PDA medium amended with different concentrations (0–500 μg L−1) of pyrrolnitrin. Each bar represents the average value for three independent repetitions. Error bars indicate the standard error of the mean.

Download figure to PowerPoint

Spore production of the five isolates was also significantly affected by pyrrolnitrin (anova, = 0·009 for BC1, = 0·011 for BC21, = 0·039 for BC25, < 0·0001 for BC26 and < 0·0001 for H6) (Fig. 2). At 50 μg pyrrolnitrin L−1, sporulation was completely inhibited for three of the isolates and for higher concentrations the inhibition was complete for all isolates (Fig. 2). A dose of 10 μg pyrrolnitrin L−1 resulted in a reduction of spore production of 93·9 ± 1·3% (average for the five isolates ± standard error) compared to the control without pyrrolnitrin. It allowed the production of 2·6 × 105 to 9·0 × 106 spores per 90-mm-diameter Petri plate, depending on the isolate (Fig. 2). This dose of pyrrolnitrin was thus selected for the production of successive generations under selection pressure.

Sensitivity to pyrrolnitrin of successive generations produced under selection pressure

For all isolates tested, there was no significant change in the level of resistance to pyrrolnitrin for the control generations G10C and G20C produced in the absence of pyrrolnitrin (anova, > 0·05). In contrast, changes in sensitivity were observed among the successive generations produced on PDA medium amended with pyrrolnitrin. The patterns of change were different among the five isolates. For isolate BC26, sensitivity to pyrrolnitrin was significantly reduced after only four successive generations (Fig. 3). The phenomenon was identical for each of the three independent replicate lineages of this isolate and no significant differences among these lineages were observed in any of the sensitivity tests conducted for generations G4P, G5P, G7P or G10P (anova, > 0·05). The estimated EC50 of this isolate remained similar between G5P and G10P, with a corresponding RF of c. 11 (Fig. 3, Table 2). For isolates BC1, BC25 and H6, the level of resistance was almost unchanged after 10 generations on 10 μg pyrrolnitrin L−1 (Fig. 3, Table 2). In contrast, the sporulation of all three replicate lineages of isolate BC21 was so reduced after seven generations on pyrrolnitrin-amended medium that it was impossible to produce a generation G8P with this isolate (Fig. 3). Work was thus pursued with only isolates BC1, BC25, BC26 and H6.

image

Figure 3.  Evolution of resistance to pyrrolnitrin for five isolates of Botrytis cinerea maintained for 20 successive generations on PDA amended with pyrrolnitrin. From parental generation G0, 10 successive generations (G1P–G10P) were produced in the presence of 10 μg pyrrolnitrin L−1. Using the spores of generation G10, another 10 successive generations were produced in conditions of increasing selection pressure. The concentration of pyrrolnitrin in the PDA medium was progressively increased from 100 μg L−1 for the production of the 11th generation to 4000 μg L−1 for the production of generation G20. The 50% effective concentration (EC50) was estimated for the generations G0–G5, G7, G10, G12 and G20. Values of EC50 correspond to the average of three independent lineages and error bars indicate the standard error of the mean.

Download figure to PowerPoint

For these four isolates, spores from the 10th generation (G10P) were transferred to fresh PDA containing a higher concentration of pyrrolnitrin (Fig. 1). All replicate lineages for each isolate were able to grow and sporulate on 100 μg pyrrolnitrin L−1, producing an 11th generation (G11P). This contrasted with the total inhibition observed with such a concentration for the parental generation G0 of all five isolates (Fig. 2). However, to obtain a sufficient number of spores for the production of the next generation, the incubation of generation G11P was extended from 14 to 30 days (Fig. 1). Following this step, the next generations (G12P to G20P) were all obtained after 14 days of incubation, despite the gradual increase in the concentration of pyrrolnitrin (Fig. 1). After a total of 20 successive generations, the last 10 of which were produced on increasing doses of pyrrolnitrin, high levels of resistance were obtained and the RFs were greater than 1000 for the four isolates (Table 2). As indicated by the small standard errors, similar values were obtained for the three lineages of each isolate (Table 2). The RF did not increase from the 12th to the 20th generation (Fig. 3). This pattern of change in resistance was similar for the four isolates, even if they differed significantly in the maximum EC50 values achieved after 12 generations (anova, = 0·005, Table 2).

Stability of pyrrolnitrin-resistant variants

Using pyrrolnitrin-resistant generation G20P to produce 10 additional generations on pyrrolnitrin-free medium did not lead to any decrease in sensitivity to the antibiotic for any of the isolates. In all cases, there was no significant difference between the EC50 values for generations G20P and G30 (Table 2).

In vitro interaction with pyrrolnitrin-producing bacteria

In vitro confrontations of pyrrolnitrin-producing strain ChPhzS24 of P. chlororaphis with parental generation G0 of any of the isolates resulted in a measurable inhibition zone and substantial growth inhibition of B. cinerea (Fig. 4; Table 3). Inhibition also occurred with generation G20P, but for all isolates it was significantly reduced in comparison with G0. For generations G10P, there was no significant difference in inhibition by P. chlororaphis in comparison with generation G0, even of isolate BC26, which showed moderate resistance to pyrrolnitrin (Table 3).

image

Figure 4. In vitro sensitivity of Botrytis cinerea isolate BC1 to the pyrrolnitrin-producing Pseudomonas chlororaphis strain ChPhzS24. The pyrrolnitrin-sensitive parent isolate (G0) and the pyrrolnitrin-resistant variant (G20P) were compared 5 days after inoculation. Control plates (a) and (b) show mycelial growth of G0 and G20P on tryptic soy agar medium without the bacterium. The effect of the bacterium on G0 and G20P is exemplified by plates (c) and (d).

Download figure to PowerPoint

Table 3. In vitro sensitivity of pyrrolnitrin-sensitive and -resistant variants of Botrytis cinerea to the pyrrolnitrin-producing bacterial strain Pseudomonas chlororaphis ChPhzS24. G0, G10P and G20P are, respectively, the wild-type parental isolates, the 10th generation and the 20th generation obtained in presence of pyrrolnitrin in the medium. Inhibition of mycelial growth index and width of the inhibition zone were estimated 5 days after inoculation. Data are means of three independent tests
Isolate of Botrytis cinereaGeneration testedPyrrolnitrin (RF)aEffect of Pseudomonas chlororaphis ChPhzS24
Growth inhibition (%)bInhibition zone (mm)
  1. aRF = resistance factor calculated by dividing the EC50 value of each of the generations tested by the EC50 value of the sensitive parent of the considered isolate (G0). For each isolate, means within a column followed by the same letter are not significantly different (anova, α = 0·05; Newman–Keuls test).

  2. bGrowth inhibition = [1 − (radius of mycelial growth on control plates in mm/radius of mycelial growth in presence of the bacteria in mm)] × 100.

BC1G0 74·8a18·3a
G10P1·772·6a19·3a
G20P170036·5b7·0b
BC25G0 66·7a19·5a
G10P1·863·7a20·0a
G20P18006·5b7·5b
BC26G0 74·6a16·6a
G10P1169·2a13·0ab
G20P123017·9b8·5b
H6G0 64·4a23·2a
G10P5·458·9a20·9a
G20P354034·3b3·5b

In vitro fitness cost for pyrrolnitrin-resistant variants

No significant differences were observed between the pyrrolnitrin-sensitive parents G0 and the resistant generation G20P for germination rate on PDA medium (anova, > 0·05, Table 4). There was no significant lineage effect for mycelial growth and sporulation on PDA medium of any of the isolates (> 0·05). On the other hand, for every isolate the mycelial growth of the resistant variants G12P and G20P was significantly reduced (P < 0·002) compared with the pyrrolnitrin-sensitive parent G0 and the control generation G20C (Table 4, Fig. 4a and b). Spore production of the pyrrolnitrin-resistant variants G12P and G20P was also significantly reduced (< 0·009) for every isolate compared with the pyrrolnitrin-sensitive parents G0 and the control generation G20C (Table 4).

Table 4.   Comparison of fitness between pyrrolnitrin-resistant generations, control generations and wild-type parental isolates (G0) of Botrytis cinerea. The different components of fitness evaluated were: hourly germination rate between the third and sixth hours of incubation (% per hour), colony diameter 3 days after inoculation (mm), daily radial growth rate between the first and fourth days after inoculation (mm per day), spore production on PDA medium 14 days after inoculation (number of spores produced per Petri plate) and aggressiveness on tomato plants and apple fruits
Isolate of Botrytis cinereaGeneration testedHourly germination ratea (% per h)MyceliumaSporulationaAggressiveness (audpc)a
Colony diameter (mm)Daily radial growth (mm day−1)×106 spores per plateTomato plantApple
Petiole stubStem
  1. aData are means of the three lineages. For each isolate, means within a column followed by the same letter were not significantly different (anova, α = 0·05; Newman–Keuls test).

  2. bG0 is the wild-type parental isolate, G20C is the 20th generation produced on PDA medium without pyrrolnitrin (control). G10P, G12P and G20P are, respectively, the 10th, 12th and 20th generations produced on PDA amended with pyrrolnitrin (10, 100 and 4000 μg L−1, respectively).

  3. c– = not measured or not relevant.

BC1G0b15·0a76·3a23·0a156a82·2a184·1a
G20Cc79·0a25·1a157a68·3a139·0a
G12P38·3b12·0b66b10·3b20·0b
G20P16·5a36·7b12·4b40b7·3b19·3b
BC25G012·3a67·3a22·3a130a126·7a
G20C66·7a23·4a98a124·1a
G12P41·0b13·5b47b54·3b
G20P9·2a40·3b14·6b34b49·1b
BC26G022·0a58·5a23·6a155a66·6a10·4a114·0a
G20C55·7a23·7a66b83·1a16·5a111·1a
G10P56·5a21·0a87·4a9·0a97·1b
G12P38·0b14·0b32c24·7c
G20P19·7a35·0b14·2b32c9·1b0·0b19·3c
H6G020·0a54·3a19·7a105a89·5a
G20C50·3a18·3a92a94·3a
G12P18·0b6·0b46b28·7b
G20P17·3a15·0b6·7b16c22·0b

Level of aggressiveness of pyrrolnitrin-resistant variants on plant tissues

There was no significant lineage effect for the aggressiveness of any of the isolates (> 0·05). For each of the four isolates tested, the pyrrolnitrin-resistant variants G12P and G20P were significantly less aggressive on apple fruits than the parental generation G0 and control generation G20C (Table 4). On tomato plants, a decrease of aggressiveness was also observed with generations G12P and G20P of isolates BC1 and BC26 compared with the parental generations G0 (Table 4).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The present study provides evidence for the first time that a fungal plant pathogen is able to gradually build up resistance to an antibiotic produced by a biological control agent. A high level of resistance to pyrrolnitrin appeared when B. cinerea was continuously exposed to the antibiotic pyrrolnitrin, indicating its capacity of adaptation in laboratory conditions. This adaptation appeared rapidly after the 10th generation following an increase in the concentration of pyrrolnitrin used in the medium. It was observed for all the isolates tested regardless of their original phenotypic differences, suggesting an absence of isolate effect in the buildup of resistance. It was not possible to assess in planta the efficacy of biocontrol by the pyrrolnitrin-producing bacterium P. chlororaphis ChPhzS24 against the pyrrolnitrin-resistant generations G20P, because of the loss of aggressiveness of these generations. Nevertheless, in vitro tests revealed that the resistance of B. cinerea to pyrrolnitrin was associated with a decrease in the antagonistic effect of the bacterium.

However, the small remaining inhibition zones observed around the bacterial colonies in confrontation tests suggest that the pyrrolnitrin-resistant variants were not completely resistant to the bacterium. A possible explanation for this phenomenon may be related to the concentration of pyrrolnitrin present in the vicinity of the bacterium in the medium. Although the quantity of pyrrolnitrin produced by strain ChPhzS24 of P. chlororaphis on PDA medium is not known, one may hypothesize that resulting concentrations could be greater than those used in the tests here to determine the level of resistance of B. cinerea to pyrrolnitrin in vitro. Another explanation could be that the variants of B. cinerea obtained on pyrrolnitrin-amended medium were probably not resistant to another antibiotic, phenazine-1-carboxylic acid, possibly also produced by strain ChPhzS24 of P. chlororaphis (Schoonbeek et al., 2002).

Spore production in the presence of a high concentration of pyrrolnitrin (100 μg L−1) was possible for the G10 variants of four isolates, whereas it was not possible for the sensitive parental generation G0. This suggests that isolates already exposed to pyrrolnitrin may be better prepared to adapt to pyrrolnitrin-rich environments than isolates that have never been exposed to this antibiotic. Luo & Schnabel (2008) proposed that the various levels of resistance to fungicides generated for different isolates of Monilinia fructicola may be linked to their original frequency of exposure to the fungicides under study. Therefore, this fungus may have reacted to its selective environment by accumulating some adaptive mutations, which may not necessarily include mutations having a direct consequence on resistance to the antibiotic (Rosenberg, 2001).

Previous studies have shown that the resistance of a fungus to a fungicide may predispose it to resistance to other fungicides and evidence of such predisposition has been documented for M. fructicola (Luo & Schnabel, 2008) and Venturia inaequalis (Koeller & Wilcox, 2001). The present study did not find a similar phenomenon for the resistance of B. cinerea to pyrrolnitrin. When B. cinerea isolates were subjected to a stable and relatively low dose of pyrrolnitrin before the 10th generation (10 μg L−1), susceptibility to the antibiotic significantly evolved only for one of the five isolates tested (BC26). For this isolate, no evidence of predisposition was established, as its fungicide-resistance pattern was not different from that of another isolate tested (BC1). Moreover, the speed of selection by pyrrolnitrin was not accelerated for the four other isolates tested, whatever their pattern of fungicide resistance. Therefore, it appears that the development of pyrrolnitrin resistance in B. cinerea is independent of its original profile of resistance to fungicides.

Fitness is a key parameter for the buildup of resistance in the field. In this study, the resistance of B. cinerea variants to pyrrolnitrin was systematically associated with reduced growth potential (mycelial growth and spore production) and decreased aggressiveness on plants or fruits (reduction of lesion growth). This fitness cost was similarly high for the four isolates tested. Based on these results, it is postulated that traits involved in resistance may be genetically linked to fitness cost and work is in progress to test this hypothesis. One practical consequence of this phenomenon is that if pyrrolnitrin resistance occurred in the field, the associated detrimental effect on the variant and competition from pyrrolnitrin-sensitive strains would limit the risk of complete loss of efficacy of pyrrolnitrin-producing biological control agents.

Several experimental studies have evaluated the relationship between fitness and fungicide resistance for B. cinerea, but results are quite contradictory to each other (Vignutelli et al., 2002; Moyano et al., 2004; Ziogas et al., 2005; Bardas et al., 2008). For example, Bardas et al. (2008) showed that the resistance of B. cinerea to the anilinopyrimidine fungicides did not affect their fitness compared to the sensitive isolates. However, the results of the present study are consistent with those reported by Ziogas et al. (2005) concerning the resistance of B. cinerea to fludioxonil, a fungicide whose structure is close to that of pyrrolnitrin (Leroux et al., 2002). A comparison of fitness among wild-type isolates and fludioxonil-resistant mutants of B. cinerea showed that all fludioxonil-resistant mutants had significant reductions in the characteristics that determine saprophytic fitness, such as mycelial growth, sporulation, conidial germination and sclerotial production (Ziogas et al., 2005). According to these authors, the polygenic control of resistance to fludioxonil may be responsible for the reduced fitness of fludioxonil-resistant isolates of B. cinerea obtained in the laboratory (Ziogas et al., 2005). Concerning biological control, experimental evidence of fitness costs associated with resistance has only been reported for pests. For instance, selection experiments showed that resistance of Drosophila melanogaster to parasitoids imposed fitness costs (Kraaijeveld & Godfray, 1997). The present study is apparently the first report of such a phenomenon for a plant pathogen.

The extent to which an organism is selected in a population depends on the advantages that the resistance provides but also on the costs of maintaining the resistance in the population in the absence of selection pressure. Ziogas et al. (2005) showed that fludioxonil-resistant variants of B. cinerea reverted to decreased resistance when they were grown in the absence of the fungicide. In contrast, in the present study, the level of resistance to pyrrolnitrin remained stable for 10 successive generations produced on pyrrolnitrin-free medium. This trait may constitute a risk factor in the field, if subsequent mutations could improve the fitness of the resistant variants, as observed for antibiotic-resistant bacterial mutants (Andersson & Levin, 1999). Furthermore, the genetic recombination associated with sexual reproduction may produce new phenotypes in the population. Population genetics studies have revealed high levels of variability in B. cinerea and a lack of linkage disequilibrium within populations from different countries, suggesting that this fungus may undergo sexual reproduction (Giraud et al., 1997; Karchani-Balma et al., 2008; Vaczy et al., 2008).

In practical terms, the development of reduced susceptibility to pyrrolnitrin in B. cinerea populations, if it should occur, may not necessarily jeopardize the efficacy of pyrrolnitrin-producing biocontrol agents. However, loss of efficacy of a biological control agent has already been described for B. cinerea: Li & Leifert (1994) showed that after 10 successive treatments on plants with the antibiotic-producing bacterium Bacillus subtilis, its efficacy against B. cinerea dropped dramatically. However, the biological control activity of a given microorganism may be associated with other mechanisms in addition to antibiosis, and it was suggested that many biological control agents suppress disease by using more than one mode of action (Elad & Stewart, 2004). Furthermore, it was suggested that resistance to complex or multiple antagonistic traits should occur only at a low frequency (Duffy et al., 2003). Nevertheless, in future studies on biological control it may be prudent to include work on the durability of efficacy of biological control agents having multiple modes of action.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank Claire Troulet, Gisèle Riqueau and François Orny for technical assistance in the plant and fruit tests and the technical and scientific assistance of Pierre Leroux, Anne Sophie Walker, Johan Confais, Christel Leyronas and Mohamed El Maataoui. This work was supported in part by the French National Research Agency (ECOSERRE project) and by a grant from INRA (SPE project). S.A. received a grant from the Syrian government to complete postgraduate studies.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Andersson DI, Levin BR, 1999. The biological cost of antibiotic resistance. Current Opinion in Microbiology 2, 48993.
  • Asser-Kaiser S, Fritsch E, Undorf-Spahn K et al. , 2007. Rapid emergence of baculovirus resistance in codling moth due to dominant, sex-linked inheritance. Science 317, 19168.
  • Bardas GA, Myresiotis CK, Karaoglanidis GS, 2008. Stability and fitness of anilinopyrimidine-resistant strains of Botrytis cinerea. Phytopathology 98, 44350.
  • Beever RE, Weeds PL, 2004. Taxonomy and genetic variation of Botrytis and Botryotinia. In: EladY, WilliamsonB, TudzynskiP, DelanN, eds. Botrytis: Biology, Pathology and Control. Dordrecht, the Netherlands: Kluwer Academic Publishers, 2952.
  • Biava M, Porretta GC, Poce G, Supino S, Sleiter G, 2007. New pyrroles with potential antimycobacterial, antifungal and selective COX-2 inhibiting activities. Synthetic methodologies. Current Organic Chemistry 11, 1092112.
  • Chernin L, Brandis A, Ismailov Z, Chet I, 1996. Pyrrolnitrin production by an Enterobacter agglomerans strain with a broad spectrum of antagonistic activity towards fungal and bacterial phytopathogens. Current Microbiology 32, 20812.
  • De Vleesschauwer D, Höfte M, 2007. Using Serratia plymuthica to control fungal pathogens of plants. CAB Reviews: Perspectives in Agriculture, Veterinary Science, Nutrition and Natural Resources 2, 112.
  • Decognet V, Bardin M, Trottin-Caudal Y, Nicot PC, 2009. Rapid change in the genetic diversity of Botrytis cinerea populations after the introduction of strains in a tomato glasshouse. Phytopathology 99, 18593.
  • Duffy B, Schouten A, Raaijmakers JM, 2003. Pathogen self-defense: mechanisms to counteract microbial antagonism. Annual Review of Phytopathology 41, 50138.
  • Elad Y, Stewart A, 2004. Microbial control of Botrytis spp. In: EladY, WilliamsonB, TudzynskiP, DelanN, eds. Botrytis: Biology, Pathology and Control. Dordrecht, the Netherlands: Kluwer Academic Publishers, 22341.
  • El-Banna N, Winkelmann G, 1998. Pyrrolnitrin from Burkholderia cepacia: antibiotic activity against fungi and novel activities against streptomycetes. Journal of Applied Microbiology 85, 6978.
  • Elmer PAG, Reglinski T, 2006. Biosuppression of Botrytis cinerea in grapes. Plant Pathology 55, 15577.
  • Giraud T, Levis C, Fortini D, Leroux P, Brygoo Y, 1997. RFLP markers show genetic recombination in Botryotinia fuckeliana (Botrytis cinerea) and transposable elements reveal two sympatric species. Molecular Biology and Evolution 14, 117785.
  • Hammer PE, Evensen KB, Janisiewicz WJ, 1993. Postharvest control of Botrytis cinerea on cut flowers with pyrrolnitrin. Plant Disease 77, 2836.
  • Holt RD, Hochberg ME, 1997. When is biological control evolutionarily stable (or is it)? Ecology 78, 167383.
  • Janisiewicz WJ, Roitman J, 1988. Biological control of blue mold and grey mold on apple and pear with Pseudomonas cepacia. Phytopathology 78, 1697700.
  • Jarvis WR, 1980. Epidemiology. In: Coley-SmithJR, VerhoeffK, JarvisWR, eds. The Biology of Botrytis. London, UK: Academic Press, 21950.
  • Karchani-Balma S, Gautier A, Raies A, Fournier E, 2008. Geography, plants, and growing systems shape the genetic structure of Tunisian Botrytis cinerea populations. Phytopathology 98, 12719.
  • Koeller W, Wilcox MH, 2001. Evidence for the predisposition of fungicide-resistant isolates of Venturia inaequalis to a preferential selection for resistance to other fungicides. Phytopathology 91, 77681.
  • Kraaijeveld AR, Godfray HCJ, 1997. Trade-off between parasitoid resistance and larval competitive ability in Drosophila melanogaster. Nature 389, 27880.
  • Leroux P, 2004. Chemical control of Botrytis and its resistance to chemical fungicides. In: EladY, WilliamsonB, TudzynskiP, DelenN, eds. Botrytis: Biology, Pathology and Control. Dordrecht, the Netherlands: Kluwer Academic Publishers, 195222.
  • Leroux P, Chapeland F, Desbrosses D, Gredt M, 1999. Patterns of cross-resistance to fungicides in Botryotinia fuckeliana (Botrytis cinerea) isolates from French vineyards. Crop Protection 18, 68797.
  • Leroux P, Fritz R, Debieu D et al. , 2002. Mechanisms of resistance to fungicides in field strains of Botrytis cinerea. Pest Management Science 58, 87688.
  • Levenfors JJ, Hedman R, Thaning C, Gerhardson B, Welch CJ, 2004. Broad-spectrum antifungal metabolites produced by the soil bacterium Serratia plymuthica A 153. Soil Biology & Biochemistry 36, 67785.
  • Li H, Leifert C, 1994. Development of resistance in Botryotinia fuckeliana (de Barry) Whetzel against the biological control agent Bacillus subtilis CL27. Zeitschrift für Pflanzenkrankheiten und Pflanzenschutz 101, 4148.
  • Luo CX, Schnabel G, 2008. Adaptation to fungicides in Monilinia fructicola isolates with different fungicide resistance phenotypes. Phytopathology 98, 2308.
  • Mazurier S, Corberand T, Lemanceau P, Raaijmakers JM, 2009. Phenazine antibiotics produced by fluorescent pseudomonads contribute to natural soil suppressiveness to Fusarium wilt. The ISME Journal 3, 97791.
  • Moyano C, Gomez V, Melgarejo P, 2004. Resistance to pyrimathanil and other fungicides in Botrytis cinerea populations collected on vegetable crops in spain. Journal of Phytopathology 152, 48490.
  • Rosenberg SM, 2001. Evolving responsability: adaptative mutation. Nature Reviews Genetics 2, 50415.
  • Schoonbeek H, Del Sorbo G, De Waard MA, 2001. The ABC transporter BcatrB affects the sensitivity of Botrytis cinerea to the phytoalexin resveratrol and the fungicide fenpiclonil. Molecular Plant-Microbe Interactions 14, 56271.
  • Schoonbeek H, Raaijmakers JM, De Waard MA, 2002. Fungal ABC transporters and microbial interactions in natural environments. Molecular Plant-Microbe Interactions 15, 116572.
  • Schouten A, Maksimova O, Cuesta-Arenas Y, Berg Gvd, Raaijmakers JM, 2008. Involvement of the ABC transporter BcAtrB and the laccase BcLCC2 in defence of Botrytis cinerea against the broad-spectrum antibiotic 2,4-diacetylphloroglucinol. Environmental Microbiology 10, 114557.
  • De Souza JT, Raaijmakers JM, 2003. Polymorphisms within the prnD and pltC genes from pyrrolnitrin and pyoluteorin-producing Pseudomonas and Burkholderia spp. FEMS Microbiology Ecology 43, 2134.
  • Vaczy KZ, Sandor E, Karaffa L et al. , 2008. Sexual recombination in the Botrytis cinerea populations in Hungarian vineyards. Phytopathology 98, 13129.
  • Vignutelli A, Hilber-Bodmer M, Hilber UW, 2002. Genetic analysis of resistance to the phenylpyrrole fludioxonil and the dicarboximide vinclozolin in Botryotinia fuckeliana. Mycological Research 106, 32935.
  • Ziogas BN, Markoglou AN, Spyropoulou V, 2005. Effect of phenylpyrrole-resistance mutations on ecological fitness of Botrytis cinerea and their genetical basis in Ustilago maydis. European Journal of Plant Pathology 113, 83100.