Assessment of the Environmental Fate of the Biological Control Agent ...

3 downloads 148 Views 169KB Size Report
Oct 17, 2005 - of Fire Blight, Pseudomonas fluorescens EPS62e, on Apple by Culture ..... ing to real-time PCR, population levels started at the same value but ...
APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Apr. 2006, p. 2421–2427 0099-2240/06/$08.00⫹0 doi:10.1128/AEM.72.4.2421–2427.2006 Copyright © 2006, American Society for Microbiology. All Rights Reserved.

Vol. 72, No. 4

Assessment of the Environmental Fate of the Biological Control Agent of Fire Blight, Pseudomonas fluorescens EPS62e, on Apple by Culture and Real-Time PCR Methods Marta Pujol,1 Esther Badosa,1 Charles Manceau,2 and Emilio Montesinos1* Institute of Food and Agricultural Technology-CIDSAV-CeRTA, University of Girona, 17071 Girona, Spain,1 and UMR 077 PaVe´, Centre INRA, 49071 Beaucouze´, France2 Received 17 October 2005/Accepted 19 January 2006

The colonization of apple blossoms and leaves by Pseudomonas fluorescens EPS62e was monitored in greenhouse and field trials using cultivable cell counting and real-time PCR. The real-time PCR provided a specific quantitative method for the detection of strain EPS62e. The detection level was around 102 cells g (fresh weight)ⴚ1 and the standard curve was linear within a 5-log range. EPS62e actively colonized flowers reaching values from 107 to 108 cells per blossom. In apple flowers, no significant differences were observed between population levels obtained by real-time PCR and plating, suggesting that viable but nonculturable (VBNC) cells and residual nondegraded DNA were not present. In contrast, on apple leaves, where cultivable populations of EPS62e decreased with time, significant differences were observed between real-time PCR and plating. These differences indicate the presence of VBNC cells or nondegraded DNA after cell death. Therefore, the EPS62e population was under optimal conditions during the colonization of flowers but it was stressed and poorly survived on leaves. It was concluded that for monitoring this biological control agent, the combined use of cultivable cell count and real-time PCR is necessary. Fire blight is a bacterial disease caused by Erwinia amylovora that carries serious economic losses in rosaceous plants production around the world (42, 44). Disease management has been focused on chemical bactericides, but alternative methods like biological control are increasingly used. Strains belonging to the bacterial species Pseudomonas fluorescens (48), Pantoea agglomerans (14), and Bacillus subtilis (1) have been the object of studies of the biocontrol of fire blight. Strains P. fluorescens A506, P. agglomerans E325 and B. subtilis QST713 are under commercialization or evaluation in the United States, and others will be registered in Europe. The research effort to identify new antagonists of E. amylovora has led to the isolation of new strains such as P. fluorescens EPS62e, which was selected for its high efficacy in controlling infections in immature pear fruit, flowers, and whole plants (6, 34). The registration of a biological control agent requires the development of monitoring methods needed for its detection and quantification in the environment. Methods of analysis are also required to study the impact of formulation, application techniques, and environmental conditions on the ecological fitness of the biocontrol strains (32, 43). However, the selectivity of the monitoring method must be at the strain level, because many biological control agents belong to species that are common inhabitants of plants (18). The use of culture-based methods to monitor biological control agents presents the limitation of the lack of specificity at the strain level. Therefore, antibiotic-resistant mutants of the wild-type strain have been used (5, 15, 27). However, antibiotic resistance traits may present pleiotropic effects, and the anti-

biotic-resistant strains may display modifications in their fitness (21). Furthermore, target bacterial population levels may be overestimated if other resident bacteria present the same resistance in the field (4). To avoid this problem with P. fluorescens EPS62e, a method of analysis was developed and validated, based on culture in selective medium and detection by PCR using primers designed in a specific sequence (34). Monitoring methods based on the cultivation of bacteria, even when coupled to PCR, may underestimate the actual population size because bacteria could enter in a viable but nonculturable (VBNC) state. The VBNC state represents a transient inability to grow on nutrient medium, on which bacteria normally grow and develop colonies, while still being metabolically active (33, 49). This state has been reported for several enteric bacteria (7, 25) and in the plant-pathogenic bacteria Ralstonia solanacearum (12), Xanthomonas campestris pv. campestris (11), and Pseudomonas syringae (47). The VBNC state has been also reported for P. fluorescens CHAO, the biocontrol agent of several soilborne diseases (22). In the phyllosphere, this state can be induced by exposure to natural environmental stress, oligotrophic conditions, or sublethal injury, due to the effect of xenobiotic agents (46, 47). It is probable that this phenomenon occurs after field release of the biological control agent P. fluorescens EPS62e and may lead to an underestimation of the effective population size when culture-based methods are used. Molecular monitoring methods based on nucleic acid targets allow the detection and quantification of microorganisms without regard to their cultivability (17). Several techniques have been used to quantify biocontrol strains such as quantitative competitive PCR (QC-PCR) (24, 35, 38) and, more recently, real-time PCR, which has been increasingly reported (2, 3, 23, 39, 41). Compared to culture-based methods, real-time PCR has the advantage of detecting cultivable and VBNC cells as

* Corresponding author. Mailing address: Institute of Food and Agricultural Technology-CIDSAV-CeRTA, University of Girona, Campus Montilivi, 17071 Girona, Spain. Phone: 34 972418427. Fax: 34 972418399. E-mail: [email protected]. 2421

2422

PUJOL ET AL.

APPL. ENVIRON. MICROBIOL.

TABLE 1. List of bacterial strains used in this work a

Species

Origin or reference

Code

Pantoea agglomerans P. agglomerans Pseudomonas corrugata Pseudomonas fluorescens P. fluorescens P. fluorescens P. fluorescens P. fluorescens Pseudomonas putida Pseudomonas syringae pv. syringae P. syringae pv. tomato

9 strains isolated from rosaceous plants 850 124T EPS62e, EPS62e Nal 64 strains isolated from rosaceous plants Q2-87 CHAO JBR1-70 385, 4064, 4584 15 strains isolated from rosaceous plants DC3000

UdG CECT CECT UdG UdG USDA IPS WAU CECT UdG 8

a

UdG, Universitat de Girona (Spain); CECT, Coleccio ´n Espan ˜ola de Cultivos Tipo (Spain); USDA, United States Department of Agriculture-Agricultural Research Service (United States); IPS, Institute of Plant Sciences (Switzerland); WAU, Wageningen Agricultural University (The Netherlands).

well (3, 13, 16). However, under certain conditions, it can also detect nondegraded residual DNA after cell death (41). The simultaneous use of molecular and culture-based methods for monitoring a biological control agent may provide important knowledge about its environmental fate. These kinds of studies have been performed with biocontrol agents of a fungal nature (2, 3, 24), but when bacteria were used, they were restricted to the soil environment (38). Neither phyllosphere bacteria nor biological control agents of fire blight have ever been objects of previous reports in which real-time PCR and culturable cell counting have simultaneously been employed. The information derived from both techniques on the presence, viability, and population size of P. fluorescens EPS62e, obtained during a period of time and under several environmental conditions, will provide useful data for improving the efficacy of biological control agents of fire blight. The aim of this work was to develop a real-time PCR for the quantification of P. fluorescens EPS62e and its evaluation, in combination with a culturable cell counting method, as a tool for monitoring its environmental fate. The study was performed under different conditions consisting of active colonization of apple flowers and epiphytic phyllosphere survival in the greenhouse and the field. MATERIALS AND METHODS Bacterial strains, growth conditions, and DNA extraction. Bacterial strains used in this work are listed in Table 1. For routine use, bacterial strains were

cultured in Luria-Bertani agar (LB) at 25°C for 24 h. A spontaneous mutant (EPS62e Nal) of the wild-type P. fluorescens EPS62e resistant to nalidixic acid was selected on LB agar supplemented with 50-mg liter⫺1 of nalidixic acid to be used for monitoring the biocontrol agent in plant tissues. The mutant retained the same phenotypic properties as the parental strain. DNA extractions were performed according to the method described by Llop et al. (19). Briefly, 1 ml of overnight culture was centrifuged at 10,000 ⫻ g for 10 min, and the pellet was resuspended in 500 ␮l of DNA extraction buffer (200 mM Tris-HCl [pH 7.5], 250 mM NaCl, 25 mM EDTA, 0.5% sodium dodecyl sulfate, 2% polyvinylpyrrolidone). Tubes were shaken for 1 h and centrifuged at 5,000 ⫻ g for 5 min, and 450 ␮l of the supernatant was transferred to a new tube where 450 ␮l of isopropanol was added for DNA precipitation. One hour later, tubes were centrifuged at 13,000 ⫻ g for 10 min, and DNA was dried and suspended in 200 ␮l of sterile ultrapure water. Purified DNA was quantified when needed at 260 nm using a spectrophotometer (NanoDrop ND-1000; NanoDrop Technologies, Wilmington, DE). Real-time PCR development. Two primer sets and two TaqMan probes were designed, based on two previously described sequence-characterized amplified regions (SCAR) (34). The sequences of primers and probes and the lengths of amplified products are listed in Table 2. For each primer set design, one original SCAR primer (34) was retained (SCAR 450R or SCAR 900F), and a new one was designed (Q 450F or Q 900R) with Primer Express software (PE Applied Biosystems, Massachusetts). For each primer set, a TaqMan probe was designed (450 PBR or 900 PBR) within the sequence delimited between primers and marked with a reporter dye. Probe 450 PBR had a VIC (PE Applied Biosystems) reporter dye at the 5⬘ end. Probe 900 PBR had 6-carboxyfluoresceine (FAM) at the 5⬘ end. Both probes contained 6-carboxytetramethylrhodamine (TAMRA) at the 3⬘ end as the quencher. The concentrations of MgCl2, primers, and probes were optimized, and the PCRs were carried in a final volume of 20 ␮l containing 1⫻ PCR TaqMan buffer A (PE Applied Biosystems), 6 mM MgCl2, 0.2 mM deoxynucleoside triphosphates, 0.3 ␮M each primer, 0.2 ␮M probe, 1 U AmpliTaq Gold DNA Polymerase (PE Applied Biosystems), and 1 ␮l of the extracted DNA. The thermocycle conditions consisted of an initial denaturation step at 95°C for 10 min, followed by 50 cycles, each consisting of 95°C for 15 s and 60°C for 1 min. Specificity of real-time PCR was assessed with extracted DNA from 67 strains of P. fluorescens and 30 strains of closely related species (P. syringae, Pseudomonas corrugata, Pseudomonas putida, and P. agglomerans) (Table 1). A negative control without DNA and a positive control with EPS62e DNA were included in each PCR run, and all reactions were performed in triplicate. Three standard curve designs were developed and compared to choose the best quantification method. The standard curve of type A was performed by mixing several concentrations of EPS62e cells (from 5 ⫻ 108 to 5 ⫻ 103 CFU ml⫺1) with plant extracts. Plant extracts were obtained from six apple blossoms that had been homogenized (IUL Instruments masticator; England) with 60 ml of extraction buffer (0.14 M NaCl, 0.26 M NaH2PO4 · 2H2O, 0.75 mM Na2HPO4 · 12H2O, 2% polyvinylpyrrolidone-10, 1% mannitol, 10 mM ascorbic acid, and 10 mM reduced L-glutathione) for 90 s. EPS62e-plant extract mixtures followed the DNA extraction by isopropanol precipitation as described above. The standard curve of type B was performed by mixing several concentrations of EPS62epurified DNA (from 7 ⫻ 106 to 7 fg ␮l⫺1) with the corresponding DNA from plant extracts. Finally, the standard curve of type C consisted of several dilutions of EPS62e-purified DNA directly in real-time PCR, without mixing it with plant extracts. Analysis of P. fluorescens EPS62e in greenhouse and field trials. Monitoring assays were performed with blossoms and leaves of the Golden Delicious apple cultivar under greenhouse and field conditions. Blossom inoculation experiments consisted of three independent trials (T1,

TABLE 2. Primers and probes designed for real-time PCR for the detection of P. fluorescens EPS62e Primers and probes

Sequence (5⬘–3⬘)

Amplified product (bp)

Source or reference

Q 450F SCAR 450R 450 PBR

AATGGGCTTGCGTCGAGTT CGGTTAGATCCGACAAGATTAGAG VIC-CCGCCACTACCAGGCTATTCAGCTGC-TAMRA

83

This work 34 This work

SCAR 900F Q 900R 900 PBR

CTCGCGTTGAGAGCAGAGAAC TGTGCCCAATTAGAAGCTGTTG FAM-CTCGATGGCCCTCACCAGGC-TAMRA

65

34 This work This work

ENVIRONMENTAL FATE OF P. FLUORESCENS EPS62e

VOL. 72, 2006 T2, and T3) performed under greenhouse conditions and one trial (T4) that was performed in an experimental orchard at Beaucouze´ (Maine-et-Loire, France). For each greenhouse trial, six branches containing 10 dormant flower buds were collected. Then, the branches were placed in containers with sand, and they were forced to bloom in 1% sucrose solution at 22°C and with natural light (31). Immediately after blossoming, flowers were artificially pollinated with pollen from a Prima apple cultivar. The flowers in T1, T2, and T3 were sprayed till runoff point with an EPS62e Nal suspension at 108 CFU ml⫺1 with a hand sprayer (Ecospray, LCF, France). For the field trial (T4), a total of nine trees were selected along a row, separated in three groups by two trees. Each of the three replicates was sprayed with a EPS62e Nal suspension at 108 CFU ml⫺1 during the bloom period by using a motorized mist blower (model SR400; Stihl, Waiblingen, Germany). For the leaf inoculation experiments, three independent trials (T5, T6, and T7) were performed under greenhouse conditions, and one trial (T8) was performed in an experimental orchard at La Meignanne (Maine-et-Loire, France). For each greenhouse trial, a set of 27 apple seedlings were sprayed with a EPS62e Nal suspension at 108 CFU ml⫺1 till runoff point. The field trial (T8) was also performed with a total of nine trees, arranged as described above, that were sprayed with a EPS62e Nal suspension at 108 CFU ml⫺1 with the motorized mist blower. Both flower and leaf trials were sampled for the first time at 12 h after inoculation, and three samples were collected from each experiment at each sampling time. Flower trials under greenhouse conditions were sampled by cutting three blossoms picked randomly among the six branches. Leaf trials under greenhouse conditions were sampled by cutting six leaves from one single seedling every sampling time. Samples were weighed and homogenized in 30 ml of the extraction buffer for 60 s. In the field trials, flowers were sampled by cutting six blossoms from three trees, and leaves were sampled by cutting 12 leaves from three trees. Samples were transported to the laboratory and weighed before being homogenized in 60 ml of the extraction buffer for 60 s. In trial T4, the protocol was changed after 33 days of sampling, due to the presence of immature fruit clusters; homogenization was replaced by shaking bags containing fruit and the extraction buffer in an orbital shaker (KS501; IKA Labortechnik, Staufen, Germany) at 200 rpm for 1 h at 4°C. Each sample extract was then analyzed twice by dilution plating and real-time PCR. At the last sampling time, a set of fruit samples was taken to analyze the distribution of EPS62 on the fruit surface. Fruits were cut to separate the calyx area containing the sepals from the rest. Both subsamples were extracted as described above. For dilution plating, 50 ␮l of serial 10-fold dilutions of the sample was dropped on LB agar plates supplemented with 50 mg liter⫺1 nalidixic acid to select for EPS62e and 50 mg liter⫺1 econazole nitrate salt to prevent fungal growth. Colonies were counted after 48 h of incubation at 25°C. For real-time PCR, DNA was extracted from 1 ml of each sample by isopropanol precipitation as described above (19), and each DNA extraction was evaluated in triplicate using the VIC TaqMan probe design for EPS62e (Table 2). Quantification was obtained by interpolating the threshold cycle values (CT) from the unknown samples against the calibrated standard curve of type A. The standard curve was adapted to the corresponding plant extracts used in each trial (nontreated flowers or leaves). Statistical analysis. The population levels were calculated and expressed either in CFU or cells per blossoms, fruit, or gram (fresh weight) of leaves. Population levels obtained from both monitoring methods were analyzed by the general linear model procedure of SPSS, version 13.0, for Windows to determine whether there were significant differences. The monitoring method and sampling time were fixed factors in the model. An interaction between both factors was included in the model. The level of significance was set at a P value of ⱕ0.05, and the normal distribution and homocedasticity of residuals from the statistical model were checked. A comparison between methods was also performed by regression analysis.

RESULTS Development of real-time PCR. At 10 ng DNA per PCR (approximately 106 CFU per PCR), successful amplification of EPS62e was achieved (CT values from 15 to 16) for both probe designs (VIC and FAM). The specificity was proven because under these conditions, no fluorescence was detected for any of the strains tested (Table 1). However, a random fluorescence signal was observed at CT values of 35 to 36 in one of the

2423

FIG. 1. Real-time PCR standard curve for the quantification of P. fluorescens EPS62e. Known amounts of EPS62e cells were diluted in plant extracts prior to DNA extraction.

three PCR replicates in 13 strains for the VIC probe and in 23 strains for the FAM probe. Using the standard curve of type A, based on dilutions of EPS62e cells mixed with plant extracts, a good linearity was obtained over a 5-log range (y ⫽ 35.25 ⫺ 3.12x) (Fig. 1). In standard curves for type B (y ⫽ 35.00 ⫺ 2.79x) and C (y ⫽ 42.33 ⫺ 3.64x), based on EPS62e-purified DNA, linearity was achieved over a 4.5-log range. In the three cases, the correlation coefficient (R2) was ⬎0.99. The comparison between the three standard curves developed led to the selection of type A for further experiments. Monitoring P. fluorescens EPS62e Nal population levels in greenhouse and field. Colonization of flowers and leaves was evaluated by both real-time PCR and plating methods under two different environmental conditions (greenhouse and field). The dynamics of population levels of EPS62e Nal in flowers is shown in Fig. 2. Levels increased or were maintained at high values upon inoculation, indicating efficient colonization of flowers. There were no significant differences between the methods used to estimate the population level (P ⫽ 0.479). Nevertheless, the interaction of method with time was significant for trials T3 (P ⫽ 0.009) and T4 (P ⫽ 0.002) (Fig. 2C and D). The significance of the interaction disappeared after repeating the analysis without the first sampling time value (P ⫽ 0.216 and 0.495, respectively). EPS62e Nal reached population levels of 5 ⫻ 106 to 5 ⫻ 107 CFU or cells per blossom under greenhouse conditions. The colonization curve of flowers in the field trial was similar to greenhouse conditions, although population levels were higher as they increased during the first 5 days during bloom from around 105 to 108 CFU or cells per blossom and then remained stable for 55 days during the fruit’s growth (Fig. 2D). After fruit set, an analysis of fruit was conducted to evaluate whether the biocontrol agent colonized the apple surface or the sepals at the calyx area. It was observed that 99% of the EPS62e Nal population remained at the sepals (data not shown). The population dynamics of EPS62e Nal in leaves is shown in Fig. 3. EPS62e Nal declined on leaves instead of increasing populations as occurred in flowers. In this case, there were significant differences between real-time PCR and cultivable cell counting methods (P ⬍ 0.05) (Fig. 3). In greenhouse trials, cultivable cell population levels started at around 5 ⫻ 107 CFU

2424

PUJOL ET AL.

APPL. ENVIRON. MICROBIOL.

creased progressively to 5 ⫻ 103 cells g (fresh weight)⫺1 by the end of the assay. To compare the two monitoring methods, the values obtained by plating were plotted against real-time PCR for the two types of experiments performed (flowers and leaves) (Fig. 4). A good correlation was obtained for flower trials (R2 ⫽ 0.8682 and P ⬍ 0.0001; log [real-time PCR values] ⫽ 1.64 ⫹ 0.78 log [CFU counts]), regardless of whether the trial was developed under field or greenhouse conditions (Fig. 4A). The origin coordinate indicated that population levels estimated by real-time PCR were around 40 times higher than those assessed by plating. Nevertheless, in leaf trials, values obtained from both techniques did not correlate as well as for flowers (R2 ⫽ 0.6854 and P ⬍ 0.0001; log [real-time PCR values] ⫽ 3.23 ⫹ 0.60 log [CFU counts]) (Fig. 4B). According to the origin coordinate, real-time PCR gave, on average, values around 1,700 times higher than those obtained from CFU counts. DISCUSSION

FIG. 2. Population dynamics estimated by real-time PCR (E) and plating (F) of P. fluorescens EPS62e Nal on apple flowers under greenhouse (A to C) and field (D) conditions. Cells were sprayed at 108 CFU ml⫺1. The standard deviation (SD) of three replicate samples is represented by a vertical bar.

g (fresh weight)⫺1 upon inoculation and decreased strongly to 104 CFU g (fresh weight)⫺1 (Fig. 3A to C). However, according to real-time PCR, population levels started at the same value but decreased slightly to 106 cells g (fresh weight)⫺1 (Fig. 3A to C). Therefore, the differences between real-time PCR and cultivable cell counting increased over time, reaching a stable value of 2 log by 7 days after treatment. In the field trial, values estimated by both techniques were very different upon inoculation but became similar 15 days after treatment (Fig. 3D). According to the plating method, a small population of cultivable EPS62e Nal remained more or less stable at 5 ⫻ 103 CFU g (fresh weight)⫺1 from the start of the experiment until 30 days later. However, according to real-time PCR, the population level started at 107 cells g (fresh weight)⫺1 and de-

Knowledge of the environmental fate of P. fluorescens EPS62e is of great importance in assuring its efficacy after field application. The use of monitoring methods targeted to cultivability and to DNA markers will allow an accurate assessment of the behavior of the population in the environment. A real-time PCR monitoring method was developed for EPS62e on the basis of two SCAR-specific markers that were developed in a previous work (34). Both SCAR markers allowed the specific amplification of EPS62e and were obtained by the selection and sequencing of differential amplified fragments from a random amplified polymorphic DNA pattern (34). These sequences were used in the present work to perform two real-time PCR designs. The two real-time PCR designs developed showed the same sensitivity and were specific for the quantification of EPS62e. In specificity assays, random amplifications with high CT values were observed for some non-target strains. However, this phenomenon, previously described in the literature as a background no-template control (20), has been considered as not important for quantification if it remains outside the range used to generate the standard curve, as occurred in the present study. Even though all standard curves developed were useful for EPS62e quantification,

FIG. 3. Population dynamics estimated by real-time PCR (䊐) and plating (■) of P. fluorescens EPS62e Nal on apple leaves under greenhouse (A to C) and field (D) conditions. Cells were sprayed at 108 CFU ml⫺1. The SD of three replicate samples is represented by a vertical bar. fw, fresh weight.

VOL. 72, 2006

FIG. 4. Relationship between culturable and real-time PCR estimates of P. fluorescens EPS62e Nal population levels in apple flowers (A) and leaves (B). Trials were performed under greenhouse (E and 䊐) and field (F and ■) conditions. fw, fresh weight.

the standard curve of type A (EPS62e cells diluted in plant extract) was chosen because it gave high amplification efficiency and good fit and because the values obtained corresponded directly to cells and not to the target DNA, as in types B and C. As in the present study, the standard curve obtained from cells diluted in the sample matrix instead of purified DNA has been commonly used for the quantification of several biocontrol agents including Pseudomonas in soil (2, 3, 23, 45). This method avoids the extrapolation of cells from DNA quantity and takes into account the efficiency of DNA extraction procedure and the presence of PCR inhibitors in the natural matrix (50). Real-time PCR monitoring methods based on a DNA target provide an estimation of the population composed of cultivable cells, VBNC cells, and residual nondegraded DNA released after cell death. In the present study, the relative contribution of VBNC cells and residual DNA on population level assessment of EPS62e was analyzed by comparing real-time PCR population estimations to those obtained by a cultivable method. Different conditions, in which increasing degrees of stress were expected on the biological control agent such as blossoms or leaves under greenhouse or field conditions, were analyzed. Part of the study was focused on blossoms because colonization of flowers is of great importance for an efficient biological control of fire blight, since they are the primary pathway for infections of E. amylovora. Flowers are described as a favorable habitat for many microorganisms because of the high

ENVIRONMENTAL FATE OF P. FLUORESCENS EPS62e

2425

levels of nutrients, with sugar concentrations ranging from 10 to 30% (9, 36). In this environment, EPS62e Nal showed an active colonization, reaching values from 107 to 108 CFU or cells per blossom. These values are almost an order of magnitude higher than those obtained for other fire blight biological control agents such as P. fluorescens A506, P. agglomerans C9-1, and P. agglomerans E325 (37). The fact that no significant differences were observed between real-time PCR and cultivable cell population estimates of EPS62e Nal in blossoms indicated that VBNC cells and nondegraded DNA released from cell death were not present or were not present in amounts sufficient to show a difference between the two methods of analysis. Therefore, under these conditions the biological control agent colonized without obvious stress. This is in agreement with a report of epiphytic populations of P. syringae, where the plate count method accurately estimated the viable bacterial population when it was in a state of active growth (47). Moreover, another report also concluded that in the absence of stress, VBNC cells of P. fluorescens CHAO in the soil were not observed, and population levels estimated by QC-PCR correlated with CFU counts (38). On leaf surfaces, EPS62e Nal cells were supposed to be stressed because of the different estimations obtained by realtime PCR and plating and the decline in cultivable population levels. The phyllosphere habitat is far different from that of flowers because the leaf surface is poor in nutrients and is exposed to large fluctuations in physical and nutritional conditions (18). Accordingly, the carrying capacity of the leaf surface is directly correlated with the amount of nutrients available (26). The differences observed in the present study between real-time PCR and cultivable cell counting methods can be attributed to the stressful conditions of the leaf environment that promote entry in a VBNC cells state of a part of the EPS62e Nal population (16, 20), the presence of cell aggregates nondispersed before plating (29, 30), or the presence of free DNA after cell death (3, 41, 45). The VBNC state might result from exposure to natural environmental stresses in the phyllosphere (7). It was reported that epiphytic populations of P. syringae in bean leaves were underestimated by plate counts from two to fourfold because of the presence of VBNC cells (47). Moreover, in the case of the biocontrol agent P. fluorescens CHAO in the soil, it has been demonstrated that under stress, culture-based methods underestimated CHAO population levels by 3 orders of magnitude, and the correlation between plating and QC-PCR was lost (22, 38). An underestimation of the epiphytic bacterial population by CFU counts might also be expected if growth in leaf surface follows aggregate patterns. The colonization of leaf surface is limited to a number of sites that may offer conditions conducive to the growth of immigrant bacteria (18, 30). The heterogeneous nature of the surface leads bacterial populations, such as P. agglomerans and P. fluorescens, to form aggregates (28, 29). However, the presence of cell aggregates of EPS62e Nal in the washing suspension was unlikely because of the high efficacy of the homogenization method used. The presence of nondegraded DNA in samples is notably the most important drawback for the detection by real-time PCR. It seems that the DNA degradation rate by nucleases after cell death strongly depends on environmental conditions

2426

PUJOL ET AL.

APPL. ENVIRON. MICROBIOL.

(41). Research has been carried out to evaluate DNA persistence in soil, but in some studies DNA was rapidly degraded (40), whereas in others DNA persisted for a long period of time (10). Further research is needed to determine DNA persistence in the phyllosphere. Nonetheless, according to the results obtained in the present work in the field trial, long stability of DNA is not expected in the phyllosphere, because a 4-log decrease in population levels of EPS62e Nal based on real-time PCR was observed 20 days after field inoculation. Different correlations between techniques were observed during leaf trials in seedlings under greenhouse and trees in field conditions. On the one hand, the nature of tree leaf surfaces compared to that of young seedling leaves is highly different and may allow a major diffusion of nutrients in the latter. On the other hand, environmental pressure on the EPS62e Nal population was weaker in the greenhouse, where temperature and relative humidity were controlled, than in the field. Upon inoculation, the cultivable population in the greenhouse was shown to be larger than in the field and more similar to real-time PCR estimations. Moreover, 3 days after inoculation under greenhouse conditions, a stabilization in cultivable and real-time PCR values, probably due to entry into the VBNC state, was observed. However, under field conditions, values obtained from real-time PCR decreased progressively, while CFU counts were stable at very low levels. The difference between methods was interpreted as reflecting DNA degradation of dead cells. Cell death started rapidly, within a few hours after EPS62e was sprayed into the trees, due to unfavorable environmental conditions. In conclusion, the combined use of real-time PCR and culture-based methods was useful to track the EPS62e population under greenhouse and field conditions and gave valuable information on population behavior, indicating the entry into the VBNC state of a part of the population or the presence of residual DNA from dead cells. ACKNOWLEDGMENTS This work was supported by the Spanish Ministry of Science and Technology (AGL2001-2349 and AGL2004-07799) and the Comissio ´ Interdepartamental de Recerca i Tecnologia of the Generalitat de Catalunya (GRC-2001SGR00293). M.P. is the recipient of a research grant (FP-2001-2130) from the Spanish Ministry of Science and Technology. REFERENCES 1. Aldwinckle, H. S., M. V. Bhaskara Reddy, and J. L. Norelli. 2002. Evaluation of control of fire blight infection of apple blossoms and shoots with SAR inducers, biological agents, a growth regulator, copper compounds, and other materials. Acta Hortic. 590:325–331. 2. Atkins, S. D., I. M. Clark, S. Pande, P. R. Hirsch, and B. R. Kerry. 2005. The use of real-time PCR and species-specific primers for the identification and monitoring of Paecilomyces lilacinus. FEMS Microbiol. Ecol. 51:257–264. 3. Atkins, S. D., I. M. Clark, D. Sosnowska, P. R. Hirsch, and B. R. Kerry. 2003. Detection and quantification of Plectosphaerella cucumerina, a potential biological control agent of potato cyst nematodes, by using conventional PCR, real-time PCR, selective media, and baiting. Appl. Environ. Microbiol. 69: 4788–4793. 4. Badosa, E., C. Moreno, and E. Montesinos. 2004. Lack of detection of ampicillin resistance gene transfer from Bt176 transgenic corn to culturable bacteria under field conditions. FEMS Microbiol. Ecol. 48:169–178. 5. Bonaterra, A., M. Mari, L. Casalini, and E. Montesinos. 2003. Biological control of Monilinia laxa and Rhizopus stolonifer in postharvest of stone fruit by Pantoea agglomerans EPS125 and putative mechanisms of antagonism. Int. J. Food Microbiol. 84:93–104. 6. Cabrefiga, J. 2004. Fire blight (Erwinia amylovora) of rosaceous plants. Pathogen virulence and selection and characterization of biological control agents. Ph.D. thesis. University of Girona, Girona, Spain.

7. Colwell, R. R., and D. J. Grimes. 2000. Nonculturable microorganisms in the environment. ASM Press, Washington, D.C. 8. Cuppels, D. A. 1986. Generation and characterization of Tn5 insertion mutations in Pseudomonas syringae pv. tomato. Appl. Environ. Microbiol. 51: 323–327. 9. De la Barrera, E., and P. S. Nobel. 2004. Nectar: properties, floral aspects, and speculations on origin. Trends Plant Sci. 9:65–69. 10. England, L. S., H. Lee, and J. T. Trevors. 1997. Persistence of Pseudomonas aureofaciens strains and DNA in soil. Soil Biol. Biochem. 29:1521–1527. 11. Ghezzi, J. I., and T. R. Steck. 1999. Induction of the viable but non-culturable condition in Xanthomonas campestris pv. campestris in liquid microcosms and sterile soil. FEMS Microbiol. Ecol. 30:203–208. 12. Grey, B. E., and T. R. Steck. 2001. The viable but nonculturable state of Ralstonia solanacearum may be involved in long-term survival and plant infection. Appl. Environ. Microbiol. 67:3866–3872. 13. Heuser, T., and W. Zimmer. 2003. Genus- and isolate-specific real-time PCR quantification of Erwinia on leaf surfaces of English oaks (Quercus robur L.). Curr. Microbiol. 47:214–219. 14. Ishimaru, C. A., E. J. Klos, and R. R. Burbaker. 1988. Multiple antibiotic production by Erwinia herbicola. Phytopathology 78:746–750. 15. Johnson, K. B., V. O. Stockwell, T. L. Sawyer, and D. Sugar. 2000. Assessment of environmental factors influencing growth and spread of Pantoea agglomerans on and among blossoms of pear and apple. Phytopathology 90:1285–1294. 16. Khan, I. U., and J. S. Yadav. 2004. Real-time PCR assays for genus-specific detection and quantification of culturable and non-culturable mycobacteria and pseudomonads in metalworking fluids. Mol. Cell. Probes 18:67–73. 17. Knight, I. T. 2000. Molecular genetic methods for detection and identification of viable but nonculturable microorganisms, p. 77–85. In R. R. Colwell and D. J. Grimes (ed.), Nonculturable microorganisms in the environment. ASM Press, Washington, D.C. 18. Lindow, S. E., and M. T. Brandl. 2003. Microbiology of the phyllosphere. Appl. Environ. Microbiol. 69:1875–1883. 19. Llop, P., P. Caruso, J. Cubero, C. Morente, and M. M. Lopez. 1999. A simple extraction procedure for efficient routine detection of pathogenic bacteria in plant material by polymerase chain reaction. J. Microbiol. Methods 37:23–31. 20. Lloyd-Jones, G., A. D. Laurie, and A. C. Tizzard. 2005. Quantification of the Pseudomonas population in New Zealand soils by fluorogenic PCR assay and culturing techniques. J. Microbiol. Methods 60:217–224. 21. Mahaffee, W. F., E. M. Bauske, J. W. L. van Vuurde, J. M. van der Wolf, M. van den Brink, and J. W. Kloepper. 1997. Comparative analysis of antibiotic resistance: immunofluorescent colony staining, and a transgenic marker (bioluminescence) for monitoring the environmental fate of a rhizobacterium. Appl. Environ. Microbiol. 63:1617–1622. 22. Mascher, F., C. Hase, Y. Moenne-Loccoz, and G. Defago. 2000. The viablebut-nonculturable state induced by abiotic stress in the biocontrol agent Pseudomonas fluorescens CHA0 does not promote strain persistence in soil. Appl. Environ. Microbiol. 66:1662–1667. 23. Massart, S., D. De Clercq, M. Salmon, C. Dickburt, and M. H. Jijakli. 2005. Development of real-time PCR using Minor Groove Binding probe to monitor the biological control agent Candida oleophila (strain O). J. Microbiol. Methods 60:73–82. 24. Mauchline, T. H., B. R. Kerry, and P. R. Hirsch. 2002. Quantification in soil and the rhizosphere of the nematophagous fungus Verticillium chlamydosporium by competitive PCR and comparison with selective plating. Appl. Environ. Microbiol. 68:1846–1853. 25. McDougald, D., S. A. Rice, D. Weichart, and S. Kjelleberg. 1998. Nonculturability: adaptation or debilitation? FEMS Microbiol. Ecol. 25:1–9. 26. Mercier, J., and S. E. Lindow. 2000. Role of leaf surface sugars in colonization of plants by bacterial epiphytes. Appl. Environ. Microbiol. 66:369–374. 27. Mercier, J., and S. E. Lindow. 2001. Field performance of antagonistic bacteria identified in a novel laboratory assay for biological control of fire blight of pear. Biol. Control 22:66–71. 28. Monier, J. M., and S. E. Lindow. 2004. Frequency, size, and localization of bacterial aggregates on bean leaf surfaces. Appl. Environ. Microbiol. 70:346– 355. 29. Monier, J. M., and S. E. Lindow. 2005. Aggregates of resident bacteria facilitate survival of immigrant bacteria on leaf surfaces. Microb. Ecol. 49: 343–352. 30. Monier, J. M., and S. E. Lindow. 2005. Spatial organization of dual-species bacterial aggregates on leaf surfaces. Appl. Environ. Microbiol. 71:5484– 5493. 31. Montesinos, E., and P. Vilardell. 1991. Relationships among populationlevels of Pseudomonas syringae, amount of ice nuclei, and incidence of blast of dormant flower buds in commercial pear orchards in Catalunya, Spain. Phytopathology 81:113–119. 32. Montesinos, E. 2003. Development, registration and commercialization of microbial pesticides for plant protection. Int. Microbiol. 6:245–252. 33. Oliver, J. D. 2000. The public health significance of viable but nonculturable bacteria, p. 277–300. In R. R. Colwell and D. J. Grimes (ed.), Nonculturable microorganisms in the environment. ASM Press, Washington, D.C.

VOL. 72, 2006 34. Pujol, M., E. Badosa, J. Cabrefiga, and E. Montesinos. 2005. Development of a strain-specific quantitative method for monitoring Pseudomonas fluorescens EPS62e, a novel biocontrol agent of fire blight. FEMS Microbiol. Lett. 249:343–352. 35. Pujol, M., D. De Clercq, S. Cognet, P. Lepoivre, and M. H. Jijakli. 2004. Monitoring system for the biocontrol agent Pichia anomala strain K using quantitative competitive PCR-ELOSA. Plant Pathol. 53:103–109. 36. Pusey, P. L. 1999. Effect of nectar on microbial antagonists evaluated for use in control of fire blight of pome fruits. Phytopathology 89:39–46. 37. Pusey, P. L., and E. A. Curry. 2004. Temperature and pomaceous flower age related to colonization by Erwinia amylovora and antagonists. Phytopathology 94:901–911. 38. Rezzonico, F., Y. Moenne-Loccoz, and G. Defago. 2003. Effect of stress on the ability of a phlA-based quantitative competitive PCR assay to monitor biocontrol strain Pseudomonas fluorescens CHA0. Appl. Environ. Microbiol. 69:686–690. 39. Rubio, M. B., M. R. Hermosa, E. Keck, and E. Monte. 2005. Specific PCR assays for the detection and quantification of DNA from the biocontrol strain Trichoderma harzianum 2413 in soil. Microb. Ecol. 49:25–33. 40. Schena, L., and A. Ippolito. 2003. Rapid and sensitive detection of Rosellinia necatrix in roots and soils by real time Scorpion-PCR. J. Plant Pathol. 85:15–25. 41. Schena, L., F. Nigro, A. Ippolito, and D. Gallitelli. 2004. Real-time quantitative PCR: a new technology to detect and study phytopathogenic and antagonistic fungi. Eur. J. Plant Pathol. 110:893–908. 42. Van Der Zwet, T., and H. L. Keil. 1979. Fire blight: a bacterial disease of rosaceous plants. U.S. Department of Agriculture, Washington, D.C. 43. Van Elsas, J. D., G. F. Duarte, A. S. Rosado, and K. Smalla. 1998. Microbiological and molecular biological methods for monitoring microbial inocu-

ENVIRONMENTAL FATE OF P. FLUORESCENS EPS62e

44. 45.

46.

47.

48.

49.

50.

2427

lants and their effects in the soil environment. J. Microbiol. Methods 32: 133–154. Vanneste, J. 2000. Fire blight: the disease and its causative agent, Erwinia amylovora. CABI Publishing, Wallingford, Oxon, United Kingdom. Wang, G., T. J. Gentry, G. Grass, K. Josephson, C. Rensing, and I. L. Pepper. 2004. Real-time PCR quantification of a green fluorescent proteinlabeled, genetically engineered Pseudomonas putida strain during 2-chlorobenzoate degradation in soil. FEMS Microbiol. Lett. 233:307–314. Wilson, M., and S. E. Lindow. 2000. Viable but nonculturable cells in plantassociated bacterial population, p. 229–241. In R. R. Colwell and D. J. Grimes (ed.), Nonculturable microorganisms in the environment. ASM Press, Washington, D.C. Wilson, M., and S. E. Lindow. 1992. Relationship of total viable and culturable cells in epiphytic populations of Pseudomonas syringae. Appl. Environ. Microbiol. 58:3908–3913. Wilson, M., and S. E. Lindow. 1993. Interactions between the biological control agent Pseudomonas fluorescens strain A506 and Erwinia amylovora in pear blossoms. Phytopathology 83:117–123. Xu, H.-S., N. Roberts, F. L. Singelton, R. W. Atwell, D. J. Grimes, and R. R. Colwell. 1982. Survival and viability of nonculturable Escherichia coli and Vibrio cholerae in the estuarine and marine environment. Microb. Ecol. 8:313–323. Zeng, Q. Y., S. O. Westermark, A. Rasmuson-Lestander, and X. R. Wang. 2004. Detection and quantification of Wallemia sebi in aerosols by real-time PCR, conventional PCR, and cultivation. Appl. Environ. Microbiol. 70: 7295–7302.