Crop Protection 20 (2001) 395}402
Bean rust biological control using bacterial agents G.Y. Yuen *, J.R. Steadman , D.T. Lindgren, D. Scha! , C. Jochum Department of Plant Pathology, University of Nebraska, Lincoln, NE 68583-0722, USA West Central Research and Extension Center, North Platte, NE 69101, USA Received 2 June 2000; received in revised form 27 September 2000; accepted 29 September 2000
Abstract Over 120 bacterial strains were evaluated in a greenhouse for control of bean rust caused by Uromyces appendiculatus. The strains, found previously to be antagonistic to some fungal pathogens, were isolated from dry edible bean (Phaseolus vulgaris) and other hosts. Only Pantoea agglomerans B1, from a bean blossom, and Stenotrophomonas maltophilia C3, a chitinolytic strain from a Kentucky bluegrass leaf, were e!ective in multiple experiments in reducing bean rust severity. The addition of colloidal chitin to C3 cell suspensions and treatment with chitin broth cultures of C3 were evaluated as methods to improve biocontrol e$cacy of C3. While chitin amendments increased rust control in the greenhouse as compared to C3 cells in bu!er, chitin broth cultures gave the highestand longest-lasting level of control. In four "eld experiments, treatments with C3 suspended in bu!er, with and without chitin amendment, reduced rust severity in only one experiment. Strain B1 was not e!ective. In three other "eld experiments, C3 chitin broth cultures were comparable to multiple applications of thiophanate methyl or thiophanate methyl combined with manganese ethylenebisdithiocarbamate (maneb) in reducing bean rust severity. 2001 Elsevier Science Ltd. All rights reserved. Keywords: Lytic enzymes; Microbial fungicides; Phylloplane colonization
1. Introduction Bean rust, caused by Uromyces appendiculatus (Pers. ex Pers.) Unger, is a chronic, yield-limiting disease of dry edible bean (Phaseolus vulgaris L.) world wide (Stavely and Pastor-Corrales, 1989). Control of bean rust depends primarily on host resistance, but because the pathogen has high genetic diversity and ability for new race development, resistance has not been long lasting. Protectant fungicides are registered in the United States for use against bean rust, but require timely and repeated applications. Fungicides, such as propiconazole, that have curative action are not labeled for dry edible bean. Biological control agents may have a role. A large number of fungi have been identi"ed as hyperparasites of rust fungi (Gowdu and Balasubramanian, 1988; Je!ries and Young, 1994; Kranz, 1981). Verticillium lecanii colonizes uredinia and penetrates urediniospores of U. appendiculatus (Allen, 1982). Although V. lecanii University of Nebraska, Agricultural Research Division Journal Number 12922. * Corresponding author. Tel.: #1-402-472-3125; fax: #1-402-4722853. E-mail address: [email protected]
showed promise in controlling bean rust in greenhouse trials, it was ine!ective in the "eld (Grabski and Mendgen, 1985), presumably because duration of moisture in the "eld was insu$cient to sustain growth of the hyperparasite on the phylloplane. There also are numerous reports, as reviewed by Gowdu and Balasubramanian (1988), of bacterial strains in the genera Bacillus, Erwinia, and Pseudomonas that were antagonists of rust fungi. In addition to ease in culturing and application, bacteria may have an advantage over fungal hyperparasites as biological control agents. Bacteria can inhibit germination of rust spores through antibiosis (Gowdu and Balasubramanian, 1988), and thus prevent infection, whereas fungal hyperparasites are more e!ective in invading uredinia and reducing secondary sporulation (Kranz, 1981). Among bacterial agents evaluated against bean rust in the laboratory and greenhouse, strains of B. subtilis were the most e!ective at inhibiting urediniospore germination and reducing uredinia formation (Baker et al., 1983; Centurian and Kimati, 1994; Mizubuti et al., 1995). One strain, APPL-1, reduced bean rust severity in "eld tests, but required three applications per week (Baker et al., 1985). Given the narrow diversity of bacterial species investigated as antagonists of bean rust, bacterial strains that
0261-2194/01/$ - see front matter 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 1 - 2 1 9 4 ( 0 0 ) 0 0 1 5 4 - X
G.Y. Yuen et al. / Crop Protection 20 (2001) 395}402
had already been shown to be antagonistic to other fungal pathogens were evaluated against bean rust in a series of greenhouse and "eld trials. Another rationale for evaluating known antagonists was that strains with potential to control multiple pathogens would be better candidates for future development and commercialization. Some of organisms were strains of Pantoea agglomerans ("Erwinia herbicola) isolated from bean blossoms. These originally were selected for inhibition of blossom colonization by Sclerotinia sclerotiorum and were e!ective colonists of bean leaves (James et al., 1995; Yuen et al., 1991; Yuen et al., 1994). Stenotrophomonas maltophilia C3, was a unique chitinolytic strain isolated from a Kentucky bluegrass leaf. Originally, selected for inhibition of Rhizoctonia solani on tall fescue leaves (Giesler and Yuen, 1998), it was e!ective in the "eld in reducing leaf spot caused by Bipolaris sorokiniana (Zhang and Yuen, 1999). The addition of chitin with cells of C3 increased "eld e$cacy against leaf spot (Zhang and Yuen, 1999). Chitinolysis was one mechanism involved in antagonism to B. sorokiniana (Zhang and Yuen, 2000b). In chitin broth culture, C3 excreted high levels of chitinase, glucanase, lipase, and protease (Zhang and Yuen, 2000a). Cell-free culture #uid applied to grass leaves inhibited infection by B. sorokiniana, and the highest degree of leaf spot control resulted from the application of C3 cells with the culture #uid (Zhang and Yuen, 2000a). One objective of this study was to evaluate bacteria from bean or other plants as biocontrol agents against bean rust under greenhouse and "eld conditions. Another objective was to determine if the e$cacy of S. maltophilia C3 against bean rust could be enhanced by incorporation of chitin or culture #uid in the application procedure.
2. Materials and methods 2.1. General bacteriological methods In addition to P. agglomerans strains B1, B346, B367, and B409 and S. maltophilia C3, which were previously reported, bacterial strains isolated from the rhizosphere of sugar beet, wheat, and bean were tested. These strains were antagonistic to Pythium ultimum and R. solani in vitro and in pot tests (Yuen, unpublished data). Unidenti"ed strains isolated at relatively high frequencies from bean leaves or turfgrass foliage also were evaluated. All strains were stored at !703C in nutrient broth amended with 10% glycerol. Fresh cultures were grown from the frozen stock before each experiment. Unless speci"ed, tryptic soy agar (TSA) (Difco, Detroit, MI) was the growth medium. Cultures were incubated at 253C for 2}3 days depending upon the strain. For application to plants, cells were suspended in 0.01 M potassium phos-
phate bu!er (PB), pH 7, or in saline (8.5 g/l NaCl). The surfactant Soydex 937 (Setre Chemical Co., Memphis, TN) was added to all suspensions at 0.25% v/v to aid cell dispersal. Cell density was determined by turbidity measured on a spectrophometer at 595 nm and then adjusted to 10 CFU/ml. In some experiments involving strain C3, the strain was cultured in a chitin broth medium containing chitin (practical grade; Sigma Chemical Company, St. Louis, MO) at 1% w/v as the carbon source (Zhang and Yuen, 2000a). Cultures in 250 ml volumes of chitin broth were incubated at 253C for 5}7 days with shaking. The cultures were "ltered through three layers of sterile cheese cloth to remove particulates before application to plants. Population numbers of strain C3 and P. agglomerans strain B1 (Yuen et al., 1994), which were e!ective in suppressing been rust in initial greenhouse experiments, were determined in subsequent trials. Spontaneous rifampicin-resistant derivatives of the two strains were used for this purpose. To quantify their numbers on treated leaves, individual trifoliolate leaves were washed in 5 ml volumes of PB. The wash was serially diluted and plate counts determined on one-tenth-strength TSA amended with cycloheximide and rifamipicin, each at 100 g/ml. 2.2. Greenhouse experiments Pinto &UI 114' was used as the host in all greenhouse experiments. Plants were grown in a pasteurized medium (Sharpsburg silty clay loam, peat, and sand in equal volumes, contained in 15-cm-diameter plastic pots) in a greenhouse until the primary leaves were fully expanded. Three sets of experiments were conducted in the greenhouse. In the "rst set, 127 bacterial strains were evaluated for e!ects on bean rust incidence. Bacterial suspensions were sprayed onto the leaves until run-o! using a handpressurized sprayer. Controls were sprayed with PB. Treated plants were inoculated with rust spores after the leaves had dried for 2 h. For inoculation, 5 mg of urediniospores of an isolate of U. appendiculatus from Scottsblu!, NE were suspended in 100 ml of distilled water with Tween 20 at 40 l/l and then applied with an aerosol sprayer to top and bottom surfaces of leaves. Inoculated plants were placed in a mist chamber for 15 h and then kept in a greenhouse for 7}10 days. Disease incidence was assessed by counting the number of uredinia within three 4-cm areas per leaf. In all experiments, there were two or three plants per treatment, and disease severity on both primary leaves of each plant was rated. Only the strains that reduced rust incidence compared to the phosphate bu!er control were retested. In the second set of experiments, the e!ects of amending cell suspensions of S. maltophilia C3 with chitin was evaluated. Colloidal chitin was prepared according to the
G.Y. Yuen et al. / Crop Protection 20 (2001) 395}402
method of Kokalis-Burelle et al. (1992). Suspensions of C3 cells grown on TSA were amended with colloidal chitin to 0, 0.2, and 1% w/v. These treatments also were compared with PB amended with colloidal chitin at the same concentrations. Inoculation with rust spores and measurement of disease incidence were as described above, except that treated plants were inoculated one day after application of bacteria. The experiment was conducted twice. In the third set of greenhouse experiments, treatment with chitin broth cultures of C3 were compared for e$cacy with treatment using C3 cells from TSA suspended in PB, with or without a colloidal chitin amendment (0.2% w/v). Five-day-old chitin broth cultures of C3 were diluted "ve-fold with PB prior to application. Treated plants were placed in the greenhouse for 1 or 7 days prior to pathogen inoculation. One week after inoculation, rust severity was rated according to the modi"ed Cobb scale (Stavely, 1985) having seven categories, with &1' being the least and &7' being the highest severity. To monitor population levels of strain C3 on leaves during the experiment, extra plants were treated and placed on the same greenhouse bench. At various intervals, treated primary leaves were collected in sterile plastic bags and later assayed as previously described. The experiment was repeated. 2.3. Field experiments Experiments were conducted from 1996 to 2000 at the West Central Research and Extension Center, North Platte, NE (NP). Experiments also were conducted in 1997 at the Panhandle Research and Extension Center, Scottsblu!, NE (PHEC) and at the Mitchell Experiment Station (MIT) near Scottsblu!. Pinto &UI 114' was planted at NP, Great Northern &Harris' at PREC, and Great Northern &Marquee' at MIT. Plots were a single row of 3 m length in 1996 and were three 6-m-long rows in all subsequent experiments. There were four replicate plots per treatment, except in 2000, in which there were three plots per treatment. The design was a randomized complete block. Bean rust occurred in the 1997 PREC experiment from natural inoculum. In all other experiments, spreader rows surrounding each experiment were inoculated with local isolates of the pathogen during the last 2 weeks of July. Unless speci"ed, all treatment applications began when rust was "rst visible in the spreader rows. Treatments were applied at 100 ml/m row in 1997 experiments and at 50 ml/m in all other years using a hand-pumped backpack sprayer. In the "rst or second week of September, rust severity was rated on the modi"ed Cobb scale. Yield of seed was determined in all experiments, except in 2000, from the middle row of each plot. In 1996, strain C3 was applied as 10 CFU/ml cell suspensions in saline. An industrial formulation of chitosan (Insectinet, Chemical Products Technology,
Cartersville, GA) was tested alone (0.2% w/v in saline solution) or as an amendment to C3 cell suspensions. For each of these treatments, four applications were made at weekly intervals. Comparisons were made with propiconazole (Tilt, 42% a.i., Novartis; 290 ml/ha), which is not currently labeled for dry edible bean, but has been used in Nebraska for bean rust on an emergency exemption (Section 18) basis. Two applications were made at a 14-day interval. There was a non-treated control. In 1997 NP and 1997 MIT, strain C3 was tested in the same manner as in 1996, except that PB was used as the suspending agent in 1997 MIT. The chitosan preparation tested in 1996 reduced viability of strain C3 in subsequent laboratory experiments, and therefore, colloidal chitin, at 0.2% w/v, was used in place of chitosan in 1997 "eld tests. In addition, P. agglomerans strains B1 was tested in the two experiments as cell suspensions containing 10 CFU/ml. There were four applications at weekly intervals for each treatment. Propiconazole was applied twice at 14-day intervals. The controls were treated four times, at weekly intervals, with the respective suspending agent. The 1997 PREC experiment was intended originally as an evaluation for e$cacy against white mold. Therefore, P. agglomerans strain B409, which was inhibitory to white mold but not to bean rust in the greenhouse, was tested in addition to strain C3. Both strains were applied as 10 CFU/ml cell suspensions in PB, with four applications at weekly intervals. Other treatments included a combination of C3, B409, and 0.2% colloidal chitin (four applications at 7-day intervals) and colloidal chitin alone (two applications at 14-day intervals). For comparison with a registered fungicide, two applications of thiophanate methyl (Topsin M 70 W, 70% a.i., Elf Atochem North America; 1.1 kg/ha) were made 7 days apart, with initial applications made at full bloom (i.e., 50% plants having open blossoms). The control was non-treated. The primary intent of the experiments conducted in 1998}2000 was to compare C3 grown in chitin broth to treatment with standard commercial fungicides. In greenhouse experiments, C3 in chitin broth culture provided greater control of bean rust than other C3 formulations, and in a previous "eld study, whole C3 chitin broth culture was more e!ective than cell-free culture #uid for controlling B. sorokiniana in tall fescue (Zhang and Yuen, 2000a). Therefore, C3 cells alone and cell-free culture #uid were not included as treatments in these "eld experiments. Whole 7-day-old cultures of C3 in chitin broth were diluted 1:1 with water and Soydex. The "nal suspensions contained approx. 5;10 CFU/ml. In 1998 and 2000, two applications of the C3 treatment were made 10 days apart. In 1999, three applications C3 in chitin broth were made at 7-day intervals. The third application was made to replace cells washed o! of bean plants by heavy rain occurring 1 h after the second application.
G.Y. Yuen et al. / Crop Protection 20 (2001) 395}402
In 1998, propiconazole was applied twice at 14-day intervals, and two applications of thiophanate methyl were made 1 week apart beginning at full bloom. In 1999, there were two applications of propiconazole. Thiophanate methyl was used in combination with manganese ethylenebisdithiocarbamate (EBDC), applied as Maneb 75 DF (Elf Atochem North America; 1.7 kg/ha). The mixture was tested as a single application at full bloom or as two applications (full bloom, full bloom plus 14 days). In 2000, fungicide treatments were a single application of propiconazole and two applications of the thiophanate methyl#EDBC combination. There were non-treated controls in the 1998 and 2000 experiments, but not in 1999. Because of the concern that heavy inoculum development in non-treated control plots could cause arti"cially high infection levels or high variability within the experiment, non-treated controls were excluded to avoid interplot interference. The biocontrol treatment was compared only to standard disease control measures, as recommended by Van der Plank (1963). To obtain an estimate of disease potential in the absence of any control measure, rust severity was measured in four areas of non-treated bean surrounding the experiment, but these data were not included in the statistical analysis. 2.4. Statistical analysis Bacterial population numbers were converted to log (CFU#1) prior to statistical analysis. The results from initial greenhouse tests of bacterial strains and from "eld experiments were subjected to analysis of variance (ANOVA) for a randomized complete block design. ANOVA for a factorial design was used for the greenhouse experiments on the e!ects of chitin amendments to C3 cell suspensions. ANOVA for a split-plot design was used for analyzing results from greenhouse experiments comparing C3 chitin broth cultures with other C3 formulations, with bacterial treatments and time of pathogen inoculation being the main and sub-plot factors, respectively. Data from repetitions of an experiment were analyzed independently and then were pooled for analysis after testing for homogeneity of variance. Duncan's multiple range test or the LSD test was used for means separation.
3. Results 3.1. Ezcacy of bacteria in the greenhouse Of 127 bacterial strains evaluated in the greenhouse, few reduced disease incidence as compared to the phosphate bu!er control. Only P. agglomerans strain B1, isolated from a bean blossom, and S. maltophilia strain C3 reduced rust in more than one experiment (Table 1).
Table 1 Inhibition of bean rust in the greenhouse by foliar applications of selected bacterial stains Species and strain
Pantoea agglomerans B1 Pantoea agglomerans B346 Pseudomonas yuorescens C Stenotrophomonas maltophilia C3 Phosphate bu!er control
Uredinia per 12 cm area of leaf Expt. A
5.9 9.5 * *
10.2 28.3 15.4 *
* * * 14.6
* * 19.7 9.9
Values represent means of three measurements per leaf on four to six replicate leaves. Source of each strain are cited as follows: B1 (Yuen et al., 1991); B346 (Yuen et al., 1994); C3 (Giesler and Yuen, 1998). Strain C was isolated from sugar beet root and has not been reported. Denote signi"cant di!erence from the control at P"0.05 according to LSD test. Denote signi"cant di!erence from the control at P"0.01 according to LSD test. Not tested.
Both strains delayed maturation of uredinia by 1}2 days (data not shown). Others, such as P. agglomerans strain B346 and P. yuorescens strain C were e!ective in initial experiments, but not in subsequent testing (Table 1). 3.2. Ewects of chitin amendment and culture yuid on ezcacy of strain C3 Addition of colloidal chitin to suspensions of strain C3 increased the e$cacy of rust control as compared to the bacterium alone (Table 2). There was no signi"cant interaction between C3 and chitin treatments. Treatments with C3 at all concentrations of colloidal chitin resulted in lower (P"0.02) infection frequency in comparison to treatments with no bacteria. Additional chitin to C3 cell suspensions or to phosphate bu!er also reduced infection frequency (P"0.04), but there was no signi"cant di!erence between 0.2 and 1.0% chitin. At the time of inoculation, one day after bacterial application, the size of C3 populations with and without the chitin amendments were the same (data not shown). In the set of experiments comparing treatments with C3 in chitin broth culture to treatments with TSA-grown C3 cells suspended in PB with and without colloidal chitin, there was a signi"cant (P(0.001) treatment by inoculation date interaction. Treatment with chitin broth cultures of C3 resulted in lower rust severity than treatment with the other C3 formulations when plants were inoculated 1 day after bacterial applications (Fig. 1A). The chitin broth culture treatment also conferred the longest duration of control. When plants were inoculated with the pathogen 7 days after spraying with bacteria, only the chitin broth culture treatment signi"cantly
G.Y. Yuen et al. / Crop Protection 20 (2001) 395}402
Table 2 In#uence of the addition of colloidal chitin to cell suspensions of Stenotrophomonas maltophilia strain C3 on the inhibition of bean rust in the greenhouse Treatment
Strain C3 Phosphate bu!er Mean
Uredinia per 12 cm leaf area No chitin
0.2% w/v chitin
1.0% w/v chitin
60 96 78 A
29 66 48 B
37 75 56 B
Values followed by the same letter are not signi"cantly di!erent at P"0.05 according to LSD test. Values represent means of two experiments, each with three measurements per leaf on four replicate leaves. Cells of strain C3 were suspended in phosphate bu!er to 5;10 CFU/ml. Denotes treatment with C3 is signi"cantly di!erent (P"0.05) from the phosphate bu!er treatment.
reduced rust severity as compared to the bu!er-treated control. Population size of C3 on leaves treated with the broth culture remained signi"cantly higher than all of the other treatments throughout the experiment (Fig. 1B). There was no di!erence in C3 population size between treatments with cells applied in PB with or without chitin amendment. 3.3. Field results In four "eld experiments conducted in 1996 and 1997, none of the treatments with bacterial strains, alone or in combination with chitinous amendments, were consistently e!ective. Strain C3 alone provided a signi"cant reduction in rust severity compared to the non-treated control in NP 1996, but not in other experiments (Table 3). The bacterial treatment was not as e!ective as propiconazole. Chitosan and colloidal chitin each reduced rust severity in one experiment, but amendment of C3 suspensions with chitosan or colloidal chitin did not improve e$cacy. In the MIT 1997 experiment, the 0.2% chitin and chitin-amended C3 treatments resulted in rust severity levels intermediate between the phosphate bu!er control and propiconazole, but the same treatments did not reduce rust ratings in comparison to the bu!er control in NP 1997. There was no di!erence in rust ratings between treatments with P. agglomerans strain B1 or strain B409 and the respective controls. Among the treatments tested in 1996 and 1997, only propriconazole in 1996 and thiophanate methyl in PREC 1997 increased yields relative to the controls (data not shown). Populations of the applied bacteria on leaves were monitored during the MIT 1997 experiment (Fig. 2). Within 6 days after each application, population levels of strain B1 declined from around 7 log CFU/leaf to
Fig. 1. E!ects of culture and application methods on (A) e$cacy of bean rust control by Stenotrophomonas maltophilia strain C3 and (B) colonization of pinto bean (&UI 114') leaves by strain C3 under greenhouse conditions. Cells of C3 suspended in phosphate bu!er were grown on TSA medium. Colloidal chitin was used at 0.2% w/v. All treatments were applied at one time and then inoculated with urediniospores 1 and 7 days later. The modi"ed Cobb scale was used for severity ratings. In (A), data points are means of two experiments with four replications each, and results from an inoculation date with the same letter are not signi"cantly di!erent at P"0.01 according to LSD test. In (B), day 0, 3, and 7 data points are means of two experiments with three replications each; day 1 and 10 values are means of three replications from one experiment. Vertical bars denote standard deviation.
around 4 log CFU/leaf. C3 populations declined more rapidly in the same time period to less than 2.5 log CFU/leaf. Application of chitin with C3 cells had no e!ect on leaf colonization by the strain. Chitin broth cultures of C3, tested in 1998}2000, were comparable to treatments with labeled fungicides for control of bean rust (Table 4). In 1998, both the C3 and thiophanate methyl treatments had less (P"0.05) rust
G.Y. Yuen et al. / Crop Protection 20 (2001) 395}402
Table 3 E!ects of bacterial and fungicide treatments on bean rust severity determined in western Nebraska "eld experiments Year and site
Number and schedule of applications
Modi"ed Cobb scale rating (1}7)
Non-treated control Stenotrophomonas maltophilia C3 C3#chitosan
4 at 7 day intervals 4 at 7 day intervals 4 at 7 day intervals 2 at 14 day intervals 4 at 7 day intervals 4 at 7 day intervals 4 at 7 day intervals 4 at 7 day intervals 4 at 7 day intervals 2 at 14 day intervals 4 at 7 day intervals 4 at 7 day intervals 4 at 7 day intervals 4 at 7 day intervals 4 at 7 day intervals 2 at 14 day intervals None
4 at 7 day intervals 4 at 7 day intervals 4 at 7 day intervals 2 at 14 day intervals 2 at 14 day intervals
Chitosan Propiconazole 1997 MIT
Phosphate bu!er control Pantoea agglomerans B1 C3 C3#colloidal chitin Colloidal chitin Propiconazole
Saline control B1 C3 C3#colloidal chitin Colloidal chitin Propiconazole
Non-treated control P. agglomerans B409 C3 B409#C3# colloidal chitin Colloidal chitin Thiophanate methyl
2.6 AB 2.5 B 1.0 C 3.2 A 2.8 AB 2.8 AB 2.2 BC 2.0 C 1.0 D
Fig. 2. Population size of Pantoea agglomerans strain B1 and Stenotrophomonas maltophilia strain C3 on leaves of Phaseolus vulgaris Great Northern &Marquee' in a "eld at PREC 1997 measured after each of three applications. Colloidal chitin was used to amend strain C3 at 0.2% w/v. N"3 for each data point. Vertical bars denote standard deviation.
2.8 A 2.8 A 2.8 A 2.5 A 2.5 A 2.8 A
Table 4 In#uence of chitin broth cultures of Stenotrophomonas maltophilia strain C3 on bean rust severity in "eld experiments in North Platte, NE Treatment
4.5 A 4.4 A 4.0 A 1.4 B
NP, MIT, and PREC denote North Platte, Mitchell, and Scottsblu!, respectively. Phosphate bu!er was used at MIT and PREC and saline was used at NP to suspend bacterial cells, chitosan (Insectinet) and chitin. Bacterial suspensions contained around 10 CFU/ml. Chitosan and chitin were added to bu!ers or cell suspensions at 0.2% w/v. Propiconazole was applied as Tilt at 290 ml/ha. Thiophanate methyl was applied as Topsin M 70 W at 1.1 kg/ha. Applications were made at 50 and 100 ml per m row in 1996 and 1997, respectively. The initial application for all treatments in 1996 NP, 1997 NP and 1997 MIT was made at the "rst sign of rust. Initial applications in 1997 PREC were made at full bloom. Values are means of four replications. Means within an experiment with the same letter are not signi"cantly di!erent at P"0.05 according to Duncan's multiple range test.
Non-treated control C3 chitin broth culture Propiconazole Thiophanate methyl Thiophanate methyl#EDBC } 1 application Thiophanate methyl#EDBC } 2 applications
Modi"ed Cobb scale rating (1}7) 1998
5.0 A 3.0 B 1.8 C 3.8 B Not tested
(4.5) 3.8 AB 2.2 C Not tested 4.2 A
2.7 A 1.7 B 1.3 B Not tested Not tested
Results with the same letter in a column are not signi"cantly di!erent at P"0.05 according to Duncan's multiple range test. Seven-day-old cultures of C3 were diluted 1:1 with water to 5;10 CFU/ml prior to spray application. In 1998 and 2000, two applications were made 10 days apart. In 1999, there were three weekly applications. Thiophanate methyl"Topsin M 70 W at 1.1 kg/ha, two applications. EBDC"manganese ethylenebisdiathiocarbamate applied as Maneb 75 DF at 1.7 kg/ha. Propiconazole"Tilt at 290 ml/ha, two applications in 1998 and 1999, one in 2000. All treatments with thiophanate methyl were initiated at full bloom. All other treatments were initiated at the "rst sign of rust. Second fungicide applications were made 2 weeks after the "rst. Applications were made at 50 ml per m row. Non-treated control data in 1999 were obtained from non-treated areas surrounding the experiment but not included in the statistical analysis.
G.Y. Yuen et al. / Crop Protection 20 (2001) 395}402
than the control. In 1999, rust severity in the C3 treatment was intermediate between that of one or two applications of the thiophanate methyl#EBDC combination. There was no non-treated control in the 1999 experiment, but rust severity in the C3 treatment was lower than that in surrounding, non-treated areas. In both years, treatment with a C3 chitin broth culture was not as e!ective as propiconazole. In 2000, disease severity in the non-treated control was low. C3 in chitin broth culture, the thiophanate methyl#EDBC combination, and propriconazole reduced disease severity to similar levels in comparison to the non-treated control. There were no signi"cant di!erences in seed yield among treatments in 1998 and 1999 experiments (data not shown).
4. Discussion There is considerable precedence for using bacteria to control rust diseases. In this study, a strain of S. maltophilia was found to have potential as a bean rust control agent, and control could be achieved with application schedules that are commercially feasible. To achieve "eld e$cacy we had to develop formulation strategies based on our knowledge of the mechanisms of action by the strain. Foliar applications of chitin to peanut were reported to enhance biocontrol of early leaf spot with a chitinolytic strain Bacillus cereus by providing a nutrient source for the applied bacterium and resident chitinolytic microbes (Kokalis-Burelle et al., 1992). Strain C3 can multiply in vitro on chitin as the sole carbon source, and application of chitin with C3 on tall fescue did improve control of Bipolaris leaf spot (Zhang and Yuen, 1999). However, foliar applications of chitin to bean in this study were not e!ective in improving control of bean rust by C3 in the "eld. As in tall fescue, application of the chitin with C3 to bean did not increase numbers of C3, leading to the conclusion that chitin was not important as a nutrient substrate for growth of C3 on the phylloplane. Chitin probably served, instead, to induce production of chitinase and other lytic enzymes by C3 that may have a!ected the rust fungus (Zhang and Yuen, 2000a). Because chitin applied alone to bean leaves provided some rust control, it also may have had a role in stimulating resident chitinolytic microorganisms, physically disrupting pathogen spore germination, or inducing host resistance, as implicated in other studies (Kokalis-Burelle et al., 1992; Pearce and Ride, 1982). Thus, the failure of the C3-chitin combination in the "eld could be attributed to poor colonization of bean foliage by C3, which the addition of chitin could not ameliorate. C3 sustained higher population sizes for longer periods in turfgrass canopies than in our bean studies (Giesler and Yuen, 1998; Zhang and Yuen, 1999). The di!erence between
bean and tall fescue as hosts for colonization by C3 is not due to a preference for one plant species over the other. When strain C3 was applied to pinto bean and tall fescue simultaneously in growth chamber experiments, there was no di!erence in colonization of the two hosts (Yuen, unpublished data). The limiting factor in bean is most likely the canopy microclimate. Canopy humidity and leaf temperatures in bean "elds in western Nebraska #uctuate considerably on a diurnal cycle (Deshpande et al., 1995), whereas turfgrass canopies have longer durations of high canopy humidity and leaf wetness, and also a narrower range of leaf temperatures (Giesler et al., 1996). By culturing strain C3 in chitin broth and applying the entire contents of the culture, propagation and application of the biocontrol agent were simpli"ed and e$cacy also was improved. Enhancement of e$cacy could be due in part to the direct antifungal e!ects of factors produced by strain C3 in broth culture, including multiple lytic enzymes (Zhang and Yuen, 2000a). Antibiotics also might be involved, as strain C3 produces low molecular weight, antifungal compounds in vitro (Zhang and Yuen, 2000b). Enzymes and antibiotics excreted by C3 into the culture #uid also could have enhanced leaf colonization by strain C3 through the conversion of complex substrates and reduction in numbers of competing resident phylloplane microorganisms. The cultures #uid also could have contained residual oligomeric products from digestion of chitin that served as immediate nutrient substrates for C3 cells deposited on leaf surfaces. The products of chitin enzymolysis in the cultures also may have elicited systemic acquired resistance in the bean plants (Vander et al., 1998). Although treatment with chitin broth cultures of C3 was not as e!ective as propiconazole in controlling bean rust under moderate disease conditions, the treatment does compare favorably with currently registered fungicides. The bene"ts of applying the bacterium in combination with these fungicides have yet to be determined. With further modi"cations, such as removal of water from broth cultures to reduce shipping volume, the C3 system might be rendered more economically practical. Before the potentials for commercialization of the strain can be ascertained, however, safety testing is needed, as strains of S. maltophilia have been isolated from the lungs of patients su!ering from cystic "brosis (Spencer, 1995). Strain C3 di!ers from clinical strains in a number of traits, including inability to grow at 373C (Yuen, unpublished data), which would suggest that it is not pathogenic to humans. Cell-free #uid from C3 chitin broth cultures also reduced the severity of Bipolaris leaf spot on tall fescue in the "eld (Zhang and Yuen, 2000a). Although cell-free #uids were not tested in this study, further research on the compounds produced by C3 in chitin broth culture might yield materials useful in controlling bean rust.
G.Y. Yuen et al. / Crop Protection 20 (2001) 395}402
Acknowledgements We thank Mischell Craig, Zhongge Zhang, Janelle Fenton, and Kathyrn Hartwig for assisting in laboratory and greenhouse experiments, and Clay Carlson for providing technical assistance in the "eld. This research was funded in part by a grant from the Nebraska Dry Bean Commission. References Allen, D.J., 1982. Verticillium lecanii on the bean rust fungus, Uromyces appendiculatus. Trans. Brit. Mycol. Soc. 79, 362}364. Baker, C.J., Stavely, J.R., Mock, N., 1985. Biocontrol of bean rust by Bacillus subtilis under "eld conditions. Plant Dis. 69, 770}772. Baker, C.J., Stavely, J.R., Thomas, C.A., Sasser, M., MacFall, J.S., 1983. Inhibitory e!ect of Bacillus subtilis on Uromyces phaseoli and on development of rust pustules on bean leaves. Phytopathology 73, 1148}1152. Centurian, M.A.P.C., Kimati, H., 1994. Controle biologico de ferrugem do feijoeiro com bacterias antagonicas. Summa-Phytopathologica 20, 179}183. Deshpande, R.Y., Hubbard, K.G., Coyne, D.P., Steadman, J.R., Parkhurst, A.M., 1995. Estimating leaf wetness in dry bean canopies as a prerequisite to evaluating white mold disease. Agron. J. 87, 613}619. Giesler, L.J., Yuen, G.Y., 1998. Evaluation of Stenotrophomonas maltophilia strain C3 for biocontrol of brown patch disease. Crop Prot. 17, 509}513. Giesler, L.J., Yuen, G.Y., Horst, G.L., 1996. The microclimate in tall fescue turf as a!ected by canopy density and its in#uence on brown patch disease. Plant Dis. 80, 389}394. Gowdu, B.J., Balasubramanian, R., 1988. Role of phylloplane microorganisms in the biological control of foliar plant diseases. Z. P#Krankh. P#Schutz 95, 310}331. Grabski, G.C., Mendgen, K., 1985. Einsatz von V. lecanii als biologisches Schandlingsbekamp!ungsmittel gegen den Bohnenrostpilz U. appendiculatus var. appendiculatus im Feld und im Gewachshaus. Phytopathol. Z. 113, 243}251. James, R.V., Stevenson, W.R., Parke, J.L., Rand, R.E., Ra!el, S.J., Yuen, G.Y., 1995. Evaluation of biological control of white mold on snap bean, 1994. Biol. Cult. Tests 10, 145.
Je!ries, P., Young, T.W.K., 1994. Interfungal Parasitic Relationships. CAB International, Wallingford, Oxon, pp. 164}168. Kokalis-Burelle, N., Backman, P.A., Rodriguez-Kabana, R., Ploper, L.O., 1992. Potential for biological control of early leafspot of peanut using Bacillus cereus and chitin foliar amendments. Biol. Control 2, 321}328. Kranz, J., 1981. Hyperparasitism of biotrophic fungi. In: Blakeman, J.P. (Ed.), Microbial Ecology of the Phylloplane. Academic Press, London, pp. 327}352. Mizubuti, E.S.G., Ma$a, L.A., Muchovej, J.J., Romeiro, R.S., Batista, U.G., 1995. Selection of isolates of Bacillus subtilis with potential for the control of dry bean rust. Fitopatol. Bras. 20, 540}544. Pearce, R.B., Ride, J.P., 1982. Chitin and related compounds as elicitors of the ligni"cation response in wounded wheat leaves. Physiol. Plant Pathol. 20, 119}123. Spencer, R.C., 1995. The emergence of epidemic, multiple-antibiotic resistant Stenotrophomonas (Xanthomonas) maltophilia and Burkholderia (Pseudomonas) cepacia. J. Hospital Infect. 30, 453}464. Stavely, J.R., 1985. The modi"ed Cobb scale for estimating bean rust intensity. Ann. Rep. Bean Improv. Coop. 28, 31}32. Stavely, J.R., Pastor-Corrales, M.A., 1989. Rust. In: Schwartz, H.F., Pator-Corrales, M.A. (Eds.), Bean Production Problems in the Tropics, 2nd Edition. CIAT, Cali, Colombia, pp. 159}194. Vander, P., Varum, K.M., Domand, A., El Guedarri, N.E., Moerschbacher, B.M., 1998. Comparison of the ability of partially Nacetylated chitosans and chitooligosaccharides to elicit resistance reactions in wheat leaves. Plant Physiol. 118, 1353}1359. Van der Plank, J.E., 1963. Plant Diseases: Epidemics and Control. Academic Press, New York, pp. 285}310. Yuen, G.Y., Craig, M.L., Kerr, E.D., Steadman, J.R., 1994. In#uences of antagonist population levels, blossom development stage and canopy temperature on the inhibition of Sclerotinia sclerotiorum on dry edible bean by Erwinia herbicola. Phytopathology 84, 495}501. Yuen, G.Y., Godoy, G., Steadman, J.R., Kerr, E.D., Craig, M.L., 1991. Epiphytic colonization of dry edible bean by bacteria antagonistic to Sclerotinia sclerotiorum and potential for biological control of white mold disease. Biol. Control 1, 293}301. Zhang, Z., Yuen, G.Y., 1999. Biological control of Bipolaris sorokiniana on tall fescue by Stenotrophomonas maltophilia C3. Phytopathology 89, 817}822. Zhang, Z., Yuen, G.Y., 2000a. E!ects of culture #uids and preinduction of chitinase production on biocontrol of Bipolaris leaf spot by Stenotrophomonas maltophilia C3. Biol. Control 10, 277}286. Zhang, Z., Yuen, G.Y., 2000b. The role of chitinase production by Stenotrophomonas maltophilia strain C3 in biological control of Bipolaris sorokiniana. Phytopathology 90, 384}389.