CONTROL OF FIRE BLIGHT, FROST INJURY, AND FRUIT RUSSET USING CULTURAL, CHEMICAL AND BIOLOGICAL CONTROLS

PRINCIPAL INVESTIGATORS:
Steven E. Lindow, Dept. of Environmental Science, Policy, and Management,
U.C. Berkeley, California 94720-3110; E-mail: icelab@violet.berkeley.edu
COOPERATORS:
Glenn McGourty, Cooperative Extension, Mendocino County
Rachel Elkins, Cooperative Extension, Lake County.

ABSTRACT

Considerable information was obtained during the 1994 growing season relating to the integrated control of fire blight disease using antibiotics and the biological control agent Pseudomonas fluorescens strain A506, as well as the cultural and chemical control of pear fruit russeting. In a large replicated trial in which the biological control agent was applied by a cooperating grower, the incidence of fire blight infections was reduced to the same extent as a full antibiotic program using a substantially reduced frequency of antibiotic sprays when the biological control agent was also periodically applied. In addition, the incidence of fire blight disease was reduced about three-fold more when the biological control agent was applied to trees which received the normal frequency of antibiotic sprays than to trees receiving antibiotics alone. These results indicate that the control of fire blight disease using the biological control agent is additive to that achieved by antibiotic sprays; reduced frequencies of antibiotic sprays still yielded acceptable fire blight control in the presence of the biological control agent. A new and simple assay for detecting the presence of the biological control agent in pear flowers was developed. Rubbing of detached flowers onto a selected culture medium permitted the rapid and simple assessment of the occurrence of this bacterium in treated orchards. This assay should facilitate the monitoring of biological control agent when it becomes commercially available in 1995. Different cover crop species were established under the trees in large replicated blocks in a commercial pear orchard by direct seeding in the fall. Substantial differences in fruit russeting were observed on trees in these blocks at harvest. The two legumes, burr clover and crimson clover were well established by seeding, whereas several perennial clovers such as berseem and red clover were not well established in these trials. The clovers generally supported substantially lower populations of bacteria of all kinds than did other weedy broadleaf plants and grasses. The population size of bacteria in pear trees was substantially higher when trees were growing over cover crops, including a substantial percentage of grass species. These results suggest that management of cover crops, either removing them by cultivation or herbicide application or establishing non-grass cover crops can substantially improve fruit finish. The severity of fruit russeting was also reduced about three-fold comparing untreated trees to trees treated three times within the month following petal fall with various nitrogenous compounds. 200 ppm to 400 ppm of either ammonium sulfate, urea, or calcium nitrate all reduced the severity of fruit russeting, apparently by inhibiting the production of the plant growth regulator 3-indoleacetic acid (IAA) by epiphytic strains of the bacterium Erwinia herbicola. The application of nitrogenous compounds to improve the finish of pear is promising, however additional work is needed to determine whether the application of nitrogenous compounds does not increase the growth of deleterious bacteria such as the fire blight pathogen.

BIOLOGICAL AND CHEMICAL CONTROL OF FIRE BLIGHT AND FROST INJURY

Considerable research effort in 1994 was directed to better understanding the additive effects of biological control of fire blight using the antagonistic bacterium Pseudomonas fluorescensstrain A506 and chemical control by streptomycin or Terramycin. Therefore, with supplemental funding from the Smith-Lever project of the California IPM program, a large replicated block was established in Scotts Valley, Lake County, in which after full bloom, the frequency of antibiotic sprays was reduced about 50% below that of the normal program used by the grower (7 sprays versus 13 sprays). In addition, trees treated with either frequency of antibiotics were also divided into two groups and received the biological control agent or not. By this design we could evaluate the relative disease control to trees which received a reduced frequency of antibiotics as well as the biological control agent, and compare that with trees receiving only a normal frequency of antibiotic sprays, or to trees receiving both a normal frequency of antibiotic sprays and the biological control agent.

The biological control agent grew well on inoculated trees. For example, while small numbers of the biological control agent established themselves on trees which received the normal frequency of antibiotic sprays but were not directly inoculated with the strain A506 (Figure 1) (presumably due to drift of spray material and/or movement via insects from the treated to the untreated area), approximately 50% of the total bacteria on antagonist-treated trees were the antagonistic bacterium (Figure 2). Similarly, the population size of the biological control agent on trees treated with a reduced frequency of antibiotics but not treated with the biological control agent was low (Figure 3), but at least 10% of the total bacteria on trees treated with strain A506 were of this antagonist (Figure 4). Clearly the biological control agent was compatible with the oversprays of the antibiotics that were applied to the trees. In addition, the grower applied various fungicides on a normal frequency for scab control to this block. The biological control agent was apparently uninhibited by the application of other chemicals to the trees.

The biological control agent itself was applied at full dose on three occasions (once at about 20% bloom, once near full bloom, and once about 10 days after full bloom) with an additional partial application during the rat-tail bloom period. It is noteworthy that the total bacteria on trees treated with the biological control agent were always at least 10-fold higher than that on untreated control trees (compare Figure 2 with Figure 1, and compare Figure 4 with Figure 3.). It thus appears that strain A506 is a superior epiphytic colonist. It appears that the normal pattern of colonization of pear flowers by bacteria in most orchards is by migration of airborne bacteria, presumably from nearby sources such as cover crops. If the bacteria are good epiphytes, such as strain A506 or the fire blight pathogen and the flowers are receptive to colonization, these migrants will grow well. Apparently many bacteria are inferior to strain A506 in their ability to grow in flowers since flowers treated with this antagonist often have far higher populations than in uninoculated flowers. The rapid growth of P. fluorescens strain A506 in inoculated flowers therefore, seems to account for its ability to inhibit the growth of the fire blight pathogen as well as other bacteria including ice nucleation-active species.

A new method to detect the presence of the biological control agent on individual flowers was developed during 1994. Since it is frequently more important to know whether an individual flower is colonized by the biological control agent than to know precisely how many bacteria are in an individual flower, we developed a "flower rub" assay to quickly determine the presence or absence of the biological control agent. In the "flower rub" assay individual flowers are detached and rubbed ( pistil side down) on the surface of a nutrient agar medium containing 100 4g/ml of rifampicin, a synthetic antibiotic to which the biological control agent is resistant but to which few, if any, other bacteria that occur in pear orchards are resistant. Using this assay we were able to determine that the incidence of colonization of pear flowers in our plots was usually in excess of 80%. On many occasions during the main bloom, nearly 100% of all the flowers of inoculated trees were colonized by strain A506. Flowers collected after petal fall tended to have lower incidence of colonization than did those collected during the main bloom period. Because up to 10 individual flowers can be assayed on an individual petri dish, and since it only takes a few seconds to rub a given flower on to a sample dish, this assay appears to have much utility in allowing PCAs and others to determine whether the biological control agent has successfully been applied to the trees and is colonizing flowers. We anticipate further demonstrating this assay during the 1995 growing season with PCAs and interested growers who will apply P. fluorescens strain A506 when it becomes commercially available.

Substantial amounts of infection by E. amylovora occurred in the plot area in which A506 was evaluated. The incidence of infection in our plot was lower than in other regions in Lake County but provided sufficient numbers of fire blight infections to evaluate the interactions of the biological control agent and chemical bactericides in controlling fire blight disease. While the incidence of fire blight infection was about three times higher on trees receiving only a reduced frequency of antibiotic sprays compared to those receiving only a normal frequency of antibiotic sprays (Table 1), the incidence of fire blight infection was reduced on trees receiving both A506 and reduced numbers of antibiotic sprays to a level similar to that of trees receiving the normal frequency of antibiotic sprays alone (Table 1). In addition, the incidence of fire blight infection was reduced approximately 3-fold by applying the biological control agent in conjunction with the normal frequency of antibiotic sprays when compared to the control on trees receiving the normal frequency of antibiotic sprays alone (Table 1). It was interesting to note that strain A506 reduced the incidence of infection about 3-fold irrespective of whether it was applied to trees receiving either full or a reduced frequency of antibiotic sprays (Table 1). This confirms our results from previous studies using smaller plots; it appears that the biological control and the chemical control of fire blight are completely additive. These results are encouraging in that they suggest that growers may be able to reduce the number of antibiotic sprays applied and still achieve acceptable levels of fire blight control by applying the biological control agent. The reduction of the number of antibiotic sprays, particularly those which might normally have been applied at a time when fire blight risk was not high, could allow cost savings.

Substantial amounts of frost entry were observed in the plot area in which A506 was evaluated in 1994. While the frost damage was quite extensive, severity of frost damage was usually quite low. The most typical symptom of frost damage in this plot area was small brown spots, sometimes sunken, near the calyx end of the fruit. It appeared that during bloom, when the fruit was extremely small, mild frost caused blistering of the epidermal cells of the young fruits. The application of A506 reduced the incidence of frost damage in this plot area about threefold compared to that to trees for which no A506 was applied (Table 2). It was interesting to note that the incidence of frost damage was not reduced on trees receiving a normal frequency of antibiotic sprays compared to those receiving a reduced frequency of antibiotics (Table 2). Our previous studies had indicated that a substantial number of the ice nucleation active strains of Pseudomonas syringae, which appear primarily responsible for causing ice formation required for frost damage in pear orchards, are resistant to streptomycin and/or oxytetracycline. Thus, as in previous years, we have demonstrated that the application of the biological control agent strain A506 can yield several benefits to fruit production. Not only can this bacterium reduce the incidence of fire blight disease but it can also reduce frost damage and, as will be shown below, the severity of fruit russeting.

In the same orchard a replicated trial was established to determine whether fungicides applied to pear trees in the early spring have the potential to inhibit the growth or survival of Pseudomonas fluorescens strain A506 on trees. In this plot, the biological control agent strain A506 was applied to trees at about 50% bloom and allowed to colonize trees for one week. One week after applying the antagonistic bacteria, trees were over-sprayed with commercial rates of several different fungicides (listed in Table 3). Pear spurs were then sampled from tree to tree one week after applying the chemicals (April 14), or two weeks following chemical application (April 21). A few fungicides such as Dithane M45, Ziram and Aliette reduced the population size of the biological control agent only transiently; population size of strain A506 was reduced on samples taken shortly after chemical sprays, but recovered similar to populations on trees treated only with the biological control agent by April 21 (Table 3). In contrast, population size of the biological control agent was substantially and continually depressed by the application of copper-containing compounds such as Kocide 101, sulfur, and oil. While we had expected that copper-containing compounds would be inhibitory to the biological control agent, and suspected that sulfur would also be inhibitory, it was curious to observe that oil sprays reduced the population size of this bacterium. Fortunately, the most commonly used fungicides, such as Dithane M45 and Ziram, appeared not to have substantial or continuing effect on the population size of the biological control agent. Thus it appears that if the application of these materials can be delayed for at least a few days after the application of the biological control agent, then they should not negatively impact its performance in the field. In contrast, copper-containing bactericides appear not to be compatible with the biological control agent. We plan to re-test spray oils to determine if they are reproducibly inhibitory to the biological control agent. In summary, the biological control agent appears compatible with most materials used in orchards. Work from researchers at Oregon State University also reveal that this biological control agent is rather robust and can tolerate the application of chemicals as long as the bacterium has established itself on trees. Therefore application of fungicides very soon after application of the bacteria may be unwise, but the application of most pesticides after the bacterium has been on the tree for a few days appears safe.

The severity of fruit russeting on trees treated with the various fungicides tested for compatibility with strain A506 was evaluated at harvest. As expected, the copper containing compound Kocide 101 increased the severity of fruit russeting compared to untreated control trees. Curiously however, Moristan also substantially increased fruit russeting. The severity of fruit russeting on trees treated with other fungicides was not significantly different from untreated trees (Table 4).

The biological control agent Pseudomonas fluorescens strain A506 significantly reduced fruit russeting on trees in all trials during 1994. The severity of fruit russeting was rather low in all plot areas in 1994. Quantitative measurements of fruit russeting however, revealed that russeting was approximately 50% less severe on trees treated with strain A506 than on trees receiving antibiotics alone in the large plot in Scotts Valley (Table 2). Similarly, in another large plot near Hopland in which large replicated blocks of trees were treated two times with strain A506, the severity of fruit russeting was also reduced approximately 50% (Table 5). These results confirm the findings over the last several years that the biological control agent commonly reduces fruit russeting. It is believed that fruit russeting is reduced by competitive displacement of strains of Erwinia herbicola which produced the plant growth hormone 3-indoleacetic acid (IAA). As was shown earlier, strain A506 colonizes treated pear trees to high levels and trees having high populations of strain A506 apparently cannot support as high a population size of IAA-producing bacteria. While strain A506 was not selected for its ability to inhibit IAA-producing bacteria, it appears to frequently do so, and increased fruit quality commonly observed on trees treated with this biological control agent can be anticipated in any orchard in which it is applied.

The ability of another biological control agent, Erwinia herbicola strain C9-1 to colonize pear flowers and control fire blight was compared to that of strain A506 and other bactericides such Aliette. Strain C9-1 established populations sizes on treated trees that was similar to that of strain A506 (compare Figures 14 and 13). The combined population size of these two strains when co-inoculated onto pear trees was similar to either strain alone (Figure 15); no strong evidence for antagonism of either strain by the other was obtained. While the population size of ice nucleation active (Ice+) bacteria on trees treated with strain A506 either alone (Figure 13) or in combination with strain C9-1 (Figure 15) was substantially reduced compared to that on untreated control trees (Figure 8), the reduction of these populations on trees treated with strain C9-1 alone was substantially less (Figure 14). While total bacterial populations on trees treated with Aliette were not reduced as in previous studies, Aliette did cause a substantial reduction in populations of Ice+ bacteria (Figure 9). The incidence of fire blight disease and frost injury in this plot area was too low for meaningful comparisons of fire blight disease or frost injury.

CULTURAL CONTROL OF PEAR FRUIT RUSSETING

Considerable work was continued during 1994 to further ascertain the role of under-tree cover crop species in contributing inoculum of bacteria that become established as epiphytic populations on pear fruit and induce fruit russeting. A large replicated trial was established in a commercial pear orchard near Hopland in which approximately three-acre blocks were planted to different cover crop species. Several different leguminous cover crop species were planted by seed with a no-till drill in the fall of 1993. Leguminous species were selected because they had exhibited relatively low epiphytic bacterial populations in previous studies. The following legumes were planted: crimson clover; a mixture of Austrian winter pea and common vetch; a mixture of annual clovers, including burr clover; red clover; a mixture of perennial clovers; and berseem clover. An untreated control was also included. By April 1, crimson clover and burr clover had become established as relatively large and dominant plants in the seeded area. As much as 90% of the soil surface was covered with either of these two plant species. In contrast, poor establishment of the mixture of pea and vetch, and perennial and berseem clovers was observed. Red clover plants were relatively small on April 1 but had increased in size and were reasonably dominant by about June 1. Because Roundup had been applied in the fall at the time of planting, few other weed species were present in any of the seeded areas; only small amounts of pepper cress, malva, and groundsel occurred in the seeded areas. In contrast, in the unseeded control area substantial numbers of different weedy plant species were present, including annual bluegrass, dandelion, annual ryegrass, and various other broadleaf plant species.

A wide range of bacterial populations were observed on the different weedy plant species occurring as cover crops in the Hopland trial. Grass species including annual bluegrass, Bermuda grass, annual ryegrass and other grass species tended to have substantially higher populations of bacteria than did other broadleaf plants, particularly leguminous species. Dandelion also had relatively high bacterial population sizes relative to other broadleaf plants. The population size of bacteria of all kinds on crimson clover, red clover, burr clover, and berseem clovers were about 20-fold lower than that on annual bluegrass and other grass species (Table 6). Similarly, wide ranges of population sizes of Ice+ were also observed on these species; the population size of ice nucleation active bacteria tended to be proportional to the total bacterial populations. Plant species that had relatively high populations of total bacteria also had relatively high populations of Ice+ bacteria (Table 6), although a few exceptions were noted.

Measurable differences in the population size of bacteria of all types on pear trees occurred when trees were grown above different cover crop treatments (Figure 5). Highest bacterial populations on pear trees, particularly early in the growing season (within the first 30 days after first bloom), were found on trees growing above the control plot which included grass species and dandelions which harbored large populations of bacteria (Figure 5). In contrast, the population sizes of bacteria on pear trees are generally lower on trees growing above bare soil where cover crop species were removed by Roundup and paraquat applications and above dense stands of burr clover (Figure 5). Population size of bacteria on trees growing above bare soil or burr clover was as much as 10-fold lower than that on trees growing above a diverse collection of plant species. The population size of bacteria on trees growing above red clover, berseem clover and perennial clovers (which were not well established in the orchard) were generally intermediate between that on the control and burr clover treatments (data not shown). These results suggest strongly that bacteria in pear trees, particularly early in the growing season, come from the green cover crop plants nearby. While this replicated plot consisted of relatively large contiguous areas of a given treatment (about three acres) it also demonstrates clearly that the sources of bacteria that colonize pear trees appear to be rather local. For example, had most of the bacteria that migrate to pear trees come from relatively long distances away, we would not have expected differences in population size of bacteria on these closely-adjacent blocks of trees. Since we do see substantial differences in population size of bacteria on the trees, which correlate with the population size of bacteria on the cover crops underneath the trees, we can surmise that the bacteria have moved from the local cover crop to the pear trees.

Significant differences in fruit russeting were observed on trees grown above different cover crop treatments (Table 7). The highest severity of fruit russeting was observed on trees grown over a diverse collection of grasses and weedy plants in the control area whereas the lowest severity of fruit russeting was observed where all cover crops had been removed by Roundup treatment. It is noteworthy that fruit russeting was not significantly different from the Roundup treatment in plots in which burr clover, a mixture of pea and vetch, and crimson clover had been established (Table 7). In general, as much as a 4-fold reduction in the severity of fruit russeting was observed in plots in which cover crop species were either absent or where the cover crop lacked plant species which harbored high population sizes of bacteria.

Further evidence that bacteria had migrated from the cover crop species to the pear and established populations there was obtained by relating the numbers of bacteria on the cover crop species with the fruit russeting in the pear trees above the cover crops. In late April, detailed measurements were made of the percentage of the soil area between the drip lines of the trees that was covered with different plant species. The occurrence of each of the plant species shown in Table 6 was assessed in each of the 24 replicated blocks present in this large plot. The average population size of bacteria on each of the 23 plant species that were found as cover crop species in any of the replicated blocks was also measured as described earlier (Table 6). Therefore by multiplying the percent of the soil area covered with a given plant species by the relative number of bacteria found on that plant species, an aggregate estimate of the total bacteria that the cover crop as a whole possessed could be calculated. This estimate of total (Figure 6 horizontal axis), was regressed against the severity of fruit russeting for each of the 24 plots (Figure 6, vertical axis). A rather striking linear relationship was observed between increasing bacteria in the cover crop and increasing fruit russet severity during 1994 (Figure 6). These results show clearly that the abundance of the cover crop is not nearly as important as the species composition of the cover crop in contributing to fruit russeting. The presence of even a relatively low number of plants which have relatively high population sizes of bacteria can contribute bacteria that will lead to fruit russeting.

The severity of fruit russeting in most pear-growing areas in California was relatively low in 1994 compared to that in 1993. The relatively dry early spring in 1994 presumably accounts for much of this difference. We also observed, however, that the population sizes of bacteria on both pear trees and on cover crop species were substantially lower in 1994 than in 1993. Presumably weather conditions strongly influenced the epiphytic population sizes of bacteria. Therefore to better understand whether seasonal differences in fruit russeting are due to seasonal differences in epiphytic populations of bacteria (either on cover crop species or on the pear trees themselves) we combined the data relating aggregate population sizes of bacteria on cover crops in the Hopland plot in 1993 and 1994 with fruit russet severity in this site (Figure 7). The maximum level of fruit russeting observed in 1994 in any of the 24 plots was lower than the lowest severity of fruit russeting observed in any plot in 1993 (Figure 7). However, it is apparent that the relationship between total bacteria in the cover crops and severity of fruit russeting is a continuous relationship; the low relative severity of fruit russeting observed in 1994 were expected based on the low relative population sizes of bacteria observed in this year. Taken together, these data all strongly suggest that both seasonal and regional variations in the severity of fruit russeting are probably all caused by differences in the population size of IAA-producing bacteria on the pear trees. Population sizes of bacteria on pear trees can be influenced both by influencing the source of inoculum of these bacteria from cover crop species as well as their ability to grow on the trees. Both apparently can be influenced by weather (over which growers have little control) but the selection or maintenance of appropriate cover crop species apparently can have a large influence on the severity of fruit russeting.

CHEMICAL CONTROL OF FRUIT RUSSETING

Basic studies in our laboratory at Berkeley into the biochemical pathways by which strains of Erwinia herbicola produce the plant growth regulator 3-indoleacetic acid (IAA) reveal that the first step in its biosynthesis is the degradation of the amino acid tryptophan, whereby a nitrogen-containing group is removed from this molecule. This laboratory research further indicated that when provided with adequate amounts of nitrogen in culture media, these bacteria do not produce significant amounts of IAA. These laboratory findings suggested that IAA biosynthesis on fruit surfaces could also be inhibited if erogenous nitrogen-containing compounds were applied to trees. Very encouraging results were obtained in 1994 to suggest that this conjecture is true. In a plot in Lake County Erwinia herbicola strain 299R which produces IAA was established on trees. Trees were then sprayed weekly with either a solution of 500 ppm urea or 300 ppm ammonium sulfate. A total of five nutrient sprays were applied. Nutrient sprays were initiated at about full bloom. Compared to the control where Erwinia herbicola strain 299R alone was applied, the severity of fruit russeting was reduced about 4-fold on trees to which either urea or ammonium sulfate were applied (Table 8). These results suggest strongly that the IAA production by this strain, as in culture media in a laboratory, can be inhibited on trees by the application of nitrogenous compounds.

To determine whether IAA-producing bacteria on pear trees commonly are inhibited in IAA synthesis by application of nitrogenous compounds, and to further determine the amounts of nitrogenous compounds required to inhibit IAA biosynthesis, replicated trials were established in Mendocino County in which different amounts of urea, calcium nitrate, and ammonium sulfate were applied to trees in a commercial orchard. All rates of these materials significantly reduced fruit russeting in this plot by from about 4 to 5 fold compared to untreated control trees (Table 9). No significant differences were observed in fruit russeting on trees treated with any of these materials. Trees treated with urea at either rate had numerically more fruit russeting than trees treated with the other materials, but these differences were not statistically significant. These results were similar to those observed during 1993. The application of nitrogenous compounds did not cause phytotoxicity to the trees and the cost of the materials applied was relatively small. This approach therefore looks promising at this time for the control of fruit russeting. The potential exists, however, that nitrogenous compounds could increase the growth of deleterious bacteria including the fire blight pathogen. Considerable evidence indicates that the growth of microorganisms on plants is limited by the amount of carbon-containing compounds they need for growth and energy sources. If this is always the case, then application of nitrogenous compounds should not lead to increased growth of the microorganisms. If, however, plants were limited by the amount of nitrogenous compounds, then application of these materials might lead to increased microbial growth. While no evidence for this was observed during trials in 1993 and 1994 (compare total bacterial populations in Figure 10 with those in Figure 11 and Figure 12), more extensive studies of this phenomenon will be needed before recommendations can be made to apply nitrogenous compounds to pear during bloom periods.

>> Plant Pathology Research Reports