Evaluation of pruning therapies in apple trees with fire blight

Ten experiments were conducted to evaluate the efficacy of methods to therapeutically remove fire blight cankers from apple trees with different age, vigor, training system and rootstock combinations in Washington, Oregon, Pennsylvania and New York, United States. Removal methods included pruning at varying distances from visible symptoms, utilization of a sanitizing agent and the use of a stub cut to minimize cankers on structural wood. Fire blight removal resulted in reductions of rootstock blight and tree death. Removal treatments resulted in fewer additional cankers compared to the no-treatment control in most trial sites and years. The standard best management practice (BMP) for removing fire blight was branch removal at 30 cm below the proximal edge of visibly cankered tissue with sanitized loppers. This practice significantly reduced the number of new systemically-caused symptoms compared to the no-treatment control in seven of nine experiments with significant reductions in five experiments. Aggressive removal, defined as BMP but branches removed approximately 76 cm from the proximal edge of visibly cankered tissue, generally did not improve the standard practice. Elimination of cutting tool sanitation did not result in more cankers than the standard practice. Breaking off diseased branches by hand provided a rapid removal method, but it resulted in a greater number of cankers in the orchard at the end of the season. In one of five experiments cutting which left a 13 cm stub distal to structural wood significantly reduced the number of cankers on structural wood compared to flush cut or 4 cm stubs.


Introduction
Fire blight, Erwinia amylovora, is the most destructive bacterial disease in apple orchards in the United States and many other countries (Vanneste 2000).In the United States the disease has been known as a serious problem in apple and other pome fruits for over 200 years (Baker 1971;Van Der Zwet and Keil 1979).Losses to fire blight and cost of control in the United States are estimated to reach over $100 million annually (Busdieker-Jesse et al. 2016;Norelli et al. 2003aNorelli et al. ). 1979).The pathogen may then progress to the supporting branches and advance further to the main limbs, trunk, and rootstock collar and crown (Gowda and Goodman 1970;Momol et al. 1998;Schroth et al. 1974;Zamski et al. 2006).
Current fire blight management tactics focus on the minimization of inoculum in the orchard through cultural practices and the use of spray treatments to prevent infection (Norelli et al. 2003a).Cultural practices include planting less susceptible cultivars and resistant rootstocks, management of the orchard environment (e.g., proper fertilization, irrigation, and tree training), and orchard sanitation (e.g., cutting and removing infected tissue).Antibiotics are the most common control material applied, often with aid of information from models that predict infection risk from temperature and humidity (Biggs et al. 2008;Billing 1999;Norelli et al. 2003a;Smith 1998;Steiner 1989;Thomson et al. 1982;Turechek and Biggs 2015).In response to increasing antibiotic resistance and the need for certified organic management, biological and biopesticide materials are increasingly used (Johnson and Temple 2013;Johnson et al. 2022;Stockwell et al. 1996;Temple et al. 2020;Wilson and Lindow 1993).In addition, plant defense triggering chemicals, such as acibenzolar-s-methyl (ASM) are used both to reduce flower infections and as therapeutics in combination with fire blight removal (Johnson and Temple 2017).
Timely cutting of infected host tissue soon after E. amylovora infections occur is currently recommended to reduce the spread of the pathogen throughout the orchard and to reduce the advance of the disease in the tree which can lead to plant death.Current in-season fire blight management recommendations generally suggest managers cut at a minimum 20 to 30 cm below the noticeably infected area into two-year or older wood, using cutting tools surface disinfested between each cut.While cutting strategies have been demonstrated in several studies as effective measures to reduce the diseased tissue and to prevent systemic spread and tree loss (Covey and Fischer 1990;Van Der Zwet and Keil 1979), other studies have shown that such practices resulted in inconsistent and even disappointing results (Clarke et al. 1991;Lake et al. 1975;Shtienberg et al. 2003).Clarke et al. (1991) reported that even though cutting was performed at the normally recommended 20-25 cm beyond visible symptoms, 12% of the cut stubs on pruned trees still harbored E. amylovora.Moreover, Shtienberg et al. (2003) found that cutting did not result in successful eradication of the pathogen, despite the fact that cutting was carried out as recommended to growers.The systemic movement of the pathogen within the infected plants either through vascular tissues (Aldwinckle and Preczewski 1979;Crosse et al. 1972;Gowda and Goodman 1970;Momol et al. 1998) or intercellular space (Zamski et al. 2006), as well as the capability of E. amylovora to colonize tissues symptomlessly beyond the visible disease (Ge and Van der Zwet 1996;Hickey et al. 1998;Keil and Van Der Zwet 1972;Steiner 2000;Tancos et al. 2017) have been frequently associated with the failure in eradicating the pathogen and frequent development of the disease after removal of blighted parts by cutting (Keil and Van Der Zwet 1972;Shtienberg et al. 2003;Steiner 2000;Thomson 2000).
The objective of this study is to evaluate fire blight removal strategies to provide improved best management strategies to growers that reduce tree loss, the number of times a grower has to perform costly cutting management, and the risk of symptom re-occurence through improved sanitation.Cutting strategies included varying length (distance) of cutting sites from the visible symptoms, utilization of a sanitizing agent on the cutting tools, and application of a plant defense elicitor (acibenzolar-s-methyl).Cutting method success was evaluated based on its ability to prevent tree death, minimize new cankers, eliminate overwintering cankers, and minimize cankers on structural wood, the removal of which reduces fruiting area and corresponding profits.

Experimental site and design
Trials were conducted in Washington, Oregon, New York and Pennsylvania, United States.Site 'Washington 2019 Yarlington Mill' consisted of a 0.5-acre commercial orchard of 105 4-yr-old naturally infected apple trees cv.Yarlington Mill, rootstock Red Delicious interstems, in Wenatchee, WA.The experiment was arranged in a randomized, complete block design with 15 replications of 7 treatments applied to single tree plots where each tree had 1 to 14 naturally infected strikes (symptomatic flower infections) per tree.Site 'New York 2019 Evercrisp' consisted of a 0.75-acre commercial orchard of 120 4-year-old apple trees cv.'EverCrisp', rootstock G.41 in Geneva, NY.The experiment was arranged in a randomized, complete block design with 10 replications of 6 treatments applied to single tree plots where each tree had 10 to 20 strikes per tree.Inoculation with E. amylovora strain Ea273 was conducted at 80% bloom with concentration 1 × 10 6 CFU/ml using a Solo 425 sprayer (Solo Inc, Newport News, VA, Unites States).Site 'New York 2019 Idared' consisted of a 1.2-acre research orchard planting of 150 7-year-old apple trees cv.'Idared', rootstock B.9 in Geneva, NY.The experiment was arranged in a randomized, complete block design with 10 replications of 5 treatments applied to single tree plots where each tree had 5 to 20 strikes per tree.Inoculation with E. amylovora strain Ea273 was conducted at 80% bloom with concentration 1 × 10 6 CFU/ml.Site 'Pennsylvania 2019 Gala' consisted of a 0.07-acre research orchard planting of 36 4-year-old apple trees cv.'Gala', rootstock M7 in Biglerville, PA.The experiment was arranged in a randomized, complete block design with 4 replications of 6 treatments applied to single tree plots where each tree had 2 to 5 strikes per tree.Inoculation with E. amylovora strain Ea273 was conducted at 80% bloom with concentration 1 × 10 6 CFU/ ml.Site 'Washington 2020 Cripps Pink' consisted of 1-acre commercial orchard planting of 1400 14-year-old apple trees cv.'Cripps Pink', rootstock M9.337 in Benton City, Washington.The experiment was arranged in a randomized, complete block design with 15 replications of 7 treatments applied to single tree plots where each tree had 1 to 15 strikes per tree (2.5 average).E. amylovora infection had naturally occurred.Site 'Oregon 2020 Gala' consisted of a 0.2-acre research orchard planting of 36 3-year-old apple trees cv.'Gala', rootstock M9.337 in Corvallis, Oregon.The experiment was arranged in a randomized, complete block design with 6 replications of 6 treatments applied to single tree plots where each tree had 4 to 23 strikes per tree (average 8.8).Inoculation with E. amylovora strain concentration 9 × 10 8 CFU/ml was conducted on April 24, 2020 at 100% bloom.Site 'Pennsylvania 2020 Gala' consisted of a 0.07-acre research orchard planting of 3-year-old apple trees cv.'Gala', rootstock M9.337 in Biglerville, PA.The experiment was arranged in a randomized, complete block design with 8 replications of 7 treatments applied to single tree plots where each tree had 1 to 2 strikes per tree.Inoculation with E. amylovora strain Ea273 was conducted at 80% bloom with concentration 1 × 10 6 CFU/ml.Spring weather was unseasonably cool through June week 3 followed by high temperatures and dry conditions in July.Experiment 'Washington 2021 Cripps Pink' consisted of a 0.5-acre commercial orchard planting with 57 4-year-old naturally infected apple trees cv.'Cripps Pink', rootstock M9.337 in Patterson, Washington.The experiment was arranged in a randomized, complete block design with 13 replications of 7 treatments applied to single tree plots where each tree had 1 to 7 symptoms per tree.E. amylovora infection had naturally occurred.'New York 2021 Ruby Frost' consisted of a high-density tall spindle planting (1200 trees per acre) 3-yr-old planting of cv.Ruby Frost on G.41 rootstock.'New York 2021 Gala' was a vertical axis planting at 300 trees per acre of cv.Gala on B.9 rootstock.NY 2021 plantings consisted of 10 replications of single tree plots where on June 1, 2021 two actively growing shoots per tree were inoculated with E. amylovora at 1 × 10 6 CFU/ml using scissor cut inoculation.Plots were monitored every 3-days and as soon as symptoms were apparent treatments were imposed.Sites varied in training system, vigor, cultivar and rootstock susceptibility (Table 1).

Treatments
Seven fire blight removal treatments were employed (Fig. 1).Best Management Practice (BMP): Branches were removed approximately 30 cm (12 inches) from the proximal edge of the canker tissue in 2-year or older wood with cutting tool dis-infested with a 0.5% sodium hypochlorite solution (10% bleach).BMP NO-sanitize: BMP but without tool sanitization.Aggressive: BMP but branches removed approximately 76 cm (30 inches) from the proximal edge of the cankered tissue.Long stub: Canker removal cuts were made to leave a 13 cm (5 inch) stub of branch from the central leader or main structural branch using loppers sanitized with a 0.5% sodium hypochlorite solution.If the canker redevelops at the pruning cut, a long stub is hypothesized to keep the symptoms from expanding to the structural wood of the tree and allow for easy removal of any potential overwintering cankers during winter pruning.Short stub/ flush cut: Same as LONG Stub but canker removal cuts left a 0 to 5 cm (0 to 2 inch) stub of branch from the central leader or main structural branch.Breaking: Limbs with symptoms were broken back by hand, snapping the wood at the joint between the current-year growth and the second-year growth.No-treatment control: In research orchard trials, no-treatment control trees received no cutting and the fire blight symptoms remained in the tree all season.For the control treatment in commercial orchards, branches were cut at the margin of the canker and symptomless wood in order to remove active ooze, which may create secondary infections in adjoining trees.The 2020 Oregon trial site also included an additional treatment.BMP + Actigard: Symptoms were removed using the BMP treatment followed by an application of concentrated acibenzolar-s-methyl (Actigard 50 WG, Syngenta, Oregon, United States) to a 61 to 91 cm (2 to 3 foot) section of the central leader of the tree using a 1-liter spray bottle at 30 g per liter (1 oz per quart) with a silicon-based surfactant.For up to eight weeks, trees were re-evaluated several times after the initial cutting treatments were imposed to count and remove any additional cankers.For secondary canker removal, the method of removal was consistent with the initial cutting treatment.The autumn after cutting treatments

Impact of removal strategy on tree death prevention
Timely cutting of fire blight infected material in-season is essential to prevent systemic movement of the bacteria through the tree, which can result in rootstock infections, and tree death (Fig. 2).In six of ten experiments rootstock blight or tree death occurred.In the six experiments where rootstock blight or tree death occurred, cutting treatments reduced the number of trees that died or acquired rootstock blight with significant differences in five experiments (Table 2).In experiments 'New York 2019 Ever Crisp', 'New York 2019 Idared', 'New York 2021 Ruby Frost', and 'New York 2021 Gala', 100% of trees that received no cutting treatment died while zero trees receiving a fire blight removal treatment died.Experiments where tree death occurred were primarily in sites where trees were highly susceptible due to cultivar genetics and vigorous growth owing to fertility and abundant rainfall: 'New York 2019 were imposed, trees were evaluated for the quantity of diseased tissue remaining in the tree measured as mm of canker; the number of cankers that progressed to structural wood; and the incidence of rootstock blight and tree death.Trees were evaluated the following spring for incidence of rootstock blight and tree death.Cumulative number of cankers, the average length of canker, the percentage of cankers which progressed to structural wood, and the incidence of rootstock blight was analyzed using an analysis of variance using a generalized linear model using in SAS 9.4.Post hoc pairwise comparisons between treatments were determined using Fishers Least Significant Difference (LSD) between pairs of least-squares means.A P-value of 0.05 was used to identify significant differences.Vigor as denoted is a function of fertility and irrigation where high vigor sites received high nitrogen fertilization (approx.100 lbs/acre).New York sites received irrigation via rainfall and 66 lbs per acre of nitrogen (ammonium nitrate at 300 lbs per acre) u Infection incidence often exceeds 90%

Prevention of new symptoms
Erwinia amylovora is known to move systemically within an infected tree from the initial infection site through vascular tissues (Aldwinckle and Preczewski 1979;Crosse et al. 1972;Gowda and Goodman 1970;Momol et al. 1998) or intercellular space (Zamski et al. 2006) and create new symptoms in other susceptible tissues (e.g., young shoot tips).In addition, E. amylovora can colonize symptomless tissues beyond visible cankers (Ge and Van der Zwet 1996;Hickey et al. 1998;Keil and Van Der Zwet 1972;Steiner Ever Crisp', 'New York 2019 Idared', 'New York 2021 Ruby Frost', and 'New York 2021 Gala'.Tree death prevention is a critical goal of fire blight removal due to the devastating economic losses that can occur.For example, in 2018 after a severe fire blight risk period in Washington, 74 orchardists surveyed reported removing 230,000 trees resulting in over $4 million in losses in tree replacement and labor costs alone.3).The exceptions were the site 'Washington 2020 Cripps Pink' where trees were 14-years old, very low vigor and initial cutting was performed more than two weeks after symptoms first became noticeable and 'Pennsylvania 2020 Gala' where weather conditions were not conducive to disease development.In other experiments where BMP was lower but not significantly different than controls, trees were young, highly-susceptible cultivars with high vigor and none of the treatments significantly reduced the number of new symptoms and cankers that re-formed after initial pruning.For example, 3-yr-old 'New York 2021 Ruby Frost' trees still had an average of more than four new 2000; Tancos et al. 2017).The recommended best management practice is to cut 30-45 cm beyond visible symptoms to remove bacterial cells concentrated in asymptomatic wood and lessen the likelihood that there are sufficient remaining cells to re-develop the canker at the cut site, or to migrate internally and cause new symptoms in young tender shoot tips.The treatment best management practice (BMP), where branches were removed approximately 30 cm from the proximal edge of the canker tissue in 2-year or older wood with sanitized cutting tools, reduced the number of new symptoms compared to the no-treatment control in seven of nine experiments with significant reductions in five

Stub cuts
When cankers caused by fire blight infections reach central leaders and main structural branches, growers face the decision to either prune out the canker, reducing the fruitbearing capacity of the tree, or to leave cankers that are a source of new inoculum the following spring.Some recommendations suggest an 'ugly stub cut' where growers make cuts in two-year-old wood 20 to 30 cm proximal to visible symptoms (where resistance is greater due to carbohydrate reserves (Suleman and Steiner 1994) while leaving a 10 to 12 cm naked stub protruding from the main branch (Fig. 4).While small cankers will re-form on many of these cuts, these cankers can be eliminated during winter pruning (Steiner 2000) (Fig. 5).Consequently, we tested this recommendation as a hypothesis by leaving a 10 to 13-cm stub from the central leader or structural branch when cutting blight.In two of five experiments where Long Stub and Short or no-Stub treatments were compared ('Washington 2020 Cripps Pink', 'Washington 2021 Cripps Pink'), a Long Stub reduced the number of cankers on structural wood with significant reductions in 'Washington 2020 Cripps Pink' (Table 4).In experiments with low host susceptibility ('Washington 2019 Yarlington Mill' Red Delicious interstems) and where fire blight pressure was low due to environmental conditions ('Pennsylvania 2020 Gala'), few cankers progressed to structural wood and stub cuts did not reduce cankers on structural wood.

Breaking
Some commercial orchard managers have been employing removal of diseased limbs by hand breaking wood at the joint between current-and second-year growth, rather than cutting to remove symptoms.This practice is designed to be fast and avoid the use of loppers, and is used primarily in V-trellis training systems, where limbs are trained and tied to a wire.In these orchard systems, cutting 30 to 45 cm from the proximal edge of visible symptoms is time consuming (as ties which secure branches to the wire must be removed) and can result in major damage to tree structure.In experiment 'Washington 2021 Cripps Pink' where 4-yr-old trees were trained to the wire, treatment Breaking resulted in significantly more new cankers than other cutting treatments, similar to the no-treatment control (Table 3).In three of ten experiments Breaking resulted in significantly more canker tissue left in the tree at the end of the season compared to BMP (Table 5).These remaining cankers provide a source of inoculum in the following year.In Washington 2021 'Cripps Pink' Breaking also resulted in significantly higher numbers of cankers on structural wood than BMP, and cankers on structural wood were numerically higher than BMP symptoms occur after initial cutting compared to eight in the no-treated control (Table 3).
We hypothesized that aggressive cutting may further reduce re-development of symptoms by removing more bacterial cells that had diffused into symptomless tissue.Reducing or eliminating re-development of cankers helps to limit the number of costly passes orchard crews must make through the orchard.For example, in a 2017 survey in Washington, orchardists invested 2 to 64 hours per acre cutting blight after a severe infection event at a cost of $27 to $864 per acre.In treatment Aggressive, branches were cut back approximately 76 cm from the proximal edge of the canker tissue with dis-infested tools.Treatment Aggressive had fewer new cankers than BMP at 'Washington 2019 Yarlington Mill' but the number of new cankers was not significantly different than BMP at other sites (Table 3).In site 'Washington 2019 Yarlington Mill' treatment Aggressive often resulted in stumping the infected leader as such no tissue that could acquire new symptoms remained.In sites 'Oregon 2020 Gala' and 'Pennsylvania 2020 Gala', no cankers re-developed in either BMP or Aggressive potentially due to low fire blight pressure and in the case of Oregon moderate vigor (Table 1).Similarly, in site "New York 2021 Gala', the number of new cankers was low in both BMP and Aggressive cutting treatments, potentially due to tree age (18-yr-old).Even with aggressive removal in vigorous trees multiple passes can be necessary to remove symptoms.For example, site 'New York 2019 Evercrisp' was composed of highly vigorous young trees where both BMP and Aggressive had three times fewer new cankers than the no-treatment control but still resulted in four to five new cankers per tree.At site 'Washington 2021 Cripps Pink', Aggressive cutting resulted in excessive new growth and provided abundant susceptible tissues for symptom re-occurrence (Fig. 3).As above, 'Washington 2021 Cripps Pink' trees that received Aggressive cutting were often effectively stumped which sometimes led to tree death.Keil and Zwet (1972) found that E. amylovora is predominantly present in the stem immediately below the blighted portion and continues to be detected approximately 9 to 18 cm below visible blight symptoms.Interestingly, in younger, more highly vigorous blocks the distance needed to be cut was not greater than in older less vigorous blocks.Our research suggests that cutting more aggressively (76 cm below proximal edge of canker) generally does not result in fewer systemically generated new symptoms than the standard best management practice of cutting 30 to 45-cm below the point proximal edge of the canker.

Sanitizing tools
Although sanitizing pruning shears has been long considered important to prevent dissemination of fire blight infections (Van Der Zwet and Keil 1979), in multiple studies sterilizing shears made no difference in preventing new cankers after initial symptoms (Toussaint and Philion 2008;Travis and Kleiner 1997).In nine of nine experiments, BMP NO-sanitize did not show more cankers after initial removal compared to BMP (or Short stub and Short stub no-sanitize) (Table 3).The average number of cankers at the end of the season was also not different between BMP and BMP NO-sanitize (Table 5).In one of nine experiments, BMP No-sanitize had a greater number (but not significantly different) number of trees develop rootstock blight than BMP (Table 2).Consequently, our hypothesis is that any viable cells of the pathogen on a cutting blade is insignificant compared to the migration of the pathogen up to several meters (Suleman 1992) proximal to the margin of visible symptoms.Non-congruity between our results and the conclusion in five out of nine experiments (Table 4).While Breaking provides a fast fire blight removal method it can leave many cankers in the orchard which provide a source for infection in subsequent years.

Removal strategy efficacy as a sanitation measure
Cutting treatments generally resulted in smaller quantities of canker at the end of the season compared to no-treatment control in most experiments (Table 5).Breaking had more end of season cankers than BMP, Aggressive, BMP NOsanitize, Short Stub, and Long Stub in 'Washington 2019 Yarlington Mill' (Table 5).At the end of the season no significant difference in the number of cankers was observed among BMP, Aggressive, BMP-NO sanitize, Short Stub and Long Stub in most sites and years (Table 5).No significant difference in the length of end season canker was observed between BMP and BMP + ASM in the 'Oregon 2020 Gala' site (Table 5).Likewise, no significant difference in end of season canker number was observed between BMP and BMP NO-sanitize in all trial sites and years (Table 5).Similarly, elimination of cutting tool disinfestation did not result in more cankers than the standard practice.Breaking off diseased branches by hand was a rapid removal method, but resulted in a greater number of cankers in the orchard at that tool dis-infestation may be because in other studies, fire blight was transmitted by dipping pruning shears in an inoculum suspension or the cuts were made through active ooze (Kleinhempel and Nachtigall 1987), rather than the recommended 30 cm from the margin of visibly cankered tissue.

Conclusions
Timely summer cutting of fire blight symptoms can significantly reduce the number of trees that develop rootstock   *Means within a column followed by the same letter do not differ significantly (P = 0.05).

Fig. 1
Fig. 1 Fire blight removal methods.Black line indicates location of pruning cut.a: Best management practice (BMP) and BMP No-sanitize, b: Aggressive, c: Breaking, d: Long Stub, e: Short Stub, f: No-treatment control

Fig. 2
Fig. 2 Rootstock blight symptoms on an untreated tree

Fig. 3
Fig. 3 Excessive new growth on trees treated with the Aggressive cutting method Washington 2021 'Cripps Pink' and resulting new symptoms in young susceptible tissue

Fig. 5 Fig. 4
Fig. 5 When fire blight is cut leaving a short stub (left) reignition of symptoms can result in a canker on structural wood.Leaving a long stub (right) can help keep re-development on the stub where it can be later removed in winter pruning Fig. 4 Fire blight symptoms (black area) (a) removed with a cut close to the central leader (b) can result in canker re-development (c) on structural wood (d).Fire blight symptoms (e) removed leaving a long stub (f), followed by canker re-development (g) removed during winter pruning (h, i) grafts, y high-density tall spindle, high vigor x tall spindle, low vigor, w tall spindle, low vigor, u Auvil V trained to the wire, moderate vigor, t Vertical axe, high vigor, s % Cuts progressing through previous season's growth *Means within a column followed by the same letter do not differ significantly (P = 0.05) ---Indicates treatment was not included at study location Best management practice (BMP), no treatment control (NTC), acibenzolar-s-methyl (ASM)Table 5 Effect of treatment on the average length (cm) of cankers left in trees per strike at the end of the grafts, y high-density tall spindle, high vigor x tall spindle, low vigor, w tall spindle, low vigor, u Auvil V trained to the wire, moderate vigor, t Vertical axe, high vigor.

Table 3
Effect of treatment on the number of new cankers after initial cutting and removal of fire blight symptoms

Table 4
Effect of treatment on the percent of cankers progressing to structural wood