Experimental orchard and treatments.

The experiment was conducted from 2022 to 2023 on ‘Simmons Gala’/‘M9-T337’ maintained at the Mountain Horticultural Crops Research and Extension Center in Henderson County, NC, USA (lat. 35.428079°N, long. 82.563295°W, elevation 649 m). Two orchards were established in 2021 to complete concurrent projects. An orchard designated for horticultural (HORT) use was maintained without any fire blight inoculations, and measurements were taken to determine horticultural impacts of P-Ca and ASM (Vogel et al. 2025). A second orchard was established for the purpose of inoculation and disease evaluation (PATH), which is addressed in this work. Unless specified otherwise, the following methods apply to the PATH orchard only.

Trees were planted on 31 Mar 2021 with 0.9- × 4-m tree-row spacing in rows oriented from north to south. The graft unions were approximately 15- to 18-cm aboveground. On planting day, latex paint with 5000 ppm of 6-benzylandenine (6-BA) (Exilis 9.5 SC; Fine Americas, Inc., Walnut Creek, CA, USA) was applied to the central leader of trees to enhance lateral branching. Trees were trained in a high-density central leader system with conduit support. Commercial management practices for fertility, pesticide, crop load management, and herbicide applications adhered to local recommendations in all years of the study. Antibiotics were not applied leading up to and following inoculation until the termination of sampling.

Experimental units consisted of five-tree plots. Treatments were implemented in a complete randomized design and replicated five times. Single trees served as a buffer between treatments. The treatments were applied twice in the spring. The first application occurred at petal fall, and the second application occurred 10 to 14 d later. Application dates were 20 May and 1 Jun 2021, 28 Apr and 8 May 2022, and 4 and 16 May 2023. The factorial treatment structure was composed of three rates of P-Ca (Kudos; Fine Americas, Inc.) and two rates of ASM (Actigard 50WG; Syngenta, Greensboro, NC, USA). The rates are as follows:

P-Ca = 125, 42.5, and 0 mg·L⁻¹ applied with 0.125% (v:v) nonionic surfactant (Regulaid; Kalo, Inc., Overland Park, KS, USA) and 0.39% (v:v) water conditioner (Choice Trio; Loveland Products, Inc., Loveland, CO, USA)

ASM = yes or no (37.5 and 0 mg·L⁻¹, respectively).

Meteorology.

Weather conditions were monitored with an on-site weather station (lat. 35.42721°N, long. 82.55888°W), and data were retrieved from the Environment and Climate Observing Network database (ECONet; https://econet.climate.ncsu.edu/). Daily temperature, relative humidity, precipitation, and leaf wetness were reported.

Height and trunk cross-sectional area.

Two uniform representative trees from each five-tree plot were chosen for annual measurements. Tree height (cm) and trunk circumference (cm) were measured on 13 May 2021 and 13 May 2022. Tree height was a measure of length from the ground to the terminal bud of the leader. Trunk circumference was measured 30 cm above the graft union, and trunk cross-sectional area (TCSA) was calculated using Eq. [1]:$$\mathit{TCSA}\hspace{0.17em}=\pi ·{\left(\frac{\mathit{Avg}\mathit{circ}}{2\pi }\right)}^2$$ [1]

Tree height and TCSA were expressed as an average of the two-tree sub-sample.

Inoculation.

Three days before inoculation, the E. amylovora strain Ea110, with naturally occurring resistance to the antibiotic rifampicin, was recovered from a freezer (−80 °C) and plated onto a 100 μg·mL⁻¹ rifampicin and 50 µg·mL⁻¹ cycloheximide-amended Luria-Bertani agar plate. Using a sterile loop, one colony of Ea110 was added to each of five flasks of Luria-Bertani broth with 100 µL·mL⁻¹ rifampicin 1 d before inoculation. Flasks were placed on an incubator shaker (I2500 series; New Brunswick Scientific, Edison, NJ, USA) at 28 °C and 200 rpm in the dark. On the day of inoculation, a spectrophotometer (LKB Biochrom Ultrospec II; Biochrome Ltd., Boston, MA, USA) was used to measure the optical density of cultures, which is the absorbance of the sample at 600 nm. Inoculum was diluted to 10⁷ cfu·mL⁻¹ into 1% phosphate buffer and divided into multiple 50 mL centrifuge tubes (Falcon™ 50 mL High Clarity Conical Centrifuge Tubes; Corning, Corning, NY, USA). Nine actively growing shoots on one representative tree in each five-tree plot were chosen for inoculation. Scissors were used to inoculate shoots by inserting them into tubes of inoculum and horizontally bisecting three leaves at each shoot apex and reinserting scissors between each cut. In 2022 and 2023, inoculation occurred on 12 May and 11 May, respectively. The P-Ca and ASM treatments were applied 28 Apr and 8 May 2022 and 4 and 16 May 2023.

Disease ratings.

Before each sampling date, ratings of disease incidence and severity were recorded. The nine selected shoots were evaluated for disease incidence by counting the number of inoculated leaves that had a necrotic midline divided by the total inoculated leaves and multiplying by 100 to obtain a percentage. Once necrosis began to move down the shoot, the length of the lesion and total length of the shoot were measured to calculate percent disease severity on a per-shoot basis. Formulas for incidence and severity are shown in Eqs. [2] and [3].$$\mathit{incidence}\mathit{nec}\mathit{midline}\hspace{0.17em}=\hspace{0.17em}\frac{\mathit{count}\mathit{nec}\mathit{midline}}{\mathit{count}\mathit{inocula}\text{t}ed\mathit{leaves}}\times 100$$ [2] $$\mathit{severity}\hspace{0.17em}=\hspace{0.17em}\frac{\mathit{lesion}\mathit{length}(cm)}{\mathit{shoot}\mathit{length}(cm)}\times 100$$ [3]

Sampling.

Shoot sampling occurred on three dates each year. In 2022, the intervals were 4, 13, and 20 d after inoculation (DAI). In 2023, the intervals were 6, 15, and 22 DAI. On each date, two inoculated shoots were sampled at random from each tree. Shoots were kept in sealable plastic bags in a cooler until destructively sampled. Each shoot was surface-sterilized by spraying with a bleach solution with 0.74% NaClO and dipping in sterilized deionized water before dissection. Tissue samples were cut into 1 cm sections at four points along each shoot, the lesion, above the bud scar (BSA), below the bud scar (BSB), and at the junction of the shoot to the central leader (CL) (Fig. 1). The lesion was defined as the section where visible symptom development stopped and healthy tissue began. On the final sampling day, entire trees were removed from the orchard and an additional sample of the scion tissue 30 cm below the lowest infected shoot was collected. Tissue sections were weighed, chopped into small pieces, and placed in 1 mL of 1% phosphate buffer in a 1.5 mL centrifuge tube (Fisherbrand™ Premium Microcentrifuge Tubes; Fisher Scientific, Waltham, MA, USA). Tubes were sonicated (Branson M3800 ultrasonic cleaner; Branson Ultrasonics, Brookfield, CT, USA) for 7 min. Serial dilutions were performed to 10⁻⁶.

View Figure

• Download as Powerpoint
• Download Figure

Population density.

Serial dilution samples were plated on Luria-Bertani agar amended with 100 μg·mL⁻¹ rifampicin and 50 µg·mL⁻¹ cycloheximide. From each dilution sample, three replications of 10 µL were pipetted onto corresponding quadrants of the plates. Plates were stored in an incubator in the dark at 28 °C for 24 to 48 h. Once colonies were visible, plates were removed and cfu were recorded at the dilution level in which 30 to 100 colonies could be counted for each replicate. The three replicates were averaged and multiplied by 10 to the power of the dilution to calculate population. Population density was calculated as the log of the quotient of population and tissue weight. Methods were adapted based on previous research of P-Ca and ASM and their influence on E. amylovora (Slack et al. 2025).

Statistical analysis.

Statistical computation was performed using statistical software (JMP Pro version 17; SAS, Cary, NC, USA). A standard least squares model was used to evaluate the effects and interactions of treatments using a two-way analysis of variance (ANOVA) for tree height, TCSA, incidence wilt and necrotic midline, disease severity, and bacterial density in different tissue types. When P-Ca was the only significant main effect, generalized linear regressions were performed to elucidate the effect of the P-Ca rate on each variable. For the main effect of ASM, significance was determined with Student’s t test (α ≤ 0.05). For interactions, significance (α ≤ 0.05) was determined with Tukey’s honestly significant difference. All data were analyzed within the time point collected (e.g., “year” was not used as a model effect for data in tables).

Height and trunk cross-sectional area.

Tree size was compared between the two experimental orchards to assess similarity of trees for later determination of relationships between horticultural impacts and disease responses to treatments. In 2021, tree height was consistent among trees in both the HORT and PATH designated orchards. The TCSA was 16.0% higher among PATH when compared with HORT (Table 1). In 2022, height was 6.8% higher in HORT relative to PATH, and TCSA remained 18.1% higher in PATH relative to HORT. The separate orchards were grown on different systems with the horticulture plot on a trellis system, while the pathology plot was trained to the central leader with a conduit. The lesser support available to the pathology trees could have resulted in reduced vertical growth and increased girth to protect the trees from wind damage (Wang et al. 2022). Although the differences may be statistically different, there was not much practical difference in tree size between the two research orchards.

Table 1.Comparison of ‘Simmons Gala’ (Malus ×domestica) tree sizes in two plots at the Mountain Horticultural Crops Research and Extension Center in Mills River, NC, USA.

View Fullscreen

The effects of P-Ca and ASM alone and in combination on tree size were minimal in PATH (Table 2). Only the initial height of trees in 2021 had significant differences, but variation was only related to differences in nursery tree quality. Historically, the use of P-Ca in young orchards has been discouraged because growth suppression would interfere with the level of canopy establishment necessary to reach full production quickly (Costa et al. 2004; Norelli and Miller 2004; Norelli et al. 2003). However, young orchards are especially vulnerable when infected with fire blight because of tree size. The maintenance of tree size regardless of the P-Ca rate confirmed that, in North Carolina, additional late-season flush of growth (Unrath 1999) is sufficient for canopy infill even when P-Ca is used. Therefore, the age of the orchard should not preclude the use of P-Ca as a disease control agent.

Table 2.Main effects and interactions of acibenzolar-S-methyl (ASM) and prohexadione-calcium (P-Ca) rates on tree size in ‘Simmons Gala’ (Malus ×domestica). Trunk cross-sectional area (TCSA) and tree height are an average of two trees that were measured in the spring of each year.

View Fullscreen

Conditions and disease evaluation.

On the days of inoculation, 12 May 2022 and 11 May 2023, average temperatures were 17.4 °C and 18.3 °C and high temperatures were 24 °C and 24.5 °C, respectively (Table 3). In 2022, from the day of inoculation to the first sampling day, 1.38 cm of precipitation accumulated. Between the following two sampling periods, 9.70 cm and 6.79 cm of rainfall occurred. In 2023, 1.02 cm of precipitation accrued between inoculation and the first sampling date, and the following two periods had 0.13 cm and 5.97 cm of rainfall. In 2022 and 2023, from inoculation until the final sampling day, there were 131 h and 56 h of leaf wetness (>435 mV), respectively. Early in the incubation period in 2022, from 0 to 4 DAI, the average temperature was 18.1 °C. From 0 to 6 DAI in 2023, the average temperature was 19.2 °C. As the incubation periods extended, average temperatures increased between sampling dates in 2022 to 19.6 °C and 19.9 °C, while average temperatures decreased in 2023 to 16.9 °C and 16.1 °C. Higher precipitation and more than twice as many hours of leaf wetness could contribute to improved survival and spread of fire blight in 2022 when compared with those in 2023. Higher average temperatures throughout the incubation period in 2022 could increase the proliferation of the bacteria, whereas the lower temperatures of 2023 could slow bacterial growth.

Table 3.Weather data from critical inoculation and sampling timings in 2022–23 provided by the Environment and Climate Observing Networkⁱ.

View Fullscreen

Disease incidence and severity were evaluated on three dates in both 2022 and 2023. On 16 May 2022, 4 DAI, P-Ca and ASM had main effects on the disease incidence. As the rate of P-Ca increased, there was a reduction in disease incidence (P < 0.0001) (Fig. 2A). Compared with no ASM, ASM reduced disease incidence (P = 0.0130). There was no interaction at 4 DAI. At 13 DAI, on 24 May, P-Ca and ASM had a significant interaction, whereas ASM alone did not influence disease incidence; however, when used in combination with P-Ca, there was a negative curvilinear relationship between the P-Ca rate and incidence (P = 0.0050) (Fig. 2B). There was also a main effect of P-Ca, resulting in a negative curvilinear relationship between the P-Ca rate and disease severity on that date (P < 0.0001). On 31 May, 20 DAI, the P-Ca rate and ASM interacted to decrease disease incidence and severity. When ASM was applied, the P-Ca rate and disease incidence and severity had a negative curvilinear relationship (P = 0.0091 and P = 0.0323, respectively) (Fig. 2C).

View Figure

• Download as Powerpoint
• Download Figure

On 17 May 2023, at 6 DAI, P-Ca had a main effect on disease incidence whereby the increased P-Ca rate resulted in a negative curvilinear relationship between rate and incidence (P < 0.0001) (Fig. 3A). At 15 DAI, on 26 May, the P-Ca rate had a negative curvilinear relationship with disease incidence and severity (P < 0.0001 and P < 0.0001, respectively) (Fig. 3B). The relationship remained at 22 DAI, on 2 Jun, when the increased P-Ca rate decreased disease incidence and severity (P < 0.0001 and P < 0.0001, respectively) (Fig. 3C). There were no main effects of ASM or interactions between P-Ca × ASM in 2023.

View Figure

• Download as Powerpoint
• Download Figure

Previous research has shown improved fire blight management with the application of P-Ca (Agnello et al. 2019; Aldwinckle et al. 2000; Costa et al. 2001; Schupp et al. 2002; Wallis and Cox 2020; Yoder et al. 1999). The proposed explanations for the efficacy of P-Ca include thickening of cell walls (McGrath et al. 2009; Sundin 2014; Wallis and Cox 2020), reduction of pathogen movement between cells (Spinelli et al. 2005), and reduction of vigor. Some found that a single petal-fall application at a high rate was effective (Aldwinckle et al. 2000; Costa et al. 2001), while others found that lower rates and earlier timings were able to control blossom and shoot blight (Wallis and Cox 2020). Vigor and fire blight incidence have been correlated because of the assumption that trees with more actively growing shoots would be more likely to experience shoot blight, and that actively growing shoots are more susceptible relative to those in which terminal buds have set. Van der Zwet et al. (2012) found that increased growth coincided with increased fire blight incidence. This study supported this observation because using P-Ca to manage growth in part I of this series resulted in improved control of shoot blight (Vogel et al. 2025).

Plants’ resistance to diseases such as fire blight can increase with ASM (Aćimović et al. 2015, 2019; Brisset et al. 2000). When ASM is applied, pathogenesis-related proteins are upregulated, leading to a resistance response (Aćimović et al. 2015; Brisset et al. 2000; Maxson-Stein et al. 2002; Yuan et al. 2023). Studies have shown that shoot blight severity and canker size can be reduced by ASM (Johnson and Temple 2016; Maxson-Stein et al. 2002) or the combination of P-Ca and ASM (Aćimović et al. 2021). Blossom blight can be reduced with applications of ASM in the spring, and this reduction in inoculum can reduce shoot blight later in the season. However, the use of ASM alone is often insufficient for complete management of fire blight (Johnson and Temple 2016; Maxson-Stein et al. 2002).

Our results confirmed that P-Ca, even at low rates, has an impact on fire blight disease development. In part I of this series, P-Ca reduced the relative growth rate and tree size across multiple years (Vogel et al. 2025). Therefore, vigor reduction and a disease symptom reduction happened concurrently; however, numerous factors are at play, and it is unclear whether disease reduction can be solely attributed to vigor reduction. In 2022, the addition of ASM to a P-Ca program aided in fire blight management; however, its lack of effect in 2023 could be attributable to differing environmental conditions that impact bacterial survival, production, and spread. Additionally, ASM could be beneficial in years when disease pressure is particularly high, such as 2022. Regardless of year, P-Ca reduced disease symptom development, and higher rates resulted in better disease control.

Population density.

Across three dates in 2022, P-Ca and ASM had no practical effects or interactions on bacterial density in different tissue types (Table 4). A statistical difference was found at 13 DAI, where both lesion and central leader tissues were impacted by the interaction or the effect of P-Ca, respectively. The relationship between the P-Ca rate and bacterial density at the CL was positive curvilinear. The bacteria likely had progressed within the shoot faster than the symptoms progressed on the exterior. It is unlikely that P-Ca was responsible for this effect and, rather, it was ineffective at slowing bacterial systemic movement in 2022. By the final sampling date, bacteria were found within the scion 30 cm below the lowest infection point. Neither P-Ca nor ASM was able to reduce bacterial movement through shoots, and barriers such as the bud scar and central leader junction did not impact bacterial density. The weather in 2022 was particularly conducive to the spread and movement of fire blight, with rainy, warm conditions frequent in the spring and consistent leaf wetness from inoculation through sampling (Table 3). Entire orchard blocks can be decimated in one season (Norelli et al. 2003; van de Zwet et al. 2012); with the level of infection present in this study, the entire second leaf orchard likely would have been removed. If allowed to persist, or if infected trees were asymptomatic, then further losses and outbreaks would occur in the following years. Momol et al. (1998) found that E. amylovora had traveled from shoot tips to the rootstock by 21 DAI, showing no symptoms along the trunk.

Table 4.Main effects and interactions of acibenzolar-S-methyl (ASM) and prohexadione-calcium (P-Ca) rates on bacterial systemic movement in ‘Simmons Gala’ (Malus ×domestica) in 2022 and 2023. Population densities in various tissue types are represented as the log-transformed cfu per gram of tissue.

View Fullscreen

In 2023, P-Ca and ASM had main effects, reducing bacterial density in multiple tissue types across all dates (Table 4). At 6 DAI, there was an interaction between P-Ca and ASM whereby ASM alone was effective at reducing bacterial density in BSA tissue, but the main effect of P-Ca was improved with the addition of ASM. Essentially, all treatments except for the control were effective at reducing bacterial density by 68.2% to 86.5% on that date. Furthermore, P-Ca had a main effect on bacterial density in all tissue types on the second sampling day (15 DAI). There was a negative linear relationship between the P-Ca rate and bacterial density at the lesion. Bacterial densities at BSA, BSB, and CL decreased in a negative curvilinear fashion as the rate of P-Ca increased. Additionally, ASM had a main effect on BSA bacterial density, resulting in a 41.2% decrease relative to no ASM. On the final sampling date, only scion tissues were impacted by P-Ca and ASM. The rate of P-Ca and bacterial density in the scion had a negative curvilinear relationship (Fig. 4). Moreover, ASM reduced bacterial density in the scion by 61.7% (Table 4). On that same date, data points for BSA and BSB were missing because the lesion had moved past the bud scar on numerous shoots, especially control trees. Therefore, statistical power was lost.

View Figure

• Download as Powerpoint
• Download Figure

In 2023, P-Ca effectively reduced bacterial density across all tissue types, and fewer bacteria were present in tissues further from the lesion. Environmental conditions were less conducive to the spread of fire blight in 2023, with drier weather and reduced leaf wetness compared with 2022 (Table 3). Regardless of these benefits, the amount of bacteria present in these tissues is enough to be pathogenic (Slack et al. 2017). Our study, however, included an artificially inoculated environment with high-concentration inoculum applied to nine shoots on each tree. In contrast, in the noninoculated orchard, we saw very few losses caused by fire blight; over the course of 4 years, only four out of 175 trees were completely removed. Ideally, the natural level of infection in a typical orchard would not override the potential benefits of P-Ca and ASM. Nevertheless, finding bacteria in the scion 30 cm below diseased shoots revealed that, in young trees, it may be improbable to prune out enough tissue to save a tree when strikes are low in the canopy, especially if there is a possibility of external reinfection.