Assessing the impact of plant growth regulators on anther retention and Fusarium head blight in spring wheat (Triticum aestivum L.) infected by Fusarium graminearum in field conditions

Abstract Fusarium head blight (FHB) resistance in wheat is often associated with undesirable agronomic traits such as tall plant height and a propensity for lodging. Plant height in wheat is genetically controlled by some semi-dwarfing alleles that alter the plant's sensitivity to gibberellins (GAs). The presence of semi-dwarfing alleles increases the frequency of anther retention, which may contribute to FHB susceptibility by providing an initiation site for infection. The application of plant growth regulators (PGRs) may enable farmers to grow the most resistant cultivars while controlling plant height to minimize lodging risk. In this study, five spring wheat cultivars that differed in level of FHB resistance, height, and semi-dwarfing alleles were tested to determine the effect of PGRs, specifically Manipulator™ and Ethrel™, on plant height, anther retention, and FHB resistance level and the interactions between them in Winnipeg and Carman, Manitoba in 2019 and 2020. Combined field results showed that Ethrel™ significantly reduced plant height. Both PGRs did not affect the anther retention or FHB resistance levels of the tested cultivars under dry conditions. There were significant interactions between variables, but they were relatively small compared to the main treatment and cultivars. Based on the results of this study, producers could benefit from the higher levels of FHB resistance often associated with tall cultivars and use PGRs to manage plant height and lodging without increased risk of FHB.


Introduction
Wheat (Triticum aestivum L.) is an important staple food crop and is grown worldwide.Fusarium head blight (FHB) is a fungal disease mainly caused by Fusarium species.In North America, Fusarium graminearum Schwabe [teleomorph Gibberella zeae (Schwein.)Petch] is the predominant FHB pathogen on wheat (Gilbert and Haber 2013).Infection by F. graminearum causes economic losses through a reduction in yield and enduse quality of the grain and the accumulation of mycotoxins such as deoxynivalenol (DON) that pose a threat to food and feed safety (Buhrow et al. 2016).Breeding for resistance to FHB has increased the number of wheat cultivars with moderate-to-intermediate resistance to FHB (Seed Manitoba 2018).Nevertheless, most cultivars with better FHB resistance tend to be taller, which makes them more difficult to adopt in high-moisture regions with higher lodging potential.Generally, growers prefer shorter cultivars, since shorter cultivars are not prone to lodging and tend to produce a higher yield under intensive management.However, research has shown that short plants tend to have higher levels of FHB than taller plants.There are two possible explanations for this problem.One explanation is that shorter plants are closer to the inoculum source from crop residues and are thus exposed to higher levels of inoculum and humidity that promote disease development (Hilton et al. 1999;Buerstmayr et al. 2009).Despite this explanation, research using spray inoculation directly on the spike shows that the association between higher FHB and shorter plants exists even when all spikes receive the same quantity of inoculum (Mwaniki 2017).Therefore, proximity to inoculum and humidity cannot fully explain why shorter plants are more susceptible to FHB.Another explanation for the higher FHB susceptibility of short plants is that they carry one or more semi-dwarfing alleles.The Rht-B1b and Rht-D1b semi-dwarfing alleles affect the plant's sensitivity to gibberellins (GAs) and shorten the plant by affecting cell elongation, including shortening the filament of the anthers.This may result in anther retention during flowering.Retained anthers between the lemma and palea provide a surface for the fungus to land on and nutrients to support fungal growth, which allows the fungus to enter Table 1.List of five spring wheat cultivars used in field experiments conducted in Winnipeg and Carman, Manitoba, in 2019 and 2020, with wheat end-use class, height, FHB resistance level, and the presence of semi-dwarfing alleles.anthers constitute a preferred target for initial infection, leading to higher FHB (Buerstmayr and Buerstmayr 2015).
There are two main FHB resistance types that can be evaluated in the field: type I and type II.Type I resistance represents resistance to initial infection (Bai and Shaner 2004;Foroud et al. 2019).Type II resistance indicates resistance to the spread of infection within the infected spike.Since retained anthers facilitate the initial establishment of the fungus, it reduces type I resistance.Anther retention is a quantitative trait (Skinnes et al. 2010;Muqaddasi et al. 2017;Steiner et al. 2017;Xu et al. 2019) and has been highly correlated with FHB susceptibility (Buerstmayr andBuerstmayr 2015, 2016;He et al. 2016).Quantitative trait loci for anther retention on chromosomes 4DS and 4BS overlap with Rht-B1 and Rht-D1 (Xu et al. 2019).This indicates that the higher FHB susceptibility of semi-dwarf plants may be due to both the effect of short plant height and anther retention.
More than 70% of current wheat cultivars grown in the world have one or both of the semi-dwarfing alleles Rht-B1b and Rht-D1b (Buerstmayr and Buerstmayr 2016;Mo et al. 2018).The semi-dwarfing alleles Rht-B1b and Rht-D1b have contributed to a huge increase in grain yields by substantially reducing height, thereby increasing harvest index and reducing lodging under intensive management.
Plant growth regulators (PGRs) are natural or synthetic compounds used to control or modify the growth of plants by altering the plant's hormonal status.PGRs are used to reduce plant height and increase stem thickness, thus improving lodging resistance in cereal crops.Also, reducing plant height through the application of PGRs makes harvesting tall cultivars easier and increases the harvest index.Since cereal yields can be reduced from 7% to 35% from lodging (Strydhorst et al. 2018), the application of PGRs can have benefits for maximizing yield potential through improving crop standability.PGRs for cereal crops exploit either GA or ethylene (ET) pathways.Reduction of plant height is achieved by either of two main groups of PGRs: GA biosynthesis inhibitors and an ET-releasing compound.Manipulator™ contains the active ingredient chlormequat chloride (CCC), which is one of the GA biosynthesis inhibitors.It blocks an early phase in GA biosynthesis, resulting in less cell elongation and cell division in plants (Rademacher 2000;Hedden and Sponsel 2015).Manipulator™ affects the apical dominance of the plant, thus making plants produce shorter, thicker, and stronger stems for improved lodging resistance.Ethrel™ contains ethephon as an active ingredient.As an ET-releasing compound, ethephon elevates the level of ET in the plant.Ethrel™ influences plant growth by accelerating fruit ripening and maturity and reducing lodging in spring and winter wheat.The effects of both PGRs are specific to plant species, cultivars, and environmental conditions (Strydhorst et al. 2018).Also, the optimal timing for application is critical for ensuring successful results.
Theoretically, both Manipulator™ and Ethrel™ reduce plant height by reducing cell elongation.If filaments are shortened through the use of PGRs and anther retention is increased, it is possible that these PGRs may lead to an increase in FHB infection in wheat.If PGRs do not affect anther retention and have no impact on FHB infection, breeders can have more options to breed FHB-resistant wheat, and growers can utilize PGRs with FHB-resistant cultivars to maximize productivity and reduce the potential risk of FHB infection.Thus, the roles of both PGRs on FHB in wheat need to be determined.The objectives of this study were to determine the effect of the PGRs, Manipulator™ (GA biosynthesis inhibitor) and Ethrel™ (ET-releasing compound), on five spring wheat cultivars that differ in their levels of FHB resistance and height, and to determine the effect of PGRs on anther retention and FHB in spring wheat cultivars under field conditions.

Materials and methods
Five commercial spring wheat genotypes (AAC Brandon, AAC Cameron, AAC Penhold, AAC Tenacious, and Prosper) were chosen to represent a range of end-use classes and differences in height, level of FHB resistance, and semi-dwarfing alleles (Table 1).
Field trials were conducted in Winnipeg and Carman, Manitoba, in the growing seasons of 2019 and 2020.Each field trial was a split-plot design with four replicates.The main plot effect (main treatment) was the combination of PGR treatments and FHB inoculation (Table 2).The cultivar was the subplot effect.
Each plot represented a single cultivar combined with a PGR and FHB treatment.A total of 120 plots (6 treatments × 5 cultivars × 4 replicates) were at each experimental site in each growing season.Each main plot was separated with a buffer plot (3 m × 6 rows spaced 17 cm apart) of a tall wheat cultivar, Amazon, to prevent wind drift of PGR and FHB applications onto non-target plots.
In Soil samples were collected at each field site for soil testing.To determine the available nitrogen, composite samples were taken from 0 to 15 cm and 15 to 60 cm.At seeding, 11-52-0 fertilizer was applied to the seed at a rate of 35 kg/ha.Additional N was applied as broadcast fertilizer without incorporation using 46-0-0 to achieve a target of 150 kg of nitrogen/ha for each site in 2019 and 2020.
Seedling numbers for each plot were determined by counting the number of plants in 1 m sections of the two middle rows in each plot before seedlings started tillering.The number of seedlings counted in the two rows was used to calculate the plant density as the plant number per square meter using the following equation: Commercially available PGRs Manipulator™ (620 g/L of CCC), developed by Taminco US LLC and Ethrel™ (240 g/L of Ethephon), manufactured by Bayer CropScience Inc. were used in field experiments.For plots treated with Manipula-tor™ (620 g/L of CCC), it was applied to the plants at the Zadoks GS 30 at the recommended rate of 1.8 L/ha as a single application.For plots treated with Ethrel™ (240 g/L of Ethephon), it was applied at Zadoks GS 37 to 45 for plants at the recommended rate of 1.25 L/ha.A CO 2 -powered back-pack sprayer with a six-nozzle boom at 30 psi air pressure was used for PGR applications.
Four different isolates of F. graminearum were obtained from Dr. Maria Antonia Henriquez (Morden Research and Development Center of Agriculture and Agri-Food Canada): the 3 acetyl-deoxynivalenol chemotype for isolates HSW-15-39 and HSW-15-87, and the 15 acetyl-deoxynivalenol chemotype for isolates HSW-15-27 and HSW-15-57.Macroconidia for each isolate were generated in the lab using the techniques described by Amarasinghe et al. (2013).
Spray inoculum was prepared by combining equal quantities of macroconidia from each isolate into a 1 L bottle and adding distilled water to produce a final concentration of 50 000 macroconidia/mL.Tween 20 (VWR International, Edmonton) was added as a surfactant to each bottle at a rate of 4 ml/L.
All FHB-inoculated plots were inoculated twice.Plots were inoculated with 1 L of inoculum when 50% of the plants in the plot were at anthesis (Zadoks GS 65) and then three days later.After each inoculation, overhead mist irrigation was used to maintain humidity and promote the development of FHB.
Prior to collecting samples for anther retention, spike numbers in the same 1 m sections of the two middle rows used for plant counts were counted.The number of spikes counted in the two rows was used to calculate the spike density per meter squared using the following equation:

Spike density =
No. of spikes in the two 1m rows × 9 4 m 2 Twenty spikes per plot were collected 5-7 days postanthesis and stored at −20 • C until data could be collected for anther retention.Samples were defrosted before counting the retained anthers.Any anthers that were found to be located within florets or trapped between the lemma and palea were considered retained anthers.The retained anthers in the primary and secondary florets of four spikelets in the central portion of the spike were counted (Buerstmayr and Buerstmayr 2015).The percentage of anther retention was calculated as the number of retained anthers divided by the maximum number of anthers times 100%.
The plot was visually evaluated for FHB disease incidence and severity 18-21 days after the first inoculation.Disease incidence was measured as the percentage of infected spikes in the plot, and disease severity was measured as the percentage of infected spikelets within infected spikes.These values were used to calculate the FHB index.The FHB index was calculated by incidence times severity and dividing by 100.
Plant height for each plot was measured using a 2 m ruler prior to harvest.Five different plants within a plot were randomly selected and measured from the soil surface to the tip of the spike without the awn.The mean of the five height measurements for each plot was used for analysis.
Lodging was rated before harvesting, and the lodging rating was calculated from the percentage portion of the plot affected by lodging multiplied by the lodging scale of 1-9, in which 1 = no lodging and 9 = completely lodged.Lodging occurred only in the Carman 2020 trial after high winds and severe rainfall on 30 June 2020.
Plots were harvested with a small plot combine (Classic Plot combine, WINTERSTEIGER., Saskatoon) after the wheat was physiologically matured.In 2019, plots in Carman and Winnipeg were harvested on 19 and 30 August, respectively.In 2020, the Carman trial was harvested on 24 August and the Winnipeg trial was harvested on 26 August.To prevent the loss of Fusarium damaged kernels (FDK), the wind speed of the combine was reduced from normal by 30%.The grain yield for each plot was measured.A seed counter (Model U, International Marketing and Design Corp., USA) was used to count one thousand kernels from each plot, which were then weighed to determine the thousand kernel weight (TKW).The TWT for each plot was measured using a standardized halflitre cylinder, a filling hopper, and stand, and weighing the sample.As per the protocol from the Canadian Grain Commission, the weight in grams from the 0.5 L was converted to kg/hL (Canadian Grain Commission 2019).
To determine FDK and DON for all plots, a 50 g subsample of seed from each plot was sent to SGS Biovision in 2019 and Central Testing Laboratory Ltd. in 2020.For FDK analysis, the samples were divided into a working portion using a Boerner divider, and a minimum of 10 grams of grain was examined to remove possible Fusarium-affected kernels.All the FDK were inspected under 10× magnification to check for fibrous mold.Fusarium-damaged kernels were measured as a percentage of the total sample by weight.For DON analysis, each 50 g sample was ground and thoroughly mixed.Five grams of ground sample were added to 100 ml of distilled water.Using ultra-turrax, the sample was blended and filtered through a Whatman No.1 filter.Fifty microlitres of the filtrate per well was used for the quantitative determination of DON content using enzyme-linked immunosorbent assay (ELISA) with the RODASCREEN FAST DON kit (R-Biopharm AG., Germany).
To determine protein content in the grain, a 400 g sample from each plot was evaluated by Central Testing Laboratory Ltd. using near-infrared analyzer (FOSS., Denmark) on the whole grain.A scoop of each grain sample was poured into the top of the hopper and levelled with the top of the hopper.The analyzer ran 10 determinations and detected the moisture and protein content of the sample.The protein content was reported on a dry matter basis.
Weather data from Winnipeg were taken from the Point Weather Station at the University of Manitoba.Carman's weather data were taken from the Environment Canada weather station at the Ian N. Morrison Research Farm, Carman.Weather data included daily average temperatures ( • C) and precipitation (mm) for the 2019 and 2020 growing seasons from May to August.From the data, line and bar graphs of the daily average temperature ( • C) and precipitation (mm) for each location in 2019 and 2020 were created in Microsoft Excel 2016.Long-term (23 and 24 years) average data of growing seasons (from May to August) for monthly average temperature and total precipitation were obtained from Environment Canada historical data (from 1996 to 2019) at Carman and Winnipeg stations.Weather data from 2019 were compared with the average monthly temperature and total precipitation of the growing seasons from 1996 to 2018.Weather data from 2020 were compared with the average monthly temperature and total precipitation of growing seasons from 1996 to 2019.

Data analysis
Statistical analyses were conducted using SAS software version 9.4 (SAS institute Inc., USA).An analysis of variance (ANOVA) was conducted on all response variables: plant density, spike density, height, anther retention, protein, TWT, TKW, yield, FHB index, incidence, severity, FDK, and DON using PROC Mixed.
ANOVA were performed on all response variables for each site year as well as the combined site years.In the single-siteyear model, the treatment, genotypes, and treatment × genotype were fixed effects.Rep and treatment × rep were listed as random effects.
The data from four site years were evaluated using PROC Univariate to see whether data from the different site years could be combined.A visual inspection of the normality of the distribution of the residuals suggested that the four site years of data could be combined.The model statement listed the treatments, genotypes, and the interaction of treatment × genotype as fixed effects.Environment, rep(environment), environment × treatment, environment × genotype, and environment × treatment × genotype were considered random effects.A Tukey means comparison test was used for comparisons of means for main effects and their interactions.
Eta squared was calculated as described by Brown (2008) to determine the proportion of the variation for each response variable, treatments, genotypes, and the interaction of treatment × genotype, environment, rep(environment), environment × treatment, environment × genotype, and environment × treatment × genotype.Contrasts between main effect treatments were done using PROC Mixed.
Pearson's correlation coefficients between response variables such as yield, TKW, TWT, anther retention, height, FHB index, incidence, severity, FDK, and DON were generated using PROC Corr.

Weather
Generally, total precipitation in June 2019 was lower than in June 2020; Carman 2019 had 37.9 mm, Winnipeg 2019 had 47.23 mm, Carman 2020 had 70.7 mm, and Winnipeg 2020 had 59.18 mm (Figs.1-4).Inoculation started in early July in 2019 and 2020.Over the month of July in 2019, the total precipitation was higher in Winnipeg (101.86 mm) than Carman (57.4 mm), while total precipitation was slightly higher in Carman (54 mm) than in Winnipeg (44.33 mm) in 2020.Overall, Winnipeg had slightly higher average temperatures than Carman in July, with 2020 having higher temperatures than 2019.
According to the long-term average (LTA) data from Environment Canada, the Carman 2019 trial had 70% of the normal total precipitation based on the 23-year average of growing seasons (from May to August).The Winnipeg 2019 had 72% LTA precipitation.The Carman 2020 trial had 64% LTA precipitation.The Winnipeg 2020 trial had 69% LTA precipitation.May 2020 in Winnipeg was the driest month, and July 2019 in Winnipeg was the wettest month among the growing seasons.
It was hotter in 2020 than in 2019.Based on historical data for the LTA, June and July in 2019 and 2020 at both locations recorded at least 1%-11% higher temperatures than the LTA.Among the experimental seasons, June 2020 in Winnipeg had the hottest month, followed by June 2020 in Carman and in Winnipeg.Generally, Winnipeg 2020 was hotter than other experimental environments.Analysis of variance for agronomic traits ANOVA showed there were significant differences among the main treatments for height, protein content, TWT, TKW, and yield (Tables 3 and 4).The main treatment did not affect anther retention.Cultivars had a significant effect on all variables except TKW and yield.Treatment × cultivar interactions were significant for anther retention, protein content, TWT, TKW, and yield.There were significant environmental effects on all agronomic traits except anther retention and TKW.Significant environment × treatment × cultivar interaction was observed for height, anther retention, TWT, TKW, and yield.

Effect of main plot treatments on agronomic traits
There was a significant difference in height for PGR treatment (Table 5).Ethrel™ treatment reduced height significantly compared to the control.Plants treated with Manipu-  5).
Inoculation led to a higher protein content than the uninoculated treatments (Table 5).There were no differences in protein content associated with PGR treatment.
Inoculated treatments showed significantly lower TWT, TKW, and yield compared to uninoculated treatments (Table 5).There were no differences in these traits associated with PGR treatment.In addition, there was no main treatment effect on anther retention (Table 5).

Effect of cultivars on agronomic traits
Plant density and spike density had similar trends among cultivars (Table 6).AAC Tenacious had the highest plant and spike density.AAC Penhold had the lowest plant density and spike density.AAC Tenacious was the tallest cultivar (91.3 cm), followed by AAC Cameron (86.5 cm), Prosper (75.3 cm), and AAC Brandon (71.4 cm).AAC Penhold was the shortest cultivar (64.9 cm) (Table 6).
The percentage of anther retention for AAC Brandon was 54.5%, which was the highest percentage of anther retention (Table 6).AAC Tenacious had the lowest percentage of anther retention (20.3%).Percentage anther retention for Prosper, AAC Cameron, and AAC Penhold was lower than that for AAC Brandon and higher than AAC Tenacious, but they were not statistically different from AAC Brandon and AAC Tenacious.
AAC Tenacious had the highest TWT (80.8 kg/hL) (Table 6).The TWT for AAC Brandon (79.5 kg/hL) was the second highest, but it was not statistically different from the TWTs for AAC Tenacious, AAC Cameron (78.4 kg/hL), and Prosper (77.9 kg/hL).AAC Penhold had the lowest TWT, which was 77.7 kg/hL, but was not significantly different from AAC Cameron and Prosper.

Analysis of variance for disease traits
Combined ANOVA showed that there were significant differences among main treatments, cultivars, main treatment × cultivar interactions, and all interactions with the environment for all disease variables: FHB index, incidence, severity, FDK, and DON (Table 7).Environment effects were significant for FHB index, incidence, severity, and DON.

Effect of main treatments on disease traits
Disease traits were higher for inoculated plots regardless of PGR treatment (Table 8).All uninoculated plots showed little disease for all FHB variables.There was a trend that treatments that were inoculated with FHB combined with the Ma-nipulator™ application showed a higher FHB index (42.9%),disease severity (53.5%),FDK (5.88%), and DON (9.75 ppm) than those inoculated without any PGR control treatment.
The Ethrel™ plus inoculation treatment showed a tendency to lower the FHB index (34.6%),disease incidence (64.4%), disease severity (47.7%),FDK (5.39%), and DON (8.90 ppm) compared to the inoculated without any PGR control treatment.However, PGR treatments did not statistically differ from no PGR treatment for disease traits within the inoculated and uninoculated treatments.

Effect of cultivars on disease traits
Fusarium head blight resistance cultivar AAC Tenacious consistently gave the lowest values for all disease traits measured and was significantly different from the other cultivars for FHB index, disease incidence, FDK, and DON (Table 9).Prosper had the numerically highest FHB index, followed by AAC Penhold, AAC Cameron, and AAC Brandon, but these were not significantly different from each other.The

Effect of Interactions
There were significant environment × treatment × cultivar interactions for height, anther retention, TWT, TKW, yield, and all disease trait variables (Tables 3, 4, and 7).These interactions for height and anther retention were likely caused by different environmental conditions, such as drought and growing temperatures.Examination of the significant threeway interactions for TWT, TKW, yield, and all disease traits revealed that the interactions were mainly due to differences in disease pressure among the environments.Furthermore, most interactions between treatment and environment are mainly due to differences in magnitude between treatments across environments.In some environments, there were no significant differences, but when there were significant differences, the trends were similar across the environments.

Proportion of total variation
Generally, the main treatment effect contributed to the highest proportion of variation for all disease traits: DON, FDK, severity, incidence, FHB index, and sample weight variables; yield, TKW, and TWT (Fig. 5).Comparisons of main treatment contrasts (data not shown) revealed that the presence or absence of inoculation strongly affected disease traits and sample weight variables, while PGR treatment had little effect on these traits.Cultivar contributed to the highest proportion of variation for protein, anther retention, and height.Therefore, protein, anther retention, and height were influenced mostly by genotypes.Although treatment × genotype interactions from combined data were statisti-cally significant for all variables, the proportion of variance attributed to the interactions was relatively small compared to the main effects of treatment and cultivars.More than 10% of total variation was caused by environmental effects in height, anther retention, protein content, TWT, and yield.Especially, approximately 33% of the variation in protein content was affected by the environment.

Correlation between the measured variables
Most of the measured variables were significantly correlated with each other except height and TKW, anther retention and yield, and height and DON content (Table 10).Height was negatively correlated with anther retention, severity, incidence, and FHB index.Significant negative correlations were observed between anther retention and test weight (TWT), and anther retention and TKW.Anther retention was positively correlated with the FHB index, severity, incidence, FDK, and DON.TWT, TKW, and yield were positively correlated with all sample weight variables and negatively correlated with all disease variables.In Carman 2020, lodging showed a significant positive correlation only with height.

Discussion
The average monthly temperatures in June and July 2019 and 2020 at Winnipeg and Carman were at least 1% and as high as 10% higher than the LTA of the growing seasons (from May to August).Total precipitations during the growing seasons were 28% (2019) to 36% (2020) lower than the LTA of the growing seasons.These weather data indicated that the two experimental seasons of 2019-2020 were hot and dry.In general, plants were shorter than expected, especially during the 2019 growing season, and the effect of PGRs was less than expected due to hot and dry weather.After inoculation, field plots were under mist irrigation to provide favourable conditions for FHB development.The development of FHB was successful on all inoculated plots.Uninoculated plots had a low level of FHB from the natural inoculum.This is due to a natural background level of FHB infection in Manitoba fields.
Ethrel™ and Manipulator™ did not affect plant density and spike density.Lack of effect on plant density is expected as PGRs had not been applied prior to measurement of plant density.Since spike density was highly dependent on plant density, a lack of effect from PGR on spike density was Fig. 5. Proportion of total variation allocated to the main treatments and cultivars and their interactions for each response variable.Eta squared was calculated by adding all the sums of squares, then dividing the sums of squares for each of the effects, interactions, and residuals by that total to indicate the relative proportion of total variation explained by each factor in the model (Brown 2008).
expected.Differences in plant density and spike density were mainly determined by cultivars.Their relative ranking in plant density and spike density were similar.This suggests the choice of cultivar is more important than relying on PGRs to influence spike density.
In this study, PGRs were applied in June 2019 and 2020.Combined data showed Ethrel™ significantly reduced plant height by about 5 cm, while Manipulator™ slightly reduced plant height by approximately 2 cm.The response from PGRs can be affected by other factors such as weather conditions, rate and timing of application, and wheat cultivars (Strydhorst et al. 2018).Labelling from both PGRs indicated that they should not be applied under stress conditions such as drought and excessive heat.According to the LTA for growing season precipitation, total precipitation for May and June in 2019 was only about half of normal precipitation or less, which is considered as severely dry.It was not as dry in June 2020 compared to June 2019, but both locations in June 2020 had lower total precipitation than the LTA of growing seasons for total precipitation.In addition to that, the average temperatures in June 2019 and 2020 were 3%-11% higher than the long-term June average temperature.These dry and hot conditions might explain why PGRs did not reduce plant height as much as expected.Height data in 2019 showed PGR treatments had less effect on plant height than in 2020.The ex-perimental mean of plant height in Carman 2019 was about 20 cm shorter than in Carman 2020, and plant height in Winnipeg 2019 was about 8 cm shorter than in Winnipeg 2020.Overall plant height in 2019 was shorter than in 2020.With dryer conditions in 2019, plants were already compact and short.A combination of unsuitable environmental conditions for PGRs could have contributed to poor height reduction.Ethrel™ was more effective in reducing plant height than Ma-nipulator™ across cultivars.The effectiveness of height reduction by PGRs depended on cultivar (Clark and Fedak 1977;Strydhorst et al. 2018).Because three of the cultivars have the GA insensitive semi-dwarfing allele, inhibiting GA production by Manipulator™ to decrease plant height might be unsuccessful compared to Ethrel™.The active ingredient of Ethrel™, ethephon, enhances levels of ET in plants.Ethylene modulates DELLA proteins to be more resistant to the effect of GA (Achard et al. 2003).This stabilization of DELLA proteins by ethylene prevents the plant from growing (Achard et al. 2003;Dugardeyn et al. 2008;Iqbal et al. 2017).Also, ET affects growth hormone auxin signaling by inhibiting plant growth by simulating the breakdown of apical dominance (Wiersma et al. 2011;Strydhorst et al. 2018;Vaseva et al. 2018).Since ET inhibits plant growth through both the GA pathway and the auxin pathway, Ethrel™ could be more effective in reducing plant height than Manipulator™.Cultivar height means across treatments showed the same relative ranking reported in Seed Manitoba 2018.AAC Penhold, AAC Brandon, and Proser were shorter than AAC Tenacious and AAC Cameron because AAC Penhold, AAC Brandon, and Prosper had semi-dwarf alleles.Lack of treatment by cultivar interaction suggests that all cultivars responded to PGRs in a similar manner.
Throughout the 2019 and 2020 growing seasons in Winnipeg and Carman, PGRs did not show any significant effect on anther retention.The most significant variation in anther retention was due to cultivars.Semi-dwarfing cultivars, which are GA-insensitive, AAC Brandon, Prosper, and AAC Penhold, showed higher anther retention.Since GAs regulate cell elongation, semi-dwarf plants may have shorter anther filaments, which enhances anther retention (Buerstmayr and Buerstmayr 2016).The loci for the semi-dwarfing alleles Rht-B1b and Rht-D1b are linked with QTL for anther retention on chromosomes 4B and 4D, respectively (Xu et al. 2019).In this study, anther retention was negatively correlated with height (Table 10).This result agreed with previous studies that showed that semi-dwarf cultivars have higher anther retention than taller cultivars (Buerstmayr and Buerstmayr 2016;He et al. 2016;Steiner et al. 2017;Xu et al. 2019).The tallest cultivar, AAC Tenacious, had the lowest anther retention, while the second tallest cultivar, AAC Cameron, showed slightly higher anther retention than AAC Tenacious and similar levels of anther retention with AAC Penhold and Prosper.The lowest anther retention was for the highly FHB-resistant cultivar AAC Tenacious, indicating that low anther retention could be one of the factors that contributes to FHB resistance in AAC Tenacious.Furthermore, there were positive correlations between anther retention and disease variables, which indicated that anther retention plays a role in FHB susceptibility (Table 10).Disease incidence and anther retention had a higher correlation (r = 0.2) compared to other combinations of disease variables and anther retention.This would be expected because retained anther gives a surface for fungal pathogens to easily invade the floret, consequently increasing initial infection (Strange et al. 1972;Buerstmayr and Buerstmayr 2015;Buerstmayr and Buerstmayr 2016).
Protein content was affected by the main plot treatment in the combined data.The evaluation of this effect showed that the main reason for the differences was due to differences between the inoculated and uninoculated treatments.The protein content of wheat grain is represented as a percentage of the total grain.Thus, protein content in infected kernels is higher than in healthy kernels because infected kernels are smaller, as confirmed by the TKW and TWT values.Therefore, even if there is the same total protein in infected kernels and healthy kernels, % protein is higher in the smaller FHB-infected grains.Cultivars had a significant effect on protein content.This may be mainly due to the different end-use wheat classes of the cultivars.The Canadian Grain Commission groups wheat cultivars depending on their functional characteristics.In this study, the five cultivars were from three different end-use classes: Canada Western Red Spring (CWRS), Canada Prairie Spring Red (CPSR), and Canada Northern Hard Red (CNHR).The Canada Western Red Spring class is known for its high protein content (Canadian Cereals 2019).Canada Prairie Spring Red has a medium protein content.Canada Northern Hard Red has a broad range of protein content between the CWRS and CPSR classes.Protein results from this study corresponded to the end use class of the cultivars, such that CWRS wheat cultivars AAC Brandon and AAC Cameron had higher protein content than CPSR cultivars AAC Tenacious and AAC Penhold.The protein content of Prosper from the CNHR class was in the range of the CWRS and CPSR cultivars.
The main plot treatment had a significant effect on sample weights represented by TWT, TKW, and yield.Simple contrasts determined that the significant treatment effect was due to differences between inoculated and uninoculated treatments and not because of PGR applications.Since Fusarium head blight-infected kernels are smaller and lighter than uninfected kernels (McMullen et al. 2012;Gilbert and Haber 2013), samples from FHB-inoculated plots had lower TWT, TKW, and yield.PGRs are primarily for improving lodging resistance.When PGRs prevent wheat from lodging, yields could be higher than non-PGR-treated wheat.However, because of hot and dry conditions, plants were shorter than usual and were not at risk of lodging from wind and rain events, except in the Carman 2020 trial.It has been reported that PGRs did not affect grain yield in the absence of lodging (Wiersma et al. 2011).In this study, although there was severe wind and rainfall on 30 June 2020, which caused lodging in the Carman 2020 trial, the correlation between lodging and yield in the Carman 2020 data did not show a significant relationship (Table 10).According to lodging data from Carman 2020 (data not shown), plants without PGR treatment showed higher lodging ratings because PGRs reduced plant height and might thicken the stem of the plant, thereby lowering the chance of lodging.Stem lodging is often caused by weather events such as high winds and heavy rain.Since taller plants have longer stems, wind and rain place more force on the stems, which can lead to stem breakage or buckling (Berry et al. 2003;Rademacher 2018).Therefore, shorter plants are less prone to lodging.The correlation between height and lodging supports the fact that there was a significant positive relationship between them (Table 10).Cultivar was the main factor contributing to lodging because all tested cultivars had different heights.As explained earlier, taller cultivars such as AAC Cameron had a higher lodging rating, and the shortest cultivar AAC Penhold had the lowest lodging rate (data not shown).
There were significant differences for TWT among the cultivars.The highest TWT was observed in the FHB-resistant cultivar AAC Tenacious.The moderately FHB-resistant cultivar AAC Penhold had the lowest TWT.In this study, this phenomenon was associated with FHB susceptibility in wheat cultivars.The relative ranking of tested cultivars for TWT was the same as for FDK and similar to other disease traits.This is because FDK is light-weighted; thus, when density of grain is measured, more FDK in a specific volume results in lower TWT.Moreover, the correlation between TWT and all disease variables, including FDK, showed a highly significant negative relationship (Table 10).Therefore, AAC Penhold, which was the most susceptible cultivar to FHB in this study, had the lowest TWT.
Combined data for disease traits showed that PGRs did not affect FHB index, incidence, severity, FDK, and DON.Based on contrasts, the difference between the main treatments was caused mainly by differences between inoculated and uninoculated treatments and not PGR treatment.Ethrel™ slightly reduced all disease traits compared to the control, but this was not statistically significant.This result agrees with findings from Sun et al. (2016) and Fauzi and Paulitz (1994).Sun et al. (2016) found that the application of ethephon did not have any impact on the FHB-resistant level of wheat in greenhouse conditions.Fauzi and Paulitz (1994) reported that the application of ethephon showed no effect on the incidence of spikelet infection and seed infection in dry and wet field conditions when spray inoculation was used.However, when there was wet weather and infested corn was used for inoculation, ethephon increased the incidence of spikelet infection (Fauzi and Paulitz 1994).Unlike Ethrel™, Manipulator™ caused a slight increase in all disease traits except incidence, but it was not significantly different from the control.This result is consistent with findings from Fauzi and Paulitz (1994) that showed that application of CCC did not affect the incidence of seed infection and spike infection in dry or wet weather when spray inoculation was used.These results suggest that Ethrel™ and Manuipulator™ did not change the FHB susceptibility in wheat.
As the most FHB-resistant cultivar, the mean of AAC Tenacious across all treatments showed the lowest values for all disease variables over different environments, indicating that AAC Tenacious has stable performance.AAC Brandon is rated as moderately resistant to FHB but has the highest incidence.This could be because of anther retention, where AAC Brandon had the highest anther retention.Anther retention affects type I resistance for initial FHB infection (Lu et al. 2013;Buerstmayr and Buerstmayr 2015;He et al. 2016) because anthers trapped between the lemma and palea enhance fungal growth and initial infection (Buerstmayr and Buerstmayr 2016).For severity, AAC Brandon ranked as the second lowest among tested cultivars.Previous research demonstrated that semi-dwarfing alleles, Rht-B1b and Rht-D1b, significantly decreased resistance to initial infection but Rht-B1b significantly increased type II resistance (Gosman et al. 2009;Buerstmayr and Buerstmayr 2016).Moreover, Saville et al. (2012) suggested that lines with DELLA accumulation are more susceptible to the initial infection, but more resistant to spreading infection within the spike.In the present study, results from the moderately resistant cultivar AAC Brandon agreed with the findings that Rht-B1b contributes to type II resistance but hinders type I resistance.Although AAC Brandon had a higher disease incidence, it may have strong type II resistance to overcome weak type I resistance.In addition to that, one of AAC Brandon's parents "ND 744", which has high levels resistance to FHB, provides a good genetic background for FHB resistance in AAC Brandon (Mergoum et al. 2005).Except for AAC Brandon, cultivars with semi-dwarfing alleles, AAC Penhold and Prosper, were generally more susceptible to FHB in the present study.Although AAC Penhold and Prosper were not significantly different from other cultivars except AAC Tenacious on all of the disease variables, they had slightly higher values for all of the disease variables than other tested cultivars.For Prosper, incidence was higher than severity, which could be due to the semi-dwarfing allele Rht-B1b.AAC Penhold, which has Rht-D1b as the semi-dwarfing allele, also showed higher incidence than severity, but it showed less reduction of severity compared to semi-dwarf cultivars with Rht-B1b.Even though AAC Brandon and AAC Penhold are at the same FHB resistance level (Seed Manitoba 2018), AAC Penhold appeared to be more susceptible to FHB than AAC Brandon based on higher disease severity, FDK, and DON than AAC Brandon.This suggested that the semidwarfing alleles, Rht-B1b and Rht-D1b, have different effects on FHB.Buerstmayr and Buerstmayr (2016) indicated the different degrees of disease severity caused by both semi-dwarfing alleles may be because Rht-D1b had a stronger impact on anther retention than Rht-B1b.However, in this study, AAC Brandon had higher anther retention than AAC Penhold.
AAC Penhold is rated as moderately resistant to FHB.However, in this study, it tended to have higher disease values than AAC Cameron, which is rated as an intermediateresistant cultivar (Seed Manitoba 2018).AAC Penhold was registered at the Canadian Food Inspection Agency in 2014, and it is derived from the cross "5700PR/HY644-BE//HY469" (Cuthbert et al. 2017).HY644-BE is the source of FHB resistance, and it has a moderate resistance level (Cuthbert et al. 2017).According to cultivar descriptions for AAC Penhold and AAC Cameron, they were evaluated for FHB at Carman in 2011 and 2012 (Fox et al. 2016;Cuthbert et al. 2017).In 2011, AAC Cameron was rated as susceptible and AAC Penhold was rated as intermediate, while in 2012, AAC Cameron and AAC Penhold were rated as moderately resistant and intermediate, respectively.This suggests both cultivars were vulnerable to environmental effects, and it can be speculated that AAC Penhold and AAC Cameron may have similar levels of FHB resistance depending on environmental conditions.In addition to that, there have been some changes in how cultivars are classified in terms of determining resistance levels of FHB since 2016 (A. Brûlé-Babel (personal communication, 2021, University of Manitoba)).With the new FHB evaluation sys-tem, cultivars are rated not only on index but also DON.The combined rating of index and DON accumulation may affect the overall cultivar rating.For example, AAC Cameron might have been rated as a moderately resistant cultivar based on index, but it accumulated more DON, so it might be registered as an intermediate resistance level of FHB.Also, Fusarium graminearum isolates for FHB testing have changed since 2018 due to the update of the standard FHB screening across western Canada.AAC Penhold might be rated as moderately resistant based only on disease symptoms with old isolates.How FHB ratings have been applied through the registration system may affect the rankings today with the use of modern isolates of F. graminearum and with DON included in the analysis.Therefore, during the 2019 and 2020 field seasons in Winnipeg and Carman, AAC Penhold showed higher FHB disease levels than the intermediate cultivar AAC Cameron, suggesting that AAC Penhold may not be moderately resistant to new populations of F. graminearum, thereby new evaluations may be required to accurately report FHB resistance levels.
The present study clearly showed that height, anther retention, protein content, TWT, and yield are quantitative traits that are affected by not only genotypes but also the environment.Each environment was different in terms of temperature and precipitation during the experiments (Figs.1-4), and mean plant heights in 2019 were generally shorter than those of 2020.Experimental means for anther retention, TWT, and yield varied depending on the environment.Particularly, protein content was highly influenced by the environment (Fig. 5).Although protein content depended primarily on genotype in this study, it is well known as one of the traditional traits that is strongly influenced by environmental factors such as drought and post-anthesis temperature (Triboi et al. 2003).Therefore, the significant environmental effect was caused by big differences in weather and disease levels between Winnipeg and Carman in 2019 and 2020.
This research determined how Manipulator™ and Ethrel™ affect FHB, plant height, and yield of spring wheat, as well as the interaction between semi-dwarfing alleles and PGRs.In this study, application of PGRs did not affect FHB in all tested spring wheat cultivars that differed in the level of resistance to FHB and plant height.Ethrel™ significantly reduced plant height, and Manipulator™ slightly reduced plant height.Both PGRs did not have a significant effect on anther retention, TWT, TKW, and yield.Cultivars with semi-dwarfing alleles tended to have higher anther retention.There was a different impact on FHB resistance depending on the type of semi-dwarfing allele.Cultivars with Rht-B1b showed stronger resistance to FHB than the cultivar with Rht-D1b at the same FHB resistance level from Seed Manitoba (2018).The level of disease variables such as FHB index, incidence, severity, FDK, and DON in tested cultivars showed the same rankings as reported in Seed Manitoba (2018), except for AAC Penhold.
This research provides important information for producers, especially in the higher moisture regions of the prairies, where FHB epidemics are most prevalent.These are the regions that have the highest yield potential and risk of lodging under intensive management and would benefit most from PGR applications.Since there was no significant effect of both PGRs on FHB, growers could choose to grow the taller FHB resistance cultivars and control height with PGRs to reduce the risk of yield losses due to lodging.However, growers will need to weigh the cost of PGRs against the potential gain in yield and quality in terms of value when using PGRs in tall FHB-resistant cultivars.It may be possible for growers to benefit from the genetics of FHB resistance and use intensive management methods without sacrificing potential yield due to lodging.Implementation of PGRs to control plant height would also allow breeders to select the lines with the highest FHB resistance without focusing so closely on plant height.
2019 and 2020, field trials were located at the University of Manitoba's Ian N. Morrison Research Station in Carman, Manitoba and the University of Manitoba's Fort Gary campus (the Point), Winnipeg, Manitoba.In 2019, the trial in Carman was sown on 8 May, and the trial in Winnipeg was sown on 17 May.In 2020, the Carman and Winnipeg field trials were sown on 6 May and 19 May, respectively.

Table 3 .
Combined ANOVA for agronomic traits: plant density, spike density, height, anther retention, and protein content from four test environments (Carman and Winnipeg for 2019 and 2020).

Table 4 .
Combined ANOVA for agronomic traits: test weight, thousand kernel weight, and yield from four test environments (Carman and Winnipeg for 2019 and 2020).

Table 5 .
Least Means followed by the same letter in a column are not significantly different at P = 0.05 based on the Tukey means comparison test.
square means for agronomic traits: height, protein, test weight, thousand kernel weight (TKW), and yield for the different main treatments tested across cultivars from pooled data from four test environments (Carman and Winnipeg for 2019 and 2020).Note:

Table 6 .
Least squares means for agronomic traits: plant density, spike density, height, anther retention, protein, and test weight for cultivars tested across different main treatments from pooled data from four test environments (Carman and Winnipeg for 2019 and 2020).Means followed by the same letter in a column are not significantly different at P = 0.05 based on the Tukey means comparison test. Note:

Table 8 .
Least square means for disease traits: Fusarium head blight (FHB) index, incidence, severity, Fusarium damaged kernel (FDK), and deoxynivalenol (DON) content for the different main treatments tested across cultivars from pooled data from four test environments (Carman and Winnipeg for 2019 and 2020).
Note: Means followed by the same letter in a column are not significantly different at P = 0.05 based on the Tukey means comparison test.

Table 9 .
Least squares means for disease traits: Fusarium head blight (FHB) index, incidence, severity, Fusarium damaged kernel (FDK), and deoxynivalenol (DON) content for cultivars tested across different main treatments from pooled data from four test environments (Carman and Winnipeg for 2019 and 2020).
There were no significant differences in FDK between AAC Brandon, AAC Cameron, and Prosper.Prosper had the numerically highest DON content, followed by AAC Penhold, AAC Brandon, and AAC Cameron.There were no statistical differences between them.
Note: Top row indicates correlation coefficients (r), second row indicates possibility, and last row indicates sample size (n) in the cell.a Represents only Carman 2020 trial.