Interactions in the Brassica napus–Pyrenopeziza brassicae pathosystem and sources of resistance to P. brassicae (light leaf spot)

This is an open access article under the terms of the Creat ive Commo ns Attri bution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2021 The Authors. Plant Pathology published by John Wiley & Sons Ltd on behalf of British Society for Plant Pathology School of Life and Medical Sciences, Centre for Agriculture, Food and Environmental Management Research, University of Hertfordshire, Hatfield, Hertfordshire, UK

lineage from the isolates found in other geographic regions (Carmody et al., 2020). P. brassicae is a polycyclic pathogen and can be present throughout the growing season, affecting crops at all plant growth stages (Gilles et al., 2001). Severe LLS can cause up to 30% yield reduction of winter oilseed rape (Agriculture and Horticulture Development Board, 2021). In vegetable brassicas, apart from yield losses, quality of the fresh produce can also be severely affected, reducing their market value (Karandeni Dewage et al., 2018).
Currently, LLS is considered as the most economically damaging foliar disease of oilseed rape in the UK. Severity of LLS epidemics has increased in recent years and it has replaced phoma stem canker (Leptosphaeria maculans) as the main disease of winter oilseed rape in the UK (Karandeni Dewage et al., 2018). Therefore, LLS has become one of the main targets for fungicide applications and for oilseed rape breeding programmes. However, various constraints associated with fungicide applications, such as poor timing (Gilles et al., 2000), development of fungicide insensitivity within P. brassicae populations (Carter et al., 2013) and legislative and environmental concerns (Hillocks, 2012) make chemical control of LLS unreliable. Therefore, the use of resistant cultivars, which is cost-effective and environmentally safe, remains an important alternative to fungicide application. However, limited understanding of the mechanism of host resistance against P. brassicae hinders the potential success of the deployment of cultivar resistance to control LLS. Often, oilseed rape breeders have to rely on sources of host resistance without knowledge of their genetic background when breeding for resistant cultivars.
Cultivation of resistant cultivars without knowledge about their genetic background can lead to growing of cultivars with the same resistance gene/s over a long period of time. This can ultimately exert selection on local pathogen populations, leading to the build-up of virulent pathogen races that reduces the efficacy of corresponding resistance genes over time (McDonald & Linde, 2002). For example, a sudden change in cultivar Bristol from resistance to susceptibility has been reported following its extensive cultivation across the UK (Boys et al., 2007). Additionally, there have been recent records of the breakdown of resistance in some cultivars with good resistance ratings (based on UK Agriculture and Horticulture Development Board (AHDB) recommended list [RL] ratings) for the north region of the UK (Karandeni Dewage et al., 2018). Therefore, it is important to identify and differentiate cultivar resistance against P. brassicae and to monitor the frequency of virulent pathogen races. This enables the spatial and temporal deployment of different sources of cultivar resistance to increase their durability.
Host resistance against P. brassicae can be either isolatespecific (qualitative) or nonspecific (quantitative). There have been several studies that focused on analysing the genetic basis of resistance in oilseed rape against P. brassicae. Some of these studies have used pathogen populations, in the form of natural ascospore (sexual spore) inoculum or conidia (asexual spores) of isolate mixtures taken from diseased leaf samples (Boys et al., 2012;Bradburne et al., 1999;Pilet et al., 1998). A very few studies have been performed with single-spore isolates of P. brassicae, including the assessment of seedling resistance of B. napus (Bradburne et al., 1999;Karolewski, 1999) and cross-infectivity of P. brassicae between different host Brassica species (Boys, 2009;Maddock et al., 1981). Simons and Skidmore (1988) have provided experimental evidence for the presence of differential interactions (specific resistance and corresponding virulence) between Brassica oleracea and P. brassicae. However, there have been no further studies on isolate-specific resistance against the LLS pathogen since then. The limited number of experiments with individual isolates limits the knowledge of the genetic basis of resistance against P. brassicae, which is important for successful deployment of cultivar resistance.
One of the challenges associated with investigating isolatespecific interactions in the Brassica napus-P. brassicae pathosystem is the detection and quantification of phenotypic changes related to P. brassicae infection. In general, LLS severity assessments are based on the amounts of P. brassicae acervuli (asexual sporulating structures) on the plants, measured as the percentage area covered with sporulation or using disease severity assessment scales (Boys et al., 2012;Bradburne et al., 1999;Karolewski, 1999;Pilet et al., 1998).
Although some other changes linked with P. brassicae infection such as necrotic responses (Boys et al., 2012;Bradburne et al., 1999), leaf deformations (e.g., leaf curling, leaf distortion, petiole elongation), and stunting of the plants (Ashby, 1997(Ashby, , 2000 have been reported, these features have not been directly accounted for in current LLS assessments. It is also not clear if and how these changes are linked with each other and whether they reflect different defence mechanisms against P. brassicae. Therefore, it is essential to investigate these resistance/susceptibility phenotypes to gain a basic understanding of B. napus-P. brassicae interactions and to improve cultivar resistance through selective breeding for effective management of LLS.
The experimental work described in this paper, which used single-spore isolates of P. brassicae, aimed to investigate specific host-isolate interactions and different phenotypic changes associated with LLS to aid the identification and selection of different sources of host resistance. A better understanding of these aspects can lead to more effective management of LLS.

| MATERIAL S AND ME THODS
To investigate B. napus cultivar/line interactions with single-spore P. brassicae isolates, two glasshouse experiments were performed.
The experiments had four replicates and were arranged in splitplot designs. In the first experiment, there were 10 B. napus cultivars/lines tested against eight P. brassicae isolates. In the second experiment, the same eight isolates were tested against another eight cultivars/lines together with four control cultivars. In both experiments, there were the same resistant and susceptible control cultivars/lines, so that the results of both experiments could be analysed together.

| Selecting P. brassicae isolates
The single-spore (single-conidial) isolates of P. brassicae were obtained from oilseed rape crops. Diseased B. napus leaves were collected and incubated at 4℃ for 5 days in the dark to induce P. brassicae asexual sporulation. After incubation, single acervuli were harvested from leaves using a sterile needle and placed in a 0.5 ml microfuge tube containing 30 μl of sterilized distilled water.
Tubes were vortexed briefly to liberate the conidia and the spore suspension was placed onto a potato dextrose agar plate and spread across the plate using a sterile L-shaped plastic spreader (Sterilin).
Inoculated plates were incubated at 15℃ in the dark for 2 days and observed for spore germination. Plates were further incubated for 1 week and single-spore isolates were selected for subculturing. A mycelial plug was taken from each selected P. brassicae isolate and placed in a 2 ml Eppendorf tube containing 200 μl of sterilized distilled water. The mycelial plug was ground using a sterilized plastic pestle and macerated mycelia were inoculated onto malt extract agar plates (100 μl inoculum per plate). Inoculated plates were incubated at 15℃ in the dark for 3 weeks for spore (conidia) production. Spore suspensions were prepared by adding 10 ml of sterilized distilled water onto each of the P. brassicae culture plates. All the mycelia were scraped off and macerated in water using a sterile L-shaped plastic spreader and then left for 5 min to liberate spores. Spore suspensions were filtered through sterile Miracloth (Calbiochem) into 50 ml Falcon tubes and a haemocytometer (Bright-Line; Hausser Scientific) was used to determine the spore concentration. Spore suspensions from single-spore isolates were diluted to obtain the required concentration of 10 5 spores/ml and stored at −20℃ until needed. A total of 18 single-spore isolates of P. brassicae were obtained from diseased leaf samples collected in 2014/2015 or 2015/2016 cropping seasons and eight isolates were selected based on their cultivar of origin and geographic location in the UK (Table 1).

| Selecting B. napus cultivars/lines
A set of 18 oilseed rape cultivars/lines was selected. Selection of commercial oilseed rape cultivars was based on the resistance rating against P. brassicae in AHDB RL trials. Cultivars Bristol and Marathon were added as the susceptible controls and cultivars Imola and Cuillin were resistant controls, included in both experiments. In addition to cultivar Imola, which harboured a major gene locus for resistance, three lines were selected from a doubled haploid (DH) population and parents of two other DH populations with diverse genetic backgrounds were also included. Detailed descriptions of each cultivar/ line, along with specific reasons for selecting it, are given in Table S1.

| Plant growth and inoculation
Seeds of the selected B. napus cultivars/lines were pregerminated at 20℃ in the dark for 48 h in a Petri dish containing a damp filter paper. Pregerminated seeds were then sown in 40-cell seed trays filled with compost prepared by mixing Miracle-Gro allpurpose compost (The Scotts Company) and John Innes No. 3 (LBS Horticulture) in a 1:1 ratio. After sowing, the compost surface was covered with a thin layer of vermiculite to maintain its moisture.
Seedlings were then transplanted into 9 cm diameter pots containing the same compost mixture and transferred to a temperatureregulated glasshouse at 16℃ day/14℃ night temperatures with natural daylight supplemented by a 12 h photoperiod. Plants were arranged in a split-plot design generated using Experiment Design Generator and Randomiser (EDGAR) (Brown, 2004). P. brassicae isolates were randomly assigned to main plots (blocks) and cultivar/

| LLS assessment
Plants were monitored regularly and the timing and appearance of P. brassicae infection-related changes were recorded. At 24 days postinoculation (dpi), plants were destructively harvested by cutting at the stem base above the compost surface, individually placed in polyethylene bags with a dampened paper towel and incubated at 4℃ for 5 days to induce sporulation. The final disease assessment was made at 29 dpi. Disease severity was measured as percentage leaf area covered with P. brassicae sporulation (acervuli) and plants were also given a LLS score using a 1-6 scale, where 6 was most susceptible; Table S2). In addition, the number of leaves with deformations (leaf curling, leaf distortion, petiole elongation) and the presence of necrotic responses were recorded for each plant.

| Data analysis
Analysis of variance (ANOVA) was based on a standard split-plot design in which the isolate treatments were the main plots and the cultivar/line treatment was randomly arranged in each main plot. The residual error from the main plot (i.e., error A) was used to test the effect of isolate and to calculate the least significant difference (LSD) to compare the differences between isolates. The residual error from the subplot (i.e., error B) was used to test the effect of cultivar/line and the interaction of isolate × cultivar/line, and to calculate the LSD to compare the differences between cultivars/lines and between combinations of the two-way interactions. These data analyses on different trait measurements were completed using GENSTAT 18th edition statistical software for Windows (Payne et al., 2011). LLS severity (percentage leaf area covered with P. brassicae sporulation) and leaf deformation (percentage leaves deformed, calculated using the number of leaves with deformations and the total number of leaves per plant) data were transformed by taking the arcsine of the square root of the proportion value (Sokal & Rohlf, 1995), so that the variance was more homogeneous across treatments and measurements were normally distributed, before the ANOVA was done. If the F test showed a significant effect of any factor, the standard error of the difference and the LSD were calculated and presented at a probability level of 5% (p ≤ 0.05). For analysis of the relationships between different measures (LLS score, percentage leaf area covered with P. brassicae sporulation, and percentage leaves deformed), simple linear regression analyses were done using calculated means for different cultivars/lines. Differences between different cultivars/lines or different isolates were tested using comparative analysis of position and parallelism of linear regression.
Two parameters were used to assess LLS severity: percentage leaf area with P. brassicae sporulation (acervuli) and disease score on 1-6 scale. There was a positive correlation between the two parameters ( Figure S1). However, percentage leaf area with acervuli appeared to be a better quantitative measure than disease score when assessing the differences in disease severity, especially at the high disease severity levels, because acervuli can appear without any lesions associated with them. Therefore, the percentage leaf area with sporulation was taken as the preferred parameter for an- Interactions between B. napus genotypes and P. brassicae isolates were further analysed using the linear regression analysis described by Ghazvini and Tekauz (2008) to identify differential responses.
Resistant/susceptible responses of the B. napus genotypes (different cultivars/lines) measured quantitatively as the percentage leaf area with P. brassicae sporulation were analysed using the regression model expressed as: where, in this study, Y ij is the mean percentage leaf area with sporulation of the ith cultivar on the jth isolate, μ i is the mean percentage leaf area with sporulation of the ith cultivar/line over all isolates, b i is the regression coefficient that measures the response of the ith cultivar/line to varying levels of virulence in isolates, I j is the isolate virulence index, which is defined as the mean percentage leaf area with sporulation of all cultivars/lines with the jth isolate minus the grand mean, and δ ij is the sum of deviations from regression of the ith cultivar/line with the jth isolate. Regression slope parameter and deviation mean squares for each cultivar/line were taken as measures of disease severity response towards increased isolate virulence and the specificity of the resistance/susceptibility of the cultivars/lines, respectively.

| Symptoms of P. brassicae infection on host plants
According  Parents of the TN DH population, Tapidor and Ningyou7, also differed in their interactions with some of the isolates, even though there was no statistically significant difference in the overall mean LLS severity between the two cultivars (Table 2).
Moreover, we detected significant differences in virulence between P. brassicae isolates (p < 0.01) on different oilseed rape cultivars/lines; isolate mean ranged from 12.3% to 37.0% leaf area with sporulation (  Note: Values in bold numbers indicate interactions that produced necrosis (black flecking phenotype). Details of all the cultivars/lines and isolates used in this study are given in Table S1 and Table 1, respectively. S, susceptible control cultivars.
Values marked with a common lower case letter do not differ at p ≤ 0.05. Abbreviation: LSD, least significant difference. season; this may be useful for farmers due to the delay in appearance of other symptoms, such as acervuli or leaf lesions, after infection. There have been previous studies that also reported various symptoms such as stunting of plants, stem elongation, leaf curling, and green island formation with P. brassicae infection (Ashby, 1997(Ashby, , 2000 and they have also been commonly observed in field crops infected with P. brassicae.
Leaf deformations are considered to be indicative of plant growth regulator imbalance in infected plants (Ashby, 1997(Ashby, , 2000. During the early endophytic phase, the pathogen has to rely on living host tissues to obtain nutrients, maintaining a fine balance in host-pathogen interactions to avoid significant tissue damage that could activate host resistance mechanisms. It has been suggested that in the case of the P. brassicae-B. napus interaction, provision of nutrients is facilitated by cytokinins that alter host metabolism and translocate nutrients to the site of infection (Murphy et al., 1997).
Experimental evidence suggests that in resistant interactions, the host recognition occurs at a later stage of P. brassicae colonization and does not prevent early colonization (Boys et al., 2012); this might provide a possible explanation for the appearance of leaf deformation in resistant as well as susceptible cultivars/lines. However, some host genotypes may be less sensitive to hormonal changes than others, as leaf deformations appeared to be more prominent on some cultivars/lines than others.
Our results suggest that formation of black necrotic flecking and limitation of P. brassicae asexual sporulation (acervuli) are generally indicative of host resistance, and these two phenotypes coincided most of the time. This observation was consistent with previous findings that reported a host resistance response against P. brassicae associated with black necrotic flecking in oilseed rape cultivar Imola, limited growth of the pathogen during the subcuticular growth phase, and no asexual sporulation (Boys et al., 2012). Furthermore, an earlier report referred to black flecking with limited asexual sporulation (Bradburne et al., 1999). However, for the first time our study enabled comparisons to be made between different oilseed rape cultivars/lines that generate a black necrotic flecking phenotype and their disease response.
Our investigation of this pathosystem revealed significant differences between cultivars/lines, between isolates, and between cultivar/line-by-isolate interactions. The analysis also suggested the presence of isolate-specific resistant interactions. Resistance against P. brassicae is often measured quantitatively based on the extent of pathogen sporulation (Boys et al., 2007;Bradburne et al., 1999 Hence, percentage leaf area covered by acervuli appeared to be a more reliable measure of host resistance.

F I G U R E 4
Differential phenotypes produced on the parents of the DY (Darmor × Yudal) doubled haploid (DH) population of oilseed rape. In a glasshouse experiment, the parental lines of the DY DH population produced differential phenotypes when sprayinoculated with single-spore isolates of Pyrenopeziza brassicae: (a) P. brassicae sporulation (S) on cv. Darmor-susceptible phenotype, (b) black necrotic flecking (F) on cv. Yudal-resistant phenotype.
Pictures were taken at 29 dpi. dpi, days postinoculation F I G U R E 5 Relationships between mean percentage leaf area with Pyrenopeziza brassicae sporulation (mean disease severity) and the regression parameters calculated for the interactions between 18 oilseed rape cultivars/lines and eight P. brassicae isolates. (a) regression coefficient, (b) deviation mean squares (MS) Cultivar Imola, which was identified as the most resistant in this study, has been previously described as containing a major gene for resistance against P. brassicae (Boys et al., 2012). Even though Imola is not in commercial use, it is likely that there are other oilseed rape cultivars that carry this source of resistance. Additionally, most of the DH lines screened in this study showed a greater level of resistance to P. brassicae than commercial cultivars. More importantly, the type of host resistance carried by Imola and those DH lines appeared to be nonspecific and less sensitive towards increased virulence of the isolates tested. Therefore, these could be exploited as effective sources of resistance in oilseed rape breeding programmes. Secondary spread of the disease through conidia substantially contributes to the widespread disease occurrence (Evans et al., 2003;Fitt et al., 1998;Gilles et al., 2001). Conidia dispersed by rain-splashes may also contribute to the spread of the pathogen up crop canopies to infect upper canopy leaves, flowers, and pods (Boys et al., 2007). Therefore, host resistance that prevents or reduces the production of acervuli is a valuable resource for the control of LLS. Bristol broke down in the early 1990s following its commercial deployment (Boys et al., 2007). The present study also suggested the existence of isolate-specific interactions for cv. Bristol. In the past, breeding for cultivar resistance in oilseed rape has mainly focused on phoma stem canker (L. maculans), which was the major disease problem on oilseed rape in the UK for many years, with occurrence of frequent epidemics before LLS become dominant. However, cultivars with good resistance against one pathogen but with poor resistance against other major pathogens are of little value to growers. This presents plant breeders with the challenge of equipping cultivars with resistance against several pathogens. Therefore, it is necessary to evaluate the cultivars with good resistance against L. maculans for resistance against P. brassicae and vice versa. Our study included three cultivars (Excel, Harper, and Hearty) with good resistance scores for phoma stem canker. These cultivars are known to carry the Rlm7 gene for resistance against L. maculans that is highly effective for controlling phoma stem canker in the UK (Mitrousia et al., 2018). Remarkably, all three cultivars scored some of the greatest overall LLS severities with no resistant interactions. There has been no evidence of a direct effect of the Rlm7 gene on susceptibility to P. brassicae. However, it is important to investigate this further to see whether these three Rlm7 cultivars have originated from a background susceptible to P. brassicae.
The isolates used in this study represent three different oilseed rape-growing areas in the UK: Herefordshire, Cambridgeshire, and Norfolk. There were significant differences between different isolates. In a preliminary experiment with P. brassicae populations (conidial suspensions collected from diseased leaves from oilseed rape fields), we observed occasional acervuli on cultivar Imola, in contrast to Boys et al. (2012), who reported no acervuli observed on Imola at any time. It was speculated that acervuli observed on this cultivar in our preliminary experiment might be an indication of the breakdown of resistance and therefore, it was anticipated that single-spore isolates 17WOSR-I1 and 17WOSR-I4 obtained from cv.
Imola, would not produce a necrotic response. However, these two isolates induced black flecking on Imola, Yudal, and the DH line Q83.
Four of the isolates, including 17WOSR-I1, were able to produce a small number of acervuli along the midribs of Imola leaves. This suggests that the resistance in cv. Imola limits P. brassicae asexual sporulation rather than completely eliminating it.
This study provided evidence that different types of resistance against P. brassicae are present in different oilseed rape genotypes.
We have demonstrated the possibility of using single-spore isolates of P. brassicae for pathogenicity assays and for identification of cultivar-by-isolate interactions using statistical methods. This knowledge can be extended to characterize the population dynamics of P. brassicae. More importantly, we have identified several sources of resistance that are valuable for oilseed rape breeding programmes.
Further investigation of isolate-specific interactions between B. napus and P. brassicae using a large panel of host plant material and P. brassicae isolates, especially those representing different oilseed rape-growing regions, may provide valuable information for effective deployment of host resistance to manage LLS.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.