A phylogenetically distinct lineage of Pyrenopeziza brassicae associated with chlorotic leaf spot of Brassicaceae in North America

Abstract Light leaf spot, caused by the ascomycete Pyrenopeziza brassicae, is an established disease of Brassicaceae in the United Kingdom (UK), continental Europe, and Oceania (OC, including New Zealand and Australia). The disease was reported in North America (NA) for the first time in 2014 on Brassica spp. in the Willamette Valley of western Oregon, followed by detection in Brassica juncea cover crops and on Brassica rapa weeds in northwestern Washington in 2016. Preliminary DNA sequence data and field observations suggest that isolates of the pathogen present in NA might be distinct from those in the UK, continental Europe, and OC. Comparisons of isolates from these regions using genetic (multilocus sequence analysis, MAT gene sequences, and rep‐PCR DNA fingerprinting), pathogenic (B. rapa inoculation studies), biological (sexual compatibility), and morphological (colony and conidial morphology) analyses demonstrated two genetically distinct evolutionary lineages. Lineage 1 comprised isolates from the UK, continental Europe, and OC, and included the P. brassicae type specimen. Lineage 2 contained the NA isolates associated with recent disease outbreaks in the Pacific Northwest region of the USA. Symptoms caused by isolates of the two lineages on B. rapa and B. juncea differed, and therefore “chlorotic leaf spot” is proposed for the disease caused by Lineage 2 isolates of P. brassicae. Isolates of the two lineages differed in genetic diversity as well as sensitivity to the fungicides carbendazim and prothioconazole.

Dispersal of P. brassicae inoculum during the growing season in areas where this pathogen is established is considered mainly to be by short distance splash-dispersal of asexual conidia, with multiple (polycyclic) rounds of host infection (Gilles et al., 2001;Karandeni Dewage et al., 2018). In addition, wind-dispersed ascospores are released into the air forcibly from apothecia that form on infected host debris, typically in late summer and autumn (Cheah et al., 1982;Gilles et al., 2001). Ascospores are thought to act as primary sources of inoculum that initiate light leaf spot outbreaks in the UK and continental Europe (Karolewski et al., 2012).
Sexual reproduction by P. brassicae has long been documented in the UK and continental Europe (Lacey et al., 1987) as well as OC (Cheah et al., 1982). Isolates of complementary MAT1-1 and MAT1-2 types are required for sexual reproduction (Ilott et al., 1984;Foster et al., 2002). Apothecia have not been found in association with outbreaks of light leaf spot in NA, and it is not known whether a sexual cycle occurs in NA. However, this information is important to underpin management strategies for light leaf spot, as populations with both sexual and asexual reproduction tend to have greater evolutionary potential than those that are exclusively asexual (McDonald and Linde, 2002). Such populations also present a greater risk of failures in disease management strategies, for example, if strains of the pathogen overcome host resistance genes (Boys et al., 2007) or develop resistance to fungicides commonly used in brassica crops, as has occurred in the UK and continental Europe (Carter et al., 2013(Carter et al., , 2014.
Effective management of light leaf spot in areas where this disease has established has necessitated the integration of planting cultivars with resistance to the disease, applying fungicides with efficacy against the pathogen, and implementing cultural practices such as incorporation of infected crop residues into the soil and/ or crop rotation (Karandeni Dewage et al., 2018). Host resistance alone has been insufficient to control economically damaging outbreaks of light leaf spot in B. napus crops as there are no fully resistant commercial cultivars currently available (Boys et al., 2007(Boys et al., , 2012. Thus, management of this disease in conventional crops has depended on applications of fungicides, including methyl benzimidazole carbamates (MBCs, Fungicide Resistance Action Committee [FRAC] Group 1) and azoles (sterol 14α-demethylation inhibitors [DMIs], FRAC Group 3; Carter et al., 2013Carter et al., , 2014. However, reduced sensitivity to these fungicides has been reported for some UK and continental European isolates of P. brassicae, and the molecular mechanisms of resistance have been characterized (Carter et al., 2013(Carter et al., , 2014. Genotypic and phenotypic data on fungicide sensitivity of NA isolates of the light leaf spot pathogen are needed to monitor the current and future potential efficacy of fungicide applications for control of this disease in NA.

Given the increasing losses associated with light leaf spot in areas
where this disease is well established, and preliminary evidence of genetic differentiation of isolates of the fungus causing this disease in NA from isolates in the UK and continental Europe, there is a need to characterize these pathogen populations. The primary objective of this study was to compare isolates of the light leaf spot pathogen from regions where P. brassicae has long been established, i.e., the UK and continental Europe and OC (Majer et al., 1998), with isolates from NA, where light leaf spot was found recently. The isolates evaluated in this study were obtained from a range of Brassicaceae genera and species, and compared using the consolidated species concept (CSC) by combining morphological, ecological, biological, and genetic (phylogenetic) data (Crous et al., 2015).

| Pyrenopeziza isolates and herbarium specimens
Details of the light leaf spot fungal isolates used in this study, including isolates and herbarium specimens of infected leaves submitted to the Westerdijk Fungal Biodiversity Institute (WFBI) in the Netherlands, isolates deposited in the International Mycological Institute (IMI) collection in the UK, and GenBank accession numbers for fungal DNA sequences, are listed in Table 1. The GenBank accession numbers listed in Table 1 were all generated as part of this study except for the following: OC isolates were obtained from the WFBI (CBS157.35) or the IMI herbarium (IMI233715 to IMI233717), and the older isolates from the UK or continental Europe (preceding 2000) were obtained from collectors or programmes listed in Table 1.
For each of the UK or continental Europe isolates generated in this study, infected leaves from a collection at Rothamsted Research were examined with a stereomicroscope, and a single pustule was placed into a drop of sterile distilled water (SDW) using a sterilized needle. The conidial suspension was spread onto a plate of 3% malt extract agar (MEA) using a sterilized disposable loop, and incubated at 15 °C for 10 days. Single colonies were then used to establish single-spore cultures. For each NA isolate, small pieces (up to 5 mm 2 ) of leaf and stem tissue with symptoms were surface-sterilized in 1.2% NaOCl for up to 2 min, and rinsed three times in SDW; or sterilized in 70% ethanol for 5 s, dried on sterilized blotter paper, and plated onto clarified V8 (cV8) agar amended with chloramphenicol (100 mg/L; Carmody, 2017

| Genus confirmation and multilocus sequence analysis
To verify identity of the genus of the NA isolates as Pyrenopeziza, phylogenetic analyses were completed for the partial ITS rDNA of 30 isolates of the light leaf spot pathogen (12 from NA, 13 from the UK, 4 from continental Europe, and 2 from OC) along with ITS rDNA sequences of isolates of 57 related fungi, including sequences available in GenBank for seven other Pyrenopeziza species (P. ebuli, P. eryngii, P. petiolaris, P. plicata, P. revincta, P. subplicata, and P. velebitica), nine Cadophora species, two Graphium species, Hormodendrum pyri, two Hymenoscyphus species, Leptodontidium orchidicola, five Mollisia species, three Oculimacula species, four Phialophora species, two Phialocephala species, two Rhynchosporium species, and Tapesia cinerella (Table 1; Table S1; Figure 1a). The ITS rDNA sequence obtained from a genome of Botryosphaeria dothidea served as the outgroup (Table S1). In addition, the β-tubulin and translation elongation factor 1-α (TEF1-α) genes were amplified from the same 30 isolates of P. brassicae isolates from the UK and continental Europe, OC, and NA, as well as from closely related fungi (Table 1; Table S1), for completing individual phylogenetic analyses of each DNA region as well as multilocus sequence analysis (MLSA) of concatenated sequences of the three DNA regions. Relevant sequences from B. dothidea served as outgroups for these analyses (Crous et al., 2003;Table S1;Figure 1b-d).
Primers used for the amplification of various DNA sequences are detailed in Table 2. The ITS rDNA was amplified as described by Bakkeren et al. (2000) in a total reaction volume of 30 μl that included 1× buffer (Invitrogen Life Technologies), 1.5 mM MgCl 2 , 0.2 mM of each dNTP, 0.4 mM of each primer, 1.5 U Taq DNA polymerase (Invitrogen Life Technologies), and 2 μl genomic DNA. The β-tubulin gene was amplified as detailed by Einax and Voigt (2003)   β-tubulin = β-tubulin gene; TEF1-α = translation elongation factor 1-α gene; MAT = mating type genes of the light leaf spot pathogen (Ilott et al., 1984;Foster et al., 2002). All sequences with accession numbers in this table were generated in this study. b Isolates confirmed as MAT1-1 or MAT1-2 type using the multiplex PCR assays of Foster et al. (2002). All mating type sequences with accession numbers in this table were generated as part of this study. c S, isolates from continental Europe and UK (n = 10) inoculated onto Brassica rapa 'Hakurei' to compare symptomology with that caused by North American isolate Cyc001, as detailed in the main text. M = isolates from continental Europe and UK (n = 4) compared with isolates from North America (n = 10) for morphology on malt extract agar, as detailed in the main text. CO, isolates used to compare conidial morphology in vitro and in vivo, as detailed in the main text.
d Type specimen of P. brassicae examined in the form of apothecia in dried culture (Rawlinson et al., 1978). Only a partial ITS rDNA sequence (MN028386) could be amplified from the herbarium specimen.
F I G U R E 1 Phylogenetic trees from Bayesian analysis of multiple gene sequences obtained from Pyrenopeziza brassicae isolates from the United Kingdom (UK), continental Europe (EU), North America (NA), and Oceania (OC), as well as other fungal genera and species. Trees were constructed with partial sequences from (a) the internal transcribed spacer (ITS) region of ribosomal DNA (rDNA); (b) the β-tubulin gene; (c) the translation elongation factor1-α (TEF1-α) gene; and (d) the concatenated sequences from all three regions. Bayesian posterior probabilities are indicated at the nodes (BPP). The outgroup sequence used for each analysis was from Botryosphaeria dothidea. Refer to Table 1 and Table S1 for details of the isolates and sequences polymerase, and 1 μl genomic DNA. The TEF1-α gene was amplified using the protocol described by Taşkin et al. (2010)  and 72 °C for 5 min for TEF1-α amplification.
After running the amplified products on 1.5% agarose gels to confirm single bands, PCR products were cleaned using an ExoSAP-IT kit (ThermoFisher Scientific) and sent to Elim Biopharmaceuticals, Inc. for bidirectional sequencing. Primers used for PCR amplification were also used in the sequencing reactions ( Table 2). The DNA sequences were processed using MEGA v. 7 (Kumar et al., 2016), and deposited in GenBank (Table 1).

| Phylogenetic analysis
Partial sequences from the ITS rDNA region, β-tubulin gene, and

| Mating type screening, distribution, and phylogeny
Sequences of the MAT1-1 and MAT1-2 genes were amplified from 40 isolates of P. brassicae (Table 1) to enable phylogenetic analyses of these mating type genes. Sequences were obtained from the isolates F I G U R E 1 Continued (Table 1) (Table 2), with each primer at a final concentration of 0.5 µM; 5 µl PCR grade water; and 2 µl unquantified DNA extract.
Amplicons were resolved on a 2% agarose gel and sent to MWG Eurofins for sequencing with primer Mt3.

| Rep-PCR DNA fingerprinting
Rep-PCR fingerprinting of a selection of nine isolates of the light leaf spot pathogen from NA and 10 isolates from the UK, continental Europe, and OC was done using the protocols and primers described by Versalovic et al. (1994). Each reaction was completed in a 20 µl volume containing 10 µl JumpStart REDTaq ReadyMix (Sigma Aldrich), 2-4 µl of each primer (see details below), 6 µl PCR-grade water, and 2 µl DNA (20 ng total per reaction). Three variants of rep-PCR fingerprinting were done: (a) BOX PCR for which each reaction included 4 µl of primer BOXAIR at a 1 µM final concentration; (b) ERIC PCR for which each reaction included 2 µl each of primers  (Tables 1 and 3). After 6 weeks, 1 ml SDW water was added to the surface of each stock plate and the colonies agitated using a sterilized bent glass rod. The conidial suspension was filtered through a double layer of sterilized cheesecloth and adjusted to 10 6 conidia/ ml. A 40 µl aliquot of conidial suspension from each of the two isolates used for each attempted sexual cross was placed onto a plate of 3% MEA and the two aliquots spread across the agar surface using a sterilized bent glass rod. Plates were sealed with Parafilm and incubated for a further 9 weeks in the dark at 18 °C, after which plates were examined microscopically at weekly intervals for the presence or absence of apothecial initials, mature apothecia, and asci with ascospores (the latter determined microscopically from thin apothecial sections examined at ≤100× magnification). Each sexual cross was attempted using three replicate plates of MEA.

| Morphological comparison
Light leaf spot isolates from NA and from the UK and continental Europe were compared morphologically in vitro and in planta (

| Fungicide sensitivity testing and molecular analyses
Ten isolates of the light leaf spot pathogen, including four reference UK and continental Europe isolates with different sensitivity profiles to carbendazim and prothioconazole, and six NA isolates that had not previously been tested for sensitivity to these fungicides (Tables 1 and 4

| Multilocus sequence analysis
Bayesian phylogenetic analyses of the ITS rDNA (Figure 1a), β-tubulin (Figure 1b), and TEF1-α sequences (Figure 1c), as well as TA B L E 3 Attempted sexual crosses of isolates of Pyrenopeziza brassicae (Lineage 1) from the United Kingdom and continental Europe (EU) with isolates (Lineage 2) from North American (NA) associated with light leaf spot, using isolates of opposite mating (MAT) type paired on 3% malt extract agar As 3 As 2 As 3 As 2 As 2 Ai − -Ai − E3A As 3 As 3 Ai 1 , As 2 As 2 Isolates were confirmed as either MAT1-1 or MAT1-2 types using the multiplex PCR assays of Foster et al. (2002).
b Three replicate pairings were established for each attempted sexual cross. The superscript number denotes the number of replicate plates on which apothecial initials (Ai), apothecia (Ap), or asci and ascospores (As) were observed. '−' indicates no sexual structures were observed. Results shown were after the isolates had been paired on 3% malt extract agar for 9 weeks. Refer to Table 1 for details of each isolate.
the concatenated sequences (Figure 1d) all revealed the UK, continental European, and OC isolates of P. brassicae formed a genetically distinct lineage, henceforth referred to as Lineage 1, from the NA isolates, henceforth referred to as Lineage 2. Maximumlikelihood analyses of the same sequences (ITS rDNA in Figure S1a, β-tubulin in Figure S1b, TEF1-α sequences in Figure S1c, and the concatenated sequences in Figure S1d) gave very similar results.
Both Bayesian and maximum-likelihood analyses supported two distinct lineages that were defined solely by geographic origin, with no evidence for additional grouping based on the Brassica or Raphanus species from which the isolates originated. These two lineages were more similarly related to each other than to sequences of any other related fungal genera examined for all DNA regions evaluated (Figure 1; Figure S1). The partial ITS rDNA sequence (GenBank accession no. MN028386) obtained from the type herbarium specimen of P. brassicae (IMI81823) showed this isolate grouped into Lineage 1.

| Mating type screening, distribution, and phylogeny
All of the light leaf spot isolates produced a single amplicon when screened with the multiplex mating type diagnostic PCR assay devel- isolates also clearly resolved the two lineages, with 90.36% similarity for MAT1-1 isolates and 93.24% for MAT1-2 isolates (data not shown).
Lineage 2 isolates (Figure 3). Evidence for high genotypic variability was also observed for the ERIC and GTG 5 data, with unambiguous bands scored as present/absent for each isolate (Figure 3 bands scored with arrows). Based on scoring of bands, 3 of 10 Lineage 1 isolates (30%), and 7 of 9 Lineage 2 isolates (78%) had unique genotypes.

| Pathogenicity of Lineage 2 isolates
The 17 isolates from Lineage 2 that were tested for pathogenicity on the turnip (cv. Hakurei) and mustard (cv. Caliente 199) plants all caused chlorotic, rapidly expanding, foliar lesions on both hosts (Figure 2c). Symptoms were not observed on SDWtreated control plants of either species. Data met assumptions for parametric analysis in pathogenicity tests 1 and 2, but data for pathogenicity test 3 had to be square root-transformed to meet assumptions of equal variance. Based on the ANOVAs, significant differences in disease severity were detected 21 dai between the turnip and mustard plants (p = .0004, p < .001, and p < .001 for tests 1, 2, and 3, respectively). The turnip plants developed more severe symptoms (100%, 99.7 ± 0.3%, and 84.1 ± 3.8% of the leaf area with symptoms in tests 1, 2, and 3, respectively) than the mustard plants (84.8 ± 3.7%, 77.0 ± 4.0%, and 21.5 ± 2.9% severity, respectively). In addition, turnip plants developed symptoms earlier than mustard plants, with pale brown streaks on the stems and veinal browning on the leaves that darkened over time. Veinal browning was followed by development of small (<5 mm diameter), chlorotic leaf spots, that became diffuse and expanded rapidly, coalescing and covering most of the leaf surface by 21 dai (Figure 2c). Symptoms were similar but developed more slowly on mustard leaves (3-5 days more slowly). Hyaline, smooth, cylindrical, mostly aseptate and eguttulate conidia were observed on short, non-branching conidiophores in pale acervuli (Figure 2d) on leaves with symptoms from plants inoculated with each of the Lineage 2 isolates.
The white, subcuticular conidiomata described by Rawlinson et al. (1978) and Fitt et al. (1998)  isolates (data not shown).  Table 1 for isolate details). Geographic origin of the isolates (EU/OC = continental Europe, UK, and Oceania; NA = North America) is noted at the base. Lanes 1-10 = Lineage 1 isolates, lanes 11-19 = Lineage 2 isolates, lane L = Hyperladder 1 (Bioline), and lane W = no-template water (control) sample. Differences between the two groups of isolates based on DNA fingerprint bands are indicated with white arrowheads Lineage 2 isolate, Cyc001, by 28 dai (4.50-5.75 necrotic leaves per plant, p > .05; Figure S2a). Only isolate 2016-5 caused fewer necrotic leaves (4.50 per plant) than that caused by Lineage 2 isolate Cyc001. The control plants averaged 2.50 ± 0.29 necrotic leaves per plant, which was less than that of any of the inoculated plants.

| Comparative symptomology caused by isolates of the two lineages
In the repeat test, the main effect of isolates was again significant (p < .0001). The Lineage 2 isolate Cyc001 caused the greatest number of necrotic leaves (4.00 ± 0.41 per plant), followed by the Lineage 1 isolate 2016-34 (2.75 ± 0.63 necrotic leaves per plant).
Three of the Lineage 1 isolates and the control plants all had <1 necrotic leaf per plant.
The main effect of isolates also significantly affected the number of chlorotic leaves per plant (p = .012 in Trial 1). Lineage 2 isolate Cyc001 caused the greatest number of leaves to turn chlorotic by 28 dai (1.8 ± 0.3 and 2.5 ± 0.7 leaves per plant in Trials 1 and 2, respectively; Figure S2b). However, this did not differ significantly from that caused by four Lineage 1 isolates in the first trial and two Lineage 1 isolates in the repeat trial (means separation based on nonparametric rank analyses). All other Lineage 1 isolates caused fewer chlorotic leaves to develop per plant than that caused by Lineage 2 isolate Cyc001 in both trials.

| Sexual compatibility testing
In vitro crosses on plates of 3% MEA between Lineage 1 isolates of P. brassicae of MAT1-1 and MAT1-2 types resulted in mature apothecia developing for 22 of the 25 crosses (88%; Table 3). Asci and ascospores were subsequently confirmed in 19 of these 25 crosses (76%) after 9 weeks. By contrast, attempts at inducing sexual reproduction under similar conditions were unsuccessful between Lineage 2 isolates of opposite MAT1-1 and MAT1-2 types, and between Lineage 1 and Lineage 2 isolates of opposite MAT types. Structures that appeared to be apothecial initials were observed in some crosses of Lineage 1 × Lineage 2 isolates but none of these developed into mature apothecia with ascospores (Table 3). Apothecial initials did not develop in any of the attempted MAT1-1 and MAT1-2 crosses among Lineage 2 isolates.

| Morphological analysis
Considerable colony variation was evident among the 10 Lineage 2 isolates of the light leaf spot pathogen, with diverse pigment colours (black, brown, grey, pink, red, and yellow; Figure 4a). For all Lineage 2 isolates examined (except Cyc023A), the observed phenotype was consistent among the three replicate cultures on MEA. Additional comparisons of the 10 Lineage 2 isolates with four representative Lineage 1 isolates revealed no obvious differences in colony phenotype that distinguished isolates from the two major geographic regions (Figure 4a,b).
Examination of conidia produced in vitro by colonies growing on 3% MEA for 6 weeks revealed it was not possible to distinguish be- In contrast, when conidia were washed directly from leaves of the turnip cv. Hakurei with symptoms, 28 dai of the plants with 10 Lineage 1 isolates and 10 Lineage 2 isolates, significant differences were observed in morphology of conidia produced by isolates from the two major geographic regions. A single septum was observed in some conidia collected from leaves inoculated with most (9 of 10) Lineage 2 isolates but only from leaves inoculated with 1 of the 10 Lineage 1 isolates. The number of conidia with a septum averaged 5.3 ± 1.1 for 60 conidia measured per isolate for the 10 Lineage 2 isolates compared to 0.1 ± 0.1 for 60 conidia per isolate for the Lineage 1 isolates (p < .0001). Conidial width did not differ significantly (p = .1300, R 2 = .39) among all 20 isolates, but was significantly greater for the 10 Lineage 1 isolates (average of 4.41 ± 0.02 µm) than for the 10 Lineage 2 isolates (3.14 ± 0.17 µm; p < .0001, R 2 = .60). Conidial length differed significantly among the 20 isolates (p = .0135, R 2 = .47), and between the 10 Lineage 1 isolates compared to the 10 Lineage 2 isolates (p < .0001, R 2 = .60, respectively). Conidial length averaged 10.08 ± 0.07 µm for the 10 Lineage 2 isolates versus 11.70 ± 0.06 µm for the 10 Lineage 1 isolates. In summary, the 10 Lineage 2 isolates produced slightly shorter and narrower conidia in planta than the 10 Lineage 1 isolates, and 90% of the Lineage 2 isolates produced a few septate conidia in planta, whereas only one of the 10 Lineage 1 isolates formed septate conidia in planta.

| Fungicide sensitivity testing and molecular analysis
In vitro testing showed the six Lineage 2 isolates to be very sensitive to carbendazim, as no fungal growth was observed on any of the agar plates amended with 0.39 μg/ml carbendazim (Table 4).
This contrasted with Lineage 1 isolates of P. brassicae known to be moderately and highly resistant to carbendazim, UK73 and 8CAB, respectively. Subsequent inspection of the β-tubulin amino acid sequences from 12 Lineage 2 isolates revealed none contained the E198A, E198G, F220Y, or L240F substitutions that have been associated with MBC resistance in some UK P. brassicae isolates (Carter et al., 2013). Additional sensitivity testing revealed the six Lineage 2 isolates to be sensitive to prothioconazole, as no fungal growth was observed on agar medium amended with 1.56 μg/ml, with the exception of one replicate plate of Lineage 2 isolate Cyc013A, on which a single colony <1 mm in diameter was observed. This contrasted with the growth observed for UK isolates UK73 and 8CAB, for which EC 50 values had previously been determined to be ≥1.23 μg/ ml (Carter et al., 2014).

| D ISCUSS I ON
In this study, isolates of the light leaf spot pathogen from three major geographic regions were resolved into two closely related but genetically distinct phylogenetic lineages. The first (Lineage   tan to light brown acervuli formed in the chlorotic and necrotic leaf tissue, in which conidia were observed when examined microscopically. In contrast, the 10 Lineage 1 isolates resulted in formation of white conidiomata on otherwise "healthy" green leaves, followed by rapid leaf necrosis (sometimes with leaf distortion and crinkling, but never with bright yellow chlorotic lesions). Overall, these results are consistent with the different symptoms observed on naturally infected plants under field conditions on the continents from which the original fungal isolates were obtained (Carmody, 2017;Karandeni Dewage et al., 2018).
Isolates of MAT1-1 and MAT1-2 types were found for both Lineage 1 and Lineage 2. In vitro crosses between Lineage 1 isolates of MAT1-1 and MAT1-2 types resulted in development of mature apothecia with asci and ascospores for a majority of the crosses (76%) within 9 weeks of pairing the isolates, which is consistent with previous studies (Ilott et al., 1984 (Milgroom, 1996). Secondly, the Lineage 2 isolates exhibited high genotypic (based on rep-PCR DNA fingerprinting) and phenotypic (based on colony morphology on 3% MEA) diversity, as is usually observed with sexually outcrossing populations (McDonald and Linde, 2002). The Lineage 2 isolates appeared more diverse (seven of nine isolates had a unique rep-PCR genotype) than the Lineage 1 isolates (3 of 10 isolates had a unique genotype). Further work is required to investigate possible cryptic sexuality in Lineage 2 isolates, including more extensive attempts at sexual crossing, e.g., in planta on senescing host debris (Gilles et al., 2001). The presence of a sexual cycle in Lineage 2 could affect pathogen dispersal and, potentially, increase the risk of breakdown in effectiveness of some disease management strategies, e.g., from development of fungicide resistance and/or the presence of virulence genes in the pathogen population that overcome host plant resistance (McDonald and Linde, 2002).
Morphologically, it was possible to distinguish between conidia of Lineage 1 and 2 isolates produced on infected B. rapa plants.
By contrast, no differences in conidial dimensions or colony colour were observed between the Lineage 1 and 2 isolates when grown on 3% MEA. Isolates from both lineages formed a range of black, brown, grey, pink, or yellow pigmentation on this medium. The difference in spore dimensions observed for spores of Lineages 1 and 2 generated in vitro versus in vivo could reflect the well-documented potential impact of substrate (3% MEA vs. live plants in this case) on spore production by many fungi. However, the measurement of spores produced in vitro was done at Rothamsted Research whereas the measurement of spores produced in vivo was done at WSU, which confounded any potential effects of the location and method with differences in spore dimensions among isolates. Given these difficulties with morphological discrimination in vitro between isolates of the two lineages, specific PCR assays have since been designed by King and West at Rothamsted to enable rapid lineage discrimination (data not shown). Such PCR assays could be used to differentiate isolates of the two lineages, including isolates of the two lineages present in infected leaves and seed.
The first report of light leaf spot in NA was in Oregon in 2014, with subsequent widespread distribution of the disease discovered across western Oregon and, more recently, in three counties in Washington State, which suggests fairly rapid spread of the causal agent within the Pacific Northwest USA. Indeed, based on the Lineage 2 isolates evaluated in this study, the pathogen was confirmed as far north as Whatcom Co., WA and as far south as Douglas Co., OR. The geographic origin of Lineage 2 isolates in the USA remains unclear. However, based on this study, Lineage 2 isolates appear not to have originated from the UK, continental Europe, or OC, as isolates from those regions were in the genetically distinct Lineage 1. One possible source of Lineage 2 isolates is Asia. Light leaf spot outbreaks have been reported in Japan and Thailand (Rawlinson et al., 1978;CABI, 2015). Future work to characterize Asian isolates should provide insight on a more global scale of the potential origin of the NA isolates.
Currently, the two lineages appear to be restricted geographically to either the UK, continental Europe, and OC (Lineage 1) or to NA (Lineage 2). Therefore, appropriate precautions are needed to prevent movement of isolates from the different lineages between regions and to other parts of the world. This includes transfer of potentially infected plants or seed (Carmody and du Toit, 2017) on which the pathogen might be present, either with or without symptoms. More comprehensive testing of the responses of B. napus, B. oleracea, B. rapa, and other Brassicaceae germplasm to isolates from the two lineages is needed to assess potential differences in susceptibility of plant germplasm (Boys et al., 2012). Although this study indicated that isolates from Lineages 1 and 2 are sexually incompatible, there remains a risk of hybridization or somatic recombination between isolates of the two groups. Given the recent rapid spread of Lineage 2 across western Oregon and western Washington, there is also a risk of spread into Canada, the world's third largest producer of canola (B. napus), and other regions of the USA, as well as Mexico.
Management of light leaf spot in the UK and continental Europe is based primarily on timely applications of efficacious fungicides.
Prior to this study, data were not available on the sensitivity of Lineage 2 isolates of the light leaf spot pathogen to fungicides used to control this disease in the UK and continental Europe. Phenotypic screening of six Lineage 2 isolates revealed all to be sensitive to both carbendazim and prothioconazole. Examination of the β-tubulin amino acid sequences of Lineage 2 isolates revealed 100% identity to that of a UK isolate previously classified as sensitive to MBC fungicides (KC342227; Carter et al., 2013), with no evidence for the key substitutions (e.g., E198A or L240F) that have been correlated with MBC resistance in Lineage 1 isolates (Carter et al., 2013). Although more isolates should be tested, it appears likely that Lineage 2 isolates might be controlled effectively with applications of MBC and DMI fungicides, as demonstrated recently with MBC and DMI fungicide seed treatments evaluated with a mustard seed lot infected with a Lineage 2 isolate (Carmody and du Toit, 2017). However, given the emergence of resistance to both fungicide groups in some Lineage 1 isolates (Carter et al., 2013(Carter et al., , 2014, implementation of fungicide resistance management strategies by NA brassica growers will be important to extend the effective life of these fungicides against the pathogen (e.g., using mixtures or rotations of fungicides with different modes of action).
In conclusion, based on the CSC that combines morphological, ecological, biological, and genetic (phylogenetic) data (Crous et al., 2015), convincing evidence was generated in this study for two genetically distinct evolutionary lineages of P. brassicae, with Lineage 1 comprising isolates from the UK, continental Europe, and OC, including the type specimen, IMI81823 (Rawlinson et al., 1978); and Lineage 2 (Claassen, 2016;Carmody, 2017), we propose the common name "chlorotic leaf spot" be used to describe the disease caused by Lineage 2 isolates in order to differentiate this disease from classic light leaf spot symptoms caused by isolates of Lineage 1 of P. brassicae.

ACK N OWLED G EM ENTS
The WSU authors acknowledge funding from the Clif Bar Family