A New Subclade of Leptosphaeria biglobosa Identified from Brassica rapa

Blackleg (Phoma stem canker) of crucifers is a globally important disease caused by the ascomycete species complex comprising of Leptosphaeria maculans and Leptosphaeria biglobosa. Six blackleg isolates recovered from Brassica rapa cv. Mizspoona in the Willamette Valley of Oregon were characterized as L. biglobosa based on standard pathogenicity tests and molecular phylogenetic analysis. These isolates were compared to 88 characterized L. biglobosa isolates from western Canada, 22 isolates from Australia, and 6 L. maculans isolates from Idaho, USA using maximum parsimony and distance analysis of phylogenetic trees generated from the ITS rDNA (internal transcribed spacer rDNA) sequence, and the actin and β-tubulin gene sequences. The L. biglobosa isolates derived from B. rapa collected in Oregon formed a separate subclade based on concatenated gene sequences or a single gene sequence, regardless of the analyses. Pathogenicity tests showed that these isolates failed to infect either resistant or susceptible B. napus cultivars, but caused severe symptoms on three B. rapa cultivars (Accession number: UM1113, UM1112, and UM1161), a B. oleracea var. capitata (cabbage) cultivar (Copenhagen Market), and two B. juncea cultivars (CBM, a common brown Mustard, and Forge). These findings demonstrated that the L. biglobosa isolates derived from a B. rapa crop in Oregon were genetically distinct from existing species of L. biglobosa, and constitute a new subclade, herein proposed as L. biglobosa ‘americensis’.


Introduction
Leptosphaeria maculans and L. biglobosa are two closely related fungal species that together form a species complex that causes blackleg or Phoma stem canker of crucifers, including Brassica napus, B. juncea, B. oleracea, and B. rapa [1,2]. L. maculans and L. biglobosa were previously described as virulent and weakly-virulent, respectively, with genetic differences and distinct phenotypic (disease) expression on oilseed rape (B. napus) leaves or stems [2]. During infection of B. napus, L. maculans produces large grey/green leaf lesions, and then grows down the vascular tissue to the stem, where the fungus causes necrotic stem cankers [2]. In contrast, L. biglobosa causes small, dark leaf lesions and typically is restricted to infection of the upper stem [3]. Consequently, L. maculans has been considered much more damaging than L. biglobosa.
β-tubulin sequences in western Australia [7]. The phylogeny of L. biglobosa isolates on the American continent has also been analyzed by using ITS rDNA, actin, and β-tubulin sequences [9]. In addition, other important host-pathogen systems, such as fusarium and wheat, where F. graminearum causes the fusarium head blight (FHB) disease, conserved gene sequences, i.e., ITS, histone H3, elongation factor 1-α, and β-tubulin, have been widely used to identify species of Fusarium [13,14]. More recently, 16 monophyletic species have been identified within the Fusarium graminearum species complex using a high-throughput multilocus assay of portions of housekeeping genes [15][16][17]. L. maculans has spread recently into many areas where only L. biglobosa was present, suggesting that L. biglobosa may have evolved earlier than L. maculans from a common ancestor [18].
In 2014, a vegetable seed grower in the Willamette Valley in Oregon, USA submitted samples from a certified organic seed crop of the B. rapa vegetable cv. Mizspoona. The plants had typical foliar and stem symptoms of blackleg. Six isolates of Leptosphaeria were recovered from the Brassica rapa seed crop, Mizspoona, and were characterized using standard pathogenicity tests and sequence analysis of the ITS rDNA and actin and β-tubulin genes to determine the species of these isolates.

Pathogenicity Tests
Small, dark brown, necrotic lesions without pycnidia were observed on cotyledons of the B. napus differential cultivars or lines, including the universal susceptible cultivar, Westar, when inoculated with each of the six isolates (Phl002 (Phoma lingam002) to Phl007) of Leptosphaeria derived from a B. rapa vegetable seed crop in the Willamette Valley of Oregon ( Figure 1). Conversely, all six isolates caused susceptible reactions on the B. juncea cultivars, Forge and CBM; and on the three B. rapa lines, UM1161, UM1113, and UM1112, by 11 dpi (Figure 1 and Table 1). Six additional isolates (Phl010 to Phl015) derived from B. napus infected plants collected from Idaho (Table S1) were all virulent on Westar, as expected, and showed avirulent or virulent reactions on a range of cultivars of B. napus differing in resistance genotypes. There were no disease symptoms or a hypersensitive reaction on the Brassica germplasm inoculated with water. reaction on the Brassica germplasm inoculated with water.
Since all the B. rapa-derived isolates were avirulent on all the B. napus cultivars and lines tested, but virulent on several B. juncea (CBM and Forge) and B. rapa (UM1113, UM1112, and UM1161) cultivars and lines, it was deemed necessary to assess whether these isolates were L. maculans or L. biglobosa. Therefore, the B. rapa-derived isolates from Oregon were subjected to a PCR (Polymerase chain reaction) assay of the ITS rDNA, and the amplified sequences used for species identification.  Table S2 for details of cultivars/lines), with isolate Phl004 of Leptosphaeria biglobosa obtained from the Willamette Valley of Oregon (see Table S1 for details of isolates). Seedlings of the B. napus cv., Westar, inoculated with water were used for the negative treatment. Small, dark brown, necrotic lesions without pycnidia were observed on cotyledons of the B. napus differential cultivars/lines.  Table S2 for details of cultivars/lines), with isolate Phl004 of Leptosphaeria biglobosa obtained from the Willamette Valley of Oregon (see Table S1 for details of isolates). Seedlings of the B. napus cv., Westar, inoculated with water were used for the negative treatment. Small, dark brown, necrotic lesions without pycnidia were observed on cotyledons of the B. napus differential cultivars/lines. Since all the B. rapa-derived isolates were avirulent on all the B. napus cultivars and lines tested, but virulent on several B. juncea (CBM and Forge) and B. rapa (UM1113, UM1112, and UM1161) cultivars and lines, it was deemed necessary to assess whether these isolates were L. maculans or L. biglobosa. Therefore, the B. rapa-derived isolates from Oregon were subjected to a PCR (Polymerase chain reaction) assay of the ITS rDNA, and the amplified sequences used for species identification.

Characterization of B. rapa-Derived Isolates from the Willamette Valley of Oregon
PCR amplification of the six Oregon isolates of Leptosphaeria obtained from a B. rapa seed crop, using the PN3 and PN10 primers, illustrated that each of the isolates, Phl002 to Phl007, produced amplified DNA bands matching the size of the L. biglobosa DNA band ( Figure 2). Evaluation of the mating type alleles indicated that none of the six isolates could be positively amplified and characterized to either of the two known mating types of L. maculans, i.e., the isolates could not be categorized as L. maculans, which supported their possible assignment to L. biglobosa based on the ITS rDNA size.
The ITS rDNA sequences and the actin and β-tubulin gene sequences were obtained from B. rapa-derived isolates, Phl002 to Phl007, and compared with sequences deposited in the NCBI (National Center for Biotechnology Information) database. The B. rapa-derived isolates (Phl002 to Phl007) had 99.00% or 99.01% ITS rDNA sequence similarities (deposited in Genbank: MG321243) to the ITS rDNA sequence from L. biglobosa 'brassicae' DQ133890, compared with only 95.59% to 96.01% for that of L. maculans JX648199. Similarly, the actin gene sequences (deposited in Genbank: MG282088) of the B. rapa-derived isolates showed 99.00% to 99.02% and 93.56% to 93.58% similarities to actin sequences of the L. biglobosa 'brassicae' group strain (AY748949) and L. maculans 'brassicae' group strain (AY748971), respectively. In addition, the B. rapa-derived isolates had 99.00% to 99.02% β-tubulin gene (deposited in Genbank: MG282089) similarities to the L. biglobosa 'canadensis' group strain, AY749006 compared with only 92.45% or 92.46% for that of L. maculans 'brassicae' group strain AY749018, in the NCBI database (Table S3). In summary, the six B. rapa-derived isolates from Oregon were categorized as L. biglobosa through sequencing of the ITS rDNA and the actin and β-tubulin genes, and based on pathogenicity tests.

Identification of a New Subclade of L. biglobosa by Phylogenetic Analysis
All 88 isolates from western Canada were characterized as L. biglobosa 'canadensis', while 12 isolates from Australia had been characterized as L. biglobosa 'canadensis' and six as L. biglobosa 'occiaustralensis'. Since 88 isolates from western Canada and 12 isolates from Australia were characterized as the same subspecies, L. biglobosa 'canadensis', 9 and 6 isolates were randomly selected from Canada and Australia, respectively, for phylogenetic analysis. Deposited sequences of
Alignment of the ITS rDNA sequences resulted in a dataset of 573 characters of which 43 (7.5%) were considered to be informative characters. The actin sequence alignment resulted in a dataset of 477 characters of which 43 (9.0%) were parsimony informative. The β-tubulin sequence alignment resulted in 501 characters of which 26 (2.6%) were parsimony informative. Finally, a set of data with concatenate sequences of these three conserved genes resulted in a dataset of 1551 characters of which 112 (7.2%) were parsimony informative for phylogenetic tree construction (CI (Consistency index) = 0.84, RI (Retention index) = 0.91). A total of six subclades contained all the L. biglobosa isolates by concatenated ITS rDNA, actin, and β-tubulin sequences, which were differentiated clearly from the two L. maculans subclades ( Figure 3). As expected, the isolates, ICBN89 and IBCN91, grouped in the L. biglobosa 'brassicae' and L. biglobosa 'australensis' clades, respectively. The L. biglobosa 'thlaspii' isolates were differentiated from a strongly supported clade that included all other subclades, and was most closely aligned with the L. maculans subclades. L. biglobosa 'australensis' and L. biglobosa 'erysimii' isolates formed one branch based on the ITS rDNA sequence ( Figure S4). The close phylogenetic relationship observed between L. biglobosa 'canadensis' and L. bilobosa 'occiaustralensis' was similar to that published by Vincenot et al. (2008) [7]. Interestingly, the six B. rapa-derived L. biglobosa isolates from Oregon formed a distinct subclade most closely related to L. biglobosa 'brassicae' isolates, and were clearly separated from the outgroup of L. maculans isolates (Figure 3). The distance and parsimony analyses on a concatenated set of three gene sequences resulted in threes of similar topology to those analyses on a single sequence of ITS rDNA or the actin gene ( Figures S4 and S5). Phylogenetic trees constructed using the maximum likelihood, minimum evolution, and unweighted pair group method with arithmetic mean (UPGMA) were very similar to those obtained using the neighbour joining and parsimony analyses, and showed that the B. rapa-derived isolates, Phl002 to Phl007, clustered into a unique subclade distinct from other L. biglobosa isolates and from L. maculans isolates (data not shown).
The six B. rapa-derived isolates formed a distinct subclade using the β-tubulin gene sequences. While the ITS rDNA and actin sequences placed the B. rapa-derived isolates closest to the L. biglobosa 'brassicae' subclade, the β-tubulin gene sequences of the B. rapa-derived isolates showed the greatest identity with L. biglobosa 'canadensis' isolates ( Figure S6). This probably reflected the greater level of sequence polymorphism detected in the β-tubulin sequences between B. rapa-derived isolates and L. biglobosa 'brassicae' isolates compared to those of the L. biglobosa 'canadensis' isolates ( Figure S3). Since the B. rapa-derived isolates fell into a distinct clade, based on the conserved gene regions and concatenated sequences analyzed, a new subclade, L. biglobosa 'americensis', is proposed. bilobosa 'occiaustralensis' was similar to that published by Vincenot et al. (2008) [7]. Interestingly, the six B. rapa-derived L. biglobosa isolates from Oregon formed a distinct subclade most closely related to L. biglobosa 'brassicae' isolates, and were clearly separated from the outgroup of L. maculans isolates (Figure 3). The distance and parsimony analyses on a concatenated set of three gene sequences resulted in threes of similar topology to those analyses on a single sequence of ITS rDNA or the actin gene ( Figures S4 and S5). Phylogenetic trees constructed using the maximum likelihood, minimum evolution, and unweighted pair group method with arithmetic mean (UPGMA) were very similar to those obtained using the neighbour joining and parsimony analyses, and showed that the B. rapa-derived isolates, Phl002 to Phl007, clustered into a unique subclade distinct from other L. biglobosa isolates and from L. maculans isolates (data not shown).

Differences in the Pathogenicity of Isolates of L. biglobosa 'americensis' and Other L. biglobosa Subspecies
To determine whether isolates of this proposed new subspecies, L. biglobosa 'americensis', have similar disease profiles to isolates of the other L. biglobosa species, pathogenicity tests were carried out with representative isolates from L. biglobosa 'canadensis, L. biglobosa 'brassicae', and L. biglobosa 'occiaustralensis'. Compared to the severity of cotyledon symptoms caused by isolates of the other subclades of L. biglobosa, the L. biglobosa 'americensis' isolates caused significantly larger lesions on cotyledons of the cultivar, Westar (Figures 4 and 5a), although these lesions were still significantly smaller than the lesions caused by the L. maculans control isolate. Alternatively, on the cultivar, Jet Neuf, no significant differences in lesion size were detected among the subclades of the L. biglobosa isolates (Figures 4 and 5a). In the pathogenicity test of the two B. juncea cultivars (CBM and Forge), lesion sizes on CBM inoculated with L. biglobosa 'americensis' isolates were all significantly larger than those on plants inoculated with the other L. biglobosa isolates, which were all rated in the susceptible or intermediate categories (Figure 5b). Most of the L. biglobosa 'americensis' isolates caused significantly larger lesions on Forge seedlings compared to lesions caused by isolates of L. biglobosa 'canadensis' and 'occiaustralensis', but not of 'brassicae' (Figure 5b and Figure S7). The lesion sizes were significantly larger on B. rapa (cultivar 'Mizspoona', where L. biglobosa 'americensis' isolates identified from) and B. oleracea plants inoculated with the L. biglobosa 'americensis' isolates than on plants inoculated with isolates of the other subclades of L. biglobosa (Figure 5c). These results suggest that the L. biglobosa 'americensis' isolates have a significantly different disease profile to the other L. biglobosa subspecies tested.
smaller than the lesions caused by the L. maculans control isolate. Alternatively, on the cultivar, Jet Neuf, no significant differences in lesion size were detected among the subclades of the L. biglobosa isolates (Figure 4 and 5a). In the pathogenicity test of the two B. juncea cultivars (CBM and Forge), lesion sizes on CBM inoculated with L. biglobosa 'americensis' isolates were all significantly larger than those on plants inoculated with the other L. biglobosa isolates, which were all rated in the susceptible or intermediate categories (Figure 5b). Most of the L. biglobosa 'americensis' isolates caused significantly larger lesions on Forge seedlings compared to lesions caused by isolates of L. biglobosa 'canadensis' and 'occiaustralensis', but not of 'brassicae' (Figures 5b and S7). The lesion sizes were significantly larger on B. rapa (cultivar 'Mizspoona', where L. biglobosa 'americensis' isolates identified from) and B. oleracea plants inoculated with the L. biglobosa 'americensis' isolates than on plants inoculated with isolates of the other subclades of L. biglobosa (Figure 5c). These results suggest that the L. biglobosa 'americensis' isolates have a significantly different disease profile to the other L. biglobosa subspecies tested.

Discussion
This study describes isolates of a proposed new subclade of L. biglobosa obtained from a certified organic B. rapa vegetable seed crop of the cv., Mizspoona, grown in the Willamette Valley of Oregon, USA in 2013 to 2014 that had developed classic symptoms of blackleg. Pathogenicity and molecular

Discussion
This study describes isolates of a proposed new subclade of L. biglobosa obtained from a certified organic B. rapa vegetable seed crop of the cv., Mizspoona, grown in the Willamette Valley of Oregon, USA in 2013 to 2014 that had developed classic symptoms of blackleg. Pathogenicity and molecular characterization clearly identified these isolates as belonging to a distinct subclade compared to isolates of the six known subclades of this species. We propose naming the new subclade L. biglobosa 'americensis'.
Relationships between members of the two Leptosphaeria species, L. maculans and L. biglobosa, based on three conserved DNA regions yielded fairly consistent phylogenetic trees and were extremely comparable to previously published studies [6,7,11]. For example, previous studies found that isolates of the subclade, L. biglobosa 'occiaustralensis', were most closely related to isolates of L. biglobosa 'canadensis', while isolates of L. biglobosa 'australensis' and 'erysimii' were closely related to each other [6,7]. Indeed, the topologies of these subclades are consistent with the data generated in this study. In addition, L. biglobosa 'thlaspii' isolates were most closely related to L. maculans isolates in the phylogenetic trees constructed previously and in this study. The distance and parsimony analyses on concatenated sequences or a single gene sequence with high bootstrap value supported that subclades including the six L. biglobosa isolates is indeed a new subspecies. L. biglobosa 'canadensis' was close to L. biglobosa 'occiaustralensis' through concatenated/single gene sequence analysis. The six B. rapa derived isolates formed an independent group and was mostly close to L. biglobosa 'brassicae' when using concatenated sequences of three conserved genes and consistent with those trees constructed by single sequence of ITS rDNA or actin. The inconsistent tree obtained by phylogeny analysis using the β-tubulin sequence, which showed the new six isolates as closer to L. biglobosa 'canadensis', reinforced that these six isolates cannot be categorized into either 'brassicae' or 'canadensis' and formed a distinct group. To eliminate the bias of the phylogenetic tree constructed by the single gene sequence, concatenated sequences of three conserved genes were applied for parsimony analyses, the result being that these six isolates formed a separate subclade that was closer to L. biglobosa 'brassicae'. More genomic or transcription data will provide further insights on the characterization of the new subclade in the future for investigation of the difference between this new subclade and other subclades of L. biglobosa. Trees obtained through parsimony analyses were similar to those obtained by the NJ method, and sufficient bootstraps supported the hypothesis that the B. rapa-derived isolates formed a monophyletic clade. Phylogenic analysis using NJ, parsimony, maximum likelihood, and minimum evolution analyses resulted in trees of similar topologies, with greater bootstrap values for the NJ trees compared with the other trees. It remains unclear why the different subspecies of L. biglobosa are so distinct from one another. With the exception of L. biglobosa 'brassicae' and L. biglobosa 'canadensis', generally, few isolates have been cultured from each of the subspecies, making population surveys difficult. Therefore, limited information is known regarding gene flow across geographical regions.
Isolates of L. biglobosa, which can coexist on host plants with isolates of L. maculans, typically are associated with upper stem lesions on infected Brassica plants, and are considered weakly virulent or even avirulent. L. biglobosa isolates have only caused significant losses in areas with high summer temperatures, e.g., in Poland [19]. Vincenot et al. (2008) demonstrated that L. biglobosa 'occiaustralensis' isolates were more virulent than L. biglobosa 'canadensis' or 'australensis' isolates, and L. biglobosa 'brassicae' isolates produced large leaf lesions on B. napus plants [7]. The pathogenicity of L. biglobosa 'americensis' isolates on Brassica spp. in this study indicated that these isolates were intermediate in virulence on B. napus and caused more severe disease on B. juncea and B. rapa. Van de Wouw et al. (2008) reported that L. biglobosa 'canadensis' isolates induced similar size lesions on B. juncea cotyledons as L. maculans isolates [10]. In contrast, L. biglobosa 'americensis' isolates in this study caused significantly larger lesions than L. biglobosa 'canadensis' isolates on both resistant and susceptible lines of B. napus, and large lesions on B. juncea and B. rapa lines. These results were confirmed by the accumulation of lignified material observed in stained, infected cotyledons, with more diffuse and less lignified material in the cotyledons caused by L. biglobosa 'americensis' and L. maculans isolates than isolates of other subclades of L. biglobosa. It remains unclear why the different subspecies of L. biglobosa show different reactions on different Brassica species, however, it raises the question of whether effector genes are present in these L. biglobosa subspecies and whether they may play a role in the pathogenicity of the various Brassica species. The genome sequencing of a subset of L. biglobosa species has shown different suites of effector-like proteins, thus supporting this hypothesis, however, the functions of these effector-like proteins remain unknown in these subspecies [4].
In contrast to isolates of other L. biglobosa subclades, L. biglobosa 'americensis' isolates characterized in this study represent the first L. biglobosa subclade derived from B. rapa that caused severe symptoms on B. rapa, B. oleracea, and B. juncea cultivars. The evolution of Leptosphaeria isolates appears to be influenced by internal and external forces, including mutation, reproduction, gene flow, genetic drift, natural selection (e.g., temperature and geography) [20], and host specificity [21]. Unlike the well understood gene-for-gene relationship of R genes and corresponding avirulence genes in the B. napus-L. maculans pathosystem [22], it remains unclear whether a gene-for-gene interaction exists among L. biglobosa isolates and Brassica species. The results of this study validated and enriched the known diversity of L. biglobosa isolates, and provide a resource not only to help determine whether the evolutionary process of L. biglobosa isolates is influenced by geographic location and host selection pressures, but also to underline the possible existence of plant defense responses among Brassica species and L. biglobosa isolates. As reflected by the more severe disease symptoms associated with the L. biglobosa 'americensis' isolates on cultivars of three Brassica spp. evaluated in this study compared to isolates of other subclades, further research is needed to investigate potential differences in resistance to L. biglobosa 'americensis' isolates vs. isolates of other subclades, and to find sources of resistance to isolates of the different subclades of L. biglobosa.

Isolate Collection
A total of 122 Leptosphaeria isolates (12 from the USA, 88 from western Canada, and 22 from Australia) were included in this study (Table S1). For the USA isolates, six were cultured from stems and roots of canola crop residues from northern Idaho (Phl010 to Phl015), and six were cultured from B. rapa stems of six plants of the cv., Mizspoona, sampled from a certified organic seed crop grown in the Willamette Valley, Oregon (Phl002 to Phl007) (Table S1). Of the 88 isolates from western Canada, 67 were sampled from infected stems or stubble sections of canola crops (B. napus) grown in 2010, 2011, 2012, 2013, or 2014. The remaining 21 isolates were sampled from the dockage of canola (B. napus) seeds. All Canadian and USA isolates were cultured by plating small pieces of surface-sterilized stubble, stem, or seed onto potato dextrose agar (PDA) medium amended with chloramphenicol (100 mg/L medium) (for USA isolates) or V8 agar amended with 0.35% streptomycin sulfate (for Canadian isolates) ( Table S1). The individual isolates were then obtained by streaking pink cirrhi that developed from pycnidia on the infected tissues, followed by hyphal tip transfers from single colonies that germinated from the pycnidiospores separated on the agar surface. The isolates were then subjected to a second round of single-spore isolation. The 22 Australian isolates were collected from infected canola stems in field trials at the end of the 2006 and 2014 growing season. The stems were left to mature over the summer, and once pseduothecia (sexual fruiting bodies) formed the following year, ascospores were obtained as described previously [23]. In summary, to obtain single-spore cultures, individual ascospores were collected using a dissecting microscope with a sterilized needle. L. maculans 'brassicae' isolate (06LM) collected from Canada was used as a control isolate throughout this study, and is hereafter referred to as L. maculans. All isolates were maintained on V8 agar medium.

Initial Pathogenicity Testing of the Oregon B. rapa-Derived Leptosphaeria Isolates on Brassica Species
Pycnidiospore inoculum of each isolate of Leptosphaeria was harvested by flooding 8-to 11-day-old cultures of the appropriate single-spore isolate using distilled water (2 mL/plate). The concentration of spores was adjusted to 2 × 10 7 spores/mL for the cotyledon inoculation test. The seed of 10 B. napus differential cultivars or lines carrying different L. maculans related resistance genes, eight B. juncea cultivars or lines, and eight B. rapa cultivars or lines (Table S2) were each sown into 96-cell flats filled with Pro-Mix BX (Premier Tech, Rivière-du-Loup, QC, Canada) and placed in a growth chamber at 16 • C (night) and 21 • C (day) with a 16-h photoperiod/day. The cotyledons of each 7-day-old seedling were each punctured and inoculated with a 10 µL droplet of inoculum at each of two wound sites/cotyledon (four wound sites/seedling) as previously described [24]. As a negative control treatment, the cotyledons of seedlings of each cultivar of each Brassica species were punctured and inoculated with a droplet of water instead of a spore suspension. A rating scale of 0 to 9, based on lesion size, chlorosis or necrosis, and the presence of pycnidia, was used to evaluate the interaction phenotype 11 to 14 dpi as described by Zhang et al. (2016) [24]. For each isolate screened, an average score was calculated from the 24 inoculation sites (four wound sites/plant for each of six plants): A mean of 6.1 to 9.0 was considered a susceptible (S) reaction, 4.6 to 6.0 an intermediate (I) reaction, and ≤4.5 a resistant (R) reaction [24,25].

PCR Identification of Isolates
A mixture of hyphae, pycnidia, and pycnidiospores was collected from a V8 agar plate of each isolate of Leptosphaeria, and stored in 1.5 mL micro-centrifuge tubes at −20 • C before DNA extraction. DNA extraction was performed using the cetyl trimethylammonium bromide (CTAB) method described previously by Calderon et al. (2002) with some minor modifications [26].
To confirm the species of the B. rapa-derived isolates from Oregon and six isolates from B. napus infected tissues in Idaho, the ITS rDNA was amplified using ITS F (PN3) and ITS R (PN10) primers designed from the 18S rDNA and 28S rDNA of Saccharomyces cerivisiae, respectively (Table 2) [6].
This primer set generates a 555 to 560 bp fragment for L. maculans and a 580 to 588 bp fragment for L. biglobosa [6]. A multiplex PCR assay developed by Cozijnsen and Howlett (2003) was employed to characterize the mating type of those isolates [27]. Therefore, we were able to further clarify whether these isolates can be categorized as L. maculans or not. A 686 bp band was amplified from all MAT1-1 isolates, while a 443 bp band was amplified from all MAT1-2 isolates (Table 2) [27]. The annealing temperatures for primers used in this study are presented in Table 2.

Conserved Gene Sequencing and Sequence Alignment
For sequencing the ITS rDNA, actin gene, and β-tubulin gene of each of the 88 isolates from western Canada and 12 isolates from USA, PCR amplification was performed as described above. The ITS rDNA, actin, and β-tubulin PCR products were then sent to AGTC (Advanced Genetic Technologies Center, University of Kentucky, Lexington, KY) for sequencing in both directions. Of the total 122 isolates, determined and available ITS rDNA, actin gene, and β-tubulin gene sequences of 22 isolates from Australia were included as different subspecies of L. biglobosa. In addition, 28 previously published sequences of these conserved genes available in GenBank were also employed as references for the analysis (Table S1). For the conserved gene similarity analysis, an L. biglobosa (Table S3) DNA sequence similarity analysis was performed using the BLAST search tool [28].

Phylogenetic Analysis
Initially, sequence alignments were carried out on the ITS rDNA, actin, and β-tubulin sequences from the six B. rapa-derived isolates compared to ITS rDNA, actin, and β-tubulin sequences from L. biglobosa isolates recovered from different Brassica spp. (Table S1) using Clustal Omega [29] with default parameters (https://www.ebi.ac.uk/Tools/msa/clustalo/) and Genedoc [30]. Since the Canadian isolates were characterized to be L. biglobosa 'canadensis' and to eliminate the genetic bias of using gene sequences of the same isolates in the phylogenetic analysis, the sequences from the six Oregon isolates were compared to randomly selected and representative isolates of the L. biglobosa-L. maculans complex using ITS rDNA (31 determined sequences and 8 deposited sequences from Genbank), actin (24 determined sequences and 7 deposited sequences from Genebank), and β-tubulin (20 determined sequences and 6 deposited sequences from Genbank) (Table S1). Therefore, we were able to determine whether the six B. rapa-derived isolates formed a separate and consistent species clade compared to previously characterized L. biglobosa subspecies. Analyses of sequences and phylogenetic relationships were calculated using MEGA6 [31]. The phylogenetic tree was reconstructed using the neighbor-joining method by the Kimura two-parameter model, with bootstrap calculation using 1000 replications. Phylogenetic analyses were also conducted on the conserved gene sequences using maximum parsimony (MP) analysis with the subtree-pruning-regrafting algorithm and maximum likelihood (ML) with the Kimura two-parameter distance method using MEGA6 [31].