Leptosphaeria maculans is the most serious disease of canola (Brassica napus) globally (Fitt et al. 2006; Van de Wouw and Howlett 2020). The canola industry needs new resources or ways to combat blackleg disease. Unfortunately, our understanding of the genes underlying the biology of L. maculans is limited, as highlighted by how few pathogenicity genes have been identified in this fungus (these genes are listed in Table S1). Hence, building upon over two decades of using insertional mutagenesis for gene discovery in L. maculans, new pathogenicity factors were sought using this approach.

The L. maculans wild type strain D5 (also known as M1 or IBCN18) was the basis for transformation, and it was cultured on 10% cleared V8 juice and 2% agar, pH 6.

The CSBX series of plasmids previously generated (Urquhart et al. 2018) was primarily used for T-DNA insertional mutagenesis of L. maculans. These plasmids feature unique 20 bp barcodes at the right border side of the T-DNA, designed towards the high-throughput identification of insertion sites in pooled DNA sample sequencing. Additionally, plasmid CSBXG418 was used in the production of strain M20W. This plasmid is similar to the CSBX series but uses G418 (geneticin) rather than hygromycin for selection of the transformed strains. It was generated, as illustrated in Fig. S1, by using the oligonucleotide AU257 as a template for PCR with primers AU258 and AU259 (sequences are in Table S2) that was cloned with the NEBuilder® Gibson assembly kit (New England Biolabs) into the PmeI restriction enzyme site of plasmid pMAI2 (Idnurm et al. 2017).

The CSBX series, CSBXG418 and plasmids used for complementation were electroporated into Agrobacterium tumefaciens strain EHA105. The T-DNA contained within these plasmids was integrated into the genome of L. maculans strains by A. tumefaciens-mediated transformation (Gardiner and Howlett 2004).

Approximately 800 T-DNA insertion transformants were generated and screened for their pathogenicity on canola using a standard cotyledon assay. Brassica napus cv. Westar plants were grown in a growth cabinet with a 12 hour day/night cycle. Conidia of the L. maculans strains were scraped off an agar plate into sterile water, filtered through miracloth and washed once with water before being resuspended at 106 spores per ml. 10 μl of spore solution was inoculated onto cotyledons, which had been wounded using a needle as two inoculation points per cotyledon (i.e. four per plant). Disease severity was assessed at 14 days post inoculation. Each transformant was screened once, and then any with reduced disease severity was tested independently for pathogenicity a minimum of three times, with in the final experiment at least 17 inoculation points examined.

Seven strains were identified in which pathogenicity was highly reduced. This represents <1% percent of transformants, which is lower than the 3.8% frequency previously reported from the largest scale study of T-DNA insertions in L. maculans (Blaise et al. 2007).

Genomic DNA was prepared from L. maculans freeze-dried mycelia using a CTAB buffer extraction protocol (Pitkin et al. 1996). The locations of the T-DNAs in the genomes of the seven strains were identified from Illumina sequencing reads of individual strains (Chambers et al. 2014), with the exception of M20W which was sequenced as part of a pool with other strains. That is, whole genome sequences were obtained from an Illumina HiSeq 2500 run with 125 nucleotide length paired reads. The reads were mapped onto the sequence of the CSBX T-DNA, and in cases onto the whole plasmid where the entire plasmid was integrated into the genome, in Geneious Prime. Sequencing reads with regions flanking the T-DNA were used to determine the insertion site in the L. maculans genome using BLAST.

The location of the T-DNA in mutant M20W was obtained using a sequencing approach on a pool of DNA samples (Urquhart et al. 2018). The T-DNA was localised in the L. maculans spt8 homolog with the T-DNA inserted with the left and right borders preserved (Fig. S2). A three-nucleotide insertion of TTT was present at the right border of the T-DNA. The remaining T-DNAs were located in the genomes using Illumina sequencing reads generated separately for each strain (GenBank BioProject PRJNA636281) mapped back onto the T-DNA sequence. This revealed a suite of non-canonical integrations (Fig. S2, Table S3).

Similar complex T-DNA induced rearrangements in L. maculans were previously observed (Chambers et al. 2014). Identification of such integration events would have been challenging using the traditional PCR techniques, such as inverse PCR or thermal asymmetric interlaced (TAIL) PCR (Liu and Whittier 1995). Whole genome sequencing may be an indispensable component of the T-DNA mutagenesis pipeline in L. maculans. The rationale for developing the CSBX series of plasmids was the ability to pool DNA from multiple strains before such a sequencing approach. Each plasmid contains a specific 20 bp barcode close to the right border of the T-DNA. However these findings, i.e. of T-DNA truncations, insertion of additional DNA and chromosomal rearrangements observed in the T-DNA insertional mutants of the fungus L. maculans in the study indicate that such a DNA pooling approach may lead to confounding interpretations in this species.

Another limitation with Agrobacterium-mediated insertional mutagenesis in L. maculans is the emergence of phenotypes not associated with T-DNA insertion in high rates (Blaise et al. 2007). We selected three of the mutants to test the role of the underlying gene in disease. In one of the three mutants, the disrupted gene was linked to known pathogenicity genes in other fungi, while in two other mutants the disrupted genes had no link to known pathogenicity genes (and future research could explore the other mutants).

For complementation testing, genomic regions corresponding to genes affected by T-DNA integrations were amplified using high-fidelity Q5® polymerase (New England Biolabs) with the primers listed in Table S2 from wild type genomic DNA. Two mutants, 437 and 605, were complemented using a construct based on plasmid pMAI2 which contains a G418 resistance cassette (Idnurm et al. 2017). Gene regions were cloned into the pMAI2 vector linearised with EcoRV and XbaI, using Gibson assembly. The Lema_T048720.1 gene (named cab3) was amplified as a single fragment (2.6 kb). The size of the region (9.7 kb) needed to complement the Lema_T006490.1 gene (named hrc1) meant that it could not be efficiently amplified as a single piece. Thus, a two-stage approach was employed in which a partial fragment was cloned into pMAI2 (Idnurm et al. 2017), then this construct was digested with XbaI and the remaining section of the gene introduced. The gene region for Lema_T009520.1 (named spt8; 3.8 kb) was cloned into the EcoRV-XbaI site of plasmid pPZPHygHindX (Elliott and Howlett 2006). As the regions cloned were large the constructs were not verified by Sanger sequencing, and instead independent clones were used to circumvent any PCR-induced mutations. These constructs were then introduced into the respective T-DNA mutants by Agrobacterium-mediated transformation, using G418 or hygromycin as the selective agent.

The pathogenicity defects of these three mutants could be restored using the wild type copy of the gene affected by the T-DNA integration (Fig. 1). Two mutants grew similarly to the wild type and complemented strains in vitro (Fig. S3) suggesting that both these genes are true pathogenicity genes as opposed to genes essential for overall growth rate. The M20W strain has a colony morphology change and reduced asexual sporulation, which were also restored through complementation with the wild type copy of the gene (Figs. S3, S4).

Fig. 1
figure 1

Pathogenicity of wild type strain D5, and T-DNA insertional mutant strains M20W, 437 and 605 and their complemented (comp.) counterparts. (A) Representative images of cotyledons 14 days post inoculation. (B) Quantification of the lesion sizes obtained for multiple cotyledons. Statistical significance in lesion sizes between mutants and their corresponding complemented strain were determined using a t-test

Full size image

One strain has a mutation in a gene that is relatively well-characterised in other fungi, i.e. encoding Spt8 that is part of the SAGA (from Spt-Ada-Gcn5 acetyltransferase) complex (Eisenmann et al. 1994). The SAGA complex, which in yeast Saccharomyces cerevisiae is composed of 19-20 subunits, is involved in transcriptional activation of genes by modification of histones through nucleosomal histone H3 acetyltransferase (HAT) and histone H2B deubiquitinase activities (Koutelou et al. 2010; Han et al. 2014). In yeast the SAGA complex regulates the transcription of approximately 10% of genes, with enrichment for stress-induced genes (Huisinga and Pugh 2004).The spt8 homolog is required for wheat disease caused by Fusarium graminearum (Gao et al. 2014). As in L. maculans (Fig. S4), the loss of spt8 causes reduced conidiation in other fungi including F. graminearum (Gao et al. 2014) and Aspergillus fumigatus (Chen et al. 2022).

The functions of the remaining two pathogenicity genes identified in this study are more difficult to predict. The gene impaired in strain 437 encodes a putative flavoprotein enzyme (Lema_T048720.1). A BLAST analysis indicated that this enzyme shows similarity to the coenzyme A biosynthesis protein 3 (Cab3) of S. cerevisiae. In S. cerevisiae, Cab3 is essential in the biosynthesis of coenzyme A (Olzhausen et al. 2009; Giaever et al. 2002). Strain 605 has a mutation in a gene encoding a large HEAT-repeat-containing protein (Lema_T006490.1). This protein shows only weak similarity (28% amino acid identity) to the Utp20 protein of S. cerevisiae. Utp20 is involved in biogenesis of the 18S rRNA in S. cerevisiae (Dez et al. 2007), so this gene may also have a role in RNA processing in L. maculans.

The discovery of pathogenicity factors in plant pathogenic fungi has been a long-term goal towards finding new control strategies and to understand the basis for how fungi cause diseases in plants. In contrast to the progress in genomics or in the identification of the avirulence genes in L. maculans (Rouxel and Balesdent 2017; Rouxel et al. 2011; Dutreux et al. 2018), there has been limited progress on the discovery of additional pathogenicity factors: i.e. 16 genes are now known to be required for full disease (Table S1). Examining these identified pathogenicity factors reveals few common properties between them, as there is no emergent pattern of the types of genes and pathways essential for blackleg disease.