Leptosphaeria maculans

Introduction

A severe attack of Leptosphaeria maculans causing intensive lodging in an oilseed rape field (credits: CETIOM)

Fungi are organisms of primary ecological, biotechnological and economic importance. Many fundamental biological processes (cell cycle, circadian rhythms, ageing, cytoskeleton dynamics, regulation of gene expression, cell differentiation) are shared by human and fungi and are currently studied in fungi as model systems, due to their experimental tractability and their small genome size. Fungi are important for food production (edible mushrooms, fermentable products) and for the industrial production of enzymes and metabolites. These microbial eukaryotes cause important diseases in humans, particularly in immunocompromised patients, and are emerging as a major threat since they are very difficult to control with current fungicides. Fungi are also responsible for many diseases of plants, causing large losses in crop production and contaminating harvested crops with mycotoxins dangerous for health. As such they are the target of large-scale fungicide use, with the corresponding environmental and health hazards.

Of the ca. 250 000 fungal species known to date, ca. 75% are Ascomycetes, of which at least three classes include specialized plant pathogens. Of these three classes, the main one is Dothideomycetes. This class encompasses genera Stagonospora, Ascochyta, Venturia, Mycosphaerella, Alternaria, Cochliobolus and Leptosphaeria, representing the most important pathogens of wheat, legumes, barley, apples, bananas, brassicas, etc.

The complex parasitic strategies of Leptosphaeria maculans

A Leptosphaeria maculans ascus containing 6-celled ascospores (credits: D. Zickler, IGM, Orsay)

Leptosphaeria maculans is the cause of “stem canker” (also termed “Blackleg”) of oilseed rape (or canola) (Brassica napus). In France, the disease is responsible for 5-20% average yield losses due to drastic lodging of the crop (see the picture on the effect in the field). In addition, L. maculans is a model fungus, since it exemplifies Dothideomycete infection strategies, but nevertheless shows specificities in its life traits. During its life cycle L. maculans is firstly:

1. saprophyte on stem residues. It can survive as a saprobe for many years on the debris where sexual mating takes place resulting in the generation of ascospores, the primary inoculum (see the ascii picture);
2. necrotrophic for a very short time period, causing leaf spots where asexual multiplication takes place (see picture).
3. this leaf spot stage is followed by a lengthy endophytic systemic colonisation of leaf tissues, which may last up to 9 months for winter oilseed rape, e.g., in Europe. This phase of the disease is fully symptomless, and the fungus is growing in intercellular spaces towards the crown at the basis of the stem, and the upper root. These plant tissues are a specific ecological niche where numerous L. maculans individuals are living quiescent;
4. finally, at the end of the growing season, the fungus suddenly turns necrotrophic, and destroys the crown tissues, causing the devastating crown canker responsible for lodging of the plants (see picture).

The primary symptom of the disease : the leaf spots occurring on oilseed rape leaves in Autumn (credits: T. Rouxel, INRA)

As compared to more specialized fungal pathogens, for which parasitism is linked with the loss of genes corresponding to unnecessary functions, such as sexual reproduction or competition with other microbes in the environment, Leptosphaeria maculans is unique in its ability to develop contrasting life modes. As a consequence, it probably maintains numerous genes required for saprophytic life (nutrient acquisition, competition with the soil flora, etc.), necrotrophic parasitism (toxins, degradation enzymes, etc.) and intercellular cryptic growth (suppression of recognition by the plant and/or suppression of plant defence responses). These diverse life traits also suggest a high level of plasticity in terms of nutrient acquisition and requirements. Such a versatile behaviour and the successive setting of all these biological or phytopathological programmes suggest Leptosphaeria maculans is an adequate and complete model for genome-wide based functional studies of pathogenicity, and to identify signalling and regulation processes responsible for shifts in life style. The importance of sexual recombination in the wild also renders it a very efficient model for the study of molecular evolution of genes submitted to selection pressure.

Stem canker observed on oilseed rape plants in the field (original painting by M. Huau, 1950, INRA-PMDV)

Superimposed on this behaviour versatility, Leptosphaeria maculans, as other Dothideomycetes, develops an extreme host specialization termed the gene-for-gene relationship, where the outcome of the infection (resistance or susceptibility) will depend on the presence of one major gene for resistance in the plant (R) and one corresponding “avirulence” (AVR) gene in the pathogen. This specific recognition lead to the rapid set-up of plant defence response at the onset of leaf penetration, and fully protects the plant from the disease.

The lab tractability of Leptosphaeria maculans

• L. maculans is a haploid fungus, with a small genome size, of about 34 Mb, and predicted to encode ca. 10 000 genes. Its 15-16 chromosomes, including a non-mendelian-transmitted minichromosome range between 0.7 and 3.5 Mb and are easily separated by CHEF electrophoresis, therefore allowing to allocate linkage groups or contigs to given chromosomes.
• The fungus is easily grown in vitro on a range of media. Culture conditions have been optimised for enhanced asexual sporulation. It can be maintained for years in collections; for example, some isolates of the early 1970s have retained sporulation and pathogenicity and can be compared to current populations.
• The fungus is strictly heterothallic and, in contrast to most fungal phytopathogens, it can be crossed easily on synthetic media, with an excellent level of fertility, whatever the cross undertaken, and giving rise to an abundant progeny within two months. Tetrad analysis, though being tedious, is feasible.
• Phenotyping is easy, mainly with regards to phytopathological traits (aggressiveness, i.e., the amount of disease, and virulence, i.e., the ability to cause or not disease on a given host). Miniaturized assays on cotyledons of plantlets can allow the phenotyping of up to 500 interactions per week in INRA-PMDV conditions (with 10-12 repeats per interaction). Such high-throughput screening is suitable to evaluate mutant libraries or progenies.
• L. maculans is amenable to genetic transformation, and highly efficient agrotransformation protocols have been developed. In this respect GFP and DsRed reporter gene vectors have been developed, along with a vector designed to enhance the recovery of fungal transformants having undergone homologous recombination events (thymidine kinase negative selection). Also BiBAC vectors for transformation of large pieces of DNA via agrotransformation are constructed. Preliminary findings in B. Howlett’s lab are that RNAi protocols to knock out gene expression are effective.

Available resources and tools include extensive collections of natural isolates or progeny, extensive collection of Agrobacterium-mediated tranformants, genetic maps, BAC libraries, including one 24 genome-equivalent BAC library maintained at INRA-Versailles, etc. In addition the whole genome sequence of two related species, the wheat pathogen Stagonospora nodorum (Richard Oliver team, Murdoch University, Australia) and the Brassica pathogen Alternaria brassicicola (Christopher Lawrence, Virginia Bioinformatics Institute, USA), will be completed in the coming year and made available to the scientific community for future comparative genomics of pathogenicity, and genome evolution approaches.

Biological and phytopathological process studied in Leptosphaeria maculans

Among the main research topics analysed in Leptosphaeria maculans as a model, regardless of researches performed on the plant side, one can stress:

• Host specificity linked with avirulence genes ; evolution of AVR genes and their genome environment under selection pressure (INRA-Versailles);
• Intrinsic function of avirulence genes, loss of function linked with gain of “virulence” and consequences for the fungal fitness (INRA-Versailles) ; consequences for the durable management of resistance sources in oilseed rape (main objective of the EU-funded project SECURE);
• Structure and evolution of Leptosphaeria maculans centromeres; consequences for the chromosome size polymorphism generated by meiosis (INRA-Versailles);
• Identification of fungal pathogenicity genes by forward genetics (Melbourne University, INRA-Versailles);
• Transcriptomics and proteomics of the interaction with Brassicas and Arabidopsis thaliana;
• Organisation and gene function of clusters of secondary metabolite genes involved in toxic metabolite production (univ. Melbourne ; univ. Saskatchewan);
• Detoxification of plant defense compounds (i.e., phytoalexins) and consequences on pathogenicity (University Saskatchewan).

Objectives of the sequencing project

The gene-for-gene paradigm, now established for numerous plant-pathogen interactions, has major economic and scientific consequences. Such strict host specificities led breeders to widely breed for, and deploy, the corresponding resistance, i.e., major gene resistance in crops. The plants harbouring these genes are fully immune to the disease. The unwanted counterpart is that such resistance genes exert extremely strong selection pressure on pathogen populations that usually can easily develop single mutations in their avirulence genes, which are enough to overwhelm the corresponding major resistance. For example, the large scale deployment of major resistance gene Rlm1(recognizing avirulence gene AvrLm1) in France since 1995 led, in three years only, to the surge and prevalence of populations of Leptosphaeria maculans, having lost the functional allele of AvrLm1. Consequently, the efficiency of Rlm1-harbouring varieties to control the disease was lost in three years only, whereas more that 10 years were needed for breeders to introduce this resistance source in commercial varieties.

An example of the gene-for-gene interaction as expressed on cotyledons of oilseed rape, the AvrLm1-Rlm1 incompatibility

On a fundamental point of view, there is a definite lack of data on fungal avirulence genes, which are broadly defined by the resistance phenotype they induce on resistant plants. In this respect their intrinsic function, their role in fungal biology or pathogenicity, the recognition mechanism between the R and AVR genes, and the mode of evolution under selection pressure are widely unknown, with only 10 avirulence genes in only 3 three ascomycete species currently identified at the molecular level, and a definite lack of common traits between these genes.

The INRA-PMDV group at Versailles develops map-based cloning strategies to identify such avirulence genes in L. maculans (termed AvrLm). Eight AvrLm genes are currently mapped on the fungal genetic map, and six of them are organized as two unlinked genetic clusters comprising three genes each. The main target of the Rouxel’s team at INRA-Versailles currently is the AvrLm1-AvrLm2-AvrLm6 cluster, due to (i) the wide use of the

Rlm1 resistance gene these last years, and the concomitant availability of large collections of isolates having been submitted to the selection pressure; (ii) the very different crop history of the three resistance gene corresponding to avirulence genes of this cluster, with Rlm2 having been used in the most ancient oilseed rape genotypes, whereas Rlm6 has never been used in crops in Europe; and (iii) the important size polymorphism of the chromosome harbouring this gene cluster, ranging between 1.90 Mb and 2.80 Mb depending on the isolates.

History of the sequencing project

Scanning electron microscopy of ascospores germinating on oilseed rape leaves (credits: AM Jaunet & MH Balesdent, INRA)

A formal genetic approach allowed us to genetically delineate the AvrLm1-AvrLm2-AvrLm6 cluster on both sides with two markers of the genetic map separated by 9.8 cM. These markers were used to screen the 24-X BAC library in order to physically border the gene cluster with two BAC contigs.

A preliminary collaboration between Genoscope and INRA, in 2000, provided us with the whole sequence of the 5’ side bordering BAC contig in isolate JN3 (whose genotype at the three avirulence loci is AvrLm1-avrLm2-AvrLm6). This 184-kb genomic region has peculiar features in that it is extremely A+T-rich, and that it mainly consists of mosaic of a few families of retrotransposons, degenerated following RIP (Repeat Induced Point mutations) and usually severely truncated. The 5’ contig nevertheless contains a 32-kb G+C equilibrated region, containing functional ORFs. The contig contains numerous genetic markers co-segregating with AvrLm1. Unfortunately, the very strong recombination-deficient nature of the whole genome environment was misleading in targeting AvrLm1 within the 32-kb ORF-rich region, and recent data suggested AvrLm1 was actually located 3’ of the 184-kb contig. For subsequent steps, the availability of the 184-kb contig precise sequence was instrumental to design PCR markers based on Single Nucleotide Polymorphism (SNPs) between degenerated repeats, which could be used to screen 3-D BAC pools in order to expand the BAC contig in the 3’ part. This allowed to delineate a 1 Mb region (out of the 2.80 Mb of the chromosome), containing AvrLm1,

AvrLm6, and most probably AvrLm2. The whole region structure seems to be very similar to that of the 5’ 184 kb contig, since it appears to be also A+T-rich, to contain small ORF islands drowned within oceans of degenerated retroelements, to correspond to a recombination-deficient region, and since it is under-represented in the BAC library. These characteristics are reminiscent of those of regional centromeres with a central nucleus bordered by heterochromatic regions that may expand several hundred of kb. Regional centromeres are specific of upper Eukaryotes, and remain largely unexplored, whichever the species, due to difficulties in cloning, sequencing, assembling and annotation.

Additional characterization of natural populations of L. maculans, encompassing virulent and avirulent isolates at the AvrLm1 locus, along with genetic studies furthermore pointed out:

1. one 210-kb region within which recombination events are strongly suppressed, therefore corresponding to a genetic distance of less than 1 cM, whereas the average correspondence between physical and genetic distances is 8 kb per cM in the Versailles genetic map
2. one 530-kb region surrounding AvrLm1 which is submitted to drastic deletions when the isolates are losing the avirulent phenotype. Such deletions can reach 210 kb and fully explain the chromosome size polymorphism observed (see above).

These data strongly suggest that this pericentromeric, recombination-deficient region, may constitute a “genomic refuge” for genes involved in pathogenicity or avirulence.

In 2003, Genoscope undertook the sequencing of the whole pericentromeric region encompassing AvrLm1, AvrLm6, and probably AvrLm2, and corresponding to a 1-Mb BAC contig of isolate JN3 (AvrLm1 avrLm2 AvrLm6). In parallel, BAC contigs originating from two other isolates with different avirulent allele combinations (isolate NZ-T : avrLm1 avrLm2 AvrLm6, which is a field isolate from New Zealand and isolate v29 : avrLm1 AvrLm2 avrLm6, which is a lab progeny from a cross with one Australian isolate as a grand-father) will be partially sequenced with a special focus on G+C-rich regions, likely to contain target genes. In contrast to JN3, these two isolates are virulent on Rlm1-harbouring plant genotypes, and their sequencing will provide a precise comparison of the avirulent and virulent loci.

Once the sequence is assembled, a precise inventory of the genes present in the region will be performed in order to identify both putative avirulence genes and pathogenicity genes clustered in this region. Structural features of the region will be thoroughly analysed with special focus on the degree of conservation of the mosaic of repeats, in order to better define the deletion extent and borders, and possibly to identify deletion “hot spots”. In addition to the immediate deliverables of this sequencing project, the data thus obtained will be the basis for :

• The cloning of avirulence genes AvrLm1 and AvrLm6, including functional validation and accurate analysis of their function, kinetics of expression, regulation and cell localisation of the gene product.
• A study of the molecular evolution of avirulence genes and of the events leading to virulence in natural populations of the pathogen (in collaboration with INAPG). The Genoscope sequencing will be completed by a large-scale sequencing of virulent alleles in natural population to analyse the different events leading to virulence, as a function of the intensity and duration of the selection pressure, and will allow to infer a sequential evolution towards virulence and between virulent alleles.
• Analysis of structural features specific of L. maculans genes based on the sequence of ca. 30 genes (and promoter region) present in the contig sequenced here by comparison with cDNAs analysed in parallel. This will be the base for refined annotation of the L. maculans genome when whole-genome sequencing is performed.
• The design of improved race markers for the PCR-based identification and monitoring of virulent races avrLm1 and avrLm6.