Comparative genomics of Australian isolates of the wheat stem rust pathogen Puccinia graminis f. sp. tritici reveals extensive polymorphism in candidate effector genes

Abstract

The wheat stem rust fungus Puccinia graminis f. sp. tritici (Pgt) is one of the most destructive pathogens of wheat. In this study, a draft genome was built for a founder Australian Pgt isolate of pathotype (pt.) 21-0 (collected in 1954) by next generation DNA sequencing. A combination of reference-based assembly using the genome of the previously sequenced American Pgt isolate CDL 75-36-700-3 (p7a) and de novo assembly were performed resulting in a 92 Mbp reference genome for Pgt isolate 21-0. Approximately 13 Mbp of de novo assembled sequence in this genome is not present in the p7a reference assembly. This novel sequence is not specific to 21-0 as it is also present in three other Pgt rust isolates of independent origin. The new reference genome was subsequently used to build a pan-genome based on five Australian Pgt isolates. Transcriptomes from germinated urediniospores and haustoria were separately assembled for pt. 21-0 and comparison of gene expression profiles showed differential expression in ∼10% of the genes each in germinated spores and haustoria. A total of 1,924 secreted proteins were predicted from the 21-0 transcriptome, of which 520 were classified as haustorial secreted proteins (HSPs). Comparison of 21-0 with two presumed clonal field derivatives of this lineage (collected in 1982 and 1984) that had evolved virulence on four additional resistance genes (Sr5, Sr11, Sr27, SrSatu) identified mutations in 25 HSP effector candidates. Some of these mutations could explain their novel virulence phenotypes.

Divergent and convergent modes of interaction between wheat and Puccinia graminis f. sp. tritici isolates revealed by the comparative gene co-expression network and genome analyses

“…A total of 1,509 genes (9.4%) annotated in the SCCL genome were not detected in the transcript data from either isolate. The high degree of divergence in the gene contents of RKQQC, MCCFC, and SCCL genomes is consistent with the degree of divergence observed in other Pgt isolates as well as between isolates of other fungal pathogens [14, 16, 44]. Expectedly, the highest level of divergence in the gene content was found between the Ptt genome and the genomes of Pgt isolates.…”

“…To remove contaminating sequences, the genome assemblies were compared against the NCBI non-redundant nucleotide database, and contigs that showed nucleotide similarity to fungal sequences were retained (see Materials and Methods). The final MCCFC and RKQQC genome assemblies measured 93.3 Mb (N50 7,133 bp) and 107.3 Mb (N50 6,292 bp), respectively (Additional file 1: Table S3), and were comparable with the previously reported Pgt genome sizes [16, 41]. The Pgt genome assemblies and their annotations can be downloaded from the project website http://129.130.90.211/rustgenomics/Download.…”

“…The differences in the virulence profiles of these isolates were reflected by substantial inter-isolate gene content variation consistent with that observed in other plant pathogenic fungi [14, 16, 61, 62]. Due to their direct interaction with the host factors, the patterns of inter-isolate sequence and presence/absence variation in the effector encoding genes have been one of the major focuses of recent plant-pathogen interaction studies [5, 9, 10].…”

ApoplastP: prediction of effectors and plant proteins in the apoplast using machine learning

“…Reads for germinated spores and haustorial tissue (100 bp paired‐end) were obtained from NCBI BioProject PRJNA253722 (Upadhyaya et al ., ). These were adapter‐trimmed using Trimgalore with default parameters (https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/).…”

Exploring and exploiting the boundaries of host specificity using the cereal rust and mildew models

“…Although well established for bacterial resistance very little is known for both the guard and decoy models in plant fungal interactions (van der Hoorn & Kamoun, ). It is well documented, however, that pathogen avirulence ( Avr ) gene effector products are recognized by the LRR domain of NLR proteins, and the generation of polymorphisms in either pathogen Avr or host R genes (gene‐for‐gene) via well‐characterized evolutionary mechanisms are critical for pathogen survival (Upadhyaya et al ., ). Recent studies on the isolation of rust resistance genes expressed in wheat during ETI suggest that in all cases NLR genes encoding cytoplasmic receptor proteins were responsible for the observed resistance (Periyannan et al ., ; Saintenac et al ., ; Mago et al ., ; Steuernagel et al ., ; Sánchez‐Martín et al ., ; Chen et al ., ; Salcedo et al ., ).…”