Wheat yellow rust, also known as stripe rust, is one of the most devastating diseases of wheat worldwide. It is caused by the basidiomycete fungus Puccinia striiformis Westend. f. sp. tritici Eriks. (PST), an obligate pathogen that along with the stem (black) rust fungus Puccinia graminis f. sp. tritici (PGT) threatens worldwide wheat production [1, 2]. Historically, the use of major race specific resistance (R) genes in wheat varieties has been an effective method for disease management. However, these approaches are hampered by the evolution of resistance-breaking races of PST. For example, the appearance of PST races that overcome widely deployed R genes (such as Yr2, Yr9, Yr17 and Yr27) has led to destructive pandemics . In recent years, concerns over yellow rust have increased with the emergence of new and more aggressive PST races that have expanded virulence profiles and are capable of adapting to warmer temperatures compared to most previous races . Combined with the intrinsic ability of PST for long distance spore dispersal , these new races pose an increasing threat to global wheat production and food security .
Biotrophic plant pathogens such as rust pathogens secrete an array of proteins, known as effectors, to modulate plant innate immunity and enable parasitic infection . Some of these effectors translocate inside plant cells probably through specialized infection structures known as haustoria [7–9]. Inside plant cells, effectors perturb host processes promoting pathogenesis. However, disease resistance genes in plants, known as R genes, encode immunoreceptors that recognize specific pathogen effector proteins. Once effector proteins are recognized, plants initiate an immune response to block the development of disease, which typically results in a localized hypersensitive reaction and programmed cell death [10, 11]. The identification and characterization of these effectors and their cognate R genes is an important first step to understanding the wheat-PST pathosystem and consequently, to our ability to develop sustainable and potentially more durable resistance breeding strategies.
Recent availability of rust pathogen genome sequences has enabled the first steps towards wide-scale cataloguing of putative effector proteins. For instance, Saunders et al.  and Duplessis et al.  both implemented high throughput computational methods to characterize the effector complements from the fully sequenced rust fungi PGT and Melampsora larici-populina. Recently, Cantu et al.  used next-generation sequencing (NGS) to assemble a draft genome of PST isolate 130 (PST-130), annotating 22,185 putative coding sequences and classifying 1,088 of these as predicted secreted proteins. In addition, resources such as cDNA libraries generated from urediniospores and isolated haustoria (to identify PST genes specifically expressed during pathogen infection [15–18]) are publicly available. Together, they provide the necessary tools to develop a framework for characterization of the putative effector repertoire of PST.
The rapid decrease in sequencing costs now makes it possible to re-sequence multiple PST isolates to further characterize its pathogenicity arsenal. For instance, comparative genome analyses of different isolates of Magnaporthe oryzae, the rice blast pathogen, expanded the knowledge gained from the original reference genome considerably and helped to identify new effector genes with avirulence activity . Similarly, genomic analysis of an epidemic isolate of the potato blight pathogen Phytophthora infestans provided insights into increased aggressiveness and virulence .
In this study, we re-sequenced four PST isolates with different virulence profiles and from two distinct geographical regions (the USA and the UK). We identified hetero- and homokaryotic SNPs, providing a first glimpse into PST genetic diversity on a genome wide scale. We performed independent gene discovery and annotation across all isolates to produce a combined PST secretome and identified haustoria-enriched transcripts. We validated the expression of a subset of genes during an infection time course, and revealed distinct temporal expression patterns among them. This data was then integrated using a modified version of the in silico pipeline described in Saunders et al.  to classify the putative effector repertoire of PST. Using this information, we identified putative secreted, haustoria-enriched proteins with non-synonymous polymorphisms specifically between the two UK isolates, which only differ in virulence to two known wheat differential varieties. This approach identified five effector candidates among 2,999 predicted secreted proteins that are highly expressed in haustoria and are polymorphic between the UK isolates, PST-87/7 and PST-08/21. These allelic variants are now a priority for functional validation as virulence/avirulence effectors in the corresponding wheat varieties.