The white mold fungus Sclerotinia sclerotiorum (Lib.) de Bary is a cosmopolitan necrotrophic pathogen infecting over 400 plant species. It is among the most devastating pathogens of soybean, rapeseed and sunflower, causing several hundred million dollar losses annually at the pre- and postharvest stages . S. sclerotiorum host range is remarkably broad, with fruit and vegetable productions also being severely impacted . S. sclerotiorum and its close relative the grey mould fungus Botrytis cinerea are among the few fungal pathogens considered as typical necrotrophs. As such, they derive energy to complete their life cycle mostly from dead plant cells, as opposed to biotrophs that feed on living plant cells.
There is now ample evidence that biotrophic and hemibiotrophic fungi secrete specialized effector proteins manipulating host cell physiology to obtain nutrients, suppress plant defense and ultimately promote infection . Effectors may also trigger plant defense responses, leading to plant resistance, when recognized directly or indirectly by the plant immune system in a gene-for-gene relationship. This results from a co-evolutionary arms race between pathogen effectors, their plant targets, and components of the plant immune system . Necrotrophs have long been considered as less adapted, secreting mostly degrading enzymes and toxins that unspecifically trigger programmed cell death (PCD) and dismantle plant cells.
However, host specific necrotrophs such as Cochliobolus victoriae secrete effector proteins translocated into plant cells that interact with specific corresponding host proteins to facilitate disease progression [5, 6]. This involves the activation of plant PCD instead of its suppression as in the case of infection by biotrophic pathogens. S. sclerotiorum also produces the non-proteic pathogenicity determinant oxalic acid. This molecule induces the synthesis of reactive oxygen species (ROS) and triggers plant PCD late during infection, but has the opposite effect, suppressing ROS burst and PCD induction, at the early stages of infection . The SSITL secreted integrin-like protein of S. sclerotiorum promotes virulence and delays the activation of plant defense responses, supporting the view that S. sclerotiorum secretes effectors to finely manipulate plant physiology . In addition, enzymes secreted by necrotrophs can act as virulence factors independently of their catalytic activity . Effector repertoires vary considerably, notably according to pathogens lifestyle , and it is becoming clear that interactions between necrotrophs and their host plants are considerably more complex and subtle than previously considered. What is the effector candidate repertoire associated with broad host range necrotrophy remains unclear. As a first step towards elucidating the molecular bases for colonization by S. sclerotiorum, its repertoire of effector candidates needs to be determined.
The recent release of genome sequences for a number of plant pathogenic fungi facilitated the search for effector candidates (ECs) at the genome level . Nevertheless, considering that pathogen effector repertoires are typically lineage-specific, the identification of effectors remains challenging . The analysis of S. sclerotiorum genome sequence uncovered sets of genes associated with the manipulation of redox status, including enzymes of OA biosynthesis, the degradation of plant cell wall, and 603 secreted proteins with other functions . Known hallmarks of fungal effectors such as the presence of signal peptides and absence of transmembrane domains, small size and amino-acid composition generally produce lists of hundreds of potential effectors in a given pathogen. Therefore, more sophisticated approaches are required to pinpoint the most relevant ECs for the promotion of infection in S. sclerotiorum secretome.
A limited number of known fungal effector families show conservation at the sequence level or similar predicted functions. This is notably the case for the toxin and cell death elicitor proteins of the Necrosis and ethylene-inducing Like Proteins (NLPs), the cerato-platanin, cyanovirin-N homology (CVNH) and ECP2 families [13–16]. The growing number of characterized fungal effectors suggests conservation at the biochemical function level in the overall effector repertoire of several fungal pathogens. The ability to bind chitin or other cell wall oligosaccharides, masking the presence of the pathogen or dampening damage-induced plant responses, is a feature common to effectors from multiple fungal pathogens [17–20]. Fungal effectors harboring a protease inhibitor activity are also common [21–24]. The biochemical activity of a few other fungal effectors such as M. oryzae Fungalysin metalloprotease AvrPita , U. maydis chorismate mutase cmu1  and peroxidase inhibitor PEP1  may also be part of the arsenal of effector functions in multiple fungal lineages. This hypothesis suggests that thorough annotation of protein domains and prediction of biochemical function of secreted proteins may prove useful to identify novel effectors in S. sclerotiorum.
However, a majority of effectors do not show significant similarity to known sequences in other organisms nor obvious protein domains. Yet other genomic characteristics may help identify EC genes. The rapid evolution of effector genes allows the fungi to overcome selection pressures induced by resistant plant cultivars. A high ratio of non-synonymous over synonymous substitutions (Ka/Ks) in alleles from related strains is a frequently used proxy for inferring fast gene evolution and the action of positive selection . This approach has been used to reveal ECs in several filamentous plant pathogen lineages [29–34]. Positive selection has been detected in B. cinerea genome  suggesting that it may be used to mine S. sclerotiorum genome for ECs. Second, gene duplication is another hallmark of several known fungal effector genes, such as the ToxB host specific toxin of Pyrenophora tritici-repentis [36, 37]. Third, genomic regions with high repeat and transposable element content are enriched in effector genes in several lineages of plant pathogens [37–39] suggesting that genome architecture analysis can assist in the search for EC genes. Finally, effectors can alter host cell function by mimicking plant peptides . These ECs likely elude functional annotation on the basis of primary amino-acid sequence, but may be revealed using three-dimensional structure prediction.
As opposed to Oomycete pathogen genomes in which many effector genes can be identified through conserved sequence motifs [29, 41], the use of conserved sequence motifs, such as the Y/F/WxC motif , has proven limited in revealing fungal effectors across lineages. However, the presence of a signal peptide directing protein secretion and gene expression in planta are relatively universal properties of effectors that can be exploited as first filters to narrow down the list of effector candidates in fungal genomes.
S. sclerotiorum effector proteins would be useful as probes to search for resistance components in plants and to design strategies for inhibiting infections by this devastating but poorly characterized pathogen. In this study, we report a diverse repertoire of S. sclerotiorum effector candidates revealed by an in depth analysis of its predicted secretome. We combined refined secretome annotation, phylogeny, selection and gene duplication analyses, and three-dimensional structure prediction to identify 78 ECs. Among those, we highlight a predicted subtilisin inhibitor, a xylanase, a duplicated gene of unknown function and three toxin mimics as high priority candidates for functional studies. We analyzed in planta expression pattern for 16 EC genes and revealed host-blind and host-regulated ECs.