Abstract
The wheat powdery mildew disease is caused by the obligate biotroph Blumeria graminis f. sp. tritici (Bgt) and race-specific resistance can be conferred by the Pm3 allelic series. These genes encode a coiled-coil (CC) nucleotide-binding leucine-rich repeat protein (NLR) that acts as an immune receptor and recognize sequence diverse AVRPM3 effectors having a predicted conserved RNase-like fold. So far, AVRPM3a2/f2, AVRPM3b2/c2 and AVRPM3d3 have been identified which are recognized by PM3a/f, PM3b/c and PM3d, respectively. Resistance conferred by PM3 variants can furthermore be suppressed in the presence of the Bgt effector SVRPM3a1/f1 indicating that this is a complex example of a gene-for-gene resistance model. In this study, we aim to elucidate the molecular mechanisms leading to recognition of AVRPM3b2/c2 by PM3b and the observed suppression. In the first chapter, we investigated the specific interactions between AVRPM3 effectors and found that AVRPM3b2/c2, AVRPM3a2/f2 and SVRPM3a1/f1 can all interact with each other to form homodimers and heterodimeric complexes. We propose that this interaction underlies the suppression activity of SVRPM3a1/f1 in a dose dependent manner. Furthermore, due to the interaction between AVRPM3 effectors and previously described structural similarities, we hypothesized that polymorphisms in their surface exposed residues could widen their recognition to non-corresponding PM3 variants. We tested a subset of synthetic AVRPM3a2/f2 variants and found that one of these was recognized by PM3b, further suggesting that PM3 variants have evolved to recognize AVRPM3 effectors that are highly similar to each other on the structural level. In the second chapter, we investigated the sub-cellular localization of PM3b and AVRPM3b2/c2. We found that PM3b does not conform to any typical sub-cellular localization such as the cytosol, nucleus, plasma membrane (PM), Golgi vesicles, or endoplasmic reticulum (ER), but was sequestered in immobile puncta in the cell. Upon further investigation we observed that these puncta are potentially associated with tubules and three-way junctions of the ER. Therefore, we hypothesize that these might represent immobile ER-PM contact sites. AVRPM3b2/c2 on the other hand seemed to mainly localize to the nucleocytoplasmic space and a comparison of two AVR variants showed that the one that is more strongly recognized by PM3b also had a higher fraction of the protein localizing to the nucleus. In the third chapter, we hypothesized that PM3b might directly recognize AVRPM3b2/c2 and we performed various interaction assays to try to confirm this. First, we tested PM3b and AVRPM3b2/c2 in a split-luciferase complementation assay, which gave negative results even though all proteins were present. We hypothesized that due to the large size of PM3b, there might be spatial or steric vi effects that hinder the reassociation of the split-luciferase fragments. Therefore, we tested for interactions between PM3b and AVRPM3b2/c2 in co-immunoprecipitation experiments. We found that in the presence of PM3b, AVRPM3b2/c2 protein levels are severely limited as compared to when it was expressed alone or with PM17, another resistance protein from wheat. Interstingly, co-expression of AVRPM3b2/c2 with just the LRR domain of PM3b also lead to lower levels of protein accumulation. Furthermore, we found that this difference was not due to general protein degradation as we saw no difference in rubisco or PM3b protein levels when expressed alone or with AVRPM3b2/c2. Co-immunoprecipitation experiments revealed that we could reciprocally immunoprecipitate PM3b and AVRPM3b2/c2 with each other, suggesting that they might indeed directly interact. In the fourth and final chapter of this thesis, we investigated the potential host targets of AVRPM3b2/c2. First, a yeast two-hybrid library screen against wheat was performed in which a large number of putative interactors were identified. As this list was too large to further validate, we hypothesized that by using a different method to “fish” for interactors, we could narrow down the list significantly. We therefore performed immunoprecipitation coupled to mass-spectrometry (IPMS) experiments in N. benthamiana, BLAST searched for homologues of these interactors in the wheat proteome and compared this list to the one found in the yeast two-hybrid experiment. This gave a significantly reduced set of 18 putative interactors which we gene-synthesized and proceeded with for further testing. By using split-luciferase and bi-molecular fluorescence complementation assays we could show that a set of six diverse proteins interacted in all of the assays which represent a list of putative AVRPM3b2/c2 host targets. While some details are still missing in understanding the interactions between PM3b and AVRPM3b2/c2, we have substantially increased our knowledge of this system. Furthermore, future key experiments that are needed to improve this understanding are suggested in this thesis. The findings described in this work can be used to better understand plant immunity and the wheat-powdery mildew interactions and has laid the groundwork for future studies in the field.
Isaksson, Jonatan