Comparative Genomics and Functional Studies of Wheat BED-NLR Loci

Abstract

Nucleotide-binding leucine-rich-repeat (LRR) receptors (NLRs) with non-canonical integrated domains (NLR-IDs) are widespread in plant genomes. Zinc-finger BED (named after the Drosophila proteins Boundary Element-Associated Factor and DNA Replication-related Element binding Factor, named BED hereafter) are among the most frequently found IDs. Five BED-NLRs conferring resistance against bacterial and fungal pathogens have been characterized. However, it is unknown whether BED-NLRs function in a manner similar to other NLR-IDs. Here, we used chromosome-level assemblies of wheat to explore the Yr7 and Yr5a genomic regions and show that, unlike known NLR-ID loci, there is no evidence for a NLR-partner in their vicinity. Using neighbor-network analyses, we observed that BED domains from BED-NLRs share more similarities with BED domains from single-BED proteins and from BED-containing proteins harboring domains that are conserved in transposases. We identified a nuclear localization signal (NLS) in Yr7, Yr5, and the other characterized BED-NLRs. We thus propose that this is a feature of BED-NLRs that confer resistance to plant pathogens. We show that the NLS was functional in truncated versions of the Yr7 protein when expressed in N. benthamiana. We did not observe cell-death upon the overexpression of Yr7 full-length, truncated, and 'MHD' variants in N. benthamiana. This suggests that either this system is not suitable to study BED-NLR signaling or that BED-NLRs require additional components to trigger cell death. These results define novel future directions to further understand the role of BED domains in BED-NLR mediated resistance.

Keywords: BED-NLR; DUF4413/659(-hAT); NLR; NLR-ID; Yr5; Yr7; integrated domain; plant disease resistance; wheat; zf-BED.

Figure 1

Schematic representing the Yr region in ten wheat varieties, a spelt variety and B. distachyon. The syntenic region was defined based on the eleven gene models with hits across all varieties (gene models shown as white triangles). Dark brown represents the region of chromosome 2B in synteny with the Yr region on RefSeqv1.0. Black triangles show potential canonical NLR loci annotated with NLR-Annotator and red triangles potential BED-NLR loci (Table S3). Purple triangles depict Yr7 best BLAST hit in Group 1 varieties (Yr7_L477P) and the white triangle with a blue outline shows Yr5a in spelt. The BED-NLR-rich region in the vicinity of the Yr7 is highlighted with a dashed light purple line (see Figure 2 for details). The orientation of the triangles reflects the orientation of the gene model. The orange triangle in B. distachyon represents a gene that was outside of the synteny with wheat. NLRs sharing 100% identity across 95% of their sequences are reported in Table S4.

Figure 2

Close-up on the BED-NLR-rich region in Landmark, Mace, Stanley and SY-Mattis where Yr7-L744P is located. Colours and shapes shown in this figure correspond to those used in Figure 2. We added the scaffolds of the Cadenza assembly that contained hits to the Landmark BED-NLRs, including the scaffold which contains Yr7. The orientation of the triangles reflects the orientation of the loci. We identified a BED-NLR locus in tail-to-tail orientation located 10 kb from Stanley-Yr7-L744P. In Cadenza, Landmark and Mace, we identified nlr_11a/b that was located ~4.5 kb and in tail to head orientation with Yr7/Yr7-L744P.

Figure 3

Neighbour-network of BED domains from non-NLR and NLR proteins in the Pooideae. The network was generated with SplitsTree v.4.16 [48] from a multiple sequence alignment produced with MAFFT v7.305 and the L-INS-I method [53]. Bootstrap values are reported in gold, NLR proteins in red and non-NLR proteins in black. Protein carrying a nuclear localization signal (NLS) in the vicinity of the BED domain (within 40 amino-acid up or downstream) are depicted in light blue (NLRs) and dark blue (non-NLRs), whereas proteins carrying an NLS further away from the BED domain are shown in brown (non-NLRs). The full list of NLS-containing proteins is in Table S15. Protein that were annotated as BED-DUF4413-(hAT) in RefSeqv1.0 gene models but are annotated as single-BED in RefSeqv1.1 gene models are shown in green. Asterisks indicate BED-only proteins that share sequence similarities with BED domains from BED-NLRs (Tables S6 and S8). Clusters were defined based on the presence of BED-NLRs and the closest clusters harboring non-NLR proteins (see Results).

Figure 4

Cellular localization in N. benthamiana of truncated variants of the Yr7 protein and their corresponding deletion mutants in the predicted NLS. (A) Schematic describing the different truncations analysed and their corresponding deletion mutants in the predicted NLS (purple). Yr7-AA201 does not contain the predicted NLS. Both Yr7-AA242 and AA308 contain the predicted NLS whereas Yr7-AA242-dNLS and Yr7-AA308-dNLS are mutants lacking the predicted NLS. (B) Fluorescent microscopy pictures from leaves expressing the corresponding Yr7 variant. Samples were taken at 1.5 dpi and five leaf fragments from two leaves per plant (two plants per construct) were assessed under confocal microscope (see Methods section). Single YFP was used as a negative control.