Dominant integration locus drives continuous diversification of plant immune receptors with exogenous domain fusions

Abstract

Background: The plant immune system is innate and encoded in the germline. Using it efficiently, plants are capable of recognizing a diverse range of rapidly evolving pathogens. A recently described phenomenon shows that plant immune receptors are able to recognize pathogen effectors through the acquisition of exogenous protein domains from other plant genes.

Results: We show that plant immune receptors with integrated domains are distributed unevenly across their phylogeny in grasses. Using phylogenetic analysis, we uncover a major integration clade, whose members underwent repeated independent integration events producing diverse fusions. This clade is ancestral in grasses with members often found on syntenic chromosomes. Analyses of these fusion events reveals that homologous receptors can be fused to diverse domains. Furthermore, we discover a 43 amino acid long motif associated with this dominant integration clade which is located immediately upstream of the fusion site. Sequence analysis reveals that DNA transposition and/or ectopic recombination are the most likely mechanisms of formation for nucleotide binding leucine rich repeat proteins with integrated domains.

Conclusions: The identification of this subclass of plant immune receptors that is naturally adapted to new domain integration will inform biotechnological approaches for generating synthetic receptors with novel pathogen "baits."

Keywords: Disease resistance genes; Gene fusions; NLRs; Plant immunity.

Fig. 1

Maximum likelihood phylogeny of NLRs in grasses identifies evolutionary hotspots of NLRs with integrated domains. a The maximum likelihood tree of the NB-ARC family in grasses (4184 proteins, nine species) showing occurrence of integrated domain (ID) across the phylogeny (red branches). b Close-up of MIC1 (red) as well as its outgroup clades C15 (blue), C14 (brown), and ancestral clade C13 (cyan) showing the key bootstrap support values. c Wordcloud summary of the integrated domain diversity from MIC1. E-value cut-off for presence of an ID domain, 0.001

Fig. 2

MIC1 has proliferated in grasses and continues to accumulate new domains as seen from comparison of wheat and its diploid progenitors. a Overall evolutionary relationship of grasses used in this study and corresponding number of NLR-IDs in MIC1. Key clade divergence is marked on the tree in millions of years as estimated at timetree.org. b The repertoires of IDs are different among wheat subgenomes and their progenitors suggesting continuous integration of new IDs

Fig. 3

NLRs from MIC1 are genetically linked in head-to-head pairs with NLRs from clade C7. a The diagram shows the orientation and maximum distance of 15 kb that we used to identify NLR gene pairs. b Heatmap showing numbers of tandem NLRs across different clades. c Circos plot on showing links between NLRs from MIC1 (C16) and NLRs from other clades that are oriented across the plot in the same order as in the overall NLR phylogeny. NLR gene pair tandems are color-coded according to their species as indicated in the legend on the left

Fig. 4

Orthologous NLRs from MIC1 in rice and Brachypodium show NLR gene coupled to generation of new NLR-IDs. a Phylogeny of NLRs from MIC1 in rice and Brachypodium based on the NB-ARC domain. Red boxes indicate NLRs with IDs. The links between trees highlight orthologous NLRs between rice and Brachypodium. b Microsynteny analyses between rice and Brachypodium. Blue boxes and ochre boxes represent syntenic genes in rice and Brachypodium, respectively. Red boxes indicate NLRs, purple boxes indicate NLR-IDs

Fig. 5

Close-up of MIC1 displaying rapid domain recycling. The branches of the hotspot clade, the outer clade, and the ancestral clade are shown in red, blue, and cyan, respectively. Dots on the branches indicate a bootstrap support value ≥ 85%. Alongside the tree are cartoons of each protein, annotated with the domain(s) in the position that they appear in the protein (protein backbone, gray line; NB-ARC domain, black rectangle; LRR and AAA, TIR, and RPW8 domains, orange rectangles; other domains in different colors and shapes as indicated in the key). For clarity, the domain lengths are shown in the key for B3, Exo70, GRAS, and the Myb_DNA_binding domains. E-value cut-off for presence of an ID domain, 0.001; domains with e-value > 0.001 and ≤ 0.05 are shown as gray rectangles. E-value cut-off for an LRR, AAA, TIR, or RPW8 domain, 10.0

Fig. 6

NLR-IDs from MIC1 share a protein motif at the site of domain integrations. a Distribution of CC, NB, LRR, and ID domains and motifs identified using MEME on unannotated regions of NLRs within the MIC1 clade with C-terminal ID. For every NLR, the length of the NLR was normalized to 1.0 and the midpoint of identified domains was normalized to protein length. b Sequence logo of CID domain [56]. c Domain structure of 70 NLR-IDs within the MIC1 clade that contain the CID domain. The CID domain is located immediately upstream of the site of integration

Fig. 7

NLR-WRKY/AP2a domain shuffling involved inter-chromosomal copy-and-paste of the AP2 gene. a A clade in the phylogeny of NLR-ID proteins from Fig. 5 that includes wheat. A, B and D genome homoeologs from the same genetic position. New integration is evident from homologs with AP2 and a WRKY domains (highlighted by red boxes). Dots on the tree branches indicate a bootstrap support value ≥ 85%. E-value cut-off for presence of an ID domain, 0.001; a domain with e-value > 0.001 and ≤ 0.05 is shown as a gray rectangle. E-value cut-off for an LRR domain, 10.0. b An AP2/ERF family tree (left) showing two clades that contain an NLR acceptor ID protein (NLR-AP2a and NLR-AP2b, indicated in red). The protein sequences of these clades were re-aligned and the trees re-estimated (right) to confirm the identity of the donor protein, evident from high bootstrap support values. c Protein alignment cartoon of one of the AP2 donor proteins and the acceptor protein, NLR-AP2a. By contrast, the NLR acceptor protein homoeologs contain two WRKY domains in their C-terminal ends (d) summary model illustrating the inter-chromosomal duplication of an ID domain and subsequent movement into an NLR gene

Fig. 8

Evolutionary model of NLR-ID hotspot formation and diversification. An ancestral protein underwent duplication to form the outer clade of proteins (blue). An ancestral pair between MIC1 and C7 formed as well as the CID LRR-cap motif evolved to be MIC1-specific. Together, this enabled MIC1 to be highly amenable to new gene fusions both on genetic and protein regulatory levels. Duplications of the MIC1-C7 pair created new landing pads for IDs. Some proteins have maintained the same domain (e.g. ID1) but other proteins have undergone further diversification through the exchange of the ID domain (e.g. ID3 and ID4)