Abstract
Infectious diseases drive wild plant evolution and impact crop yield. Plants, like animals, sense biotic threats through pattern recognition receptors (PRRs). Overly robust immune responses can harm plants; thus, understanding the tuning of defense response mechanisms is crucial for developing pathogen-resistant crops. In this study, we found that an inverted-repeat transposon (EFR-associated IR, Ea-IR) located between the loci encoding PRRs ELONGATION FACTOR-TU RECEPTOR (EFR) and myosin XI-k (XI-k) in Arabidopsis affects chromatin organization, promoting the formation of a repressive chromatin loop. Upon pathogen infection, chromatin changes around EFR and XI-k correlate with increased EFR transcription. Pathogen-induced chromatin opening causes RNA polymerase II readthrough, producing a longer, Ea-IR-containing XI-k transcript, processed by Dicer-like enzymes into small RNAs, which reset chromatin to a repressive state attenuating the immune response after infection. Arabidopsis accessions lacking Ea-IR have higher basal EFR levels and resistance to pathogens. We show a scenario in which a transposon, chromatin organization and gene expression interact to fine-tune immune responses, during both the course of infection and the course of evolution.
Similar content being viewed by others
Main
Like vertebrates, plants have evolved complex mechanisms in which two arms of the immune system interact to efficiently combat pathogens while minimizing collateral damage to the host1. The first layer uses pattern recognition receptors (PRRs) situated on cell surfaces to sense conserved pathogen-associated molecular patterns (PAMPs), initiating PAMP-triggered immunity (PTI)2,3. The second layer involves intracellular receptors that recognize pathogen release effectors, activating effector-triggered immunity (ETI), resulting in a hypersensitive response and programmed cell death1,2. Nucleotide-binding site leucine-rich repeat proteins (NLR or NBS-LRR) act as principal ETI immune receptors4,5. PTI and ETI often synergize to optimize plant immunity6,7,8,9,10.
During PTI, activation of PRRs initiates a series of signal transduction events leading to changes in gene expression and increased basal disease resistance against pathogens that are not adapted to the host3,11,12. Most plant PRRs are receptor kinases or receptor-like proteins with extracellular ligand-binding, transmembrane and intracellular kinase domains11. The Brassicaceae-specific PRR ELONGATION FACTOR-TU RECEPTOR (EFR) recognizes bacterial EF-Tu protein13,14,15,16,17,18. Increased pathogen resistance often correlates with elevated EFR expression19,20,21,22,23.
While it is imperative for plants to promptly activate defense responses when attacked, these must be tuned to the size of the threat to avoid collateral damage in the form of reduced growth and seed set24,25. Hence, highly effective immune alleles have an advantage in populations under high pathogen pressure but decline in nonthreatening environments. Indeed, both PRR and NLR genes are often polymorphic in the natural populations of flowering plants26,27,28,29,30. Conversely, the homeostatic function and high conservation of most PAMPs cause PRR receptors to be less genetically variable12. Nevertheless, in addition to the absence or presence of variation in specific genes or alleles, variation in transcript or protein levels can influence PRR activity and alter immune responses in natural populations31. PRR RNA levels are known to be positively regulated by ETI, which indicates that there are opportunities for fine-tuning PRR expression levels before and during pathogen attack32.
Recently, we found that the insertion of transposable element (TE)-derived inverted repeats (IRs) near genes changes the local chromatin organization, with a profound impact on the activity of neighboring loci33,34. Mechanistically, transcription of these IRs leads to the synthesis of 24-nt small RNAs (sRNAs) and the methylation of the IRs. This provides an anchor point for the formation of short-range chromatin loops (SRCLs) that influence the expression of neighboring genes35. Many IRs vary between Arabidopsis accessions, causing differences in SRCL and gene expression34. Such variation in the gene-hosted IR content is notable because it can be associated with changes in phenotypic traits34. This points to IRs as powerful evolutionary agents that can epigenetically contribute to plant adaptation. In this sense, while TEs are commonly silenced through the RNA-dependent DNA methylation (RdDM) pathway that relies on the plant-specific RNA polymerases (Pol) IV and V and the RNA-dependent RNA polymerase 2 (RDR2)36,37, IRs can trigger DNA methylation through a noncanonical pathway as Pol II transcripts of IRs can fold into double-stranded RNAs (dsRNAs), triggering sRNA production without the need for RDR2 (ref. 38). This turns IRs into autonomous elements that, when inserted near a gene, can be transcribed, triggering local RdDM, chromatin looping and changes in gene expression34. While 24-nt small interfering RNAs (siRNAs) are well-known actors of the RdDM pathway, the initiation of TE methylation also relies on 21-nt siRNA. The initiation pathway acts on transcriptionally active transposons and begins with the conversion of TE transcripts into dsRNAs by the action of RDR6 (refs. 39,40,41). These dsRNAs are processed into 21-nt siRNAs by Dicer-like 4 (DCL4) and loaded into Argonautes 1 or 6 (AGO1 or AGO6)38,41,42. IR transcripts, because of their dsRNA nature, can also be processed into 21-nt siRNAs by DCL1 in the nucleus and integrated to the initiation pathway. AGO6-bound siRNAs guide cytosine methylation at complementary DNA sequences in collaboration with Pol V and DRM2, resulting in a low-level methylation. Although this low-level methylation is insufficient for transcriptional gene silencing, it acts as a signal to recruit the maintenance pathway machinery, which relies on 24-nt siRNA and is responsible for most RdDM activity39.
Here, we show that an IR that separates EFR from the myosin XI-k gene (Xl-k), which is associated with root organogenesis, ER dynamics and fungal penetration43,44,45, controls EFR expression by shaping local chromatin topology. Mechanistically, the methylation of the IR induces the formation of a chromatin loop that represses EFR expression. Upon pathogen infection, demethylation of the TE causes the opening of the loop, which correlates with the activation of EFR transcription. Methylation of the IR is triggered upon the DCL3 or DCL1 processing of a long, IR-containing XI-k transcript. This long transcript results from a Pol II readthrough event traversing the termination site when the loop is in the open conformation, as is the case during pathogen infection. At a population level, the IR is found in only some Arabidopsis accessions. Accessions without the IR present a distinctive locus topology, higher basal levels of EFR and enhanced resistance to pathogens. Because of the central role of EFR in plant innate immunity, this naturally occurring epigenetic variation gains relevance for environmental adaptability and evolution.
Results
An EFR-linked IR causes chromatin looping and EFR repression
On the basis of the recently discovered link between structural variants involving IRs near genes and potentially adaptive phenotypic variation in Arabidopsis natural accessions34, we searched for IR-proximal genes with roles in pathogen response. We focused on IRs within 500 bp of protein-coding genes, as these IRs tend to have the most profound impact on chromatin organization and gene expression34. Among the 261 genes located ≤500 bp from an IR element, the overlap between Gene Ontology (GO) categories ‘response to stress’ and ‘response to biotic stimulus’ included 24 genes (Supplementary Fig. 1a), including the well-known PRR gene EFR (AT5G20480) and the adjacent XI-k (AT5G20490) gene, which has also been linked to pathogen-relevant biological processes. A 442-bp-long IR occupies almost the entire 615-bp intergenic region separating these two pathogen-related genes (Fig. 1a). This IR, named EFR-associated IR (Ea-IR) for simplicity, corresponds to an element of the ATMU3N1 family of DNA and MuDR transposons and is annotated as a miniature IR TE (MITE) in the P-MITE database46.
a, Diagram of the EFR and Xl-k region on chromosome 5 (yellow bars, exons; blue bars, untranslated regions). The LIMPET1 transposon and the transposon-derived Ea-IR IR are indicated by purple bars. The T-DNA insertion disrupting Ea-IR in the eair-1 mutant is also indicated. Numbers indicate chromosome 5 coordinates. b, The 24-nt siRNAs mapping Ea-IR in Col-0 wild-type, dcl234 and ddc plants. For each genotype, the results from three replicates are shown with different shading. c, Left, cytosine methylation in the CHH, CHG and CG contexts at Ea-IR (delimited by vertical gray lines) and flanking regions (700 bp) in Col-0 wild-type, dcl234 and ddc plants. The average of three individual BS-seq replicates is plotted for each genotype. Right, the average percentage of cytosine methylation in the CG, CHG and CG for each region and the same genotypes. d, Chromatin contacts from CaptureC in the EFR and Xl-k region in Col-0 wild-type, ddc and dcl234 plants. LoopEFR, which includes Ea-IR, is marked in cyan, LoopXl-k is marked in green and all other interactions are marked in purple. The region targeted by CaptureC is indicated, with all potential restriction fragments below. Fragments diagnostic for the Ea-IR region are indicated in cyan. e, Top, purple pins indicate XbaI sites used for 3C experiments. Given the resolution of 3C experiments, the anchor points for the two loops, which are also indicated, can be located anywhere within each of the ligated restriction fragments (green arrow). Bottom, quantification of LoopEFR (blue rings) and LoopXl-k (gray rings) by 3C–qPCR in Col-0 wild-type, eair-1, ddc and dcl234 plants. f, Expression of EFR (blue rings) and Xl-k (gray rings) as measured by quantitative PCR with reverse transcription (RT–qPCR) and normalized against ACT2. In e,f, values of n = 3 biologically independent mutant samples are expressed relative to the average values of Col-0 wild-type plants. Data are presented as the mean values ± s.d. P values were calculated with a two-tailed unpaired t-test with Welch’s correction.
IRs can produce 24-nt-long siRNAs that trigger DNA methylation at the IR that generates them, leading to changes in chromatin organization and the expression of adjacent genes33,34. Abundant 24-nt but not 21-nt siRNAs were detected at Ea-IR (Fig. 1b and Supplementary Fig. 1b), as was methylation in all cytosine contexts (Fig. 1c and Supplementary Fig. 1c). CHH and CHG methylation and siRNAs were significantly reduced in drm1 drm2 cmt3 (ddc) and dcl2 dcl3 dcl4 (dcl234) mutants, indicative of involvement of the RdDM pathway (Fig. 1b,c). Abundant 24-nt siRNAs and DNA methylation were also found at a 2,060-bp LIMPET1 DNA and MuDR transposon located 6,620 bp upstream of EFR (Fig. 1a and Supplementary Fig. 1c–e). Given the non-IR nature of LIMPET1 and its proximity to Ea-IR, it served as a valuable control to discern specific IR-triggered molecular effects in subsequent analyses.
To determine whether Ea-IR affected local chromatin organization, we analyzed our published CaptureC data34, which revealed an SRCL that included the entire EFR-encoding region (Fig. 1d, blue interaction). This loop, termed LoopEFR, appeared to be disrupted in ddc and dcl234 mutants, suggesting that its formation relies on Ea-IR methylation (Fig. 1d). In contrast, a second loop encompassing the XI-k gene (LoopXI-k) was only detected in the RdDM-deficient mutants (Fig. 1d, green interactions). These findings agree with a role of siRNA-induced DNA methylation in forming and stabilizing specific SRCLs.
To study LoopEFR and LoopXI-k and their dependence on Ea-IR in detail, we conducted chromatin conformation capture followed by quantitative PCR (3C–qPCR) in Col-0 wild-type, ddc and dcl234 plants and a transfer DNA (T-DNA) insertion line (WiscDsLox440E09), hereafter referred to as eair-1, which contains a large insertion in the middle of Ea-IR (Fig. 1a and Supplementary Fig. 1f). These experiments demonstrated that the chromatin loop encompassing the EFR gene in wild-type plants was dependent on an intact and methylated Ea-IR (Fig. 1e). Collectively, our data reinforce the idea of LoopEFR formation triggered by Ea-IR in combination with the RdDM pathway (Fig. 1e). The observation that the formation of LoopXI-k was enhanced in eair-1 and RdDM-deficient mutants, which lack LoopEFR, implies competition between the two loops, with LoopEFR being dominant (Fig. 1e). We repeated the 3C–qPCR analyses with different sets of restriction enzymes to more precisely delineate the interacting regions of LoopEFR and LoopXI-k. These experiments confirmed that LoopXI-k only forms in the absence of LoopEFR (Supplementary Fig. 2).
Lastly, to assess whether changes in chromatin organization caused by physical disruption of Ea-IR in eair-1 mutants or by demethylation of Ea-IR in RdDM-deficient mutants affected the expression of nearby genes, we measured EFR and XI-k mRNA accumulation. EFR but not XI-k was more expressed in the mutants tested (Fig. 1f), defining LoopEFR as a repressive loop. That XI-k expression was unaltered, indicating that LoopEFR effects on EFR are specific, as Ea-IR is similarly close to both EFR and Xl-k. A similar regulatory event was reported for the PID1 locus, where an intergenic chromatin loop controlled by the long noncoding RNA (lncRNA) APOLO modulates the gene expression in response to auxins47. Similar to the lncRNA MARS, which controls chromatin organization of the marneral cluster in response to ABA, APOLO’s effects on chromatin three-dimensional organization is attributed to its interaction with LIKE HETEROCHROMATIN PROTEIN 1 (LHP1), which in turn interacts with methylated regions at the locus47,48. Unlike these cases, when analyzing publicly available chromatin immunoprecipitation followed by sequencing (ChIP–seq) datasets of histone modification and chromatin-associated factors49, we did not observe any clear signal of specific LHP1 binding in the EFR locus (Supplementary Fig. 3). Identifying the factors allowing loop formation is a compelling question that remains to be addressed.
EFR chromatin loops are sensitive to activation of immunity
It has been reported that, upon pathogen infection, many regions in the Arabidopsis genome experience a reduction in DNA methylation levels, which is often coupled to increased production of 21-nt siRNAs50,51. Pol-V-defective but not Pol-IV-defective mutants showed enhanced disease resistance toward the bacterial pathogen Pseudomonas syringae52, suggesting that a noncanonical RdDM pathway may be involved. Conversely, reduced resistance to this pathogen was observed in mutants lacking the important cytosine demethylase ROS1 (ref. 53), supporting a role for DNA methylation during Pst response. In accordance with this, we observed that DNA methylation at the Ea-IR is reduced, especially CHH and CHG methylation, upon Pseudomonas syringae pv. tomato strain DC3000 hrcC (Pst DC3000 hrcC), a type III secretion-system-deficient strain that facilitates the detection of PTI-related phenotypes independently of the activation of ETI54 (Fig. 2a and Supplementary Fig. 4a). This reduction was not observed for LIMPET1, showing some degree of specificity over the Ea-IR (Supplementary Fig. 4b). Interestingly, the methylation of some specific cytosines did not appear to be significantly reduced (Fig. 2a). This may have been caused by trans-regulation by other copies of the Ea-IR within the Arabidopsis genome. Infected plants showed elevated levels of siRNAs, especially 24-nt siRNAs (Supplementary Fig. 4c). We posited that this pathogen-induced alteration of the epigenetic landscape could modify chromatin organization and the expression of the EFR locus through effects on Ea-IR. To investigate this hypothesis, we infected Col-0 plants using Pst DC3000 hrcC. Infection with Pst DC3000 hrcC induced the expression of EFR and the defense-associated marker PATHOGENESIS-RELATED GENE 1 (PR1) but not XI-k (Fig. 2b).
a, Cytosine methylation in the CG, CHG and CG contexts at Ea-IR (demarcated by vertical black lines) and flanking regions (700 bp) in Col-0, dcl234 and ddc mutants in mock-treated plants and 1 day after infection with Pst DC3000 hrcC. The average of three individual BS-seq replicates are plotted for each genotype. b,c, Relative expression of EFR, Xl-k and PR1 in Col-0 wild-type (b), and dcl234, ddc and eair-1 (c) plants infected with Pst DC3000 hrcC as measured by RT–qPCR and normalized against ACT2. Samples were collected at different time points after infection (days or hours after infection). RNA levels are expressed relative to the mock-treated (b) or infected (c) Col-0 plants. Values are from n = 3 biologically independent samples. HAI, hours after infection; DAI, days after infection. d,e, Quantification of the formation of LoopEFR and LoopXl-k as measured by 3C–qPCR analysis in Col-0 wild-type, eair-1, ddc and dcl234 plants mock-treated or infected for 1 day with Pst DC3000 hrcC. Values of n = 4 (d) or n = 3 (e) biologically independent samples are expressed relative to the mock-treated samples. In all plots, data are presented as the mean values ± s.d. P values were calculated with a two-tailed unpaired t-test with Welch’s correction.
Induction of EFR was less pronounced in dcl234, ddc or eair-1 mutants, indicating that the RdDM pathway and Ea-IR are both required for EFR upregulation after pathogen infection (Fig. 2c). We, therefore, assessed changes in local chromatin topology using 3C–qPCR assays, which revealed that infection of Col-0 plants with Pst DC3000 hrcC led to the opening of the LoopEFR (Fig. 2d), in support of a repressive role of the closed loop. In addition, the opening of LoopEFR correlated with the establishment of LoopXI-k (Fig. 2d), suggesting again that the formation of LoopXI-k is contingent upon the opening of LoopEFR. No significant changes in chromatin topology were observed in eair-1 mutants, as could be expected from our result given the lack of an intact Ea-IR (Fig. 2d). In dcl234 and ddc mutants, LoopEFR, which was predominantly open in mock conditions, adopted an even more relaxed configuration upon infection (Fig. 2e). This may reflect a drop in the remaining CHH and CHG DNA methylation observed in these mutants (Fig. 1c). The high basal levels of EFR in all tested mutants and their lower induction rate suggest that these plants may have a primed EFR response, which correlates with the status of LoopEFR.
Ea-IR lack reshapes chromatin enhancing pathogen resistance
Ea-IR corresponding to a transposon insertion raised the possibility that it might be polymorphic in Arabidopsis populations and influence pathogen responses in natural accessions. We first consulted published transposon insertion polymorphism data based on short sequencing reads55, which revealed that Ea-IR was partially or completely absent in nearly three quarters of the 216 accessions analyzed (Fig. 3a). Because there are trade-offs in sensitivity and precision when calling transposon insertion polymorphisms from short reads, we compared the EFR and Xl-k region in assembled long-read genomes of 149 accessions56,57. Similarly, we found that ~70% of these accessions lacked Ea-IR (Fig. 3b). The absence of Ea-IR typically coincided with the absence of LIMPET1, about 6 kb away, but the EFR-encoding sequence between Ea-IR and LIMPET1 was present in all accessions (Fig. 3b). Recombinants in which only Ea-IR or LIMPET1 was found were rare.
a, Ea-IR variation, as deduced from a published short-read based variant atlas in 216 natural accessions from the 1001 Genomes Project. b, Multiple=sequence alignment of EFR and Xl-k region in genomes assembled from long reads. Each line represents one genome. Single-nucleotide polymorphisms are indicated by short, vertical colored lines, relatively small deletions are indicated as horizontal black lines and large deletions are indicated as white spaces. LIMPET1 and Ea-IR are highlighted with blue and orange lines. The genomes are ordered according to the presence or absence of the Ea-IR as mapped to the Col-0 reference genome. c, Expression of EFR and XI-k in 216 accessions, as measured by RNA-seq from85. Each dot represents the normalized expression value for accessions with or without Ea-IR. P values were calculated with a two-tailed unpaired t-test with Welch’s correction. d, Chromatin contacts from CaptureC in the EFR and Xl-k region in Col-0, Ba-1 and Hod accessions. LoopEFR, which includes Ea-IR, is marked in cyan, LoopXl-k is marked in green and all other interactions are marked in purple. The region targeted by CaptureC is indicated, with all potential restriction fragments below. Fragments diagnostic for the Ea-IR region are indicated in cyan. e, The 24-nt siRNAs mapping uniquely over Ea-IR in Col-0, Ba-1 and Hod accessions. Three individual sRNA-seq replicates are plotted separately and shown with different color tones. f, Left, cytosine methylation in the CG, CHG and CG contexts at Ea-IR (demarcated by vertical black lines) and flanking regions (700 bp) in Col-0, Br-0 and Hod accessions. The average of three individual BS-seq replicates is plotted for each genotype. Data are presented as the mean values ± s.d. Right, the average percentage of cytosine methylation in the CG, CHG and CG for each region and the same genotypes. The absence of the Ea-IR in Hod and Br-0 rendered no informative reads (not determined, N/D) over the IR-encoding region for these specific genotypes.
There was a clear association between the absence of Ea-IR and elevated EFR expression levels in most accessions, while XI-k expression was unaffected (Fig. 3c). This result again highlights the dependency of EFR expression, as opposed to XI-k, on Ea-IR. Likewise, when comparing our CaptureC data between two accessions with Ea-IR (Col-0 and Ba-1) and one lacking Ea-IR (Hod), we observed that LoopEFR was only formed in Col-0 and Ba-1 but not in Hod (Fig. 3d). As anticipated, siRNAs and cytosine methylation associated with Ea-IR in Col-0 and Ba-1 were absent in Hod and Br-0, which also lack the IR (Fig. 3e,f and Supplementary Fig. 5). Importantly, we observed no signs of CHH or CHG methylation in the flanking regions of these accessions, showing that the epigenetic modifications causing the regulatory effects were restricted to and triggered by the Ea-IR region when present. These data support the idea that Ea-IR has a critical role in chromatin looping and EFR regulation.
To further assess the influence of structural variation at Ea-IR on LoopEFR formation and EFR expression, we first confirmed the presence of Ea-IR in the selected accessions using PCR (Supplementary Figs. 1f and 2). We then profiled and quantified the formation of LoopEFR and LoopXI-k using 3C–qPCR, finding that the presence of Ea-IR is essential for establishing LoopEFR but dispensable for LoopXI-k, formed by a region that was present in all accessions (Fig. 4a, mock-treated samples). Nevertheless, as previously observed, the absence of LoopEFR in Hod and Br-0 caused by the lack of Ea-IR enhanced the formation of LoopXI-k in these plants (Fig. 4a, mock-treated samples). The accessions with Ea-IR had markedly lower EFR expression levels, in line with the hypothesized repressive role of LoopEFR (Fig. 4b, mock-treated samples). By contrast, the levels of XI-k, which we determined to be independent of Ea-IR in Col-0, were similar among accessions (Fig. 4b, mock-treated samples).
a, Quantification of LoopEFR (left) and LoopXl-k (right) as measured by 3C–qPCR in Col-0, Ba-1, Hod and Br-0 accessions infected with Pst DC3000 hrcC (1 day after infection). Values of n = 4 biologically independent samples are expressed relative to the mock-treated samples. Data are presented as the mean values ± s.d. P values were calculated with a two-tailed unpaired t-test with Welch’s correction. b, EFR (left) and Xl-k (right) expression in the same samples as in a. RNA levels are expressed relative to the mock-treated Col-0 wild-type plants. Values of n = 3 biologically independent samples. Data are presented as the mean values ± s.d. P values were calculated with a two-tailed unpaired t-test with Welch’s correction. c, Diagram of Ea-IR controlled chromatin dynamics in different backgrounds. The EFR gene model is yellow, the Xl-k model is blue and Ea-IR is indicated by green arrows. Red pins represent cytosine methylation. Expression status of EFR (on or off) is indicated in green. Left, before infection. Right, after infection. All backgrounds converge on a similar chromatin organization. d, Growth of Pst DC3000 hrcC in accessions or mutants, expressed as log10 CFU per cm2 of infected tissue 1 day after infection. Values of n = 4 biologically independent samples. Data are presented as the mean values ± s.d. P values were calculated with a two-tailed unpaired t-test with Welch’s correction. e, Autofluorescence of Arabidopsis leaves, indicative of cell death, after treatment with elf18 peptide in mutants or accessions with Ea-IR (Col-0 and Ba-1) or without it (Hod and Br-0). f, PR1 (left) and EFR (right) expression in plants treated for 1 day with mock solution or 1 μM elf18 or flg22. RNA levels are expressed relative to the mock-treated Col-0 wild-type plants. Values of n = 3 biologically independent samples. Data are presented as the mean values ± s.d. P values were calculated with a two-tailed unpaired t-test with Welch’s correction. g, Quantification of callose deposition spots in aniline-blue-stained leaved after 1 day of treatment with mock solution or 1 μM elf18 (left) or flg22 (right). Spots were counted using an automated ImageJ script. Values of n = 10 independent leaves. P values were calculated with a two-tailed unpaired t-test with Welch’s correction.
Our next inquiry was whether structural variation at Ea-IR alters the effects of Pst DC3000 hrcC infection on chromatin organization and EFR expression. Like Col-0, infection led to the opening of LoopEFR and the concurrent emergence of LoopXI-k in Ba-1 (Fig. 4a, 1 day after infection). Unlike the situation in Col-0 and Ba-1, infection affected neither LoopEFR nor LoopXI-k in the accessions lacking Ea-IR (Fig. 4a, 1 day after infection). Concordantly, Hod and Br-0 had substantially higher levels of EFR expression even before infection. Equivalent levels were reached in the accessions with Ea-IR only late during infection (Fig. 4b), when both chromatin organization and EFR expression levels had become very similar in all tested accessions (Fig. 4a,b). The role of Ea-IR can, thus, be interpreted as reducing EFR expression in unchallenged plants. Ea-IR being relatively widespread in natural Arabidopsis populations suggests that there is a considerable selective advantage of reduced EFR expression in the absence of infection, at least in some environments (summarized in Fig. 4c).
To test whether the differences in chromatin and EFR expression dynamics translate into varying pathogen responses, we challenged the accessions with Pst DC3000 hrcC and monitored bacterial growth. As the analyzed accessions present many differences in addition to the presence of Ea-IR, we also included eair-1, ddc and dcl234 mutants in the analysis to reinforce the causality of any effect to the absence of Ea-IR. In line with the higher basal levels of EFR, the mutants and the accessions without Ea-IR, Hod and Br-0, supported less pathogen growth than accessions with Ea-IR (Fig. 4d). To directly test the levels of EFR, we treated the four accessions, along with ddc, dcl234 and eair-1 mutants, with 10 μM elf18, an EFR ligand, and measured autofluorescence as an indicator of cell death after 5 days. Signs of cell death were much more prominent in the accessions without Ea-IR, as well as in all three mutants, indicating that the increased basal levels of EFR expression are protective and confirming that the higher EFR expression levels translate into more active receptors (Fig. 4e). This result was not paralleled when the same plants were treated with 10 μM flg22, which is not an EFR ligand, showing that the effect is specific to the deregulation of EFR (Supplementary Fig. 6a). We also quantified EFR and PR1 gene expression and the callose deposition after elf18 and flg22 treatment. Induction of PR1 after elf18 and flg22 treatments confirmed that both peptides are perceived by the plants (Fig. 4f), discarding a deficient treatment as the cause for the differences in the observed autofluorescence. As would be expected by a constitutively primed EF-TU-induced defense, the three genotypes missing Ea-IR (Hod, Br-0 and eair-1) showed a stronger response to elf18 than to flg22, contrary to the other genotypes that showed an even response to both peptides (Fig. 4f). We observed higher basal EFR levels and increased callose deposition after elf18 but not flg22 treatment in Hod, Br-0, dcl234, ddc and eair-1, compared to the genotypes with the Ea-IR (Fig. 4f,g and Supplementary Fig. 6b), supporting the idea of a primed EFR response in the plants missing the Ea-IR. Although there are many differences between tested accessions in addition to the presence or absence of Ea-IR and EFR expression, the matching results observed with the eair-1 mutants strongly support the role of this element and its polymorphisms as a relevant factor during defense against Pst.
Pol II readthrough generates a long XI-k::Ea-IR transcript
Pol II commonly transcribes regulatory IRs within 500 bp of genes, generating 24-nt siRNAs through a noncanonical DCL3-dependent pathway34,38. There are at least eight other copies of Ea-IR like sequences across the Col-0 genome (Supplementary Fig. 7a), all near protein-coding genes, with most being found in other accessions (Supplementary Fig. 7b,c). However, there are uniquely mapping siRNAs that come from the Ea-IR locus between EFR and Xl-k, indicating that this element is indeed transcribed (Supplementary Fig. 7d).
To understand how Ea-IR is transcribed, we analyzed publicly available Pol IV and Pol V ChIP–seq data58 alongside Pol II native elongating transcript sequencing (NET-seq) data59. As with other transposons, the ChIP–seq data indicated that Pol IV and Pol V likely contribute to Ea-IR transcription (Fig. 5a) but the NET-seq data revealed that Pol II can also transcribe this region. The specific distribution of NET-seq reads suggested that Pol II-dependent Ea-IR transcripts also contain XI-k mRNA sequences (Fig. 5a). In animals, specific stressful conditions can trigger Pol II to read through termination sites, producing transcripts that include sequences from nearby transposons60,61,62. Similarly, we observed Pol II readthrough at the termination sites of XI-k, leading to the incorporation of Ea-IR sequences into Xl-k transcripts (Fig. 5a). Specific primers allowed us to identify and quantify the extended XI-k transcript that includes Ea-IR sequences (‘XI-k long transcript’, Supplementary Fig. 8a). Long XI-k transcript levels increased substantially upon pathogen infection (Fig. 5b,c). This observation suggests that LoopEFR, which opens after infection, impedes Pol II readthrough. To test this hypothesis, we measured the levels of the long XI-k transcript in ddc and dcl234 mutants with a constitutively open LoopEFR. Long XI-k transcript levels were significantly increased in both mutants (Fig. 5d).
a, Integrated Genomics Viewer views of Pol IV and Pol V ChIP–seq58 and Pol II nascent elongating transcripts86 in untreated Col-0 wild-type plants. The scale is normalized to total mapped reads. Note that Pol II signal extends beyond the Xl-k termination site, indicated by a vertical orange line. b–d, Expression of the canonical and the long Xl-k transcript isoforms, as measured by RT–qPCR in untreated Col-0 wild-type plants (b), in accessions with Ea-IR (Col-0 and Ba-1) after different time points of infection with Pst DC3000 hrcC (c) and in Col-0, ddc and dcl234 mutants (d). Values of n = 4 (b) and n = 3 (c,d) biologically independent samples are expressed relative to the Col-0 plants. Data are presented as the mean values ± s.d. P values were calculated with a two-tailed unpaired t-test with Welch’s correction.
Long Xl-k transcripts are processed to produce Ea-IR siRNAs
RNA hairpin structures, as in the long Xl-k transcript, can affect mRNA transcription and translation but also serve as substrates for processing into sRNAs mediated by Dicer and Drosha63,64,65,66,67. Consistent with this last notion, we found that the abundance of the long XI-k transcript was much higher in dcl1-100 mutants (Fig. 6a), indicating that this transcript with Ea-IR sequences is subject to DCL1-dependent processing and degradation. To test this directly, we used Nicotiana benthamiana to transiently express chimeric transcripts that included GFP-encoding sequences and Ea-IR sequences. Inserting Ea-IR downstream of GFP but not upstream of GFP reduced GFP fluorescence, suggesting either transcript degradation or lack of translation (Fig. 6b). The simultaneous introduction of an artificial miRNA silencing DCL1 or its HYL1 cofactor restored GFP fluorescence levels to those of controls, confirming that the reduction in GFP signal was indeed DCL1 dependent (Fig. 6b). This suggests that DCL1 can process the long XI-k transcript, likely generating epigenetically active 21-nt siRNAs. In this sense, 21-nt siRNAs derived from Ea-IR were also produced in RDR6 mutants and in Pol IV mutants as detected in publicly available sRNA-seq experiments68 (Fig. 6c). Their independence of the RDR6–DCL4 pathway differentiates them from the 21-nt siRNAs originating from nearby LIMPET1 (Supplementary Fig. 8b), which behaves in this respect similarly to other canonical transposons, where the dsRNA substrate for DCL processing into 21-nt siRNAs requires RDR6 activity41. Levels of 21-nt Ea-IR-derived siRNAs were particularly high in dcl3 mutants, with a concurrent decrease in 24-nt siRNAs (Fig. 6c), suggesting that processing of Ea-IR transcripts involves competition between DCL1 and DCL3. We introduced the GFP–Ea-IR and Ea-IR–GFP constructs into dcl234 mutants and found that the GFP fluorescence of both was much higher in the mutant plants, indicating that DCLs other than DCL1 (most likely DCL3 as suggested by the drop in 24-nt sRNAs) contribute to processing of the long XI-k transcript containing Ea-IR sequences (Fig. 6d).
a, Expression of the canonical and the long Xl-k transcript isoform in untreated Col-0 wild-type, hyl1-2 and dcl1-100 plants. Values of n = 4 biologically independent samples are expressed relative to the Col-0 plants. Data are presented as the mean values ± s.d. P values were calculated with a two-tailed unpaired t-test with Welch’s correction. b, Left, confocal microscopy of N. benthamiana leaves expressing constructs with the Ea-IR sequence inserted upstream or downstream of the GFP-encoding sequence (Ea-IR–GFP and GFP–Ea-IR). Leaves were coinfiltrated with an empty vector or artificial miRNAs targeting HYL1 and DCL1. Scale bars, 50 μm. Right, quantification of fluorescence intensity of leaves expressing from the left panel. Data are presented as the mean values ± s.d. of ten independent leaves. P values were calculated with a two-tailed unpaired t-test with Welch’s correction. c, The 21-nt and 24-nt sRNAs68 mapping to the Ea-IR in Col-0 wild-type, and rdr6, dcl3, pol IV and pol IV rdr6 plants; cpm, counts per million. d, Confocal microscopy images and quantification of GFP–Ea-IR fluorescence in transgenic Arabidopsis wild-type and dcl234 plants. Transgenic plants expressing Ea-IR–GFP were included as a positive control. Scale bars, 50 μm. Data are presented as the mean values ± s.d. of ten independent leaves. P values were calculated with a two-tailed unpaired t-test with Welch’s correction. e, The 21-nt and 24-nt sRNAs69 mapping to the Ea-IR in Col-0 wild-type samples 6 or 14 h after infection with a virulent strain (Pst DC3000), a type III secretion-system-deficient strain (Pst DC3000 hrcC) or an avirulent strain (Pst DC3000 avrRpt2).
So far, our findings show that pathogen infection causes Pol II readthrough at XI-k and generation of transcripts that contain Ea-IR sequences, which can undergo DCL1-mediated and DCL3-mediated processing into siRNAs. Exploring publicly available sRNA-seq experiments69, we confirmed that pathogen challenge causes an increase in 21-nt Ea-IR-derived siRNAs as early as 14 h after infection (Fig. 6e). This induction is specific and not seen at the nearby LIMPET1 (Supplementary Fig. 8c). This observation agrees with the opening of LoopEFR during infection, the increase in the long XI-k transcript and the DCL1-dependent processing of this transcript isoform into siRNAs. All tested P. syringae strains (Pst DC3000, Pst DC3000 hrcC and Pst DC3000 avrRpt2) induced the production of Ea-IR-derived 21-nt siRNAs (Fig. 6e), suggesting that the signal leading to the production of the long Xl-k transcript isoform is likely a PAMP signal. The analysis also revealed increased 24-nt sRNAs upon pathogen infection with these strains (Fig. 6e), confirming that Ea-IR-derived 24-nt sRNAs are not exclusively produced in response to Pst DC3000 hrcC, used for most experiments. Nevertheless, we cannot be sure whether the detected 24-nt sRNAs are those produced after DCL3 processing of the long Xl-k transcript or after a canonical DCL3 processing of Pol IV Ea-IR short transcripts. Conversely, DCL1-processed 21-nt sRNAs, which are independent of the Pol IV transcript, give us more precise information about the processing of long Xl-k transcript upon infection.
The RdDM pathway can be initiated by 21-nt siRNAs by facilitating the recruitment of Pol IV and Pol V (refs. 41,42,70,71,72). Altogether, our data may suggest that the DCL1 processing of the long Xl-k transcript that occurs upon pathogen infection may lead to an initial burst of IR methylation, triggered by 21-nt siRNAs, that help to buffer the EFR response by reestablishing LoopEFR (Fig. 7). However, our data are not sufficient to undoubtedly define the production order in which each siRNA species (long Xl-k DCL1-dependent 21-nt sRNAs, long Xl-k DCL3-dependent 24-nt sRNAs and Pol-IV-dependent DCL3-dependent Ea-IR 24-nt sRNAs) is produced upon infection, making the proposed model hypothetical and requiring it to be subjected to further investigation.
In a resting state, Arabidopsis plants with the Ea-IR insertion between EFR and Xl-k genes from the repressive LoopEFR chromatin loop, which encompasses EFR-encoding sequences. This loop dampens the background activity of the immune system and also promotes the termination of Xl-k transcription. Upon pathogen infection, the repressive LoopEFR opens, allowing for enhanced EFR transcription, which increases immunity and allows Pol II readthrough at the Xl-k termination site. The hairpin at the 3′ end of the long Xl-k transcript isoform is processed by DCL1 and DCL3 into epigenetically active siRNAs. These siRNAs can subsequently initiate the remethylation of Ea-IR sequences, promoting the reformation of LoopEFR and thereby resetting chromatin topology to a homeostatic state with a steady but reduced immune response. Open LoopEFR or the absence of Ea-IR allows for the formation of a second, apparently innocuous LoopXl-k.
Discussion
Plants are constantly exposed to a diverse range of pathogens. Swift activation of defense responses is vital for plant survival but it must be balanced with the risk of an overactive immune system, which can negatively affect growth and yield73,74. Conversely, an inadequate immune response leaves plants vulnerable to pathogens24,25,73,75. Our study uncovers a potential regulatory mechanism for the activation and deactivation of immune responses, with epigenetic control of the expression of the EFR immune receptor gene. We discovered that a transposon-derived IR, Ea-IR, located downstream of EFR, orchestrates the formation of a repressive chromatin loop that suppresses EFR transcription in plants not yet exposed to pathogens. That such a dampened immune state is often not advantageous can be inferred from the finding that Ea-IR is found in only a subset of Arabidopsis natural accessions. Plants without Ea-IR had higher basal activity of the immune system, which is likely beneficial in a pathogen-rich environment.
Activation of EFR requires the coreceptor BRI1-ASSOCIATED RECEPTOR KINASE 1 (BAK1). Because BAK1 is also involved in brassinosteroid signaling2,13, competition for this rate-limiting coreceptor contributes to defense–growth trade-offs25,76. Importantly, Ea-IR not only limits EFR expression before acute exposure to pathogens but may also help to dampen EFR expression after its induction. During infection, the open LoopEFR state not only leads to increased EFR expression but also allows Pol II readthrough at the neighboring Xl-k gene. The long Xl-k transcript isoform generated by readthrough includes Ea-IR sequences, which can be processed by DCL1 or DCL3 into epigenetically active siRNAs. These siRNAs can then trigger the methylation of Ea-IR, restoring LoopEFR and EFR repression (Fig. 7). Nevertheless, although this model is the most parsimonious interpretation of our data and the current knowledge about the RdDM initiation and maintenance pathway39,40, further investigation is required to confirm this proposed buffering system.
As a consequence of the high level of conservation of EF-Tu protein sequences across bacteria, increased levels of EFR confer resistance against a wide range of bacterial pathogens77,78. Conversely, transgenic plants expressing EFR do not show enhanced resistance to the fungal pathogen Verticillium dahliae77. Although the mechanism described was tested exclusively in response to Pst and is likely not widespread to all pathogens, it could be expected that the elevated basal levels of EFR in plants without Ea-IR enhance defense against other pathogens as well. In this sense, the constitutively higher levels of EFR in accessions missing Ea-IR could produce a primed state of the immune system through the constant interaction of plants with nonpathogenic bacteria still recognized by EFR. While EFR-induced defense may be constitutively higher in plants missing Ea-IR, a TE-induced priming of the defense through derepression of the EFR locus is possible in those accessions with the IR present. TE-induced priming of the immune response has been previously reported79,80,81,82. Whether epigenetic changes in chromatin organization at EFR locus could lead to a primed state remains to be explored.
Epigenetically active 21-nt siRNAs, especially when loaded in AGO6, can initiate RdDM42,83. While canonical RdDM pathways, led by DCL3-dependent 24-nt siRNAs, may initiate Ea-IR DNA methylation after infection, DCL1-dependent 21-nt sRNAs, produced from the long Xl-k transcript, might also participate in this process. The association of Pol IV and Pol V with the Ea-IR-encoding region suggests that the long-term maintenance of DNA methylation and, thus, LoopEFR follows a more canonical RdDM pathway. Nevertheless, the presence of additional copies of the Ea-IR in the Arabidopsis genome opens the possibility of regulatory crosstalk between them, causing the initiation or maintenance of DNA methylation. Whether the structural nature of the Ea-IR transcript is sufficient to trigger the reported regulatory mechanism independently of the specific Ea-IR sequence or whether this IR has special properties in terms of recruiting specific elements of the DNA methylation machinery remains to be explored. However, the first scenario appears more likely, as we found that many IRs, regardless of their sequence, promote chromatin reorganization when expressed34. Nevertheless, it is possible that some sequence-specific DNA methyl readers can recognize methylated Ea-IR and fine-tune its regulatory function. The fact that only the expression of EFR but not Xl-k is regulated by Ea-IR argues against a repressor binding the IR to directly regulate the locus.
Transcriptional readthrough might have an important role in small genomes like that of Arabidopsis where genes are closely spaced. Our data suggest that LoopEFR formation inhibits readthrough, probably serving as a barrier to Pol II elongation and promoting transcriptional termination. This feature of LoopEFR unveils a new mechanism of action for SRCL.
Our study uncovers a complex epigenetic regulatory mechanism where IR-derived siRNAs modulate chromatin topology and EFR expression to shape the dynamics of plant immune responses. Ea-IR is a fascinating example of a genetic allele that acts through epigenetic changes84, with the modulation of loop formation providing an elegant mechanism for rapid transcriptional responses to pathogens. It also shows that cis-regulatory variation is not restricted to differences in the recruitment of DNA-binding transcription factors. This work shows how the evolutionary loss of specific IRs can determine a plant’s ability to mount an effective immune response, highlighting the importance of chromatin organization in shaping adaptive traits. Understanding these processes also opens avenues for improving the stress and pathogen resistance of crops through manipulating IR elements without altering coding sequences.
Methods
Plant material and growth conditions
Arabidopsis thaliana and N. benthamiana plants were cultivated in soil for 3–4 weeks under long-day conditions (16 h of light and 8 h of darkness). A. thaliana accessions Col-0, Hod, Ba-1 and Br-0 were used. T-DNA insertion mutants were dcl2 dcl3 dcl4 (dcl234)87, drm1 drm2 cmt3 (ddc)88, eair-1 (WiscDsLox440E09), hyl1-2 (SALK_064863) and dcl1-100 (GABI_098F10). Transgenic plants contained the constructs 35S::GFP–Ea-IR and 35S::Ea-IR–GFP, for which an Ea-IR fragment was amplified using Phusion polymerase (Thermo Fisher Scientific), then cloned into pEntr/D-TOPO (Thermo Fisher Scientific) and subsequently recombined using LR Clonase (Thermo Fisher Scientific) into Gateway-compatible pGreen destination vectors (pCB012 and pCB011, respectively). A. thaliana plants were transformed using the floral dip method.
Transient expression in N. benthamiana
For transient expression in N. benthamiana, Agrobacterium tumefaciens GV3101 carrying GFP–Ea-IR, Ea-IR–GFP, amiR:HYL1 or amiR:DCL1 constructs was cultured for 20 h from frozen stocks on Luria–Bertani medium supplemented with antibiotics. The bacterial cultures were then resuspended in infiltration medium (containing 10 mM MgCl2, 10 mM MES pH 5.7 and 150 μM acetosyringone) at an optical density at 600 nm (OD600) of 0.5 and incubated for 3 h at room temperature. Subsequently, plants were infiltrated using equal parts of these bacterial suspensions. Samples were analyzed after 2 days. Artificial miRNA constructs were designed following the methodology outlined in a previous study89.
Comparison of EFR and Xl-k region in Arabidopsis genomes
A set of 149 Arabidopsis thaliana accessions genomes sequenced with the PacBio technology56,57 were analyzed. The obtained long reads were mapped to the The Arabidopsis Information Resource version 10 (TAIR10) genome sequence with minimap2 (version 2.24)90 using the ‘-ax asm5’ parameter. The alignments on the EFR and XI-k region were then inspected with Integrative Genomics Viewer91. The expression of EFR and XI-k in 216 accessions is expressed as the normalized gene expression level from datasets retrieved from the NCBI Gene Expression Omnibus (GEO) under accession number GSE80744 (file name: GSE80744_ath1001_tx_norm_2016-04-21-UQ_gNorm_normCounts_k4.tsv). Accessions included correspond to those analyzed in a previous study55.
Detection of Ea-IR presence and T-DNA insertions
IR presence and T-DNA insertions were assessed through PCR using DreamTaq polymerase (Thermo Fisher Scientific) with genomic DNA. For genomic DNA extraction, 100 mg of fresh plant material was ground in 700 μl of extraction buffer (comprising 200 mM Tris-HCl pH 8, 250 mM EDTA and 0.5% SDS), followed by precipitation with isopropanol. PCR was conducted using primers as detailed in Supplementary Table 1. Molecular weight marker GeneRuler 1 kb plus (Thermo Fisher Scientific) was used to infer band size.
3C assay and RT–qPCR
The 3C assay was conducted following a published procedure33. Approximately 2 g of tissue was collected per sample and crosslinked with 1% formaldehyde. Crosslinking was stopped by adding glycine to a final concentration of 0.125 M, after which nuclei were isolated.
To detect loops 1–8, including LoopEFR and LoopXI-k, DNA was digested with XbaI (Thermo Fisher Scientific). For loops 9 and 10, Bsp1407I (Thermo Fisher Scientific) was used; for loop 11, EcoRI (Thermo Fisher Scientific) was used. DNA fragments were ligated with 100 U of highly concentrated T4 DNA ligase (Thermo Fisher Scientific) at 22 °C for 5 h in a 2-ml volume. Subsequent steps included reverse crosslinking, proteinase K treatment (Qiagen) and DNA purification with the phenol–chloroform. To measure interaction frequency, qPCR was performed and results were analyzed with the 2−ΔΔCt formula using ACT2 as a control. Oligonucleotide primers are given in Supplementary Table 1. The final qPCR products were subjected to agarose gel electrophoresis (2%) and bands corresponding to the expected size were purified and sequenced using the Sanger method to confirm ligation products.
For RT–qPCR, 500 ng of total RNA underwent RT using the RevertAid RT kit (Thermo Fisher Scientific) after treatment with DNase (Thermo Fisher Scientific). Subsequently, qPCR was conducted using SYBR green (Thermo Fisher Scientific Maxima SYBR Green qPCR Master Mix (2×)). Relative expression levels were determined using the 2−ΔΔCt method with ACT2 as the housekeeping gene control. Oligonucleotide primers are given in Supplementary Table 1.
Plant inoculation and pathogen growth assays
To assess disease phenotypes, gene expression and chromatin conformation, plants were subjected to syringe infiltration with Pst DC3000 hrcC following the method described in a previous study92. For this purpose, bacteria were streaked from frozen stocks onto King’s B medium supplemented with rifampicin. After transferring the bacteria into liquid King’s B medium, they were grown at 28 °C for 4 h and then resuspended at an OD600 of 0.002 in sterile water. Fully developed leaves were infiltrated on the abaxial side using a needleless syringe. To determine bacterial growth, samples were collected at various time points. Leaf discs were extracted from eight distinct leaves on separate plants, divided into four replicates, ground, resuspended in 1 ml of sterile water, diluted and plated on agar-King’s B plates supplemented with rifampicin for colony-forming unit (CFU) counting. Alternatively, in assays involving ddc plants, samples were taken by weight.
For elf18 (Proteogenix) and flg22 treatments, plant leaves were infiltrated on the abaxial side using a needleless syringe with a 10 μM peptide solution in sterile water. After 5 days, cell death around the infiltration area was observed using an ultraviolet transilluminator. For callose deposition assays, mature leaves were infiltrated with 1 μM elf18 or flg22 and stained after 1 day with 0.01% (w/v) aniline blue in 67 mM K2HPO4 (pH 12), following the method outlined in a previous study93. Callose deposition calculations were conducted using Fiji software, as described in a previous study93. Additionally, samples were collected for total RNA extraction.
Confocal microscopy
Samples were excited at 480 nm to detect GFP and emission was collected within 503–531 nm. All assays were conducted using a TCS SP8 confocal microscope and software (Leica). Relative fluorescence units (RFU) were calculated using Fiji Software.
Bisulfite sequencing
DNA was extracted with the DNeasy Plant Mini Kit (Qiagen). Subsequently, DNA was fragmented to a size of 350 bp through Covaris ultrasonication. Libraries were constructed with the Illumina TruSeq DNA nano kit. After the incorporation of adaptors, libraries underwent bisulfite conversion using the EpiTect Plus DNA bisulfite conversion kit (Qiagen), followed by treatment with Kapa Hifi Uracil+ DNA polymerase (Kapa Biosystems). Paired-end reads (2 × 150 bp) were sequenced on a HiSeq 3000 instrument (Illumina). Sequencing reads were first quality-trimmed and filtered with Trimmomatic (version 0.36)94 and Bismark95. The Bowtie2 mapper was used for quantifying cytosine methylation. The results were analyzed in R (R core team, 2022) with the methylKit package. Only cytosines with a coverage of 4 or more were considered. Chop–qPCR assays were performed using 200 ng of restriction-enzyme-digested (‘chopped’) genomic DNA with 10 U of MspI (methylation-sensitive), as described in a previous study33. Uncut genomic DNA was used as a control.
sRNA-seq and ChIP–seq analyses
The sRNA-seq data for Col-0, BR-0, dcl234 and ddc genotypes were generated and analyzed as described in a previous study34,96. The sRNA coverage over the Ea-IR element was generated with deepTools97 using the computeMatrix and plotProfile commands. Publicly available sRNA-Seq reads were retrieved and analyzed. The sequencing data used were from previous studies68[,69,83 and obtained from the GEO under accession numbers GSE79780, GSE19694 and GSE57191. ChIP–seq data for Pol IV (PRJNA193689) and Pol V (PRJNA169238) were used in this study. Reads were quality-trimmed with Trimmomatic and mapped to the genome with STAR. The coverage over the Ea-IR locus was visualized with Gviz.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
The datasets analyzed during the current study are available from the Gene Expression Omnibus (GEO) under accession numbers GSE79780, GSE19694, GSE57191 and GSE80744, and European Nucleotide Archive (ENA) under accession numbers PRJNA193689, PRJNA169238 and PRJEB65271. For the Col-0 reference genome, TAIR10 was used (https://www.arabidopsis.org/api/download-files/download?filePath=Sequences/Assemblies/TAIR9_chr_all.fas). Source data are provided with this paper.
References
Ngou, B. P. M., Ding, P. & Jones, J. D. G. Thirty years of resistance: zig-zag through the plant immune system. Plant Cell 34, 1447–1478 (2022).
Jones, J. D. & Dangl, J. L. The plant immune system. Nature 444, 323–329 (2006).
Zipfel, C. Pattern-recognition receptors in plant innate immunity. Curr. Opin. Immunol. 20, 10–16 (2008).
Li, X., Kapos, P. & Zhang, Y. NLRs in plants. Curr. Opin. Immunol. 32, 114–121 (2015).
Kourelis, J. & van der Hoorn, R. A. L. Defended to the nines: 25 years of resistance gene cloning identifies nine mechanisms for R protein function. Plant Cell 30, 285–299 (2018).
Bernoux, M., Zetzsche, H. & Stuttmann, J. Connecting the dots between cell surface- and intracellular-triggered immune pathways in plants. Curr. Opin. Plant Biol. 69, 102276 (2022).
Yuan, M., Ngou, B. P. M., Ding, P. & Xin, X. F. PTI–ETI crosstalk: an integrative view of plant immunity. Curr. Opin. Plant Biol. 62, 102030 (2021).
Yuan, M. et al. Pattern-recognition receptors are required for NLR-mediated plant immunity. Nature 592, 105–109 (2021).
Tian, H. et al. Activation of TIR signalling boosts pattern-triggered immunity. Nature 598, 500–503 (2021).
Ngou, B. P. M., Ahn, H. K., Ding, P. & Jones, J. D. G. Mutual potentiation of plant immunity by cell-surface and intracellular receptors. Nature 592, 110–115 (2021).
Boutrot, F. & Zipfel, C. Function, discovery, and exploitation of plant pattern recognition receptors for broad-spectrum disease resistance. Annu. Rev. Phytopathol. 55, 257–286 (2017).
Boller, T. & Felix, G. A renaissance of elicitors: perception of microbe-associated molecular patterns and danger signals by pattern-recognition receptors. Annu. Rev. Plant Biol. 60, 379–406 (2009).
Zipfel, C. et al. Perception of the bacterial PAMP EF-Tu by the receptor EFR restricts Agrobacterium-mediated transformation. Cell 125, 749–760 (2006).
Howard, B. E. et al. High-throughput RNA sequencing of Pseudomonas-infected Arabidopsis reveals hidden transcriptome complexity and novel splice variants. PLoS ONE 8, e74183 (2013).
Lewis, L. A. et al. Transcriptional dynamics driving MAMP-triggered immunity and pathogen effector-mediated immunosuppression in Arabidopsis leaves following infection with Pseudomonas syringae pv tomato DC3000. Plant Cell 27, 3038–3064 (2015).
Spears, B. J. et al. Direct regulation of the EFR-dependent immune response by Arabidopsis TCP transcription factors. Mol. Plant Microbe Interact. 32, 540–549 (2019).
Tintor, N. et al. Layered pattern receptor signaling via ethylene and endogenous elicitor peptides during Arabidopsis immunity to bacterial infection. Proc. Natl Acad. Sci. USA 110, 6211–6216 (2013).
Xu, M. et al. Natural variation in the transcription factor REPLUMLESS contributes to both disease resistance and plant growth in Arabidopsis. Plant Commun. 3, 100351 (2022).
Schoonbeek, H. J. et al. Arabidopsis EF-Tu receptor enhances bacterial disease resistance in transgenic wheat. New Phytol. 206, 606–613 (2015).
Lu, F. et al. Enhancement of innate immune system in monocot rice by transferring the dicotyledonous elongation factor Tu receptor EFR. J. Integr. Plant Biol. 57, 641–652 (2015).
Piazza, S. et al. The Arabidopsis pattern recognition receptor EFR enhances fire blight resistance in apple. Hortic. Res 8, 204 (2021).
Pfeilmeier, S. et al. Expression of the Arabidopsis thaliana immune receptor EFR in Medicago truncatula reduces infection by a root pathogenic bacterium, but not nitrogen-fixing rhizobial symbiosis. Plant Biotechnol. J. 17, 569–579 (2019).
Mitre, L. K. et al. The Arabidopsis immune receptor EFR increases resistance to the bacterial pathogens Xanthomonas and Xylella in transgenic sweet orange. Plant Biotechnol. J. 19, 1294–1296 (2021).
Ning, Y., Liu, W. & Wang, G. L. Balancing immunity and yield in crop plants. Trends Plant Sci. 22, 1069–1079 (2017).
Huot, B., Yao, J., Montgomery, B. L. & He, S. Y. Growth–defense tradeoffs in plants: a balancing act to optimize fitness. Mol. Plant 7, 1267–1287 (2014).
Barragan, A. C. & Weigel, D. Plant NLR diversity: the known unknowns of pan-NLRomes. Plant Cell 33, 814–831 (2021).
Gomez-Gomez, L. & Boller, T. FLS2: an LRR receptor-like kinase involved in the perception of the bacterial elicitor flagellin in Arabidopsis. Mol. Cell 5, 1003–1011 (2000).
Pruitt, R. N. et al. The EDS1–PAD4–ADR1 node mediates Arabidopsis pattern-triggered immunity. Nature 598, 495–499 (2021).
Torres Ascurra, Y. C. et al. Functional diversification of a wild potato immune receptor at its center of origin. Science 381, 891–897 (2023).
Baggs, E., Dagdas, G. & Krasileva, K. V. NLR diversity, helpers and integrated domains: making sense of the NLR IDentity. Curr. Opin. Plant Biol. 38, 59–67 (2017).
Vetter, M. M. et al. Flagellin perception varies quantitatively in Arabidopsis thaliana and its relatives. Mol. Biol. Evol. 29, 1655–1667 (2012).
Li, B., Meng, X., Shan, L. & He, P. Transcriptional regulation of pattern-triggered immunity in plants. Cell Host Microbe 19, 641–650 (2016).
Gagliardi, D. et al. Dynamic regulation of chromatin topology and transcription by inverted repeat-derived small RNAs in sunflower. Proc. Natl Acad. Sci. USA 116, 17578–17583 (2019).
Arce, A. L. et al. Polymorphic inverted repeats near coding genes impact chromatin topology and phenotypic traits in Arabidopsis thaliana. Cell Rep. 42, 112029 (2023).
Gagliardi, D. & Manavella, P. A. Short-range regulatory chromatin loops in plants. New Phytol. 228, 466–471 (2020).
Cuerda-Gil, D. & Slotkin, R. K. Non-canonical RNA-directed DNA methylation. Nat. Plants 2, 16163 (2016).
Martins, L. M. & Law, J. A. Moving targets: mechanisms regulating siRNA production and DNA methylation during plant development. Curr. Opin. Plant Biol. 75, 102435 (2023).
Ariel, F. D. & Manavella, P. A. When junk DNA turns functional: transposon-derived non-coding RNAs in plants. J. Exp. Bot. 72, 4132–4143 (2021).
Hung, Y. H. & Slotkin, R. K. The initiation of RNA interference (RNAi) in plants. Curr. Opin. Plant Biol. 61, 102014 (2021).
Manavella, P. A. et al. Beyond transcription: compelling open questions in plant RNA biology. Plant Cell 35, 1626–1653 (2023).
Nuthikattu, S. et al. The initiation of epigenetic silencing of active transposable elements is triggered by RDR6 and 21–22 nucleotide small interfering RNAs. Plant Physiol. 162, 116–131 (2013).
Sigman, M. J. et al. An siRNA-guided ARGONAUTE protein directs RNA polymerase V to initiate DNA methylation. Nat. Plants 7, 1461–1474 (2021).
Griffing, L. R., Gao, H. T. & Sparkes, I. ER network dynamics are differentially controlled by myosins XI-K, XI-C, XI-E, XI-I, XI-1 and XI-2. Front. Plant Sci. 5, 218 (2014).
Yang, L. et al. Myosins XI modulate host cellular responses and penetration resistance to fungal pathogens. Proc. Natl Acad. Sci. USA 111, 13996–14001 (2014).
Abu-Abied, M. et al. Myosin XI-K is involved in root organogenesis, polar auxin transport, and cell division. J. Exp. Bot. 69, 2869–2881 (2018).
Chen, J., Hu, Q., Zhang, Y., Lu, C. & Kuang, H. P-MITE: a database for plant miniature inverted-repeat transposable elements. Nucleic Acids Res. 42, D1176-81 (2014).
Ariel, F. et al. Noncoding transcription by alternative RNA polymerases dynamically regulates an auxin-driven chromatin loop. Mol. Cell 55, 383–396 (2014).
Roule, T. et al. The lncRNA MARS modulates the epigenetic reprogramming of the marneral cluster in response to ABA. Mol. Plant 15, 840–856 (2022).
Liu, Y. et al. PCSD: a plant chromatin state database. Nucleic Acids Res. 46, D1157–D1167 (2018).
Dowen, R. H. et al. Widespread dynamic DNA methylation in response to biotic stress. Proc. Natl Acad. Sci. USA 109, E2183–E2191 (2012).
Hannan Parker, A., Wilkinson, S. W. & Ton, J. Epigenetics: a catalyst of plant immunity against pathogens. New Phytol. 233, 66–83 (2022).
Lopez, A., Ramirez, V., Garcia-Andrade, J., Flors, V. & Vera, P. The RNA silencing enzyme RNA polymerase V is required for plant immunity. PLoS Genet. 7, e1002434 (2011).
Yu, A. et al. Dynamics and biological relevance of DNA demethylation in Arabidopsis antibacterial defense. Proc. Natl Acad. Sci. USA 110, 2389–2394 (2013).
Block, A. & Alfano, J. R. Plant targets for Pseudomonas syringae type III effectors: virulence targets or guarded decoys? Curr. Opin. Microbiol. 14, 39–46 (2011).
Stuart, T. et al. Population scale mapping of transposable element diversity reveals links to gene regulation and epigenomic variation. eLife 5, e20777 (2016).
Lian, Q. et al. A pan-genome of 72 Arabidopsis thaliana accessions reveals a conserved genome structure throughout the global species range. Nat. Genet. 56, 982–991 (2024).
Wlodzimierz, P. et al. Cycles of satellite and transposon evolution in Arabidopsis centromeres. Nature 618, 557–565 (2023).
Law, J. A. et al. Polymerase IV occupancy at RNA-directed DNA methylation sites requires SHH1. Nature 498, 385–389 (2013).
Kindgren, P., Ivanov, M. & Marquardt, S. Native elongation transcript sequencing reveals temperature dependent dynamics of nascent RNAPII transcription in Arabidopsis. Nucleic Acids Res. 48, 2332–2347 (2020).
Pabis, K., Barardo, D., Selvarajoo, K., Gruber, J. & Kennedy, B. K.A concerted increase in readthrough and intron retention drives transposon expression during aging and senescence. eLife 12, RP87811 (2024).
Rosa-Mercado, N. A. & Steitz, J. A. Who let the DoGs out?—biogenesis of stress-induced readthrough transcripts. Trends Biochem. Sci. 47, 206–217 (2022).
Rosa-Mercado, N. A. et al. Hyperosmotic stress alters the RNA polymerase II interactome and induces readthrough transcription despite widespread transcriptional repression. Mol. Cell 81, 502–513 (2021).
Johanson, T. M., Lew, A. M. & Chong, M. M. MicroRNA-independent roles of the RNase III enzymes Drosha and Dicer. Open Biol. 3, 130144 (2013).
Rybak-Wolf, A. et al. A variety of dicer substrates in human and C. elegans. Cell 159, 1153–1167 (2014).
Lee, D. & Shin, C. Emerging roles of Drosha beyond primary microRNA processing. RNA Biol. 15, 186–193 (2018).
Kim, B., Jeong, K. & Kim, V. N. Genome-wide mapping of Drosha cleavage sites on primary microRNAs and noncanonical substrates. Mol. Cell 66, 258–269 (2017).
Saldi, T., Fong, N. & Bentley, D. L. Transcription elongation rate affects nascent histone pre-mRNA folding and 3′ end processing. Genes Dev. 32, 297–308 (2018).
Panda, K. et al. Full-length autonomous transposable elements are preferentially targeted by expression-dependent forms of RNA-directed DNA methylation. Genome Biol. 17, 170 (2016).
Zhang, W. et al. Bacteria-responsive microRNAs regulate plant innate immunity by modulating plant hormone networks. Plant Mol. Biol. 75, 93–105 (2011).
Creasey, K. M. et al. miRNAs trigger widespread epigenetically activated siRNAs from transposons in Arabidopsis. Nature 508, 411–415 (2014).
Lee, S. C. et al. Arabidopsis retrotransposon virus-like particles and their regulation by epigenetically activated small RNA. Genome Res. 30, 576–588 (2020).
Panda, K., McCue, A. D. & Slotkin, R. K. Arabidopsis RNA polymerase IV generates 21–22 nucleotide small RNAs that can participate in RNA-directed DNA methylation and may regulate genes. Philos. Trans. R. Soc. Lond. B Biol. Sci. 375, 20190417 (2020).
Brown, J. K. Yield penalties of disease resistance in crops. Curr. Opin. Plant Biol. 5, 339–344 (2002).
Liu, M. et al. Inducible overexpression of Ideal Plant Architecture1 improves both yield and disease resistance in rice. Nat. Plants 5, 389–400 (2019).
Burdon, J. J. & Thrall, P. H. The fitness costs to plants of resistance to pathogens. Genome Biol. 4, 227 (2003).
Chandran, D. et al. Atypical E2F transcriptional repressor DEL1 acts at the intersection of plant growth and immunity by controlling the hormone salicylic acid. Cell Host Microbe 15, 506–513 (2014).
Lacombe, S. et al. Interfamily transfer of a plant pattern-recognition receptor confers broad-spectrum bacterial resistance. Nat. Biotechnol. 28, 365–369 (2010).
Roux, M. et al. The Arabidopsis leucine-rich repeat receptor-like kinases BAK1/SERK3 and BKK1/SERK4 are required for innate immunity to hemibiotrophic and biotrophic pathogens. Plant Cell 23, 2440–2455 (2011).
Cambiagno, D. A. et al. Immune receptor genes and pericentromeric transposons as targets of common epigenetic regulatory elements. Plant J. 96, 1178–1190 (2018).
Harris, C. J., Amtmann, A. & Ton, J. Epigenetic processes in plant stress priming: open questions and new approaches. Curr. Opin. Plant Biol. 75, 102432 (2023).
Wilkinson, S. W. et al. Long-lasting memory of jasmonic acid-dependent immunity requires DNA demethylation and ARGONAUTE1. Nat. Plants 9, 81–95 (2023).
Wilkinson, S. W. et al. Surviving in a hostile world: plant strategies to resist pests and diseases. Annu. Rev. Phytopathol. 57, 505–529 (2019).
McCue, A. D. et al. ARGONAUTE 6 bridges transposable element mRNA-derived siRNAs to the establishment of DNA methylation. EMBO J. 34, 20–35 (2015).
Weigel, D. & Colot, V. Epialleles in plant evolution. Genome Biol. 13, 249 (2012).
Kawakatsu, T. et al. Epigenomic diversity in a global collection of Arabidopsis thaliana accessions. Cell 166, 492–505 (2016).
Gonzalo, L. et al. R-loops at microRNA encoding loci promote co-transcriptional processing of pri-miRNAs in plants. Nat. Plants 8, 402–418 (2022).
Lu, C. et al. MicroRNAs and other small RNAs enriched in the Arabidopsis RNA-dependent RNA polymerase-2 mutant. Genome Res 16, 1276–1288 (2006).
Kurihara, Y. et al. Identification of the candidate genes regulated by RNA-directed DNA methylation in Arabidopsis. Biochem. Biophys. Res. Commun. 376, 553–557 (2008).
Schwab, R., Ossowski, S., Riester, M., Warthmann, N. & Weigel, D. Highly specific gene silencing by artificial microRNAs in Arabidopsis. Plant Cell 18, 1121–1133 (2006).
Li, H. New strategies to improve minimap2 alignment accuracy. Bioinformatics 37, 4572–4574 (2021).
Robinson, J. T. et al. Integrative Genomics Viewer. Nat. Biotechnol. 29, 24–26 (2011).
Katagiri, F., Thilmony, R. & He, S. Y. The Arabidopsis thaliana–Pseudomonas syringae interaction. Arabidopsis Book 1, e0039 (2002).
Mason, K. N., Ekanayake, G. & Heese, A. Staining and automated image quantification of callose in Arabidopsis cotyledons and leaves. Methods Cell Biol. 160, 181–199 (2020).
Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114–2120 (2014).
Krueger, F. & Andrews, S. R. Bismark: a flexible aligner and methylation caller for Bisulfite-Seq applications. Bioinformatics 27, 1571–1572 (2011).
Re, D. A. et al. CURLY LEAF regulates microRNA activity by controlling ARGONAUTE 1 degradation in plants. Mol. Plant 13, 72–87 (2020).
Ramirez, F. et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res. 44, W160–W165 (2016).
Acknowledgements
We thank Nicolás Cecchini (CIQUIBIC, Córdoba, Argentina) for sharing the Pst DC3000 hrcC strain and for fruitful discussions and Matias Capela for critical reading of the manuscript. This work was supported by grants from Agencia Nacional de Promoción Científica y Tecnológica (ANPCyT) (PICT-2021-I-A-00452), Universidad Nacional del Litoral (CAI+D 2020), and Spanish Ministry of Science and Innovation (PID2022-137037NB-I00 and CNS2023-145312) to P.A.M., by DFG-TRR356 and the Max Planck Society to D.W and “Proyecto QUAL21 012 IHSM, Consejería de Universidad, Investigación e Innovación, Junta de Andalucía”. A.L.A. are members of CONICET; R.M. is a fellow of ANPCyT. P.A.M is a member of CSIC. We thank the Secretaria de Ciencia de Santa Fe and Company of Biologists for a short-term fellowship to R.M.
Author information
Authors and Affiliations
Contributions
R.M. performed all wet-biology experiments with the help of A.C., G.S. and C.H. A.L.A. performed the bioinformatic analysis with the help of W.X. R.M., A.L.A., P.A.M. and D.W. conceptualized the study. P.A.M. and D.W. supervised the work and secured project funding. R.M., D.W. and P.A.M. wrote the manuscript with help from all authors.
Corresponding author
Ethics declarations
Competing interests
D.W. holds equity in Computomics, which advises plant breeders. D.W. also consults for KWS SE, a plant breeder and seed producer. All other authors declare no competing interests.
Peer review
Peer review information
Nature Structural & Molecular Biology thanks Elena Caro, Martin Crespi and Jurriaan Ton for their contribution to the peer review of this work. Peer reviewer reports are available. Primary Handling Editors: Sara Osman and Dimitris Typas, in collaboration with the Nature Structural & Molecular Biology team.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Source data
Source Data Fig. 1
Statistical source data.
Source Data Fig. 2
Statistical source data.
Source Data Fig. 3
Statistical source data.
Source Data Fig. 4
Statistical source data.
Source Data Fig. 5
Statistical source data.
Source Data Fig. 6
Statistical source data.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Mencia, R., Arce, A.L., Houriet, C. et al. Transposon-triggered epigenetic chromatin dynamics modulate EFR-related pathogen response. Nat Struct Mol Biol 32, 199–211 (2025). https://doi.org/10.1038/s41594-024-01440-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/s41594-024-01440-1
